8,8′-Dialkyl-1,1′-biisoquinolines: Preparation, absolute configuration and unexpected racemization behaviour

Article abstract of DOI:10.1039/a905161i

A series of 8,8′-dialkyl-1,1′-biisoquinolines, in which methyl, ethyl and isopropyl groups are introduced for enhancing the transannular steric hindrance, are synthesized. The atropisomeric biisoquinolines are separated into both enantiomers, of which the absolute configurations and the optical stabilities are determined. Contrary to prior expectations, the racemization behaviour is inversely proportional to the steric size of the alkyl groups. The Royal Society of Chemistry 1999.

Full text of DOI:10.1039/a905161i

View Article Online / Journal Homepage / Table of Contents for this issue
8,8Ј-Dialkyl-1,1Ј-biisoquinolines: preparation, absolute
configuration and unexpected racemization behaviour†
Hirohito Tsue,a Hideyuki Fujinami,a Takeshi Itakura,a Ryuta Tsuchiya,a Kimiko Kobayashi,b
Hiroki Takahashib and Ken-ichi Hiraoa
a Division of Material Science, Graduate School of Environmental Earth Science, Hokkaido
University, Sapporo 060-0810, Japan
b Division of Molecular Characterization, Department of Research Fundamentals Technology,
The Institute of Physical and Chemical Research (RIKEN), Wako 351-0198, Japan
Received (in Cambridge, UK) 28th June 1999, Accepted 7th October 1999
A series of 8,8Ј-dialkyl-1,1Ј-biisoquinolines, in which methyl, ethyl and isopropyl groups are introduced for
enhancing the transannular steric hindrance, are synthesized. The atropisomeric biisoquinolines are separated
into both enantiomers, of which the absolute configurations and the optical stabilities are determined. Contrary
to prior expectations, the racemization behaviour is inversely proportional to the steric size of the alkyl groups.
N,NЈ-dioxide derivative retains enough steric hindrance to be
Introduction
resolved into both enantiomers.7 Successful enhancement of the
optical stability has been also achieved independently by both
us8 and Chelucci9 with the aid of two methyl groups introduced
at the 8- and 8Ј-position; the dimethyl derivative 1b revealed a
half-life of 17 h at 20 ЊC in toluene.10 However, 1b does show
gradual racemization at room temperature, probably due to the
small contribution of the nitrogen lone pairs to the rotational
resistance about the pivotal bond, although the optical stability
is increased in comparison with that of the parent compound
1a. This foregoing finding led us to prepare biisoquinolines 1c–e
with a series of alkyl substituents at the 8,8Ј-positions and to
study how bulky alkyl groups are required to freeze the rota-
tion. However, on the other hand, we encountered unexpected
racemization behaviour in 1b–d, such that the bulkier the alkyl
substituent, the lower the optical stability. In this paper, we
report a full account of the syntheses and the absolute con-
figurations of the biisoquinolines 1b–d. Furthermore, we report
the unexpected reversal of the optical stability observed in
1b–d, and the racemization mechanism.
Optical activity based on a high barrier to rotation about
σ-bonds is designated as ‘atropisomerism’,1 that is exemplified
by a wide range of biaryl compounds, both natural and syn-
thetic. One of the important factors making such molecules dis-
symmetric is substituents adjacent to the rotational axes, and
the steric size, shape and hybridization of these substituents
exert a large influence upon the optical stability of the mole-
cules.2 Admittedly, increasing bulkiness of substituents tends to
cause an enhancement of the configurational stability as a
result of steric hindrance. For instance, 1,1Ј-binaphthyl shows
slow racemization at ambient temperature (half-life of ca. 10
h),3 whereas the 2,2Ј-diol or 2,2Ј-diphosphine derivatives, which
act as useful chiral inducers in asymmetric syntheses,4,5 give rise
to no racemization at the corresponding temperature.
Another illustration of this kind of substituent-induced
stabilization can be seen in 1,1Ј-biisoquinoline 1a. The parent
Results and discussion
Synthesis
Preparation of biisoquinolines 1b–e was carried out as shown
in Scheme 1. Starting from the condensation reaction of
o-alkylbenzaldehydes 2b–e with aminoacetaldehyde dimethyl
acetal, the resultant imines were converted to 8-alkyl-
isoquinolines 3b–e by the application of Hendrickson’s pro-
cedure,11 i.e., treatment successively with ethyl chloroformate,
trimethyl phosphite and titanium(IV) chloride. N-Oxidation
was accomplished either by MCPBA or by hydrogen peroxide.
Acetylation of 4b–e with acetic anhydride and subsequent
hydrolysis with aqueous sodium hydroxide12 afforded iso-
quinolones 5b–e, which were then converted to 6b–d by phos-
phoryl trichloride in moderate yields with the exception of 6e
(vide infra). Final homocoupling of 1-chloroisoquinolines 6b–d
using nickel(0) complex generated in situ by reduction of nickel
chloride with activated zinc13 gave the required 1,1Ј-biiso-
quinolines 1b–d in which a series of alkyl substituents was
introduced at the 8- and 8Ј-position. The structures of the new
compounds were determined by a combination of spectral data
and an elemental analysis.
compound 1a shows quite rapid racemization owing mainly to
the very small transannular steric hindrance between H-8 (8Ј)
and N-2Ј (2), and therefore isolation as the optically active form
is known to be substantially impossible.6 By contrast, the
† CD and UV-VIS spectra, primary kinetic data of racemization and
PM3-calculations are available as supplementary data. For direct elec-
tronic access see http://www.rsc.org/suppdata/p1/1999/3677, otherwise
available from BLDSC (SUPPL. NO. 57675, pp. 12) or the RSC
Library. See Instructions for Authors available via the RSC web page
(http://www.rsc.org/authors).
