Prototype Pictet-Spengler reactions catalyzed by superacids. Involvement of dicationic superelectrophiles

Article abstract of DOI:10.1021/jo982019e

The Pictet-Spengler reaction, an acid-catalyzed intramolecular cyclization of intermediate imines of 2-arylethylamine to give 1,2,3,4- tetrahydroisoquinolines, has long been limited to active substrates which bear strongly electron-donating groups such as a methoxy or a hydroxy group on the cyclizing benzene ring. In this paper, we present superacid-catalyzed Pictet-Spengler reactions of imines of 2-phenethylamine, including the prototype Pictet-Spengler reaction of N-methylene-2-phenethylamine, to give the parent and 1-substituted 1,2,3,4-tetrahydroisoquinolines in moderate to high yields. The yields are dependent on the acidity of the media. A linear relationship was found between the rate of the cyclization and the acidity of the reaction media in kinetic studies of N-methylene-2-phenethylamine and related imines, strongly supporting the intervention of an additional protonative activation of the N-protonated imines, that is, the involvement of dicationic superelectrophiles, N,N-diprotonated imines (ammonium- carbenium dications). We further found that the prototype cyclization of the parent N-methylene-2-phenethylamine is also catalyzed by TFA to give 1,2,3,4- tetrahydroisoquinoline in good yield, although the cyclization is significantly slower than that catalyzed by superacids. The prototype Pictet- Spengler cyclization of N-methylene-2-phenethylamine can thus take place both through the monocation (the N-monoprotonated imine) and the dication (the N,N-diprotonated imine), the latter reaction being predominant in superacids.

Full text of DOI:10.1021/jo982019e

J . Org. Chem. 1999, 64, 611-617
611
P r ototyp e P ictet-Sp en gler Rea ction s Ca ta lyzed by Su p er a cid s.
In volvem en t of Dica tion ic Su p er electr op h iles
Akihiro Yokoyama, Tomohiko Ohwada,*,† and Koichi Shudo
Graduate School of Pharmaceutical Sciences, The University of Tokyo, 7-3-1 Hongo,
Bunkyo-ku, Tokyo 113, J apan
Received October 6, 1998
The Pictet-Spengler reaction, an acid-catalyzed intramolecular cyclization of intermediate imines
of 2-arylethylamine to give 1,2,3,4-tetrahydroisoquinolines, has long been limited to active substrates
which bear strongly electron-donating groups such as a methoxy or a hydroxy group on the cyclizing
benzene ring. In this paper, we present superacid-catalyzed Pictet-Spengler reactions of imines of
2-phenethylamine, including the prototype Pictet-Spengler reaction of N-methylene-2-phenethy-
lamine, to give the parent and 1-substituted 1,2,3,4-tetrahydroisoquinolines in moderate to high
yields. The yields are dependent on the acidity of the media. A linear relationship was found between
the rate of the cyclization and the acidity of the reaction media in kinetic studies of N-methylene-
2-phenethylamine and related imines, strongly supporting the intervention of an additional
protonative activation of the N-protonated imines, that is, the involvement of dicationic superelec-
trophiles, N,N-diprotonated imines (ammonium-carbenium dications). We further found that the
prototype cyclization of the parent N-methylene-2-phenethylamine is also catalyzed by TFA to give
1,2,3,4-tetrahydroisoquinoline in good yield, although the cyclization is significantly slower than
that catalyzed by superacids. The prototype Pictet-Spengler cyclization of N-methylene-2-
phenethylamine can thus take place both through the monocation (the N-monoprotonated imine)
and the dication (the N,N-diprotonated imine), the latter reaction being predominant in superacids.
The Pictet-Spengler reaction is an acid-catalyzed
intramolecular cyclization of the intermediate imine of
2-arylethylamine, formed by condensation with a carbo-
nyl compound, to give 1,2,3,4-tetrahydroisoquinoline
derivatives.1-3 The Pictet-Spengler reaction, as well as
the Bischler-Napieralski reaction,4 is one of the key
reactions for construction of the isoquinoline skeleton,
which constitutes an important motif of naturally occur-
ring bioactive substances and pharmacophores.5 For
example, unsubstituted (2a ) and 1-methyl-1,2,3,4-tet-
rahydroisoquinolines were detected in the brains of
humans, rats, and mice and are thought to be relevant
to the pathogenesis of Parkinson’s disease.6,7 These
simple endogenous tetrahydroisoquinolines are proposed
to be biosynthesized from imines derived from 2-phen-
ethylamine and a C1 or C2 carbonyl unit through a
mechanism similar to that in the Pictet-Spengler
reaction.7f The Pictet-Spengler reaction itself has long
been believed to have chemical limitations regarding the
prerequisite substituents. Although Pictet and Spengler
originally reported the formation of the parent 1,2,3,4-
tetrahydroisoquinoline by heating at reflux (150 °C) of a
mixture of 2-phenethylamine and dimethoxymethane
(i.e., an equivalent for formaldehyde) in the presence of
concentrated hydrochloric acid,2 many attempts to re-
produce the synthesis of the parent 1,2,3,4-tetrahydroiso-
quinoline in this prototype Pictet-Spengler reaction have
failed.1i,8 For example, Kondo and Ochiai8 reported the
formation of a trace amount of 1,2,3,4-tetrahydroiso-
quinoline, along with a large amount of bis(2-phenyl-
ethylamine)methane under reaction conditions similar to
those originally described.2 Thus, the general consensus
hitherto has been that the Pictet-Spengler cyclization
is very sensitive to the aromatic substituents of the
2-arylethylamine moiety of the imine 1, i.e., the prototype
Pictet-Spengler reaction of 1a is impractical, and an
electron-donating hydroxy or an alkoxy group at the para
position of the cyclization site is the minimum requisite
for a facile reaction.1,9,10 In this paper, we will describe
superacid-catalyzed prototype and related Pictet-Spen-
gler reactions of imines of 2-phenethylamine (Scheme 1)
† Present address: Faculty of Pharamceutical Sciences, Nagoya City
University, 3-1 Tanabe-dori, Mizuho-ku, Nagoya 467, J apan.
(1) For reviews, see: (a) Czerwinski, K. M.; Cook, J . M. Adv.
Heterocycl. Nat. Prod. Synth. 1996, 3, 217-277. (b) Cox, E. D.; Cook,
J . M. Chem. Rev. 1995, 95(6), 1797-1842. (c) Waldmann, H. Synlett
1995, 133-141. (d) Rozwadowska, M. D. Heterocycles 1994, 39(2), 903-
931. (e) Badia, D.; Dominguez, E.; Lete, E.; Villa, M. J . Trends
Heterocycl. Chem. 1991, 2, 1-11. (f) Ungemach, F.; Cook, J . M.
