Ruthenium-complex-catalyzed N-(Cyclo)alkylation of aromatic amines with diols. Selective synthesis of N-(ω-hydroxyalkyl)anilines of type PhNH(CH2)(n)OH and of some bioactive arylpiperazines

Article abstract of DOI:10.1021/jo972260d

A new class of well-defined neutral mono-, and dicationic ruthenium(II) complexes containing a neutral terdentate donor system [C₅H₃N(CH₂E)₂- 2,6] (E = PPh₂ (PNP) or NME₂ (NN'N) has been found effective as catalyst precursor in N-(cyclo)alkylation reactions of aromatic amines with diols Y(CH₂CH₂OH)₂ (Y = CH₂, NR). With these catalysts, N-phenylpiperidine is synthesized from aniline and 1,5-pentanediol in 85% yield (at 180 °C for 24 h in 1.4-dioxane). With neutral RuCl₂(NN'N)-(PPh₃) as a catalyst precursor, aniline can be selectively N-monoalkylated with diols of the type HO(CH₂)(n)OH (n = 4-6, 10) to give N-(n-hydroxyalkyl)anilines in 40-75 yield. To our knowledge, this represents the first useful catalytic route to this type of compounds. The new catalysts can also be used in the synthesis of arylpiperazines. For example, N-phenyl-N'-methylpiperazine is obtained from aniline and MeN(CH₂CH₂OH)₂ in yields up to 34%. N- [m(Trifluoromethyl)phenyl]-N'-methylpiperazine, TFMPMP, is successfully produced from m-(trifluoromethyl)aniline and MeN-(CH₂CH₂OH)₂ in 44% yield using monocationic [RuOTf(NN'N)(PPh₃)]OTf as the catalyst precursor. A mechanism for the N-(cyclo)alkylation reaction is proposed.

Full text of DOI:10.1021/jo972260d

4282
J . Org. Chem. 1998, 63, 4282-4290
Ru th en iu m -Com p lex-Ca ta lyzed N-(Cyclo)a lk yla tion of Ar om a tic
Am in es w ith Diols. Selective Syn th esis of
N-(ω-Hyd r oxya lk yl)a n ilin es of Typ e P h NH(CH2)n OH a n d of Som e
Bioa ctive Ar ylp ip er a zin es
Rob A. T. M. Abbenhuis, J aap Boersma, and Gerard van Koten*
Debye Institute, Department of Metal-Mediated Synthesis, Utrecht University,
Padualaan 8, 3584 CH Utrecht, The Netherlands
Received December 12, 1997
A new class of well-defined neutral mono-, and dicationic ruthenium(II) complexes containing a
neutral terdentate donor system [C5H3N(CH2E)2-2,6] (E ) PPh2 (P NP ) or NMe2 (NN′N)) has been
found effective as catalyst precursor in N-(cyclo)alkylation reactions of aromatic amines with diols
Y(CH2CH2OH)2 (Y ) CH2, NR). With these catalysts, N-phenylpiperidine is synthesized from aniline
and 1,5-pentanediol in 85% yield (at 180 °C for 24 h in 1,4-dioxane). With neutral RuCl2(NN′N)-
(PPh3) as a catalyst precursor, aniline can be selectively N-monoalkylated with diols of the type
HO(CH2)nOH (n ) 4-6, 10) to give N-(n-hydroxyalkyl)anilines in 40-75% yield. To our knowledge,
this represents the first useful catalytic route to this type of compounds. The new catalysts can
also be used in the synthesis of arylpiperazines. For example, N-phenyl-N′-methylpiperazine is
obtained from aniline and MeN(CH2CH2OH)2 in yields up to 34%. N-[m-(Trifluoromethyl)phenyl]-
N′-methylpiperazine, TFMPMP, is successfully produced from m-(trifluoromethyl)aniline and MeN-
(CH2CH2OH)2 in 44% yield using monocationic [RuOTf(NN′N)(PPh3)]OTf as the catalyst precursor.
A mechanism for the N-(cyclo)alkylation reaction is proposed.
In tr od u ction
and 1,2- and 1,3-dialcohols, respectively. Only very
recently was the first example of N-alkylation of het-
eroaromatic amines with alcohols reported.6
From a practical point of view, the reaction of amines
with 1,5-dialcohols is interesting since in this way
N-substituted saturated heterocyclic products like pip-
eridines, piperazines, or morpholines can be obtained in
one step7 (eq 1). The synthesis of arylpiperazines is of
The transition-metal-catalyzed N-alkylation reaction
of amines with alcohols as alkylating agents is a poten-
tially interesting route for the synthesis of secondary and
tertiary amines. The first catalysts reported for this type
of reactivity were metallic palladium1 and complexes of
rhodium and iridium.2 After ruthenium(II) complexes
had been recognized as active catalysts for the hydrogen
transfer from alcohols,3 a number of procedures have
been developed for ruthenium-catalyzed N-alkylation of
amines with primary alcohols.4 In this way, indoles5 and
quinolines5c have been prepared from aromatic amines
* To whom correspondence should be addressed. Tel.: +31 30
2533120. Fax: +31 30 2523615. E-mail: g.vankoten@chem.ruu.nl.
(1) Murahashi, S.-I.; Shimamura, T.; Moritani, I. J . Chem. Soc.,
Chem. Commun. 1974, 931.
(2) Grigg, R.; Mitchell, R. B.; Sutthivaiyakit, S.; Tongpenyai, N. J .
Chem. Soc., Chem. Commun. 1981, 611.
particular interest because some of these compounds find
applications as antipsychotic pharmaceuticals. A major
advantage of this catalytic method over conventional
methods lies in the fact that water instead of salts is
formed as the only side product. Typically, the reactions
are performed by heating the reaction mixture to 150-
200 °C for a few hours in a closed pressure reactor.
So far, only a few of the examples reported show
synthetic value, and also, the mechanism of this reaction
is not yet fully clarified. It is generally thought that the
reaction involves dehydrogenation of an alcohol substrate
by a ruthenium catalyst as a key step.4c-e,7a The most
active ruthenium catalyst found so far, RuCl2(PPh3)3, is
known to decarbonylate aldehyde intermediates to give
RuHCl(CO)(PPh3)3, which is ineffective as a catalyst.4e
(3) (a) Regen, S. L.; Whitesides, G. M. J . Org. Chem. 1972, 37, 1832.
(b) Sasson, Y.; Blum, J . Org. Chem. 1975, 40, 1887. (c) Tovrog, B. S.;
Diamond, S. E.; Mares, F. J . Am. Chem. Soc. 1979, 101, 5067. (d)
Tomioka, H.; Takai, K.; Oshima, K.; Nozaki, H. Tetrahedron Lett. 1981,
22, 2667. (e) Murahashi, S.-I.; Ito, K.-I.; Naota, T.; Maeda, Y.
Tetrahedron Lett. 1981, 52, 5327. (f) Matsumoto, M.; Kuroda, K. J .
Am. Chem. Soc. 1982, 104, 1433. (g) J ung, C. W.; Garrou, P. E.
Organometallics 1982, 1, 658. (h) Shvo, Y.; Blum, Y.; Reshef, D. J .
Organomet. Chem. 1982, 238, C79. (i) Blum, Y.; Shvo, Y. J . Organomet.
Chem. 1984, 263, 93. (j) Smith, T. A.; Maitlis, P. M. J . Organomet.
Chem. 1984, 269, C7. (k) Tsuji, Y.; Ohta, T.; Ido, T.; Munbu, H.;
Watanabe, Y. J . Organomet. Chem. 1984, 270, 333.
(4) (a) Watanabe, Y.; Tsuji, Y.; Ohsugi, Y. Tetrahedron Lett. 1981,
22, 2667. (b) Arcelli, A.; Khai, B.-T.; Porzi, G. J . Organomet. Chem.
1982, 235, 93. (c) Watanabe, Y.; Tsuji, Y.; Ige, H.; Ohsugi, Y.; Ohta, T.
J . Org. Chem. 1984, 49, 3359. (d) Ganguly, S.; J oslin, F. L.; Roundhill,
D. M. Inorg. Chem. 1989, 28, 4564-4565. (e) Ganguly, S.; Roundhill,
D. M. Polyhedron 1990, 9, 2517.
(6) Watanabe, Y.; Morisaki, Y.; Kondo, T.; Mitsudo, T. J . Org. Chem.