In the course of the attempted synthesis of tert-butyl deriv-
ative 1e, however, the intermediate 5e was not chlorinated, but
J. Chem. Soc., Perkin Trans. 1, 1999, 3677–3683
This journal is © The Royal Society of Chemistry 1999
3677

View Article Online
Scheme 1 Reagents and conditions: (i) H2NCH2CH(OMe)2, benzene,
reflux, 15 h; (ii) ClCO2Et, THF, Ϫ10 ЊC, 5 min; (iii) P(OMe)3, THF, rt,
20 h; (iv) TlCl4, CH2Cl2, reflux, 41 h; (v) MCPBA, CH2Cl2, rt, 4 h or
H2O2, AcOH, 80 ЊC, 12 h; (vi) Ac2O, reflux, 5 h; (vii) 1.3 M NaOH,
80 ЊC for 40 min and then rt for 12 h; (viii) POCl3, reflux, 3 h; (ix) NiCl2,
Zn, PPh3, DMF, 50 ЊC, 5 h.
Fig. 1 CD (upper panel) and UV–visible (lower panel) spectra of
biisoquinoline 1b in ethanol. Enantiomeric excesses of the samples were
78% ee for (R)-(ϩ)-1b and 68% ee for (S)-(Ϫ)-1b.
recovered almost quantitatively. Furthermore, direct dimeriz-
ation of 3e to 1e by using lithium diisopropylamide according
to Meth-CohnЈs procedure14 also failed. Considerations based
on the Corey–Pauling–Koltun molecular model suggested that
tert-butyl groups were too bulky to be incorporated into the
peri-positions of the biisoquinoline framework, and this might
be responsible for the failure of the synthesis. Indeed, this result
is consistent with the fact that the yields of the final homo-
coupling reaction 6 → 1 decreased gradually with a rise in
steric size of the alkyl substituents.
Absolute configuration
Enantiomeric enrichment of 1b–d was performed by two kinds
of well known methods, i.e., (1) high-performance liquid
chromatography (HPLC) using a chiral stationary-phase
column and (2) enantiomeric resolution through transform-
ation into a diastereomeric salt by using a chiral binuclear
palladium complex was reported by both Dai15 and Chelucci.9
All the biisoquinolines were resolved into both enantiomers
and isolated as optically active forms of 68–86% ee. This
incomplete enantiomeric resolution was caused by experi-
mental difficulties mainly due to the moderately easy racemiz-
ation of 1b–d (full particulars are given in the next section).
The absolute configurations of a series of biisoquinolines
1b–d were successfully determined by applying the exciton
chirality method.16 Almost the same CD and UV–visible
spectra were obtained for optically active 1b–d, and those of 1b
are shown in Fig. 1 as a typical example. Transition moments
along the long axes of the isoquinoline rings, which appeared as
an intense absorption at 219 nm in the UV spectrum, interact
with each other to give the exciton-split CD spectra. Cotton
effects for (Ϫ)-1b were observed positively at 235 nm and
negatively at 220 nm, indicating positive exciton chirality. On
the other hand, the mirror image was obtained for (ϩ)-1b. In
addition to the CD spectra, the specific rotations of [α]D ϩ60‡
for (ϩ)-1b of 78% ee and [α]D Ϫ53‡ for (Ϫ)-1b of 68% ee ensure
Fig. 2 ORTEP drawings of biisoquinolines 1b (top) and 1c (bottom).
that these are a pair of enantiomers. The cotton effect is known
to depend strongly on the dihedral angle between chromo-
phores; in analogous 1,1Ј-binaphthyl systems positive and
negative signs are exchanged at a dihedral angle of 110Њ.17
X-Ray structural elucidation using racemic single crystals
revealed an almost perpendicular conformation for 1b and 1c,
of which dihedral angles between the two isoquinoline rings
were found to be 102.1Њ for 1b and 93.8Њ for 1c (Fig. 2). Despite
a great deal of effort, a single crystal of 1d appropriate for
X-ray crystallography was not obtained. However, the AM1-
‡ In this paper, [α]D-values are given in units of 10Ϫ1 deg cm2 gϪ1
.
3678
J. Chem. Soc., Perkin Trans. 1, 1999, 3677–3683

View Article Online
Table 1 Racemization-rate constants at 30 ЊC and activation parameters for racemization in biisoquinolines 1b–d
Compound
krac/sϪ1
Ea/kJ molϪ1
∆H‡/kJ molϪ1
∆S‡/J molϪ1 KϪ1
∆G‡303/kJ molϪ1
1b
1c
1d
5.8 × 10Ϫ6
2.2 × 10Ϫ5
5.3 × 10Ϫ5
113(2)a
103(1)a
95(1)a
110(2)a
100(1)a
92(1)a
19(7)a
105
101
99
Ϫ3.9(3.2)a
Ϫ22(4)a
a Numbers in parentheses are standard deviations of this variable.
optimized18 geometry of 1d showed a dihedral angle of 73.3Њ,
and those of 1b and 1c were estimated to be 90.2Њ and 88.6Њ,
respectively. The X-ray and AM1 results showed slightly differ-
ent structures, but the dihedral angles for each compound were
found to be less than 110Њ. Hence, the exciton chirality method
should be applicable to 1b–d in a similar way to that for 1,1Ј-
binaphthyl systems, and led to a conclusive assignment that
all the (ϩ)-forms corresponded to (R)-configurations and all
(Ϫ)-forms to (S) ones as shown in Chart 1. Although the
Chart 1 Absolute configurations of 8,8Ј-dialkyl-1,1Ј-biisoquinolines.
methodologies are different, the assignment thus obtained was
identical with that reported by Chelucci,9 who succeeded in
determining the absolute configuration of 1b by using 1H NMR
spectroscopy.