Heterocycles 1978, 9, 1089-1119. (g) Claret, P. A. Comprehensive
Organic Chemistry; Barton, D., Ollis, W. D., Eds.; Pergamon: Oxford,
1979; Vol. 4, pp 209-211. (h) Gensler, W. J . Heterocyclic Compounds;
Elderfield, R. C., Ed.; Wiley: New York, 1952; Vol. 4, pp 353-361. (i)
Whaley, W. M.; Govindachari, T. R. Org. React. 1951, 6, 151-190.
(2) Pictet, A.; Spengler, T. Ber. Dtsch. Chem. Ges. 1911, 44, 2030-
2036.
(3) Decker, H.; Becker, P. J ustus Liebigs Ann. Chem. 1913, 395,
342-362.
(4) March, J . Advanced Organic Chemistry, Reactions, Mechanisms
and Structure, 4th ed.; J ohn Wiley & Sons: New York, 1992.
(5) Bringmann, G.; Ewers, C. T.; Walter, R. Comprehensive Organic
Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon Press: Oxford,
1991; Vol. 6, Chapter 4.2.
(6) (a) Sandler, M.; Carter, S. B.; Hunter, K. R.; Stern, G. M. Nature
1973, 241, 439-443. (b) Collins, M. A. The Alkaloids; Brossi, A., Ed.;
Academic Press: New York, 1983; Vol. 21, pp 329-350. (c) Langston,
J . W.; Ballard, P.; Tetrud, J . W.; Irwin, I. Science 1983, 219, 979.
(7) (a) Kohno, M.; Ohta, S.; Hirobe, M. Biochem. Biophys. Res.
Commun. 1986, 140, 448-454. (b) Ohta, S.; Kohno, M.; Makino, Y.;
Tachikawa, O.; Hirobe, M. Biomed. Res. 1987, 8, 453-456. (c) Nagatsu,
T.; Yoshida, M. Neurosci. Lett. 1988, 87, 178-182. (d) Tasaki, Y.;
Makino, Y.; Ohta, S.; Hirobe, M. J . Neurochem. 1991, 57, 1940-1943.
(e) Kotake, Y.; Tasaki, Y.; Makino, Y.; Ohta, S.; Hirobe, M. J .
Neurochem. 1995, 65, 2633-2638. (f) Yamakawa, T.; Ohta, S. Biochem.
Biophys. Res. Commun. 1997, 236, 676-681.
(8) Kondo, H.; Ochiai, E. Yakugaku Zasshi 1923, 495, 313-319;
Chem. Abstr. 1923, 17, 3032. See also: Clemo, G. R.; Swan, G. A. J .
Chem. Soc. 1964, 617-621.
10.1021/jo982019e CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/29/1998

612 J . Org. Chem., Vol. 64, No. 2, 1999
Yokoyama et al.
Sch em e 1
was heated at 50 °C for 1 h and then added to a large
amount of ice and water. This mixture was basified with
aqueous sodium hydroxide and extracted (see Experi-
mental Section) to afford the cyclized product 2a in 76%
yield.14
It should be noted that the cyclization reaction of the
parent imine 1a can also be catalyzed by the relatively
weak acid TFA alone, which provides a condition similar
to the original Pictet-Spengler condition,1i,2,8,9 although
the reaction is significantly slower than that catalyzed
by TFSA and TFSA-TFA (Table 1). In TFA at reflux (72
°C) for 20 h, the yield of the cyclized product was 63% in
the form of the N-benzoyl form of 2a .14 The rate of
cyclization is accelerated when the acidity of the medium
is increased by the addition of TFSA, as judged from the
yield of the cyclized product (2a or the benzoyl form of
2a after benzoylation) under unified reaction conditions
(50 °C and 1 h): in TFA (H0 ≈ -2.7), the cyclized product
(2a ) is virtually undetectable after 1 h; in the acid of H0
≈ -9.0 (1.2% w/w TFSA-98.8% w/w TFA), the cyclized
product was formed but in a small amount; in 90% w/w
TFSA-10% w/w TFA (H0 ≈ -12.5), the yield of 2a was
to give the parent and 1-substituted 1,2,3,4-tetrahy-
droisoquinoline in moderate to high yields. We also
scrutinized the prototype Pictet-Spengler reaction of 1a
and found a reproducible cyclization reaction in TFA
alone to give the parent 1,2,3,4-tetrahydroisoquinoline.
Nevertheless, the cyclization reaction of 1a in TFA is very
slow, being significantly accelerated with further increase
of the acidity of the media. In kinetic studies a linear
relationship was found between the rate and the acidity
of the medium in the superacid-catalyzed cyclization
reactions of N-methylene-2-phenethylamine 1a and the
relevant N-benzylidene-2-phenethylamines (1c-1e),
strongly supporting the involvement of an additional
proton transfer to the N-protonated imines (3) to form
superelectrophiles,11 i.e., the N,N-diprotonated imines (4).
+
76%. The pKBH values of imines 1a -1e for N-protonation
are estimated to be +6 to +8.15 Thus, 1a should be
completely protonated to form the monocation 3a even
in TFA, where the acidity (H0) was -2.7 (Scheme 2). In
1
the H NMR spectrum of 1a in TFA at -18 °C, an NH
proton was observed at 12.29 ppm and two methylene
signals were coupled with the NH signals.16 The observed
acceleration of the cyclization of 1a upon increase of the
acidity of the medium strongly suggested that an ad-
ditional proton transfer of the iminium cation 3a is
involved in the superacid-catalyzed cyclization of 1a .
Th e Su p er a cid -Ca ta lyzed P ictet-Sp en gler Rea c-
tion s To Give 1-Su bstitu ted 1,2,3,4-Tetr a h yd r oiso-
qu in olin es. The aldimines (1b-1e) bearing a carbon
substituent R1 were also prepared by condensation of
2-phenethylamine and the appropriate aldehydes (R1-
CHO).17 We found that the cyclizations of the isolated
imines 1b-1e can be catalyzed by TFSA (Table 1). In
contrast to the behavior of the parent imine 1a , the
cyclization of the imines 1b-1e does not proceed in TFA
Resu lts a n d Discu ssion
Th e P r ototyp e P ictet-Sp en gler Rea ction . We
found that N-methylene-2-phenethylamine (1a ), prepared
from 2-phenethylamine and paraformaldehyde, cyclized
in the presence of 100 equiv of trifluoromethanesulfonic
acid (TFSA) at ambient temperature to give the parent
1,2,3,4-tetrahydroisoquinoline (2a ). The maximum yield
of the cyclized product 2a was obtained when a somewhat
weaker acid, 90% w/w TFSA-10% w/w TFA (H0
≈
-12.5),12,13 and the following addition order of the acids
were used: to a solution of the imine 1a in trifluoroacetic
acid (TFA), weighed TFSA was added at ambient tem-
perature to obtain finally a 90% w/w TFSA-10% w/w
TFA acid solution (100 equiv), and the reaction solution
(14) The yield of the cyclized product was calculated after conversion
to the corresponding benzoyl derivative by benzoylation of the crude
amine products to allow convenient purification of the compound.