1996, 61, 4214.
(7) (a) Tsuji, Y.; Huh K.-T.; Ohsugi, Y.; Watanabe, Y. J . Org. Chem.
1985, 50, 1365. (b) Murahashi, S.-I.; Kondo, K.; Hakata, T. Tetrahedron
Lett. 1982, 2, 229. (c) Bitsi, G.; Schleiffer, E.; Antoni, F.; J enner, G. J .
Organomet. Chem. 1989, 373, 342. (d) Shim, S. C.; Doh, C. H.; Woo,
B. W.; Kim, H. S. J . Mol. Catal. 1990, 62, L11.
(5) (a) Tsuji, Y.; Huh, K.-T.; Yokoyama, Y.; Watanabe, Y. J . Chem.
Soc., Chem. Commun. 1986, 1575. (b) Tsuji, Y.; Huh, K.-T.; Watanabe,
Y. Tetrahedron Lett. 1986, 27, 377. (c) Tsuji, Y.; Huh, K.-T.; Watanabe,
Y. J . J . Org. Chem. 1987, 52, 1673. (d) Tsuji, Y.; Kotachi, S.; Huh,
K.-T.; Watanabe, Y. J . Org. Chem. 1990, 55, 580. (e) J pn. Kokai Tokkyo
Koho 81 61.353; Chem. Abstr. 1981, 95, 115293b.
S0022-3263(97)02260-3 CCC: $15.00 © 1998 American Chemical Society
Published on Web 06/05/1998

N-(Cyclo)alkylation of Aromatic Amines with Diols
J . Org. Chem., Vol. 63, No. 13, 1998 4283
Ta ble 1. Effect of Ca ta lyst on th e Syn th esis of
P ip er id in es fr om 1,5-P en ta n ed iol a n d An ilin ea
Moreover, RuCl2(PPh3)3 itself is not well-defined under
catalytic conditions. It shows a tendency to lose phos-
phine with formation of dimeric species.8
yieldb (%)
In a parallel project, we are investigating homogeneous
transfer-hydrogenation reactions involving well-defined
transition-metal catalyst containing monoanionic ligand
systems of type [C6H3(CH2E)2-2,6]- (E ) NMe2, [NCN]-
or PPh2, [P CP ]-).9 Recently, neutral ruthenium(II) com-
plexes containing these [ECE]- ligand systems10 have
proven to be very active catalysts in transfer-hydrogena-
tion reactions of ketones and imines11 and as such would
be promising catalysts for the direct alkylation reaction.
However, due to the thermal instability of these com-
plexes, they are not applicable in N-cycloalkylation reac-
tions for which relatively high temperatures of 140-180
°C are required. In order to establish the necessity of
the presence of a metal-carbon σ-bond in catalytically
active complexes, we studied potential ruthenium(II)
catalysts containing the neutral analogs of the [ECE]-
ligand system, i.e., the pyridine-based ligands [C5H3N-
(CH2E)2-2,6] (E ) NMe2, [NN′N] or PPh2, [P NP ]).12 We
report here on the Ru[ENE]-catalyzed N-(cyclo)alkylation
reaction of aromatic amines with (i) diols of type HO-
(CH2)nOH (n g 4) and (ii) N-substituted diethanolamines.
entry
catalyst
convnb (%)
100
6
7
1
2
RuCl3‚xH2O/3 PPh3
85
RuCl3‚xH2O/3 NN′N
3c
4d
5
RuCl2(PPh3)3
100
22
60
74
93
98
85
3
5
7
43
22
1
2
3
4
5
19
55
45
47
47
6
7
8
a
Pressure-reactor (Pmax ) 12 atm), reaction time 5 h, temper-
ature 180 °C, solvent 1,4-dioxane (50 mL), [aniline] ) 2.0 M, [diol]
) 3.0 M, [Ru] ) 0.02 M; NN′N ) 2,6-[bis(dimethylamino)meth-
yl]pyridine, P NP ) 2,6-[bis(diphenylphosphino)methyl]pyridine.
Conversion based on aniline. c See ref 7c. After 24 h.
b
d
(nbd ) 2,5-norbornadiene), [RuCl2(bpy)2] (bpy ) 2,2′-
bipyridine), or [RuCl2(terpy)(PPh3)] (terpy ) 2,2,6:2′′-
terpyridine) is catalytically active. Although the mixture
RuCl3‚nH2O/3PPh3 effectively catalyzes the reaction of
aniline with 1,5-pentanediol to give N-phenylpiperidine
in 85% yield (Table 1, entry 1),7a we found the mixture
RuCl3‚nH2O/3NN′N (Table 1, entry 2) to be ineffective.
In contrast, all complexes 1-5 showed at least some
catalytic activity (Table 1, entries 3-7). Interestingly,
the known RuCl2(PPh3)3 does not give monoalkylated 6
after 5 h at 180 °C (Table 1, entry 1), whereas a
considerable amount of this monoalkylated product is
formed in the reactions catalyzed by complexes 1-5. In
the case of complex 2, the yield of 6 is as high as 55% at
a conversion of aniline of 60%; this corresponds to a
selectivity of 92% (Table 1, entry 5). It is noteworthy that
no N,N-bis(5-hydroxypentyl)aniline is formed, which
might arise either from an amine-scrambling reaction of
the monoalkylation product 6 or from N-dialkylation of
aniline. The latter type of ruthenium-catalyzed reactivity
is well-studied.13 Obviously, in the case of complexes
1-5, amine exchange of 6 is not competing with N-
cycloalkylation to give 7. Furthermore, no formation of
1,5-dianilinopentane is observed, a product that could
result from a double monoamination of 1,5-pentanediol.
Also, no products arising from lactonization3e,14 of 1,5-
pentanediol are detected. It must be noted that all
catalytic runs give incomplete mass balances; up to 29%
of the aniline starting material could not be detected in
the products. These side products, probably arising from
polymerization reactions, could not be identified by
GCMS. The analogous reaction between more basic
benzylamine15 and 1,5-pentanediol is less effectively
catalyzed. With RuCl2(PPh3)3, full conversion of benzyl-
Resu lts
The catalysts used were the neutral, mono-, and
dicationic complexes 1-5,12 which contain the neutral
ligand system P NP or NN′N as well as one PPh3 ligand.
In these complexes, the ENE ligands act as terdentate
coordinating ligands. Complexes 1-4 are coordinatively
saturated with the ENE ligand coordinating meridionally
while complex 5 is a coordinatively unsaturated 16-
electron complex:
P ip er id in es fr om 1,5-P en t a n ed iol a n d a n Ar yl-
a m in e. The activity of complexes 1-5 and of some other
ruthenium(II) complexes was determined for a model
N-(cyclo)alkylation reaction of aniline with 1,5-pen-
tanediol (eq 2; Table 1). This reaction may give two
products, N-(5-hydroxypentyl)aniline, 6, and N-phenylpi-
peridine, 7, resulting from N-mono- and N-cycloalkylation
of aniline.
(8) Hoffman, P. R.; Caulton, K. G. J . Am. Chem. Soc. 1975, 94, 4221.
(9) (a) van Koten, G. Pure Appl. Chem. 1989, 61, 1681. (b) Rietveld,
M. H. P.; Grove, D. M.; van Koten, G. New J . Chem. 1997, 21, 751-
771.
(10) Sutter, J .-P.; Steenwinkel, P.; Karlen, T.; Grove, D. M.; Veld-
man, N.; Smeets, W. J . J .; Spek, A. L.; van Koten, G. Organometallics
1996, 15, 941.
(11) Karlen, T.; Grove, D. M.; van Koten, G. Manuscript in prepara-
tion.
(12) Abbenhuis, R. A. T. M.; Del R´ıo, I.; Bergshoef, M. M.; Boersma,
J .; Smeets, W. J . J .; Veldman, N.; Kruis, D.; Spek, A. L.; van Koten,
G. Inorg. Chem., in press.
(13) (a) Yoshimura, N.; Moritani, I.; Shimamura, T.; Murahashi, S.-
I. J . Am. Chem. Soc. 1973, 95, 3038. (b) Khai, B.-T.; Consiglio, C.; Porzi,
G. J . Organomet. Chem. 1981, 208, 249. (c) J ung, C. W.; Fellmann, J .
D.; Garrou, P. E. Organometallics 1983, 2, 1042.