Racemization
The rate constants (krac) for the conversion from one enantio-
mer to the other in 1b–d were determined in methanol over the
range of ϩ10 to ϩ50 ЊC. Optically active biisoquinolines 1b–d
of 52 to 92% ee were used for the kinetic measurements, and the
changes in the concentration of both enantiomers were fol-
lowed by using HPLC equipped with a chiral stationary phase
column. Although continuous chiroptical techniques such as
CD and optical rotatory dispersion should be superior candi-
dates for these kinetic measurements, a chiral HPLC technique
was used in the present study owing to experimental circum-
stances. Fig. 3 represents a typical example of the concen-
trational changes with the passage of time in 1b at ϩ40 ЊC.
Under the described conditions, both the enantiomers, of
which absolute configurations were identified in the preceding
section, were well resolved. The initial enantiomeric excess of
92% was gradually decreased to 14% ee after 11 h. Contrary to
prior expectations, the bulkier the alkyl substituents, the greater
the rate constants. As summarized in Table 1, the rate con-
stants were found to be largest in 1d and smallest in 1b, and
this unexpected racemization behaviour in 1b–d was clearly
reflected in the systematic decrease in activation parameters
such as the Arrhenius activation energy (Ea), the heat of acti-
vation (∆H‡) and the Gibbs energy of activation (∆G‡); these
thermodynamic parameters were inversely proportional to the
size of the alkyl groups. However, it is difficult to present a
detailed discussion on the decrease in the entropy of activation
(∆S‡) at this moment.
Fig. 3 Concentrational changes of (R)- and (S)-1b in methanol at
40 ЊC. Column Chiralcel OD; mobile phase 30% EtOH–hexane; flow
rate 0.5 ml minϪ1; λ = 280 nm.
gave only one imaginary frequency for each TS structure. The
GS and TS geometries of 1c are shown in Fig. 4 as a typical
example. As seen in Fig. 4, the two isoquinoline rings were
nearly perpendicular in the GS, but almost coplanar in the TS
though both rings were much distorted. The calculated barrier
heights (∆∆Hf), deduced from the heat of formation in the
GS and TS, were nearly in accordance with the experimental
findings as summarized in Tables 1 and 2.
Optically active rotational isomers have sometimes been
reported to be more optically labile than would be expected
from the structures.2a,20 As a good example, Fuji et al.21
recently reported this kind of behaviour in 8,8Ј-disubstituted
1,1Ј-binaphthyl compounds. From considerations based on
X-ray crystallography and theoretical calculations, they con-
cluded that the relatively easy racemization of a compound
with a bulkier substituent originated from the destabilization of
the ground state. On the theoretical side, the methodology used
in SchleyerЈs study22 is helpful to us in examining this possibil-
ity. Using the PM3-optimized GS and TS geometries of 1b–d,
their distortional energies were evaluated according to the
described procedures.22 The results of the calculations are
summarized in Table 2.
In order to clarify the unexpected reversal of the sequence in
1b–d, PM3-calculations19 were carried out for the ground state
(GS) and the transition state (TS) during the racemization pro-
cess, in which a more favourable anti-pathway was taken into
consideration between two kinds of possible pathways such as
syn and anti. Harmonic vibrational frequency calculations
What is significant in Table 2 is that the calculated dis-
tortional energies in the GS were proportional to the steric size
J. Chem. Soc., Perkin Trans. 1, 1999, 3677–3683
3679

View Article Online
Table 2 PM3-calculated activation barriers for racemization and dis-
on the transition-state stabilization which acts as an altern-
tortional energies inherent in biisoquinoline frameworks
ative mechanism to lower the racemization barrier. The
simple racemization via an unfavourable syn-pathway in the
chiral ruthenium(II) complex of the parent compound 1a is
accounted for by the metal-assisted favourable coordination in
the TS,24 but this kind of special interaction is completely
eliminated in the present system. Further work on the non-
distortional contribution may be needed to allow us to fully
understand the mechanism, nonetheless the present result
clearly demonstrates that destabilization in the GS would be the
most plausible mechanism for the unexpected racemization
behaviour observed in biisoquinolines 1b–d.
Distortional energy/kJ molϪ1
Compound
∆∆Hf/kJ molϪ1
Ground state
Transition state
1b
1c
1d
111.6
104.9
97.9
10.1
19.9
25.6
94.7
89.4
94.7
Conclusions
Summarizing the present study, we have arrived at the following
conclusions: (1) a series of 8,8Ј-dialkyl-1,1Ј-biisoquinolines, in
which methyl, ethyl and isopropyl groups are introduced, have
been prepared in nine steps from readily available o-alkyl-
benzaldehydes. (2) Enantiomeric resolution into both enan-
tiomers was achieved by two kinds of well known techniques,
and then the absolute configurations were successfully deter-
mined by application of the exciton chirality method. (3)
Although the optical stabilities of the 8,8Ј-dialkyl derivatives
were largely enhanced in comparison with that of the parent
compound, the racemization behaviour was found to be
inversely proportional to the steric size of the alkyl groups. On
the basis of PM3 and X-ray analyses, the unexpected racemiz-
ation was ascribed to destabilization of the ground state.
Although the detailed mechanism is being investigated in
our laboratory, the present work affords one example that
shows the curious relationship between molecular structure and
atropisomerism.