Under the same reaction conditions (in 90% w/w TFSA-10% w/w TFA,
50 °C, 1 h), N-benzoyl-1,2,3,4-tetrahydroisoquinoline was formed in
74% yield, along with the N-benzoyl derivative of 2-phenethylamine
(5% yield), which was generated through decomposition of the starting
imine 1a . This decomposition of the starting imine partially accounts
for the limiting conversion (for example, 1b) particularly under a longer
reaction time and a higher reaction temperature. The yield of the
cyclized product after benzoylation is comparable with that obtained
by the direct separation of 2a . Benzoylation of the crude reaction
mixture obtained in the cyclization reaction of the imine 1a allows
convenient isolation of the crude amines, and also affords good
separation of a small amount of an undefined polar byproduct.
(15) (a) Olah, G. A.; Kreienbuhl, P. J . Am. Chem. Soc. 1967, 89,
4756-4759. (b) Olah, G. A.; Donovan, D. J . J . Org. Chem. 1978, 43,
860-867. (c) Allen, M.; Roberts, J . D. Can. J . Chem. 1981, 59, 451-
458. (d) Knorr, R.; Ferchland, K.; Hoang, T. P. Leibigs. Ann. Chem.
1994, 943-948. (e) Cordes, E. H.; J encks, W. P. J . Am. Chem. Soc.
1963, 85, 2843-2848. The basicity of the related imine N-p-chloroben-
(9) Kumar, P.; Dhawan, K. N.; Kishor, K.; Bhargava, K. P.; Satsangi,
R. K. J . Heterocycl. Chem. 1982, 19, 677-679. Although the formation
of 1-phenyl-1,2,3,4-tetrahydroisoquinoline (2c) from the imine 1c in
TFA was reported, we could not reproduce the result (see Table 1).
(10) (a) Buck, J . S. J . Am. Chem. Soc. 1934, 56, 1769-1771. (b) Ide,
W. S.; Buck, J . S. J . Am. Chem. Soc. 1937, 59, 726-731.
(11) Olah, G. A. Angew. Chem., Int. Ed. Engl. 1993, 32, 767-788.
(12) The acidity (H0) of the TFSA-TFA system is described in Saito
et al. (Saito, S.; Saito, S.; Ohwada, T.; Shudo, K. Chem. Pharm. Bull.
1991, 39, 2718-2720). See also: Saito, S.; Sato, Y.; Ohwada, T.; Shudo,
K. J . Am. Chem. Soc. 1994, 116, 2312-2317, footnote 8.
(13) The acidity (stronger than H0 ) -10) of the TFSA-TFA acid
system which catalyzes the cyclization of 1 would be decreased, because
the starting imines 1 and the cyclized products, tetrahydroisoquinolines
2, are strong nitrogen bases that would be completely N-monoproto-
nated in the acid. To determine the lowering of the acidity of the acid,
we used piperidine as a nitrogen base which has a similar structure
and basicity to those of 1 and 2, because piperidine, unlike 1 and 2,
has no strong UV absorption which would interfere with the absorption
of indicators (see ref 12). The corrected acidities (H0) of the TFSA-
TFA acid in the presence of 1 mol % of piperidine are shown in Table
4. In the case of 26.2% TFSA-73.8% TFA acid (H0 ≈ -10), the decrease
of the acidity by the addition of 1 mol % of piperidine is 0.3 in terms
of H0 units, while the addition of 1 mol % of piperidine to 100% TFSA
caused the largest lowering of the acidity (by 1.4 in terms of H0).
+
zylidene-1,1-dimethylethylamine (pKBH ) 6.5) is almost equal to that
+
of N-benzylidene-1,1-dimethylethylamine (pKBH ) 6.7).
(16) In the 1H NMR spectrum of 1a in TFA at -18 °C, the two
methylene signals are observed at 8.10 ppm (J ) 8.5 Hz) and 7.87
ppm (J ) 8.3 and 17.8 Hz) as
a triplet and a quartet signal,
respectively. The splittings of these signals are due to the geminal
coupling (J ) 8 Hz) and the coupling with the NH signal.
(17) Tomaszewski, M. J .; Warkentin, J .; Werstiuk, N. H. Aust. J .
Chem. 1995, 48, 291-321. Sakamoto, M.; Tomimatu, Y. Yakugaku
Zasshi 1970, 90, 1339-1346; Chem. Abstr. 1971, 74, 53468t.

Pictet-Spengler Reactions Catalyzed by Superacids
J . Org. Chem., Vol. 64, No. 2, 1999 613
Ta ble 1. Acid -Ca ta lyzed Cycliza tion of th e Im in es of 2-P h en eth yla m in e Der iva tives
imines
R1
R2
R3
R4
acid
temp (°C)
time (h)
yield (%) of 2
1a
H
H
H
H
H
H
H
H
90% TFSA-10% TFA
TFA
TFSA
TFSA
TFSA
TFA
TFA
TFSA
TFA
TFSA
TFSA
TFA
TFA
TFSA
TFA
50
72c
150
120
120
72c
120e
120
120e
120
80
1
20
13
17
15
16
40
15
40
3
14
40
32
30
26
2
76 (74)a,b
(63)a
69
1b
1c
CH(CH3)2
C6H5
H
H
32
90
0d
0d
1d
1e
p-ClC6H4
H
H
H
H
H
H
94
0
90
p-CH3C6H4
85
120e
72c
60
0d
1f
C6H5
C6H5
C6H5
CH3
H
H
H
3f
77g
50
1g
1h
CH3
CH3
H
72c
60
TFSA
TFA
TFA
94
89
89
OCH3
OCH3
30
62
1.2
72c
a
b
The yield of N-benzoyltetrahydroisoquinoline after benzoylation of the crude amines is shown in parentheses. N-Benzoyl
2-phenethylamine (5% yield) was generated upon benzoylation of the crude amines. c At reflux. 2-Phenethylamine was formed in 10-
d
20% yield through the decomposition of the imine. e Sealed. f A single isomer (1-phenyl-6-methyltetrahydroisoquinoline). A mixture of
g
two isomers (1-phenyl-6-methyltetrahydroisoqinoline: 1-phenyl-8-methyl-tetrahydroisoqinoline ) 3:1).
Sch em e 2
even by heating at reflux or at 120 °C in a sealed tube,
and the imines were unchanged. On the other hand, in
TFSA the alkyl-substituted imine N-(2-methylpropy-
lidene)-2-phenethylamine (1b ) gave a cyclized product
(2b) in 69% yield at 150 °C for 13 h.