(14) Murahashi, S.-I.; Naota, T.; Ito, K.; Maeda, Y.; Taki, H. J . Org.
Chem. 1987, 52, 4319.
Without a catalyst, no reaction is observed. Also, none
of the known complexes [RuCl2(PMe3)4], [RuCl2(nbd)]n

4284 J . Org. Chem., Vol. 63, No. 13, 1998
Abbenhuis et al.
Ta ble 2. [Ru Cl2L(P P h 3)]-Ca ta lyzed (L ) P P h 3)2 or NN′N)
N-(Cyclo)a lk yla tion of An ilin e w ith HO(CH2)n OH (n ) 4,
5, 6, 10)a
yieldb
entry
catalyst ) L
n
convnb
9
10
1
2
(PPh3)2
NN′N
(PPh3)2
NN′N
(PPh3)2
NN′N
(PPh3)2
NN′N
4
4
5
5
6
6
10
10
100
50
100
60
100
70
100
100
95
8 (90)c
85
5 (85)c
48
42
3d
4
55
2
66
45
70
5
6
7
8
4
a
Pressure-reactor (Pmax ) 12 atm), reaction time 5 h, temper-
ature 180 °C, solvent 1,4-dioxane (50 mL), [aniline] ) 2.0 M, [diol]
b
) 3.0 M, [Ru] ) 0.02 M. In %, based on aniline. c After 24 h.
d
See ref 7c.
F igu r e 1. Effect of reaction time on the RuCl2(PPh3)3-
catalyzed N-(cyclo)alkylation of aniline with 1,5-pentanediol:
conversion of aniline (0), yield of 6 (O), and yield of 7 (4).
Aniline (2.0 M), 1,5-pentanediol (3.0 M), RuCl2(PPh3)3 (0.02
M) in 1,4-dioxane at 180 °C.
selectivity for the formation of the monoalkylated product
6 (Table 1, entry 5) in the reaction catalyzed by complex
2 prompted us to study the scope of this reaction using
higher homologs of the diol with n ) 4, 5, 6, and 10.
Again, we compared the activity and selectivity of
complex 2 with that of RuCl2(PPh3)3 (see Table 2). In
these reactions, both N-(n-hydroxyalkyl)anilines 9 and
saturated cyclic amines 10 may be formed (eq 3). The
results in Table 2 indicate that in the RuCl2(PPh3)3-
catalyzed reaction of aniline with the diols with n ) 4
and 5 the corresponding cyclic products are formed
predominantly (Table 2, entries 1 and 3). Ring-closure
reactions of both N-(4-hydroxybutyl)aniline, 9a , and N-(5-
hydroxypentyl)-aniline, 6, under the influence of RuCl2-
(PPh3)3 have been reported.17 When the alkanediyl chain
is further extended (n ) 6-10), the amount of cyclization
product decreases and considerable amounts of N-
monoalkylated products are formed (Table 2, entries 5
and 7). For hexanediol (n ) 6), the formation of small
amounts of the bis-amination product, i.e., 1,6-dianili-
nohexane, are observed. However, with catalyst 2 the
N-monoalkylated products are isolated in fair to good
yields (Table 2, entries 2, 4, and 6). For the N-monoalky-
lation of aniline with 1,10-decanediol, complex 2 leads
to a more selective reaction than RuCl2(PPh3)3 (Table 2,
entries 8 and 7, respectively). When reaction times are
extended to 24 h, full conversion of aniline and selective
cyclization of the N-(n-hydroxyalkyl)anilines also occurs
with this catalyst. The results in Table 2 show that the
N-alkylation of aniline with the higher homologs HO-
(CH2)nOH (n g 4), catalyzed by complex 2, provides an
excellent route for the synthesis of N-(n-hydroxyalkyl)-
anilines, and to our knowledge, this represents the first
useful catalytic route to this type of compounds. Con-
ventional routes to the amino alcohols with (n ) 4 and
5) involve hydrogenation of R-aminotetrahydrofurans or
-pyranes18 or N-alkylation of aniline with halo alcohols
of type Cl(CH2)nOH (n ) 4 and 5)19 and proceed with
yields of 25-45%. N-(n-hydroxyalkyl)anilines with n )
F igu r e 2. Effect of reaction time on the complex 2-catalyzed
N-(cyclo)alkylation of aniline with 1,5-pentanediol: conversion
of aniline (0), yield of 6 (O), and yield of 7 (4). Aniline (2.0
M), 1,5-pentanediol (3.0 M), complex 2 (0.02 M) in 1,4-dioxane
at 180 °C.
amine is found after 5 h at 180 °C, but only a moderate
yield of 55% N-benzylpiperidine, 8. This reaction suffers
from competing amine scrambling13 of benzylamine due
to the presence of an R-hydrogen atom, resulting in the
formation of considerable amounts (30-40%) of diben-
zylamine.
Effects of Rea ction Tim e on th e N-(Cyclo)a lk yla -
tion of An ilin e w ith 1,5-P en ta n ed iol. The effects of
the reaction time on the N-(cyclo)alkylation of aniline
with 1,5-pentanediol catalyzed by either RuCl2(PPh3)3 or
complex 2 were measured. The results are presented in
Figures 1 and 2, respectively.
The similarity of the shapes of the curves in Figures 1
and 2 suggests that both N-cycloalkylations involve the
same type of mechanism; it is also clear that complex 2
is less active. Figure 2 shows that with 2 the cycloalky-
lation of N-monoalkylated intermediate 6 proceeds much
slower than the initial N-monoalkylation.
Selective N-Mon oa lk yla tion of An ilin e w ith Diols
of Typ e HO(CH2)n )OH (n g 4). The high (92%)
(15) Basicity of amines:16 Aniline, pKa ) 4.63; o-anisidine, pKa
4.52; benzylamine, pKa ) 9.33.
)
(17) Felfo¨ldi, K.; Klyavlin, M. S.; Barto´k, M. J . Organomet. Chem.
1989, 362, 193.
(18) Glacet, G. Bull. Soc. Chim. Fr. 1954, 575.
(16) Handbook of Chemistry and Physics; Lide, R. D., Ed.; CRC
Press: Boston, 1991; pp 8-37.

N-(Cyclo)alkylation of Aromatic Amines with Diols
J . Org. Chem., Vol. 63, No. 13, 1998 4285
Ta ble 3. Ru th en iu m -Ca ta lyzed Tr a n sfer Hyd r ogen a tion
of Cycloh exa n on e a n d Im in es w ith 2-P r op a n ola
yield (%)
entry
substrate
catalyst
(time (h))
1b
2b
cyclohexanone
cyclohexanone
cyclohexane
PhNdCHPh
PhNdCHPh
RuCl2(PPh3)3
1
2-5
RuCl2(PPh3)3
87 (5)
79 (5)
85-87 (5)
88 (8)
88 (7.5)
48 (18)
41 (7.5)
3-6b
7c,d
8c
2
9c,d
10c
PhCH2NdC(Me)P
PhCH2NdC(Me)Ph
RuCl2(PPh3)3
2
a
Reaction time 5 h, reflux temperature, solvent 2-propanol (50
mL), [cyclohexanone] or [imine] ) 1.0 M, [Ru] ) 0.01 M. b 1.5 mmol
NaOH as cocatalyst. c 2.5 mmol K2CO3 as cocatalyst. See ref 23b.
d
6-10 have not yet been reported in the literature. Only
a few reports have been made on selectivity control in
the N-mono- and N-dialkylation of amines with alcohols.
An example of this is the amination of ethylene glycols.20
Only very recently, N-mono- or N-dialkylation of het-
eroaromatic amines with primary monoalcohols catalyzed
by RuCl2(PPh3)3 was reported.6
Activity of Com p lexes 1-5 in Tr a n sfer -Hyd r oge-
n a tion Rea ction s. To further compare the Ru[NN′N]
systems with RuCl2(PPh3)3, we have tested our catalyst
precursors in transfer-hydrogenation reactions of propan-
2-ol with cyclohexanone21 (eq 4) and imines22 (eq 5) (Table
3). These reactions were cocatalyzed by NaOH (in the
F igu r e 3. Reaction curves of ruthenium-catalyzed transfer
hydrogenation of cyclohexanone with propan-2-ol: RuCl2(PPh3)3
(0), RuCl2(NN′N)(PPh3) (∆), RuCl2(P NP )(PPh3) (O).
small amounts of aldol condensation products were
formed. The complex 1-catalyzed reaction gave a lower
yield of 79% (TTO 790), and more of the aldol condensa-
tion products were found. Some of the transfer-hydro-
genation reactions of cyclohexanone with propan-2-ol
were also monitored by GC (Figure 3). From the curves
in Figure 3, it can be seen that the reaction rates for the
Ru[NN′N] catalyst precursors are lower than those
observed for RuCl2(PPh3)3.