Experimental
General
Fig. 4 PM3-optimized geometries of the ground and transition states
during the racemization process in 1c.
Mps were determined on a Yanako melting point apparatus
MP-500D and are uncorrected. 1H NMR spectra were obtained
at 400 MHz with a JEOL EX-400 spectrometer for samples in
CDCl3 solution with tetramethylsilane as an internal standard.
J-Values are given in Hz. IR and mass spectra were recorded on
JEOL FT/IR-230 and JEOL JMS DX-300 spectrometers,
respectively. UV–visible spectra were recorded on a Shimadzu
UV-2400PC spectrometer at a concentration of 9.67–9.83 ×
10Ϫ6 mol dmϪ3 in ethanol at 20 ЊC. Elemental analyses were
performed on a Yanako MT-5. CD spectra were measured on a
JEOL J-720 spectrometer in a 1 cm path-length cell at a concen-
tration of 1.96–2.28 × 10Ϫ5 mol dmϪ3 in ethanol at 10 ЊC.
Optical rotations were measured in a 1 dm path-length cell on a
JASCO DIP-370 polarimeter. Merck Kieselgel #7734 was used
for column chromatography. For analytical thin-layer plates
Merck #5715 and #5721 were used. HPLC analyses were
carried out with a JASCO instrument equipped with an
870-UV detector, an 880-PU pump and a Chromatocorder 12
recorder. Shiseido Ceramospher Chiral RU-1 and Daicel Chi-
ralcel OD were used as chiral stationary-phase columns. All
chemicals were reagent grade and were used without further
purification. Organic solvents were purified by standard pro-
cedures. Compound 2b is commercially available and used
without further purification. Compounds 2c–e were prepared
according to the described procedures.25 Semiempirical calcu-
lations based on AM118 and PM321 methods were carried
out using the MOPAC97 program package implemented in
WinMOPAC Version 2.0, Fujitsu Limited, 1998.
of the alkyl groups, whereas this was not the case for the TS.
Corey–Pauling–Koltun molecular models suggested that there
exists a small space around the nitrogen lone pairs at the 2,2Ј-
positions, enough to slightly accommodate the alkyl groups in
the TS, where the alkyl substituents lie in close proximity to the
nitrogen lone pairs. This might result in little influence to effect-
ively enhance the steric hindrance, and therefore the dis-
tortional energies in the TS should become almost constant
irrespective of the alkyl groups, while those in the GS should
increase in proportion to the steric size of the substituents.
The systematic increase in the distortional energies in the GS
is evident from the X-ray structural elucidation, by which
bond angles N2–C1–C1Ј were found to be reduced to 112.1Њ for
1b and 110.9Њ for 1c due to the steric interaction between the
isoquinoline ring and the alkyl group at the peri-position
(Fig. 2). As mentioned in the preceding section, a single crystal
of 1d was not obtained, but the PM3-optimized geometries of
1b–d showed a similar decrease in the relevant angles that were
estimated to be 112.7Њ for 1b, 112.2Њ for 1c and 111.3Њ for 1d.
Here, it is worth noting that the increase in the distortional
energies in the GS (ϩ9.8 kJ molϪ1 for 1b → 1c, and ϩ15.5
kJ molϪ1 for 1b → 1d) is approximately reflected on the
decrease in the PM3-calculated barrier heights (Ϫ6.7 and Ϫ13.7
kJ molϪ1, respectively), as can be seen from Tables 1 and 2.
Equally importantly, the calculated relative values are in good
agreement with the experimental findings such as Ea and ∆H‡
(Ϫ10 kJ molϪ1 for 1b → 1c, and Ϫ18 kJ molϪ1 for 1b → 1d),
although the distortional energies were roughly estimated.
This moderate coincidence between the experimental and calcu-
lated values means that the unexpected racemization observed
in 1b–d would be caused predominantly by the destabilization
in the GS. One extensive experiment23 has so far been reported
General procedure for compounds 3b–e
8-Methylisoquinoline 3b. A solution of 2b (10.22 g, 85 mmol)
and aminoacetaldehyde dimethyl acetal (8.75 g, 85 mmol) in
dry benzene (50 ml) was refluxed for 15 h; during this period
3680
J. Chem. Soc., Perkin Trans. 1, 1999, 3677–3683

View Article Online
water was removed by using a Dean–Stark trap. After removal
of the solvent, the resultant viscous oil was dissolved in dry
THF. To the solution was added ethyl chloroformate (8.7 ml, 85
mmol) at Ϫ10 ЊC with vigorous stirring. After stirring of the
mixture for 5 min, 12 ml (100 mmol) of P(OMe)3 was added at
room temperature. The mixture was stirred for 20 h at room
temperature and was then concentrated under reduced pressure.
In order to remove trace amounts of P(OMe)3, evaporation
with toluene was repeated twice. The resulting oil was dissolved
in dry CH2Cl2 (110 ml), and 6 molar equiv. (56 ml, 0.51 mol) of
TiCl4 were added. The mixture was heated under reflux for 41 h.