8.10 ppm (J ) 10.9 Hz), supporting the essentially
complete N-protonation of 1c in TFA to give the cation
3c. When the solution of 1c in TFA was heated at reflux
(72 °C) for 16 h, no reaction took place, no cyclized
product (2c) was obtained, and 3c remained unchanged
(monitored by 1H NMR spectra) (Table 1). This result
excludes the contribution of the N-protonated imine 3c
as a reactive intermediate in the cyclization of the imine
1c. We, thus, postulate the intervention of an additional
proton transfer to the iminium cation 3c, and probably
3b, 3d , and 3e in the superacid-catalyzed Pictet-Spen-
gler reactions of the imines 1b-1e. In contrast to the
N-monoprotonated N-methylene-2-phenethylamine (3a ),
the iminium cations 3b-3e lack cyclizing ability, prob-
ably due to significant stabilization of the carbenium
center by an alkyl or a phenyl group, resulting in
deactivation of the electrophilicity of the iminium cation
3.
Kin etic Stu d ies of th e Su p er a cid -Ca ta lyzed P ic-
tet-Sp en gler Rea ction s.We measured the rates of the
cyclization of the N-benzylidene-2-phenethylamines 1c-
1e in strong acids of a variety of acidities (TFSA-TFA).
The reactions hardly proceed in acid weaker than H0 )
-10, despite complete N-monoprotonation of the imines
1 to give the cation 3. The reactions were carried out in
the presence of 100 equiv of the acid (TFSA-TFA) in a
sealed NMR tube and heated at 120 ((0.1) °C, and the
concentrations of the N-protonated imine 3 were deter-
The reaction of 1b seemed to reach a plateau at 150
°C even at long reaction time (20 h, 60% yield). At lower
temperature (120 °C) the cyclization reaction of 1b in
TFSA was retarded, and the yield of the product 2b was
decreased. When the substituent (R1) of the aldehyde was
an aryl group (1c-1e), the cyclization proceeded ef-
ficiently in TFSA at 120 °C to give the cyclized products
2c-2e in high yields (Table 1). The reaction conditions
(time and temperature) are partially optimized: for
example in the case of the imine 1c, the yield of 2c was
decreased to 60% in a shorter reaction time (3 h, at 120
°C). Heating is also necessary for a facile reaction: at 60
°C the cyclization reaction of the imine 1c in TFSA is
too slow to be detected. As judged from the reaction time,
p-methyl 1e cyclized faster than unsubstituted 1c and
p-chloro 1d . This is compatible with the measured
reaction rates of 1c-1e (vide infra).
In a manner similar to that for 1a , the imine 1c is also
completely monoprotonated in TFA to give the iminium
1
cation 3c. The H NMR spectrum of 1c in TFA at -10
°C showed a unit proton attached to the nitrogen atom
of the imine at 11.2 ppm, and the coupling of the imino-
methylene signal with an NH signal to give a doublet at

614 J . Org. Chem., Vol. 64, No. 2, 1999
Yokoyama et al.
Ta ble 2. Ra te Con sta n ts for Su p er a cid -Ca ta lyzed
Ta ble 3. Th er m och em ica l Da ta for Acid -Ca ta lyzed
Cycliza tion s of 1a in TF A a n d 1a , 1c, 1d , a n d 1e in
Su p er a cid s
Cycliza tion s of 1a a n d 1c-1ea ,b
1a c
1cd
1d d
1ed
rate
constants
105k (s-1
)
∆Gq e,h
(kcal/mol)
∆Hq f
(kcal/mol)
∆Sq g
(eu)
105k
(s-1
105k
(s-1
105k
105k
(s-1
at T (°C)d
H0
)
H0
)
H0
(s-1
)
H0
)
1a a
1a b
1cc
1d c
1ec
50.0
60.0
70.0
0.0
10.0
4.20
8.03
19.5
4.15
12.5
31.8
2.66
4.95
8.86
2.70
4.93
9.21
16.6
25.5
36.9
-12.7 4.95 -12.7
-12.5 4.52 -12.4
-12.1 2.43 -12.1
-11.5 1.10 -11.6
4.93
3.64
2.52
1.29
0.52
0.23
-12.7 25.50
-12.5 15.20
24.8
22.0
27.8
27.8
25.3
16.2
15.7
17.7
18.0
11.5
-28.7
-21.1
-33.8
-32.9
-46.3
-12.4 31.8
-12.0 24.7
-11.5 15.8
-12.1
-11.5
-11.0
-10.6
5.99
2.12
0.79
0.39
-11.0
-10.7
8.69 -11.0 0.55 -10.9
4.99 -10.6 0.29 -10.5
20.0
110.0
120.0
130.0
110.0
120.0
130.0
110.0
120.0
130.0
a
Corrected values of acidity function of the reaction media (see
b
Table 4 (Experimental Section) and refs 12 and 13). Errors of
d
rates, (2%. c At 20.0 °C. At 120 °C.
a
b
At H0 ) -2.7 (in TFA). At H0 ) -12.4 (in 90% w/w TFSA-
10% w/w TFA, the corrected value of acidity function of the
medium (Table 4)). c At H0 ) -12.7 (in TFSA, the corrected value
d
of acidity function of the medium (Table 4)). Errors in temper-
ature, (0.1 °C. e Errors ( 1.5 kcal/mol. f Errors ( 0.9 kcal/mol.
g
h
Errors ( 2.0 eu. At 25 °C (298.15 K).
of 1c, 1d , and 1e were also obtained from the rate
constants of cyclization at three different temperatures
(110, 120, and 130 °C) (Table 3).20,21 A smaller enthalpy
of activation (∆Hq, 11.5 kcal mol-1) of 1e is found as
compared with those of 1c (∆Hq, 17.7 kcal mol-1) and 1d
(∆Hq, 18.0 kcal mol-1). This magnitude is consistent with
the facile cyclization reaction of 1e, which proceeds even
at 80 °C in TFSA to give the cyclized product 2e in 85%
yield after 14 h.
We also measured the rates of the cyclization of the
N-methylene-2-phenethylamine 1a in TFA and in the
TFSA-TFA acid system (Table 3). As described, the
imine 1a was completely monoprotonated to give the
cation 3a in TFA. In TFA the cyclization of 3a showed
good first-order kinetics, and the activation parameters
for the cyclization of 3a were obtained from the rate
constants of cyclization at three different temperatures
(50.0, 60.0, and 70.0 °C);20,21 the values are shown in
Table 3. The reaction of 1a in TFA below 50 °C is too
slow to allow kinetic measurements. The cyclization of
1a in the TFSA-TFA acid system at 20 °C also showed
good first-order kinetics, e.g., in 90% w/w TFSA-10% w/w
TFA (H0 ≈ -12.5) the rate constant at 20 °C is 3.18 ×
10-4 s-1 (Table 2), and a linear relationship (r ) 0.99)
between the rate of the cyclization and the acidity of the
medium was also found, the slope being 0.45 (Figure 1).22
This linear relationship supports the postulation that the
cyclization of 1a in the superacid involves an additional
proton transfer to the iminium ion 3a , i.e., the involve-
F igu r e 1. Acidity-rate profiles of the superacid-catalyzed
Pictet-Spengler cyclization reactions.