When compared to the ruthenium-catalyzed transfer
hydrogenation of ketones, imines react much slower. This
may be due to both steric and electronic effects. The
results in Table 3 show that our catalyst precursors have
a catalytic activity comparable to that of RuCl2(PPh3)3
in transfer-hydrogenation reactions of imines with pro-
pan-2-ol. As has already been recognized by other
authors,22 aldimines generally react faster than ketimines.
The two examples in this work suggest that this is also
true for the Ru[NN′N]-catalyzed transfer-hydrogenation
reactions. When 1,5-pentanediol is reacted with RuCl2-
(PPh3)3 or complex 1 or 2, selective formation of (tetrahy-
dropyranyloxy)-1-pentanol, 11, in ∼35% yield is observed.
This complex is the product of a selective oxidative
coupling reaction of two molecules of 1,5-pentanediol with
exclusion of water. In these reactions, neither δ-valero-
lactone nor the esterification product 5-(hydroxypentyl)-
5-hydroxypentanoate is found (eq 6). These results also
show that the catalyst precursors 1 and 2 behave
similarly to RuCl2(PPh3)3.
case of cyclohexanone) or K2CO3 (in the case of imines)
using 30 equiv of base based on [Ru]. The transfer-hy-
drogenation reaction of imines with propan-2-ol has much
in common with the N-alkylation reaction of amines with
alcohols: in both cases, it is proposed (vide infra) that
after dehydrogenation of the alcohol the hydrogen pro-
duced is transferred to an imine to yield an amine. In
the case of the N-alkylation reactions, an imine is
proposed to result from an intermediate condensation
reaction of an amine substrate with the aldehyde formed.
All transfer-hydrogenations of cyclohexanol with pro-
pan-2-ol catalyzed by complexes 2-5 (Table 3) gave high
yields of cyclohexanol (86-89% after 5 h), corresponding
to total turnover (TTO) numbers of 850-870. For RuCl2-
(PPh3)3, typical turnover numbers of 890/h have been
observed.21 It must be noted that the total turnover
numbers for the Ru[NN′N] complexes are considerably
lower than those we recently observed for Ru[ECE]
complexes. For example, for the neutral complex [RuOTf-
(P CP )(PPh3)], a turnover number as high as 25 000/h
was achieved over a period of 12 h.11 During all of the
Ru[NN′N]-catalyzed transfer-hydrogenation reactions,
Ar ylp ip er a zin es fr om Ar om a tic Am in es a n d Di-
eth a n ola m in es. Classically, arylpiperazines are ob-
tained by ring closure of suitably substituted anilines and
bis(2-chloroethyl)amine hydrochloride in the presence of
base.23 In these reactions, yields of around 40% are
obtained after reaction times of 50 h or more.
(19) (a) Kon, G. A. R.; Roberts, J . J . J . Chem. Soc. 1950, 978-982.
(b) Everett, J . L.; Ross, W. J . C. J . Chem. Soc. 1949, 1972-1983.
(20) (a) Marsella, J . A. Adv. Chem. Ser. 1992, 230, 433. (b) Marsella,
J . A. U.S. Patent 4.680.393, 1987.
(21) Chowdury, R. L.; Ba¨ckvall, J .-E. J . Chem. Soc., Chem. Commun.
1991, 1063.
(22) Wang, G.-Z.; Ba¨ckvall, J .-E. J . Chem. Soc., Chem. Commun.
1992, 980.
(23) Kiritski, J . A.; Yung, D. K.; Mathony, D. E. J . Med. Chem. 1978,
21, 1301.

4286 J . Org. Chem., Vol. 63, No. 13, 1998
Abbenhuis et al.
in 84% yield7a). Interestingly, when complexes 4 and 5
were used, N-(2-methoxyphenyl)-N′-methylpiperazine was
formed, albeit in low yield (ca. 10%). These low yields
are probably caused by steric effects, the basicity of
o-anisidine being comparable to that of aniline.15 Also,
N-[m-(trifluoromethyl)phenyl])-N′-methylpiperazine
(TFMPMP, 13),27 a potent serotonine agonist, has been
synthesized from m-(trifluoromethyl)aniline and N-me-
thyldiethanolamine (eq 8 and Table 4).
Ta ble 4. Ru th en iu m Com p lex-Ca ta lyzed Syn th esis of
N-Meth yl-N′-p h en ylp ip er a zin e, 12, a n d
N-[m -(Tr iflu or om eth yl)p h en yl]-N′-m eth ylp ip er a zin e, 13a
yield (%)
entries
catalyst
12
13
1b,c
2, 3
4, 5
6, 7
8, 9
RuCl2(PPh3)3
RuCl2(PPh3)3
2
3
4
5
32
23 (23)d
24
7 (23)e
23
25
22
32
34
30
10, 11
30 (44)e
a
Pressure-reactor (Pmax ) 14 atm), reaction time 5 h, temper-
ature 180 °C, solvent 1,4-dioxane [aniline] ) 3.0 M; [N-methyldi-
b
ethanolamine] ) 2.0 M; 2 mol % [Ru] based on aniline. [Ru] ) 3
d
mol % based on aniline. c See ref 7a. After 18 h. e After 24 h.
Recently, some new methods for the synthesis of
arylpiperazines have been developed that are based on
the palladium-catalyzed coupling of aryl halides24 or on
the reaction of (η6-fluoroarene)tricarbonylchromium com-
plexes with unprotected piperazines.25 However, in these
reactions, halogen-containing substrates are still required
and salts are formed as a side product. This makes these
reactions less interesting from an industrial point of view.
In the present study, we have extended the Ru[NN′N]-
catalyzed N-cycloalkylation reaction of aromatic amines
with diols to the synthesis of N-phenyl-N′-methylpipera-
zine, 12, from aniline and N-methyldiethanolamine (eq
7 and Table 4). Excess aniline (aniline/diol ratio ) 3/2)
The results in Table 4 show that in this particular
reaction the NN′N-containing ruthenium(II) catalyst
precursors are more active than RuCl2(PPh3). With
complex 5, 13 was obtained in 44% yield (Table 4, entry
11). Recently, an alternative synthesis of this product
was reported, which involves the reaction of m-(trifluo-
romethyl)aniline with bis(2-bromoethyl)amine over a
solid alumina support (70% yield).28 To synthesize the
N-benzyl-substituted analog of 13, a reaction of m-
(trifluoromethyl)aniline with N-benzyldiethanolamine
was carried out; this gave the desired N-benzylated
product 14 in 29% yield using 5 as a catalyst. The
N-benzyl-substituted piperazine 14 is of particular inter-
est, as a benzyl group can be easily removed by a simple
hydrogenation step, giving the unprotected piperazine.
A problem is that this N-cycloalkylation suffers from a
competing amine-scrambling reaction, giving substantial
amounts of dibenzylamine. Finally, we applied our
method to the synthesis of N-methyl-N′-benzhydrylpip-
erazine (Valoid, Cyclizine, 15), a drug against travel
sickness.29 However, using our catalytic systems 1-5,
only low yields (<10%) of the desired product were
obtained. It must be noted that in none of the arylpip-
erazine syntheses reported here could unreacted aniline
starting materials be recovered. Also, we have been
unable to isolate or identify any side products.
was used as it is known that in reactions of aromatic
amines with N-substituted diethanolamines no pipera-
zines are obtained when excess N-methyldiethanolamine
is used.7a This is in contrast to the arylpiperidine
syntheses, where an excess of 1,5-pentanediol (aniline/
diol ratio ) 2/3) is favored. The arylpiperazine syntheses
are known7a to give only low yields. Interestingly, our
new catalyst precursors give comparable or higher yields
than RuCl2(PPh3)3 (entries 2, 3-10, 11). For example, 5
improves the yield of N-methyl-N′-phenylpiperazine to
34% (entry 10). During these reactions, no formation of
4-methylmorpholin-2-one from lactonization of N-meth-
yldiethanolamine is observed, a product that has been
reported to be formed under the influence of RuH2-
(PPh3)4.14
We have also applied our catalysts to the synthesis of
some bioactive arylpiperazines. First, we tried to syn-
thesize N-(2-methoxyphenyl)-N′-methylpiperazine, a com-
pound that has recently been recognized26 as biologically
active, from o-anisidine and N-methyldiethanolamine. In
the presence of either RuCl2(PPh3)3 or complex 2, no
reaction was observed under standard conditions (the
analogous reaction of o-anisidine with 1,5-pentanediol
has been reported to give N-(2-methoxyphenyl)piperidine
Discu ssion
Effect of th e Na tu r e of th e Ca ta lyst P r ecu r sor .