The reaction mixture was basified by adding 10% aq. NaOH,
whereupon TiO2 precipitated as a white solid. The mixture was
filtered though Celite and the filtrate was acidified with 3 mol
dmϪ3 HCl. After washing with CH2Cl2, the aqueous layer
was basified strongly with 10% aq. NaOH and extracted with
CH2Cl2. The organic layer was washed successively with water
and brine, dried over Na2SO4, and evaporated in vacuo to afford
3b (8.2 g, 68%) as a pale yellow oil (Found: C, 83.94; H, 6.48; N,
9.65. C10H9N requires C, 83.88; H, 6.34; N, 9.78%); νmax(neat)/
cmϪ1 1615, 1580, 1570; δH 2.78 (3 H, s, CH3), 7.38 (1 H, dd, J 1.0
and 6.8, 7-H), 7.56 (1 H, dd, J 6.8 and 7.8, 6-H), 7.63 (1 H, d,
J 5.4, 4-H), 7.66 (1 H, dd, J 1.0 and 7.8, 5-H), 8.55 (1 H, d, J 5.4,
3-H), 9.45 (1 H, s, 1–H); m/z (EI) 143 (Mϩ).
νmax(Nujol)/cmϪ1 1604, 1260; δH 1.38 (3 H, t, J 7.6, CH2CH3),
3.01 (2 H, q, J 7.6, CH2CH3), 7.46 (1 H, d, J 7.1, 7-H), 7.52
(1 H, dd, J 7.1 and 8.1, 6-H), 7.64 (1 H, d, J 8.1, 5-H), 7.67 (1 H,
d, J 7.1, 4-H), 8.14 (1 H, dd, J 1.2 and 7.1, 3-H), 8.99 (1 H, d,
J 1.2, 1-H); m/z (EI) 173 (Mϩ) (Found: Mϩ 173.0847.
C11H11NO requires M, 173.0840).
8-Isopropylisoquinoline N-oxide 4d. Yield 90%. Light yellow
oil (Found: C, 71.69; H, 7.08; N, 6.78. C12H13NOؒ0.75H2O
requires C, 71.80; H, 7.28; N, 6.98%); νmax(neat)/cmϪ1 1626,
1600, 1566, 1256; δH 1.40 (6 H, dd, J 6.8 and 2.9, 2 × CH3), 3.51
[1 H, m, CH(CH3)2], 7.52–7.57 (2 H, m, 6- and 7-H), 7.63 (1 H,
dd, J 6.7 and 2.6, 5-H), 7.67 (1 H, d, J 7.1, 4-H), 8.13 (1 H, dd,
J 7.1 and 1.7, 3-H), 9.06 (1 H, d, J 1.7, 1-H); m/z (EI) 187 (Mϩ)
(Found: Mϩ, 187.1010. C12H13NO requires M, 187.0996).
8-tert-Butylisoquinoline N-oxide 4e
A solution of MCPBA (4.23 g, 24.5 mmol) in CH2Cl2 (12 ml)
was added to 3e (1.85 g, 9.99 mmol). After stirring for 4 h, the
mixture was poured onto dilute aq. NaHSO3, and then the reac-
tion mixture was extracted with CH2Cl2. The organic layer was
dried over anhydrous Na2SO4 and evaporated under reduced
pressure. Recrystallization from EtOAc afforded 4e (1.77 g,
88%) as brown columns (Found: C, 77.55; H, 7.56; N, 6.93.
C13H15NO requires C, 77.58; H, 7.52; N, 6.96%); mp 197–
198 ЊC (from EtOAc); νmax(KBr)/cmϪ1 1179; δH 1.60 (9 H, s,
3 × CH3), 7.49 (1 H, dd, J 8.5 and 8.1, 6-H), 7.60–7.65 (2 H, m,
5-H and 7-H), 7.67 (1 H, d, J 7.1, 4-H), 8.13 (1 H, d, J 7.1, 3-H),
9.40 (1 H, s, 1-H); m/z (EI) 201 (Mϩ).
8-Ethylisoquinoline 3c. Yield 56%. Colourless oil (Found: C,
79.77; H, 7.03; N, 8.29. C11H11Nؒ0.5H2O requires C, 79.48; H,
7.28; N, 8.43%); νmax(neat)/cmϪ1 1621, 1583; δH 1.42 (3 H, t,
J 7.5, CH2CH3), 3.21 (2 H, q, J 7.5, CH2CH3), 7.42 (1 H, d,
J 7.1, 7-H), 7.60 (1 H, dd, J 7.1 and 8.3, 6-H), 7.64 (1 H, d, J 5.9,
4-H), 7.67 (1 H, d, J 8.3, 5-H), 8.53 (1 H, d, J 5.6, 3-H), 9.50
(1 H, s, 1-H); m/z (EI) 157 (Mϩ) (Found: Mϩ, 157.0905.
C11H11N requires M, 157.0891).
General procedure for compounds 5b–e
8-Methylisoquinolin-1(2H)-one 5b. A mixture of 4b (246 mg,
1.5 mmol) and Ac2O (5 ml) was refluxed for 5 h. After removal
of Ac2O in vacuo, the resulting residue was heated to 80 ЊC with
1 mol dmϪ3 NaOH (4.1 ml) for about 40 min and stored at
room temperature for 12 h. The mixture was extracted with
CH2Cl2. The organic layer was dried over Na2SO4 and evapor-
ated under reduced pressure. The residue was purified by
column chromatography on SiO2 with hexane–EtOAc–MeOH
(6:3:1). Recrystallization from MeOH gave 5b (137 mg, 56%)
as colourless plates (Found: C, 75.76; H, 5.83; N, 8.56. C10H9NO
requires C, 75.45; H, 5.70; N, 8.80%); mp 139–141 ЊC (from
MeOH); νmax(Nujol)/cmϪ1 3130, 1660, 1640, 1590; δH 2.93 (3 H,
s, CH3), 6.44 (1 H, d, J 6.8, 4-H), 7.01 (1 H, d, J 6.8, 3-H), 7.22
(1 H, dd, J 1.0 and 7.3, 5-H), 7.35 (1 H, dd, J 1.0 and 7.8, 7-H),
7.48 (1 H, dd, J 7.3 and 7.8, 6-H), 9.54 (1 H, s, NH); m/z (EI)
159 (Mϩ).