1
mined from the integral values in the H NMR spectra
at 23 °C, where the reaction was too slow to be detected.18
These reactions showed good first-order kinetics, giving
linear relationships between the rates and the acidities
(r > 0.99), the slope being 0.60 for 1c and 1d and 0.86
for 1e (Table 2 and Figure 1). The rate constant of 1c
was coincidently equal to that of 1d , and they are smaller
than that of 1e. According to the Zucker-Hammett
criteria,19,20 these linear relationships indicate the in-
volvement of additional protonation of the N-monopro-
tonated imines 3; i.e., the intervention of the dication 4
(vide infra). The activation parameters for the cyclization
(18) We confirmed that the imines 1c-1e are fully N-monoproto-
nated in acid at an acidity stronger than H0 ) -10 where the kinetic
measurements were performed, on the basis of the coupling of the
(20) The excellent linear relationships (regression coefficient r >
0.99) obtained in the estimation of activation parameters (Table 3)
support the postulation that the temperature-dependent changes of
+
imino methylene as a doublet. The spectrum of 1c in TFSA (H0
)
H0 and pKBH are canceled out in the TFSA-TFA acid system and
-12.7) gave essentially the same pattern as that in TFA, and therefore,
the concentration of the diprotonated imine (4c) seems to be too low
to allow for detection by NMR. This conclusion is consistent with the
Zucker-Hammett criteria which postulate that such a linear relation-
ship between the rates and the acidities (Figure 1) can be obtained
only when the concentration of the diprotonated species is very small.
(19) (a) Zucker, L. Hammett, L. P. J . Am. Chem. Soc. 1939, 61,
2791-2798. (b) Bonner, T. G.; Thorne, M. P.; Wilkins, J . M. J . Chem.
Soc. 1955, 2351-2358. (c) Bonner, T. G.; Barnard, M. J . Chem. Soc.
1958, 4181-4186. (d) Long, F. A.; Paul, M. A. Chem. Rev. 1957, 57,
935-1010. If the reactive electrophile in the cyclization is the mono-
protonated imine, the rates of the reaction will be saturated because
the imine 1 is fully monoprotonated even in TFA.
thus that the observed rate constant kobs is affected by kr (the rate
constant for cyclization) and not Ke (the ionization constant of the
carbodication 4) at the different temperatures. See: Ohwada, T.;
Suzuki, T.; Shudo, K. J . Am. Chem. Soc. 1998, 120, 4629-4637,
footnote 24. Therefore the activation parameters obtained from the
kinetic data should represent the thermal parameters in the rate-
determining cyclization process. A similar situation is also expected
in the case of the monocationic cyclization, i.e., the cyclization of the
imine 1a in TFA.
(21) Moore, J . W.; Pearson, R. G. Kinetics and Mechanism; J ohn
Wiley & Sons: New York, 1981.
(22) We noted a slight curvature of the plot in the highest acidity
region.

Pictet-Spengler Reactions Catalyzed by Superacids
J . Org. Chem., Vol. 64, No. 2, 1999 615
ment of the dication 4a (vide infra). That is, the accelera-
tion of the reaction of 1a in the superacid can be
attributed to dicationic cyclization. This idea is consistent
with the superacid-catalyzed cyclization reactions of the
imines 1c-1e. The rate constants of the cyclization of
1a in 90% w/w TFSA-10% w/w TFA were also measured
at three different temperatures (0.0, 10.0, and 20.0 °C),
and the activation parameters of the dicationic cyclization
were also obtained (Table 3).20,21 According to the ob-
tained linear relationship between the rate and the
acidity (H0 < -10), the reaction rate of the dicationic
cyclization of 1a at the acidity H0 ≈ -2.7 (equivalent to
the acidity of TFA) is predicted to be 1.6 × 10-8 s-1 (at
20 °C). This value is much smaller than the rate of
cyclization of 1a in TFA estimated at 20 °C (2.7 × 10-6
s-1) on the basis of the observed thermal parameters.
Therefore, the rate ratio of 1a between the dicationic
cyclization through the dication 4a (in TFSA-TFA at 20
°C) and the monocationic cyclization through the mono-
cation 3a (in TFA, the rate estimated at 20 °C) is roughly
102. Thus, this deviation also supports the idea that the
reaction of 1a in TFA is not the dicationic cyclization but
involves a different pathway, through the monocation 3a .
This situation is reminiscent of the inherent reactivity
of the acetyl cation in the acetylation of toluene.23 The
salt, acetylium hexafluoroantimonate (CH3CO+SbF6-),
can acetylate toluene slowly in a neutral solvent, sug-
gesting that the acetyl cation itself has the ability to
acetylate toluene, but only slowly. The acetylation is
accelerated as the acidity of the media is increased,
through the intervention of protonated acetyl cation.