The nature of the catalyst precursor has a critical
influence on the N-alkylation reaction. Early catalysts4
that have been used for N-(cyclo)alkylation reactions are
neutral ruthenium(II) complexes with monodentate phos-
phine ligands as well as chloride anions. Ruthenium(II)
complexes containing bidentate nitrogen or phosphorus
(24) Zhao, S.-H.; Miller, A. K.; Berger, J .; Flippin, L. A. Tetrahedron
Lett. 1996, 37, 4463.
(25) Perez, M.; Potier, P.; Halazy, S. Tetrahedron Lett. 1996, 37, 848.
(26) Verdonk, M. L. Reflections on Structural Aspects of Serotonin
5-HT2C Receptor Pharmacology. Thesis, Utrecht University, 1995; p
70.
(27) Fuller, R. W.; Maso, N. R.; Moloy, B. B. Biochem. Pharmacol.
1980, 29, 833.
(28) Mishani, E.; Dence, C. S.; McCarthy, T. J .; Welch, M. J .
Tetrahedron Lett. 1996, 37, 319.
(29) Clubiey, M.; Henson, T.; Peck, A. W.; Riddington, C. Br. J . Clin.
Pharmacol. 1977, 4, 652.

N-(Cyclo)alkylation of Aromatic Amines with Diols
J . Org. Chem., Vol. 63, No. 13, 1998 4287
Sch em e 1
ligands, or mixed nitrogen-phosphorus donor ligands
(e.g., RuCl2(bpy)2), RuCl2(dppe)2 (dppe ) 1,2-bis(diphe-
nylphosphino)ethane) or RuCl2(P N)2 (P N ) 2-(diphe-
nylphosphino)-N,N′-dimethylaniline) are ineffective in
the reaction of alkylamines with long-chain alcohols.4e
However, upon replacement of terpy by P NP (complex
1), some catalytic activity is found (Table 1, entry 4).
Replacement of the PPh2 groups in P NP by purely
σ-donating NMe2 groups to give NN′N then drastically
increases the catalytic activity, cf. complex 2 (Table 1,
entry 5). Obviously, the combination of one sp2- and two
sp3-hybridized nitrogens (as in NN′N) gives rise to active
catalysts. As had been recognized earlier, ruthenium
catalyst precursors containing basic phosphines like
PMe3 are inactive in the N-(cyclo)alkylation of aromatic
amines.7a When the chloride ligands in [RuCl2(NN′N)-
(PPh3)], 2, are replaced by acetonitrile ligands, the
resulting mono- and dicationic complexes 3 and 4,
respectively, show an increase in conversion of aniline
(Table 1, entries 6 and 7), although the selectivity of the
reaction for 6 and 7 is slightly reduced. This is the first
successful application of cationic ruthenium(II) complexes
in N-alkylation reactions. The high activity of the
cationic complexes 3 and 4 suggests that in solution free
coordination sites are made available by Ru-MeCN bond
dissociation. The coordinatively unsaturated complex 5
contains a weakly coordinating triflate ligand. The high
activity of this complex may similarly be explained by
facile dissociation of the η1-bonded OTf ligand.30 During
the reactions, no loss of P NP or NN′N is observed, which
indicates that the catalytically active species indeed
contains a Ru[ENE]-moiety.
monitored by 31P NMR spectroscopy. 1 itself shows a
characteristic pair of one-doublet and one-triplet reso-
nances due to the presence of a MX2Y system. During a
model N-cycloalkylation reaction of aniline with 1,5-
pentanediol using catalyst precursor 1, samples were
taken from the reaction mixture and their 31P NMR
spectra were recorded (see the Supporting Information).
In none of the spectra was free P NP observed, but during
the reaction a new set of a doublet and a triplet appears
at 48.6 and 57.9 ppm with a coupling constant of 30 Hz.
Such a pattern is characteristic for a complex containing
two equivalent P nuclei coupled to a third phosphorus
atom as in the [Ru(P NP )(PPh3)] moiety. During the
reaction, an increasing amount of free triphenylphos-
phine (δ ) -5 ppm) is observed. After prolonged reaction
times (>24 h), the intensities of all of the resonances in
the 31P{1H} NMR spectrum decrease in intensity and
gradually disappear completely. Along with this, a
precipitate is formed, and the catalytic activity goes back
to zero. Apparently, the catalyst decomposes.
(ii) Effect of P h osp h in e Liga n d s on Ca ta lytic
Activity. The presence of phosphine ligands is regarded
to be essential for any catalytic activity.4c-e,7a Even in
the case of our Ru[NN′N] complexes, the presence of a
triphenylphosphine ligand seems to be essential for
catalytic activity, cf. the ineffectiveness of the combina-
tion of RuCl3‚xH2O/NN′N. Indeed, in the N-cycloalky-
lation of aniline with 1,5-pentanediol catalyzed by RuCl2-
(NN′N)(PPh3), 2, a 36% increase in reactivity was found
when 3 equiv of triphenylphosphine was added to the
reaction mixture. Roundhill et al.4e reported a compa-
rable increase in catalytic activity for the RuCl2(PPh3)3-
catalyzed N-alkylation of C16H33OH with diisopropyl-
Ca ta lysts for th e P ip er a zin e Syn th eses. In con-
trast to the piperidine syntheses, in the synthesis of
piperazines our catalyst precursors 1-5 are more active
than conventional RuCl2(PPh3)3, although the overall
yields are still not satisfactory. The low activity of RuCl2-
(PPh3)3 may be due to deactivation of the catalyst by
interaction of the basic nitrogen function in the N-
substituted diethanolamines with ruthenium. Such in-
teractions are less likely in our Ru[NN′N] complexes in
which the terdentate-bonded NN′N ligand is unlikely to
be replaced by N-methyldiethanolamine. This is cor-
roborated by reactions of complex 2 with N-methyldi-
ethanolamine in which no dissociation of the NN′N ligand
was observed. Also, at room temperature, no interaction
between the ruthenium complexes 1-5 and aniline, 1,5-
pentanediol, or N-methyldiethanolamine could be de-
1
tected by H and 31P NMR and UV/vis spectroscopy.
Other authors have observed a relation between the
basicity of the amine substrate and the basicity of the
phosphine ligands31 in ruthenium catalysts, i.e., more
basic alkylamines require ruthenium complexes contain-
ing more basic alkylphosphines.7a The ineffectiveness of
complex 5 for the synthesis of N-methyl-N′-benzhy-
drylpiperazine 15 is in accord with this qualitative
relationship.
Mech a n istic Con sid er a tion s. The results obtained
with the Ru[NN′N] catalysts allow a detailed mechanism
for the N-alkylation of amines with alcohols to be
proposed (Scheme 1). The different steps in this mech-
anism will be discussed.
(i) F a te of th e Ca ta lyst P r ecu r sor . The fate of the
complex RuCl2(PNP)(PPh3), 1, during the catalysis was
(31) Basicity of phosphines:32 PPh3, pKa ) 2.73; PBu3, pKa ) 8.43.
(32) Henderson, W. A., J r.; Streuli, C. A. J . Am. Chem. Soc. 1960,
82, 5791.
(30) Lawrence, G. A. Chem. Rev. 1986, 86, 17.