8-Isopropylisoquinoline 3d. Yield 61%. Colourless oil (Found:
C, 84.08; H, 7.79; N, 8.26. C12H13N requires C, 84.17; H, 7.65;
N, 8.18%); bp 111–112 ЊC (2 mmHg); νmax(neat)/cmϪ1 1617,
1587, 1572; δH 1.43 (6 H, dd, J 6.8 and 1.0, 2 × CH3), 3.89 [1 H,
m, CH(CH3)2], 7.49 (1 H, dd, J 6.2 and 2.1, 7-H), 7.59–7.65
(3 H, m, 4-, 5- and 6-H), 8.52 (1 H, d, J 5.6, 3-H), 9.59 (1 H, s,
1-H); m/z (EI) 171 (Mϩ).
8-tert-Butylisoquinoline 3e. Yield 48%. Pale yellow oil
(Found: C, 83.47; H, 8.36; N, 7.10. C13H15Nؒ0.1H2O requires C,
83.47; H, 8.19; N, 7.49%); νmax(neat)/cmϪ1 2982, 1739; δH 1.66 (9
H, s, 3 × CH3), 7.57–7.61 (2 H, m, 7- and 6-H), 7.66–7.68 (2 H,
m, 4- and 5-H), 8.49 (1 H, d, J 5.4, 3-H), 9.94 (1 H, s, 1-H);
m/z (EI) 185 (Mϩ) (Found: Mϩ, 185.1192. C13H15N requires M,
185.1204).
8-Ethylisoquinolin-1(2H)-one 5c. Yield 82%. Colourless col-
umns (Found: C, 76.36; H, 6.45; N, 8.08. C11H11NO requires
C, 76.28; H, 6.40; N, 8.09%); mp 169–170 ЊC (from MeOH);
νmax(KBr)/cmϪ1 3162, 1642, 1598; δH 1.32 (3 H, t, J 7.3,
CH2CH3), 3.44 (2 H, q, J 7.3, CH2CH3), 6.47 (1 H, d, J 7.0,
4-H), 7.07 (1 H, d, J 7.0, 3-H), 7.26 (1 H, d, J 5.9, 5-H), 7.37 (1
H, d, J 7.8, 7-H), 7.52 (1 H, dd, J 5.9 and 7.8, 6-H), 10.82 (1 H,
br s, NH); m/z (EI) 173 (Mϩ).
General procedure for compounds 4b–d
8-Methylisoquinoline N-oxide 4b. A mixture of 3b (332 mg,
2.3 mmol) and 30% H2O2 (0.3 ml, 3.0 mmol) in AcOH (10 ml)
was stirred for 3 h at 80 ЊC. After an additional amount of
30% H2O2 (0.3 ml, 3.0 mmol) was added, stirring was continued
for 9 h. The mixture was concentrated in vacuo, diluted with
CH2Cl2, and washed with saturated aq. NaHCO3. After the
aqueous layer had been extracted again with CH2Cl2, the
organic layers were combined, and dried over Na2SO4. Removal
of the solvent gave crude 4b, which was subjected to column
chromatography on SiO2 with hexane–EtOAc–MeOH (6:3:1)
as a eluent. Recrystallization from MeOH afforded 4b (246 mg,
67%) as colourless needles (Found: C, 74.72; H, 5.78; N, 8.69.
Calc. for C10H9NO: C, 75.45; H, 5.70; N, 8.80%); mp 137–
139 ЊC (from MeOH) (lit.,9 137 ЊC).
8-Isopropylisoquinolin-1(2H)-one 5d. Yield 66%. Colourless
needles (Found: C, 77.14; H, 7.19; N, 7.46. C12H13NO requires
C, 76.98; H, 7.00; N, 7.48%); mp 112–113 ЊC (from EtOAc–
hexane); νmax(KBr)/cmϪ1 3173, 1653; δH 1.33 (6 H, d, J 6.8,
2 × CH3), 4.96 [1 H, m, CH(CH3)2], 6.47 (1 H, d, J 7.0, 4-H),
7.06 (1 H, d, J 7.0, 3-H), 7.36 (1 H, dd, J 7.8 and 1.0, 5-H), 7.47
(1 H, dd, J 7.8 and 1.0, 7-H), 7.57 (1 H, t, J 7.8, 6-H), 10.54 (1
H, br s, NH); m/z (EI) 187 (Mϩ).
8-Ethylisoquinoline N-oxide 4c. Yield 81%. Colourless plates
(Found: C, 69.61; H, 6.90; N, 7.41. C11H11NOؒ0.9H2O requires
C, 69.75; H, 6.81; N, 7.39%); mp 38–39 ЊC (from EtOAc);
8-tert-Butylisoquinolin-1(2H)-one 5e. Yield 69%. Colourless
columns (Found: C, 77.80; H, 7.60; N, 6.92. C13H15NO requires
C, 77.58; H, 7.52; N, 6.96%); mp 166–168 ЊC (from hexane);
J. Chem. Soc., Perkin Trans. 1, 1999, 3677–3683
3681

View Article Online
νmax(KBr)/cmϪ1 2952, 1636; δH 1.63 (9 H, s, 3 × CH3), 6.45 (1 H,
d, J 6.7, 4-H), 7.03 (1 H, dd, J 6.7 and 3.8, 3-H), 7.38 (1 H, dd,
J 7.8 and 1.3, 5-H), 7.52 (1 H, dd, J 7.8 and 7.9, 6-H), 7.62 (1 H,
dd, J 7.9 and 1.3, 7-H), 9.40 (1 H, s, NH); m/z (EI) 201 (Mϩ).