In deuterated trifluoromethanesulfonic acid (CF3SO3D)
at 23 °C for 24 h, the imine 1c showed no deuterium
incorporation of the imino methine proton and the
benzene protons attached to the imino group, while the
thorough deuterium exchange occurred at all of the
phenyl protons of the 2-phenethylamino moiety. Although
the C,N-diprotonated species can be formed,24,25 the
phenyl group of the R1 substituent is not essential for
the dicationic cyclization, i.e., the cyclization reactions
of the imines 1a and 1b can be promoted by TFSA. Thus,
we can postulate that the dicationic superelectrophiles
commonly involved in the superacid-catalyzed Pictet-
Spengler reaction of 1a -1e are the N,N-diprotonated
imines, i.e., the ammonium-carbenium dications 4
(Scheme 2).18,26 The carbenium center of the iminium
moiety of 4 is activated by an electron-withdrawing
ammonium group, restoring or increasing the electro-
philicity intrinsic to the methylium center of the mono-
cation 3a . Involvement of relevant ammonium-carbeni-
um dications has been shown in the acid-catalyzed
stereoisomerization of olefin moieties of cinnamaldi-
mines27 and in the Friedel-Crafts type reactions of
cinnamaldimines with benzene, catalyzed by TFSA.28
Su bstitu en t Effects in th e P ictet-Sp en gler Rea c-
tion s. We investigated the substituent effect of the
electron-donating group on the cyclizing benzene ring of
the relevant imines (Table 1). The imines 1f-1h , N-ben-
zilidene-2-arylethylamines, were prepared by the con-
densation of the corresponding 2-arylethylamine and
benzaldehyde. The imine 1f, N-benzylidene-2-(3-meth-
ylphenyl)ethylamine, which bears a monomethyl group
on the cyclizing benzene ring, requires the superacid
catalyst to undergo the facile cyclization. In TFA, the
cyclization of 1f was very slow: the yield of the cyclized
product, 6-methyl-1-phenyl-1,2,3,4-tetrahydroisoquino-
line (2f), was only 3% after 32 h at reflux (72 °C). In TFSA
the reaction was accelerated to give a mixture of 6-meth-
yl-1-phenyl-1,2,3,4-tetrahydroisoquinoline (2f) and 8-meth-
yl-1-phenyl-1,2,3,4-tetrahydroisoquinoline (2f ′) (ratio 3:1)
in a total yield of 77% even at a lower reaction temper-
ature (60 °C, 30 h). A double methyl substitution of the
aromatic ring as in N-benzylidene-2-(3,5-dimethylphe-
nyl)ethylamine 1g activated the monocationic cyclization
through the N-protonated imine 3g: in TFA the imine
1g undergoes cyclization to give the product 2g in 50%
yield after heating at reflux (72 °C) for 26 h. The
cyclization reaction is significantly accelerated in TFSA:
the yield of 2g is 94% in TFSA, even at 60 °C after only
2 h. Acceleration of the cyclizations of the imines 1f and
1g by TFSA as compared with the reactions in TFA can
be interpreted in terms of the intervention of the dica-
tionic cyclization through the corresponding dications 4f
and 4g (Scheme 2).24,25 As estimated from the reaction
time and reaction temperature, the cyclization reactions
of 1f and 1g in TFA through the monocations 3f and 3g
are slow as compared with those of 1f and 1g in TFSA,
the rate ratio being evaluated to be on the order of 102;
(25) Despite the lack of deuterium exchange of the phenyl ring
protons of 1c attached to the iminio carbon atom in TFSA-d1 at 23 °C,
we could not exclude as contribution of C,N-diprotonated species such
as 6i-6iii
as well as ipso-protonated 6iv which might be activated electrophiles
due to conjugate positive ion centers. Relevant dications can also be
formed in the cases of the imines 1f and 1g.
(23) Sato, Y.; Yato, M.; Ohwada, T.; Saito, S.; Shudo, K. J . Am.
Chem. Soc. 1995, 117, 3037-3043.
(24) The facile deuterium exchange of the phenyl protons of the
(26) The electrophilic reactivity of imines is known to be increased
by N-acylation or N-sulfonylation, so-called amidoalkylation. The
proposed intermediates of the amidoalkylation are N-acyliminium ions
or N-sulfonyliminium ions, which are activated to react with nitroben-
zene. The ammonium group of the dication 4 acts as an acyl-nitrogen
atom or a sulfonyl-nitrogen atom. (a) Venkov, A.; Lukanov, L. Synthesis
1989, 59-61. (b) Lukanov, L.; Venkov, A. P.; Mollov, N. M. Synthesis
1987, 204-206. (c) For reviews of amidoalkylations, see: Speckamp,
W. N.; Hiemstra, H. Tetrahedron 1985, 41, 4367-4416. (d) Zaugg, H.
E. Synthesis 1985, 49. (e) Sisido, K.; Inada, H.; Isida, T. Tetrahedron
Lett. 1968, 50, 5267-5269.
2-phenethylamine moiety of 1c even at 23 °C indicates
formation of dicationic species such as 5i-5iii
a facile
(27) Child, R. F.; Dickie, B. D. J . Am. Chem. Soc. 1983, 105, 5041-
5046.
(28) (a) Ohwada, T.; Yamagata, N.; Shudo, K. J . Am. Chem. Soc.
1991, 113, 1364-1373. (b) Saito, S.; Sato, Y.; Ohwada, T.; Shudo, K.
J . Am. Chem. Soc. 1994, 116, 2312-2317. (c) Saito, S.; Ohwada, T.;
Shudo, K. J . Am. Chem. Soc. 1995, 117, 11081-11084. (d) Suzuki, T.;
Ohwada, T.; Shudo, K. J . Am. Chem. Soc. 1997, 119, 6774-6780.
and probably ipso-protonated 5iv. Nevertheless, they are not substrates
for the cyclization due to positive charge repulsion between the
benzenium cation and the iminium cation. A more facile ring proto-
nation is expected in the imines 1f and 1g.

616 J . Org. Chem., Vol. 64, No. 2, 1999
Yokoyama et al.
the magnitude corresponds well to that between the
dicationic and monocationic cyclizations in the case of 1a .
Furthermore, the cyclization reactions of the methyl-
substituted imines (1f and 1g) in TFSA (at 60 °C) are
faster than that of the unsubstituted imine 1c in TFSA
(at 60 °C, practically no reaction takes place). This result
strongly supports the idea that the cyclization process is
involved in the rate-determining step of the reaction,
because the presence of one or more methyl substituent
increases the electron-donating ability of the cyclizing
benzene ring and the cyclization is accelerated. If rate-
determining protonation were involved,29 the cyclization
reaction would be independent of the aromatic substit-
uents of the phenethylamino moiety because these re-
mote substituents would not modify the basicity of the
common iminium moiety of the monocations 3c, 3f, and
3g, which is relevant to the equilibrium with the dica-
tions 4.30 The large values of negative entropy of activa-
tion (∆Sq, -33.8 eu to -46.3 eu) of the cyclization of 1c-
1e (Table 3) are also consistent with the rate-determining
cyclization.
Because a methoxy group on the aromatic ring is a
much stronger electron-donating group than the methyl
group, the imine 1h , N-benzylidene-2-(3,5-dimethoxyphe-
nyl)ethylamine, which bears two methoxy groups, can be
regarded as a truly activated substrate. This imine 1h
readily undergoes cyclization in TFA even at 30 °C (62
h) to give the corresponding 2h in a high yield (89%
yield). When the reaction was carried out at reflux (72
°C) in TFA, the reaction is rapid (1.2 h), and the yield of
2h is 89%. The imine 1h is completely N-monoprotonated
in TFA to give the iminium cation 3h in a manner similar
to the that for imine 1c, since the basicities of the imino
nitrogen atoms of 1c and 1h are identical. This facile
reaction of 1h confirms the intermediacy of the N-
monoprotonated imine 3h in the conventional Pictet-
Spengler reaction, wherein strongly activated cyclizing
benzenes are involved.31
the ammonium-carbenium dications. These findings
dispel the notion that the Pictet-Spengler reaction is
restricted to activated substrates which bear strongly
electron-donating groups on the cyclizing benzene ring.
The present chemistry of the superelectrophiles will give
a new impetus to synthetic applications of the Pictet-
Spengler reaction, and the proposed reaction mechanisms
also shed light on the mechanisms of biosynthesis of the
endogenous tetrahydroisoquinolines, which are of great
interest in neuroscience.6,7
Exp er im en ta l Section
Ma ter ia ls Trifluoromethanesulfonic acid (TFSA) was pur-
chased from Central Glass Co. (J apan) and was purified by
distillation as described previously.12 Great care was taken to
obtain anhydrous TFSA, which has an acidity of H0 ) -14.1.