4288 J . Org. Chem., Vol. 63, No. 13, 1998
Abbenhuis et al.
amine (i-Pr)2NH upon addition of 4 equiv of triphen-
ylphosphine. They suggested that triphenylphosphine
may displace the complexed aldehyde to facilitate its
reaction in solution with amine4e or that it may accelerate
the catalytic hydrogenation of imine to tertiary amine
(compare the N-alkylation of 2-aminopyridine with etha-
nol catalyzed by (η4-1,5-cyclooctadiene)(η6-1,3,5-cyclooc-
tatriene)ruthenium (Ru(cod)(cot)), which is also positively
affected by the addition of phosphine ligands6).
ment. According to the transfer-hydrogenation mecha-
nism,34 II can then be converted into an alkoxyruthenium
species that undergoes subsequent â-elimination to pro-
vide a ruthenium(aldehyde)(hydride) intermediate. Re-
action of the protonated base with the latter ruthen-
ium(aldehyde)(hydride) results in formation of dihydro-
gen.3b,35-37 Indeed, after each catalytic run, we found a
substantial amount of hydrogen gas (2-5 atm) in the
pressure reactor. Moreover, from a separate experiment
in which the reaction between aniline and 1,5-pentane-
diol was performed under 10 atm of hydrogen pressure,
we found that the activity decreases drastically: after 5
h under standard conditions, only trace amounts of
N-phenylpiperidine and monoalkylated N-(5-hydroxy-
pentyl)aniline were formed. This suggests that molecular
hydrogen is present during the reaction and acts as a
competing ligand blocking the catalytic activity. There-
fore, we propose an equilibrium between a Ru(alkoxide)-
(hydride) and a Ru(aldehyde) complex III.
(v) Step 5. Con d en sa tion of Ald eh yd e a n d Am in e
in th e Ru th en iu m Coor d in a tion Sp h er e. We propose
that in a next step the aldehyde condenses with the
amine to an imine intermediate and 1 equiv of water.4c,7a
During the reaction, no imine intermediates could be
detected by GCMS analysis. This suggests that such
intermediates are rapidly hydrogenated to the alkylated
products. Despite the comparable activity of RuCl2-
(PPh3)3 and our Ru[NN′N] complexes in transfer-hydro-
genation reactions (see Table 3), the latter were found
to be far less active in the N-cycloalkylation of aniline
with 1,5-pentanediol. Therefore, the proposed condensa-
tion reaction between aldehyde and amine (giving imine)
must be the rate-limiting step. This is also suggested
by other authors.4c,7a Furthermore, the rate differences
in catalytic activity of the Ru[NN′N] complexes indicate
that the condensation reaction takes place in the coor-
dination sphere of the ruthenium center, a feature that
has also been postulated.4c,7a Supporting evidence is the
observation that during the N-cycloalkylation reactions
no lactonization products are formed (vide supra). This
suggests that the stronger coordination of amine sub-
strates in comparison with alcohols inhibits bidentate
coordination of the dialcohol substrates and thus prevents
a ring-closure reaction. A similar effect has been found
in the oxidative transformation of dialcohols catalyzed
by RuH2(PPh3)4.14 For example, in the absence of the
strong donor acetonitrile, 1,5-pentanediol is converted in
δ-valerolactone, whereas in the presence of acetonitrile
it is converted into 5-hydroxypentyl-5-hydroxypentanoate.
The water that is formed in the condensation reaction
does not effect the N-alkylation reaction. In a separate
experiment in which 50 mmol of water was added to the
reaction mixture of aniline with 1,5-pentanediol using
complex RuCl2(NN′N)(PPh)3, 2, as a catalyst, no differ-
ences in the product distribution or reaction rate were
found.
(iii) Step 1: Activa tion of Coor d in a tively Sa tu -
r a ted Ru [NN′N] Com p lexes. We have observed that
RuCl2(NN′N)(PPh3), 2, easily loses chloride ions in MeCN
at 55 °C to form a monocationic complex that can be
formulated as [mer-RuCl(NN′N)(PPh3)-(MeCN)]Cl.12
Therefore, we propose that the first step in the catalytic
cycle consists of dissociation of one of the axial ligands.
This may give either a five-coordinate 16-electron ruthe-
nium(II) complex with a square pyramidal geometry, cf.
the structure of complex 5, or a six-coordinate 18-electron
complex. In the 16-electron complex, both in the solid
state and in solution, the triphenylphosphine occupies
the apical position of a square pyramid12 (see Scheme 1).
This can be converted into a structure I analogous to 5
via a Berry-pseudo rotation. In the six-coordinate com-
plex, the sixth coordination site is occupied by the
substrate (aniline or the dialcohol), cf. intermediate II
in Scheme 1, which was identified separately by an X-ray
determination.12 As stated earlier, whereas the presence
of excess diol is required in the synthesis of piperidines,
the syntheses of piperazines do not proceed when an
excess of diethanolamine is used. Watanabe et al.4c have
shown that for the N-monoalkylation of aromatic amines
with primary alcohols the reaction is zero order in alcohol
concentration, and they suggested that this is also the
case for the N,N-dialkylation and N-cycloalkylation reac-
tions. However, the piperazine syntheses are obviously
not zero order in diethanolamine concentration. Al-
though this implies differences in the mechanistic details,
it is difficult to evaluate them.
(iv) Step s 2-4. Coor d in a tion a n d Deh yd r ogen a -
tion of Alcoh ol Su bstr a te: Red u ctive Elim in a tion
of Dih yd r ogen . In mechanisms proposed so far for the
N-alkylation reaction catalyzed by RuCl2(PPh3)3, dehy-
drogenation of alcohol into aldehyde is a key step.4c-e,7a
For the N-alkylation of secondary amines with alcohols,
Roundhill et al.4c,d proposed the alcohol to add oxidatively
to a [RuCl2(PPh3)2] intermediate (resulting from dissocia-
tion of triphenylphosphine from RuCl2(PPh3)3) to give
[RuIVHCl2CH2R)(PPh3)2], which subsequently reductively
eliminates HCl and produces [RuIICl(OCH2CH2R)P2L] (L
) phosphine or amine). Formation of HCl is well known
in reactions of RuCl2(PPh3)3 with alcohols.33 Some
authors4c,7a have pointed out that the presence of a
chloride anion is a prerequisite for catalytic activity in
N-alkylation reactions. However, this does not apply to
our catalytic systems, as our mono- and dicationic
complexes [RuOTf(NN′N)(PPh3)]OTf, 2, and [Ru(NN′N)-
(MeCN)2(PPh3)]OTf2, 4, which both lack chloride anions,
are more active than the neutral analog [RuCl2(NN′N)-
(PPh3)]. An alcohol substrate can coordinate directly to
coordinatively unsaturated ruthenium intermediate I to
give II (see Scheme 1); on the other hand, formation of
II could also occur directly from the starting material
RuCl2(NN′N)(PPh3) via direct solvent or anion replace-
(vi) Step 6. P r od u ct F or m a tion by Hyd r ogen a -
tion of th e Un sa tu r a ted In ter m ed ia te. In a final
step, the imine intermediate was hydrogenated, releasing
the monoalkylated secondary amine product. This prod-
uct can re-enter the catalytic cycle and can be intramo-
(34) Ba¨ckvall, J .-E.; Wang, G.-Z. Perspect. Coord. Chem. 1992, 463-
486.
(35) Sasson, Y.; Blum, J . J . Chem. Soc., Chem. Commun. 1974, 309.
(36) Speier, G.; Marko, L. J . Organomet. Chem. 1981, 210, 253.
(37) Kaesz, H. D.; Saillant, R. B. Chem. Rev. 1972, 72, 231.
(33) Sasson, Y.; Rempel, G. L. Tetrahedron Lett. 1974, 3221.

N-(Cyclo)alkylation of Aromatic Amines with Diols
J . Org. Chem., Vol. 63, No. 13, 1998 4289
Gen er a l P r oced u r e for th e N-(Cyclo)a lk yla tion Rea c-
tion s. A typical reaction of aniline with 1,5-pentanediol will
be described here. A stainless steel reactor (300 mL, Parr 4600
minireactor) was charged (under a nitrogen stream) with
dioxane (25 mL), aniline (9.32 g, 100 mmol), 1,5-pentanediol
(15.62 g, 150 mmol), and 1 (0.63 g, 1 mmol). After the reactor
was sealed, a nitrogen purge was performed by three pres-
surization-depressurization sequences. Subsequently, the
reactor was heated to 180 °C in 15 min in a mantle heater
and kept at this temperature for 5 h with stirring. The
reaction was terminated by rapid cooling, and the reactor was
discharged.
lecularly cyclized into a tertiary cyclic amine. As has
been stated before, this cycloalkylation of the monoalky-
lated products proceeds slower with Ru[NN′N]-based
catalysts than with RuCl2(PPh3)3. The only important
difference between the first alkylation step and the
second is that the condensation reaction between second-
ary amine and aldehyde in the latter case should give
an iminium intermediate instead of an imine. Several
authors have proposed the second cycle to proceed via a
Ru(iminium) complex.4c,7a However, a positively charged
iminium ion is unlikely to coordinate to a cationic ruthen-
ium center. Therefore, we propose the reaction to proceed
via a ruthenium(ene-amine) intermediate. On the basis
of a series of catalyzed and uncatalyzed reactions of
valeraldehyde with diethylamine, Roundhill et al. have
also postulated an ene-imine instead of an iminium
intermediate.4d,e
Tr a n sfer -H yd r ogen a t ion R ea ct ion s. All hydrogen-
transfer experiments were carried out under
a nitrogen
atmosphere in refluxing propan-2-ol with magnetic stirring.