(2 H, dd, J 8.1 and 7.1, 6- and 6Ј-H), 7.76 (2 H, dd, J 8.1 and
1.4, 5- and 5Ј-H), 8.93 (2 H, d, J 5.5, 3- and 3Ј-H); m/z (EI) 340
(Mϩ).
Enantiomeric enrichment
General procedure for compounds 6b–d
Racemic biisoquinolines 1b–d were each partly resolved to their
enantiomers by HPLC equipped with a chiral stationary-phase
column (λ 280 nm). Ceramospher Chiral RU-1 (1.0 ml minϪ1)
was used for the preparative separation of 1b and 1d with
MeOH as eluent, and Chiralcel OD (0.5 ml minϪ1) was used for
1c with 10% EtOH–hexane. The retention times were 25 and 34
min for 1b, 18 and 24 min for 1c and 10 and 12 min for 1d. The
first eluted enantiomers in 1b and 1d were (R)-forms, while that
of 1c was the (S)-form. Enantiomeric excesses of the obtained
samples were analyzed by using Chiralcel OD (0.5 ml minϪ1,
λ 280 nm). The mobile phase for 1b, 1c and 1d was 30, 20 and
10% EtOH–hexane, respectively. Under these conditions, all the
first eluted enantiomers indicated positive specific rotation, and
the second ones negative. (R)-(ϩ)-1b (78% ee): [α]D23 ϩ60 (c 0.10
in CHCl3); CD (EtOH) [∆ε]318 ϩ11.4, [∆ε]235 Ϫ134 and [∆ε]219
ϩ184. (S)-(Ϫ)-1b (68% ee): [α]D24 Ϫ53 (c 0.10 in CHCl3); CD
(EtOH) [∆ε]318 Ϫ6.9, [∆ε]235 ϩ92.9 and [∆ε]220 Ϫ125. (R)-(ϩ)-1c
(86% ee): [α]D26 ϩ58 (c 0.25 in CHCl3); CD (EtOH) [∆ε]319 ϩ11.5,
[∆ε]235 Ϫ135 and [∆ε]220 ϩ190. (S)-(Ϫ)-1c (81% ee): [α]D26 Ϫ55
(c 0.51 in CHCl3); CD (EtOH) [∆ε]319 Ϫ9.85, [∆ε]235 ϩ141 and
[∆ε]220 Ϫ186. (R)-(ϩ)-1d (82% ee): [α]D24 ϩ67 (c 0.10 in CHCl3);
CD (EtOH) [∆ε]321 ϩ9.6, [∆ε]238 Ϫ72.1 and [∆ε]222 ϩ162. (S)-
(Ϫ)-1d (71% ee): [α]D25 Ϫ56 (c 0.10 in CHCl3); CD (EtOH) [∆ε]321
Ϫ10.4, [∆ε]238 ϩ92.6 and [∆ε]222 Ϫ188.
1-Chloro-8-methylisoquinoline 6b. A mixture of 5b (6.58 g, 43
mmol) and POCl3 (100 ml, 1.07 mmol) was refluxed for 3 h.
After the excess of POCl3 was evaporated in vacuo, saturated
aq. NaHCO3 and CH2Cl2 were added. The organic layer was
dried over Na2SO4 and evaporated under reduced pressure.
Column chromatography on SiO2 with CH2Cl2 and recrystal-
lization from MeOH gave 6b (5.68 g, 74%) as brown prisms
(Found: C, 67.72; H, 4.61; N, 8.00. Calc. for C10H8ClN: C,
67.62; H, 4.54; N, 7.89%); mp 144–146 ЊC (from MeOH) (lit.,9
78 ЊC); m/z (EI) 177 (Mϩ).
1-Chloro-8-ethylisoquinoline 6c. Yield 97%. Colourless oil
(Found: C, 69.04; H, 5.47; N, 7.25; Cl, 18.37. C11H10ClN
requires C, 68.94; H, 5.26; N, 7.31; Cl, 18.50%); νmax(Nujol)/
cmϪ1 1609, 1591, 1558; δH 1.36 (3 H, t, J 7.5, CH2CH3), 3.48
(2 H, q, J 7.5, CH2CH3), 7.46 (1 H, d, J 7.1, 7-H), 7.52 (1 H, d,
J 5.4, 4-H), 7.56 (1 H, dd, J 7.1 and 8.1, 6-H), 7.64 (1 H, d, J 8.1,
5-H), 8.18 (1 H, d, J 5.4, 3-H); m/z (EI) 191 (Mϩ) (Found: Mϩ,
191.0484. C11H10ClN requires M, 191.0501).
1-Chloro-8-isopropylisoquinoline 6d. Yield 81%. Colourless oil
(Found: C, 70.13; H, 6.00; N, 6.86; Cl, 17.32. C12H12ClN
requires C, 70.07; H, 5.88; N, 6.81; Cl, 17.24%); bp 111–112 ЊC
(2 mmHg); νmax(neat)/cmϪ1 1610, 1591, 1556; δH 1.39 (6 H, d,
J 6.0, 2 × CH3), 4.79 [1 H, m, CH(CH3)2], 7.54 (1 H, d, J 5.4,
4-H), 7.61–7.68 (3 H, m, 5-, 6- and 7-H), 8.19 (1 H, d, J 5.4,
3-H); m/z (EI) 205 (Mϩ).