Trifluoroacetic acid (TFA) was also purified by distillation as
described previously.12 Piperidine was purified by distillation
and stored in glass ampules. The imines are prepared as
previously described.15,17,31-34.
Cycliza tion of 1a -1e in TF SA. Gen er a l P r oced u r e. The
imine 1 (1.0 mmol) was dissolved in 100 equiv of TFSA (15.2
g, 100 mmol) at ambient temperature, and the mixture was
stirred at a specified temperature for a specified time shown
in Table 1. Then the solution was poured into ice-water (100
mL), and the mixture was basified with 5 N NaOH (30 mL)
and extracted with CH2Cl2 (100, 50, and 50 mL). The combined
organic layer was dried over anhydrous Na2SO4, and the
solvent was evaporated. The residue was flash-chromato-
graphed (CH2Cl2, then MeOH:CH2Cl2 ) 1:20, or MeOH:ethyl
acetate ) 1:20) to give the cyclized product, the 1,2,3,4-
tetrahydroisoquinoline derivative 2.
Acid -Ca ta lyzed Cycliza tion of N-Meth ylid en e-2-p h en -
eth yla m in e 1a to 1,2,3,4-Tetr a h yd r oisoqu in olin e 2a . (1)
Dir ect Sep a r a tion of 2a . TF SA. To well-stirred TFSA (15.11
g, 101 mmol) was added the imine (135 mg, 1.01 mmol) at
ambient temperature. The solution was stirred at ambient
temperature for 3 h. Then the acid solution was poured into
100 mL of ice and water, followed by addition of 30 mL of 5 N
NaOH. The whole was extracted with methylene chloride, and
the organic layer was dried over Na2SO4. The solvent was
evaporated to give the residue, which was flash-chromato-
graphed (aluminum oxide, acetonitrile/methylene chloride) to
give 55 mg (41%) of 2a . The amine 2a obtained above was
identical with authentic 1,2,3,4-tetrahydroisoquinoline in terms
of the 1H NMR spectrum.
(2) Sep a r a tion a fter Ben zoyla tion . In the case of the
imine 1a , the crude extract of the reaction mixture was
subjected to benzoylation (benzoyl chloride in dry CH2Cl2 and
triethylamine at 0 °C), followed by the separation. The
benzoylation of authentic 1,2,3,4-tetrahydroisoquinoline pro-
ceeded quantitatively (98.2% yield).
TF SA-TF A. To a well-stirred solution of the imine 1a
(532.6 mg, 4 mmol) in TFA (5.86 g) was added weighed TFSA
(52.7 g) at ambient temperature to give a 90% w/w TFSA-
10% w/w TFA solution (100 equiv with respect to 1a ). The
solution was heated at 50 °C with stirring for 1 h. After cooling
in an ice-water bath, the whole was poured into 200 mL of
ice and water, followed by basification with 5 N aqueous
NaOH. The whole was extracted with methylene chloride and
dried over Na2SO4. The solvent was evaporated to give the
crude amine residue (455.8 mg). To a solution of the crude
amine in 3 mL of dry CH2Cl2 and triethylamine (1 mL) was
added a solution of benzoyl chloride (484.6 mg) in 1 mL of dry
CH2Cl2 at 0 °C. After 1 h, water was added; then the whole
was extracted with CH2Cl2 (200 mL), and the organic layer
was washed with brine and dried over Na2SO4. The residue
obtained after evaporation of the solvent was flash-chromato-
graphed (silica gel, acetone/n-hexane 1:8) to give 705.2 mg
(74%) of N-benzoyl-1,2,3,4-tetrahydroisoquinoline (the N-ben-
zoyl form of 2a ) and 45.1 mg (5% yield) of N-benzoyl-2-
phenethylamine.
Con clu sion
Our study showed that the prototype Pictet-Spengler
reaction of N-methylene-2-phenethylamine 1a can be
catalyzed by TFA to give 1,2,3,4-tetrahydroisoquinoline
2a in good yield. This result shows that cyclization
through the N-protonated iminium monocation 3a can
intervene in the prototype Pictet-Spengler reaction and
also confirms the original observation of Pictet and
Spengler, approximately 90 years after their publication.
We also found that a superacid can catalyze the Pictet-
Spengler reactions of the imines (1a -1e) of 2-phenethy-
lamine, including 1a . The acidity dependence of the
reaction suggested that N-protonated imines 3b-3e
(except 3a ) are not reactive intermediates in the Pictet-
Spengler reactions. Our kinetic studies of the cyclizations
revealed that the true electrophiles are dicationic super-
electrophiles, the N,N-diprotonated imines 4a -4e, i.e.,
(29) Ritchie, C. D.; Lu, S. J . Am. Chem. Soc. 1989, 111, 8542-8543.
(30) Even when the N,C-diprotonation to give the dication 6 is
involved, the basicities of the benzene carbon atoms of the substituent
R1 of the monocations 3c, 3f,and 3g would be similar to each other. If
the N,C-protonation process is rate-determining, the cyclization rates
of 1c, 1f, and 1g would not be modified. This is in disagreement with
the observed acceleration, 1g > 1f > 1c.
(31) The reaction of 1h in TFA at 72 °C is significantly faster than
those of 1a , 1f, and 1g under the same reaction conditions (Table 1).
Thus, in the monocationic cyclization in TFA, the cyclization process
is also involved in the rate-determining process.

Pictet-Spengler Reactions Catalyzed by Superacids
J . Org. Chem., Vol. 64, No. 2, 1999 617
Ta ble 4. Cor r ected Va lu es of Acid ity F u n ction of th e
Acid Med ia in th e P r esen ce of 1 m ol % P ip er id in e
which was sealed tightly with a Teflon cap. The same acid,
containing piperidine, but free from the indicator, was used
as a reference in the UV measurements. The value of the
acidity function was determined from the absorption of the
indicator as described previously12
b
c
acida
uncorrected H0
corrected H0
100% TFSA
-14.1
-13.3
-12.7
-12.2
-10.8
-12.66
-12.42
-12.10
-11.52
-10.54
89.0% TFSA-11.0% TFA
78.5% TFSA-21.5% TFA
59.9% TFSA-40.1% TFA
26.2% TFSA-73.8% TFA
Kin etics of th e Cycliza tion of 1c-1e a n d 1a . Samples
of the imines 1c-1e were prepared at ambient temperature.