To solid 1 (40.2 mg, 0.0641 mmol), after evacuation and
purging with nitrogen (3×), was added propan-2-ol (32.0 mL),
and this mixture was heated at 82 °C for 10 min. Cyclohex-
anone (6.28 g, 64.0 mmol) dissolved in propan-2-ol (16.5 mL)
was added dropwise to the refluxing mixture. The resulting
reddish brown mixture was stirred for 10 min, and then a
solution of NaOH (77.0 mg, 1.925 mmol) in propan-2-ol (15.6
mL) was added dropwise. The reaction mixture rapidly turned
into a clear dark-red colored solution after addition of the base.
Hexanol was used as an internal standard during the GC
analyses of the reaction mixtures.
N-(5-Hyd r oxyp en tyl)a n ilin e (6). The reaction mixture
was extracted with water (100 mL) to remove remaining diol.
Column chromatography of the evaporated reaction mixture
on silica gel (Merck, 60-230 mesh) with a hexane/ethyl acetate
mixture as eluent (gradient from 98:2 to 2:1 in volume) gave
N-(5-hydroxypentyl)aniline. The product was purified by
further Kugelrohr distillation and was obtained as a colorless
oil: pot temperature 105 °C (0.01 mmHg); 1H NMR (300 MHz)
(CDCl3) δ 7.23 (t, 2H), 6.75 (t, 1H), 6.65 (d, 2H), 3.63 (t, 2H),
3.13 (t, 2H), 1.68-1.60 (m, 4H), 1.50-1.48 (m, 2H), OH and
NH not observed; 13C (75.47 MHz) (CDCl3) δ 148.57, 129.30,
117.28, 112.93, 62.50, 44.00, 32.46, 29.3, 23.45. MS m/ z 179
(M+). Anal. Calcd for C11H17NO: C, 73.70; H, 9.50; N, 7.88.
Found: C, 73.61; H, 9.56; N, 7.81.
(vii) Dea ctiva tion of Ru th en iu m [NN′N]-Ca ta lysts.
After some of the catalytic runs, light-brown ruthenium-
containing deposits were recovered from the reaction
mixtures. Infrared spectra of these residues show a
strong absorption in the ν(COstretch) area at 1944 cm-1
,
pointing to the formation of ruthenium-carbonyl com-
plexes during the reaction. Decomposition of ruthenium-
aldehyde intermediates has been reported to occur during
N-alkylation reactions of long-chain aliphatic amines
with primary alcohols catalyzed by RuCl2(PPh3)3.4e In
these reactions, decarbonylation leads to the formation
of RuHCl(CO)(PPh3)3, which is inactive as a catalyst.
Obviously, the carbonyl complexes formed during our Ru-
[NN′N]-catalyzed reactions are not active anymore. At-
tempts to characterize them failed because of their
insolubility. In separate experiments, we have isolated
a number of Ru[NN′N] carbonyl complexes.12
Exp er im en ta l Section
N-P h en ylp ip er id in e (7). Pure product was obtained by
distillation of the evaporated reaction mixture as a colorless
oil: bp 86 °C (1 mmHg); 1H NMR (300 MHz, CDCl3) δ 7.47 (t,
2H), 7.15 (d, 2H), 7.05 (t, 1H), 3.36 (t, 4H), 2.00-1.82 (m, 4H),
1.82, 1.75 (m, 2H); 13C NMR (300 MHz, CDCl3) δ 152.54,
129.21, 119.37, 116.76, 50.87, 26.19, 24.66; MS m/ z 161 (M+).
N-Ben zylp ip er id in e (8). Pure product was obtained by
distillation of the evaporated reaction mixture as a colorless
oil: bp 66 °C (0.5 mmHg); 1H NMR (300 MHz, CDCl3) δ 7.35-
7.21 (m, 5H), 3.49 (s, 2H), 2.42-2.38 (m, 4H), 1.62-1.57 (m,
4H), 1.46-1.44 (m, 2H); 13C NMR (300 MHz, CDCl3) δ 138.74,
129.90, 128.09, 126.80, 63.92, 54.54, 26.05, 24.45; MS m/ z 175
(M+). Anal. Calcd for C12H17N: C, 82.23; H, 9.78; N, 7.99.
Found: C, 82.11; H, 9.69; N, 8.06.
All boiling points are uncorrected. 1H and 13C NMR spectra
were recorded on 200 or 300 MHz spectrometers. Elemental
analyses were performed by Dornis und Kolbe, Microanalytis-
ches Laboratorium, Mulheim a.d. Ruhr, Germany. GC analy-
ses were performed using a DB-17 capillary column in
combination with a flame ionization detector. Conversions and
yields were determined by the internal standard method
according to calibration curves obtained for each product in
separate experiments. MS analyses were made on a Unicam
Automass System 2. N-Phenylpyrrolidine,38 N-phenylpiperi-
dine,39 N-benzylpiperidine,40 N-(4-hydroxybutyl)aniline,18 N-(5-
hydroxypentyl)aniline,18 N-phenyl-N′-methylpiperazine,41 and
N-[m-(trifluoromethyl)phenyl]-N′-methylpiperazine27 are known
compounds. 1H and 13C NMR data for these compounds are
cited below. Aniline, alkanediols, N-substituted diethanola-
mines, and solvents were commercial materials and were
purified by distillation under a dry nitrogen atmosphere before
use. N-Benzyldiethanolamine,42 [RuCl2(nbd)]n,43 RuCl2(PMe3)4,44
RuCl2(PPh3)3,45 RuCl2(bpy)2,46 RuCl2(terpy)(PPh3),47 and the
complexes 1-512 were prepared according to published pro-
cedures.
N-(4-Hyd r oxybu tyl)a n ilin e (9a ). The reaction mixture
was extracted with water (100 mL) to remove remaining diol.
Column chromatography of the concentrated reaction mixture
on silica gel (Merck, 60-230 mesh) with a hexane/ethyl acetate
mixture as a eluent (gradient from 98:2 to 2:1 in volume) gave
N-(4-hydroxybutyl)aniline. The pure product was obtained by
further Kugelrohr distillation as a colorless oil: pot temper-
1
ature 122 °C (0.01 mmHg); H NMR (300 MHz) (CDCl3) 7.20
(t, 2H), 6.75 (t, 1H), 6.64 (d, 2H), 3.66 (t, 2H), 3.15 (t, 2H),
1.72-1.67, 4H2), OH and NH not observed; 13C NMR (75.47
MHz) (CDCl3) 148.40, 129.28, 117.46, 113.01, 62.51, 43.93,
30.35, 26.09; MS m/ z 165 (M+). Anal. Calcd for C10H15NO:
C, 72.69; H, 9.15; N, 8.48. Found: C, 72.52; H, 8.94; N, 8.58.
N-(6-Hyd r oxyh exyl)a n ilin e (9b). The reaction mixture
was extracted with water (100 mL) to remove remaining diol.
Column chromatography of the evaporated reaction mixture
on silica gel (Merck, 60-230 mesh) with a hexane/ethyl acetate
(38) Signaigo, F. K.; Adkins, H. J . Am. Chem. Soc. 1936, 58, 709.
(39) Horning, C. H.; Bergstrom, F. W. J . Am. Chem. Soc. 1945, 67,
2110.
(40) Staple, E.; Wagner, E. C. J . Org. Chem. 1949, 14, 559.
(41) Forsee, W. T., J r.; Pollard, C. B. J . Am. Chem. Soc. 1935, 57,
1788.
(42) Saavedra, J . E. J . Org. Chem. 1985, 50, 2271.
(43) Abel, E. W.; Bennett, M. A.; Wilkinson, G. J . Chem. Soc. 1959,
3178.
(44) Singer, H.; Hademer, E.; Oehmichen, U.; Dixneuf, P. J . Orga-
nomet. Chem. 1979, 178, C13.
(45) Holm, R. Inorg. Synth. 1970, 12, 238.
(46) Sullivan, B. P.; Salmon, D. J .; Meyer, T. J . Inorg. Chem. 1978,
17, 3334.
(47) Sullivan, B. P.; Calvert, J . M.; Meyer, T. J . Inorg. Chem. 1980,
19, 1004.

4290 J . Org. Chem., Vol. 63, No. 13, 1998
Abbenhuis et al.
mixture as an eluent (gradient from 98:2 to 2:1 in volume) gave
6-N-(6-hydroxyhexyl)aniline. The pure product was obtained
by crystallization from benzene/pentane: white crystals; mp
43 °C; 1H NMR (300 MHz) (CDCl3) δ 7.19 (t, 2H), 6.71 (t, 1H),
6.62 (d, 2H), 3.64 (t, 2H), 3.12 (t, 2H), 1.67-1.56, and 1.42-
1.44 (m, 4H), OH and NH not observed; 13C NMR (75.47 MHz)
(CDCl3) δ 148.54, 129.25, 117.28, 112.79, 62.67, 43.95, 32.69,
12H); 13C NMR (300 MHz, CDCl3) δ 98.84, 67.50, 62.50, 62.29,
32.44, 30.68, 29.40, 25.41, 22.45, 19.59. Anal. Calcd for
C
10H20O2: C, 63.80; H, 10.71. Found: C, 63.68; H, 10.65.
N-Phenyl-N′-methylpiperazine (12). Pure product was ob-
tained by vacuum distillation of the evaporated reaction
mixture as a colorless oil: bp 105 °C (0.4 mmHg); 1H NMR
(300 MHz, CDCl3) δ 7.28 (t, 2H), 6.94 (d, 2H), 6.66 (t, 1H),
3.22 (t, 4H), 2.58 (t, 4H), 2.36 (s, 3H); 13C NMR (300 MHz,
CDCl3) δ 151.30, 129.14, 119.10, 116.06, 55.20, 49.12, 46.23;
MS m/ z 176 (M+). Anal. Calcd for C11H16N2: C, 74.96; H,
9.15; N, 15.89. Found: C, 74.88; H, 9.26; N, 16.04.
N-[m-(Trifluoromethyl)phenyl]-N′-methylpiperazine (13). Col-
umn chromatography of the concentrated reaction mixture on
silica gel (Merck, 60-230 mesh) with a hexane/ethyl acetate
mixture as an eluent (gradient from 98:2 to 2:1 in volume) gave
the crude product as a light yellow oil. The pure product was
obtained by further Kugelrohr distillation as a colorless oil:
pot temperature 75 °C (0.45 mmHg); 1H NMR (300 MHz)
(CDCl3) δ 7.32 (t, 1H), 7.11 (s, 1H), 7.07-7.03 (m, 2H), 3.24 (t,
4H), 2.55 (t, 4H), 2.34 (s, 3H); 13C NMR (75 MHz, CDCl3) δ
151.37, 131.42 (q, 2J CF ) 32 Hz), 129.51, 124.35 (q, 1J CF ) 272
Hz), 118.59, 115.70, 112.04, 54.90, 48.56, 46.07; MS m/z 244
(M+). Anal. Calcd for C12H15N2F3: C, 59.01; H, 6.19; N, 11.47.
Found: C, 58.85; H, 6.25, N, 11.56.
29.56, 27.00, 25.64; MS m/ z 193 (M+). Anal. Calcd for C12H19
-
NO: C, 74.57; H, 9.91; N, 7.25. Found: C, 74.38; H, 9.86; N,
7.22.
N-(10-Hyd r oxyd ecyl)a n ilin e (9c). The reaction mixture
was extracted with water (100 mL) to remove remaining diol.
Column chromatography of the evaporated reaction mixture
on silica gel (Merck, 60-230 mesh) with a hexane/ethyl acetate
mixture as an eluent (gradient from 98:2 to 2:1 in volume) gave
N-(10-hydroxydecyl)aniline. The pure product was obtained
by crystallization from benzene/pentane: white crystals; mp
39 °C; 1H NMR (300 MHz) (CDCl3) δ 7.20 (t, 2H), 6.72 (t, 1H),
6.63 (d, 2H), 3.63 (t, 2H), 3.12 (t, 2H), 1.67-1.50 (m, 4H), 1.49-
1.30 (m, 12H), OH and NH not observed; 13C NMR (75.47 MHz)
(CDCl3) δ 148.60, 129.23, 117.13, 112.79, 62.93, 44.06, 32.81,
29.61, 29.57, 29.47, 27.10, 25.80; MS m/ z 249 (M+). Anal.
Calcd for C16H27NO: C, 77.06; H, 10.91; N, 5.62. Found: C,
77.22; H, 10.84; N, 5.75.
N-P h en ylp yr r olid in e (10a ). Pure product was obtained
by distillation of the evaporated reaction mixture as a colorless
N-Benzyl-N′-[m-(trifluoromethyl)phenyl]piperazine (14). Low-
boiling impurities were removed by Kugelrohr distillation.
Flash distillation gave the pure product as a light yellow oil:
1H NMR (300 MHz, CDCl3) δ 7.37-7.30 (m, 6H), 7.12 (s, 1H),
7.07 (t, 2H), 3.59 (s, 2H), 3.26 (t, 4H), 2.63 (t, 4H); 13C NMR
1
oil: bp 81 °C (0.5 mmHg); H NMR (300 MHz, CDCl3) δ 7.33
(t, 2H), 7.76 (t, 1H), 6.66 (d, 2H), 3.37 (t, 4H), 2.08 (q, 4H); 13
C
NMR (300 MHz, CDCl3) δ 148.09, 129.21, 115.47, 111.75,
47.65, 25.56; MS m/ z 149 (M+).
2
(75 MHz, CDCl3) δ 159.49, 137.92, 131.43 (q, J CF ) 32 Hz),
N-Phenylhexahydroazepine (10b). Pure product was ob-
tained by distillation of the evaporated reaction mixture as a
colorless oil: bp 117 °C (2.5 mmHg); 1H NMR (300 MHz,
CDCl3) δ 7.42 (t, 2H), 6.90 (d, 2H), 6.85 (t, 1H), 3.64 (t, 4H),
2.00-1.97 (m, 4H), 1.77-1.73 (m, 4H); 13C NMR (300 MHz,
CDCl3) δ 149.10, 129.46, 115.42, 111.4, 49.30, 28.04, 27.40;
MS m/ z 175 (M+). Anal. Calcd for C12H17N: C, 82.23; H, 9.78;
N, 7.99. Found: C, 82.06; H, 9.82; N, 8.06.
(Tetr a h yd r op yr a n yloxy)-1-p en ta n ol (11). The crude
reaction mixture was flash-distilled to remove ruthenium
complexes and subsequently dissolved in water (100 mL).
Extraction with Et2O (50 mL, 5×) and removal of the solvent
under reduced pressure afforded the pure product as a
colorless oil: bp 92 °C, 0.1 mmHg; 1H NMR (300 MHz, CDCl3)
δ 4.51 (t, 1H), 3.81-3.30 (m, 6H), 2.36 (s, 1H), 1.80-1.34 (m,
1
124.37 (q, J CF ) 272 Hz), 129.51, 129.15, 128.32, 127.21,
118.64, 115.68, 112.10, 62.99, 52.88, 48.69; MS m/ z 320 (M+).
Anal. Calcd for C18H19N2F3: C, 67.49; H, 5.98; N, 8.74. C,
67.42; H, 6.04; N, 8.71.
Ack n ow led gm en t. R.A.T.M.A. thanks the Dutch
Department of Economic Affairs in terms of the Innova-
tion Oriented Research Program Catalysis (IOP-Katal-
yse) for financial support. Michel M. Bergshoef, Arne´
Dijksman, J oris van Slageren, and Ageeth Bol are
gratefully acknowledged for their experimental con-
tributions.
J O972260D