An imine (1c-1e) (1 mmol) and a trace amount of dry CH2Cl2
(2-3 mg, as an internal reference for the 1H NMR measure-
ments) were dissolved in a premixed TFSA-TFA acid (ca 1.0
g, 0.1 mol, 100 equiv) under a dry nitrogen atmosphere in an
AtmosBag (Aldrich). A 0.5-mL aliquot of the solution was
transferred into an NMR tube, which was then sealed. A dozen
sealed samples were prepared similarly. A set of the samples
sealed in the NMR tubes, containing the same acid solution
of the imine 1, was immersed in a silicon oil (KF-96-10, Shin-
Etsu, J apan) bath heated at a specified temperature (100, 110,
and 120 °C). The temperature was controlled by a thermostat-
controlled circulator (DCS-B3, Haake, Germany). The error
in temperature is within (0.1 °C. At regular intervals, a tube
was withdrawn from the heating bath and cooled in an ice-
water bath to quench the reaction, and the process of the
reaction was monitored by measuring in the 1H NMR spectrum
at 30 °C (at 30 °C practically no cyclization reaction takes place
for the imines 1c-1e). The disappearance of the starting
material was monitored in terms of the integrated signal
intensity with respect to the signal of the acid as an integration
reference with a GX 400-MHz NMR spectrometer. We also
measured the rates of the formation of the product and
confirmed that the rate of disappearance of the starting imine
(for example 1c) is identical with the rate of appearance of
the cylized product.
a
b
Weight percent of the composite acids. Reference 12. c See
also ref 13.
TF A. To 31 mL of TFA was added the imine 1a (531.7 mg,
4 mmol) at ambient temperature. It was completely dissolved
in 15 min at ambient temperature. The solution was heated
at 50 °C with stirring for 20 h. After cooling in an ice-water
bath, the whole was poured into 200 mL of ice and water,
followed by basification with 5 N aqueous NaOH. The whole
was extracted with methylene chloride (500 mL), washed with
brine, and dried over Na2SO4. The solvent was evaporated to
give the crude amine residue (445.1 mg). The residue was
subjected to benzoylation as described in (2), and the residue
obtained was flash-chromatographed (silica gel, acetone/n-
hexane 1:8) to give 510.5 mg (54%) of the N-benzoyl form of
2a and 183.2 mg (20% yield) of N-benzoyl-2-phenethylamine.
Acid ity Va lu es H0 of TF SA-TF A in th e P r esen ce of 1
m ol % of P ip er id in e. Acid mixtures, TFSA-TFA, used in
the determination of the acidities (H0) and in the kinetic
studies were prepared in a polyethylene glovebox (AtmosBag,
Aldrich) under a dry nitrogen atmosphere. UV spectra were
recorded with a Shimadzu UV-200S using matched quartz
glass 1-cm cells sealed with a Teflon stopper.
The solution of 1a in the TFSA-TFA acid was prepared as
follows: to the weighed imine 1a was added weighed TFA at
ambient temperature in an AtmosBag (Aldrich). This TFA
solution was cooled to 0 °C in an ice-water bath after a good
solution had been obtained. To this TFA solution was added a
precooled weighed aliquot of TFSA under a dry nitrogen
atmosphere in an AtmosBag (Aldrich). A portion of the final
solution was transferred to an NMR tube, which was sealed.
The 1H NMR spectra were recorded at a specified temperature.
The NMR probe temperature was controlled to within (0.1.
Rates of cyclization of 1a in TFA were measured at 50.0, 60.0,
and 70.0 °C, and those of 1a in 90% w/w TFSA-10% w/w TFA
were obtained at 0.0, 10.0, and 20.0 °C. Temperatures were
calibrated with ethylene glycol (in the range of 50.0-70.0 °C)
and with methanol (in the range of 0.0-20.0 °C) as usual.36
The disappearance of the starting material was monitored at
regular intervals. We obtained good first-order kinetics in the
cyclization reactions of the imines 1a and 1c-1d . Errors of
rates, derived from the signal integration, are within (2%.
In strong acid of more than H0 ) -10, wherein the
cyclization can take place, the imines 1 and the resultant
cyclized product, 1,2,3,4-tetrahydroisoquinolines 2, behave as
strong nitrogen bases because of complete N-monoprotonation,
thus reducing the acidity of the reaction medium. We therefore
estimated the lowering of the acidity of the reaction medium
by measuring the acidity of TFSA-TFA (100 equiv with
respect to piperidine) in the presence of piperidine at acidity
levels stronger than H0 ≈ -10 (Table 4).12,13
Determination of the acidity of the acid mixtures was
conducted by means of the usual indicator method with UV
spectroscopy at 25 °C.12 Acid mixtures, TFSA-TFA, were
prepared in an AtmosBag (Aldrich) under a dry nitrogen
atmosphere at ambient temperature. The acidity of acids
stronger than 59.9% w/w TFSA-40.1% w/w TFA, including
100% TFSA (H0 ) -11.52 to -12.66), was determined by using
4-fluoronitrobenzene as an indicator.12 In the determination
of the acidity of 26.2% w/w TFSA-73.8% w/w TFA acid, 2,4,6-
trinitroaniline was used as an indicator.12
The corrected values of the acidity function were determined
as follows. For example, 59.9% w/w TFSA-40.1% w/w TFA
acid in the presence of 1 mol % piperidine: the acid mixture
(59.9% w/w TFSA-40.1% w/w TFA) was prepared by mixing
TFSA (30.052 g) and TFA (20.138 g) at ambient temperature.
To the weighed acid mixture (50.190 g, 377 mmol, 102 equiv
with respect to piperidine) was added piperidine (315 mg, 3.70
mmol). To an aliquot of the above acid solution (9.873 g) was
added a weighed indicator, 4-fluoronitrobenzene (6.18 mg), to
obtain a good solution. A small portion of this solution (0.819
g) was transferred into a 10-mL volumetric flask and diluted
to 10 mL with the stock acid solution (the final concentration
of 4-fluoronitrobenzene was 3.63 × 10-4 mol/L in this case). A
portion of this solution was transferred into a 1-cm quartz cell,
Su p p or tin g In for m a tion Ava ila ble: General experimen-
tal details and compound characterization data (6 pages).
Ordering information is given on any current masthead page.
J O982019E
(32) Reduction of nitriles to the corresponding amines was carried
out in the presence of a Lewis acid. See: Nystrom, R. B. J . Am. Chem.
Soc. 1955, 77, 2544-2545.
(33) Schonne, A. Bruylants, A. Bull. Soc. Chim. Belg. 1953, 62, 155-
171; Chem. Abstr. 1954, 48, 9341b.
(34) Belleau, B. Can. J . Chem. 1957, 35, 651-662.
(35) Ginos, J . Z.; Cotzias, G. C.; Tolosa, E.; Tang, L. C.; LoMonte,
A. J . Med. Chem. 1975, 18, 1194-1200.
(36) Van Geet, A. L. Anal. Chem. 1970, 42, 679-680.