Kinetic measurements
A solution of 1b–d in methanol (7.9–9.4 mol dmϪ3) was heated
or cooled at a given temperature in a thermostat-controlled
bath. Uncertainty of temperature was 0.1 ЊC. Initial enantio-
meric excesses were 92% ee for (S)-(Ϫ)-1b, 82% ee for (R)-(ϩ)-
1c and 52% ee for (R)-(ϩ)-1d. The changes in the concentration
of both enantiomers were followed at appropriate intervals by
using HPLC (Chiralcel OD, 0.5 ml minϪ1, λ 280 nm). The
mobile phase for 1b, 1c and 1d was 30, 20 and 10% EtOH–
hexane, respectively. First-order rate constants were obtained
by analyzing 10–20 concentration data for each sample, and the
correlation factors were >0.977. The rate constants were as
follows. 1b: 5.8 × 10Ϫ6 sϪ1 at 30 ЊC, 2.4 × 10Ϫ5 sϪ1 at 40 ЊC and
9.1 × 10Ϫ5 sϪ1 at 50 ЊC. 1c: 5.5 × 10Ϫ6 sϪ1 at 20 ЊC, 2.2 × 10Ϫ5 sϪ1
at 30 ЊC and 8.3 × 10Ϫ5 sϪ1 at 40 ЊC. 1d: 3.7 × 10Ϫ6 sϪ1 at 10 ЊC,
1.5 × 10Ϫ5 sϪ1 at 20 ЊC and 5.3 × 10Ϫ5 sϪ1 at 30 ЊC.
Preparation of 8,8Ј-dimethyl-1,1Ј-biisoquinoline 1b. Typical
procedure
Zinc powder (25 mg, 0.37 mmol) was added to a stirred solution
of NiCl2ؒ6H2O (90 mg, 0.37 mmol) and PPh3 (390 mg, 1.5
mmol) in DMF (6 ml) under argon atmosphere at 50 ЊC. After
stirring for 1 h, 6b (66 mg, 0.37 mmol) was added to the solu-
tion. After 5 h, the mixture was poured onto 28% aq. NH3 and
extracted with CH2Cl2. The organic layer was washed succes-
sively with water and brine, dried over Na2SO4, and evaporated.
Column chromatography on SiO2 with hexane–EtOAc–MeOH
(5:4:1) as eluent gave crude 1b, which was recrystallized from
MeOH to yield 1b (32 mg, 81%) as colourless prisms (Found: C,
84.65; H, 5.67; N, 9.76. Calc. for C20H16N2: C, 84.48; H, 5.67;
N, 9.85%); mp 210.0–213.5 ЊC (from MeOH) (lit.,9 207–210 ЊC);
m/z (EI) 284 (Mϩ).
Crystal-structure determination§
Crystal data for compound 1b. C20H16N2, M = 284.35, tri-
¯
gonal, space group R3 (no. 148), hexagonal cell constants
8,8Ј-Diethyl-1,1Ј-biisoquinoline 1c. Yield 53%. Colourless
needles (Found: C, 84.75; H, 6.67; N, 8.95. C22H20N2 requires C,
84.58; H, 6.45; N, 8.97%); mp 125–126 ЊC (from EtOAc–
hexane); λmax(EtOH)/nm 220, 319 and 330 (ε/dm3 molϪ1 cmϪ1
71800, 8930 and 10400); νmax(KBr)/cmϪ1 1613, 1594, 1557;
δH 0.90 (6 H, t, J 7.3, 2 × CH2CH3), 2.14 (2 H, m, CH2CH3),
2.27 (2 H, m, CH2CH3), 7.43 (2 H, d, J 7.1, 7- and 7Ј-H), 7.65
(2 H, dd, J 7.1 and 7.8, 6- and 6Ј-H), 7.74 (2 H, d, J 5.6, 4- and
4Ј-H), 7.79 (2 H, d, J 7.8, 5- and 5Ј-H), 8.52 (2 H, d, J 5.6,
3- and 3Ј-H); m/z (EI) 312 (Mϩ).
a = 31.695(2), c = 7.7985(7) Å, V = 6784.7(8) Å3, Z = 18,
Dx = 1.253 g cmϪ3, F(000) = 2700, µ = 0.074 mmϪ1. Specimen:
colourless hexagonal prisms, 0.43 × 0.50 × 0.54 mm, 2355
reflections measured (4.0 р 2θ р 51.5Њ), 2215 unique reflections
with |Fo| у 4σ|Fo|, h,k,l 0 → 38, 0 → 38, 9, R = 0.044,
Rw = 0.041. Residual extrema, 0.17 and Ϫ0.17 e ÅϪ3.
Crystal data for compound 1c. C22H20N2, orthorhombic,
space group Pbca (no. 61), cell constants a = 21.343(3), b =
20.316(2), c = 7.5398(7) Å, V = 3269.3(5) Å3, Z = 8, Dx = 1.269 g
cmϪ3, F(000) = 1328, µ = 0.075 mmϪ1. Specimen: colourless
prisms, 0.79 × 0.32 × 0.37 mm, 7392 reflections measured
(5.0 р 2θ р 55Њ, h,k,l), 3748 unique reflections with I > 3σ|I|,
h,k,l 26, Ϫ27 → 0, 0 → 9, R = 0.037, Rw = 0.047. Residual
extrema, 0.14 and Ϫ0.16 e ÅϪ3. Lattice constants and intensity
8,8Ј-Diisopropyl-1,1Ј-biisoquinoline 1d. Yield 35%. Colour-
less prisms (Found: C, 84.65; H, 7.20; N, 8.04. C24H24N2
requires C, 84.67; H, 7.10; N, 8.23%); mp 211–212 ЊC (from
EtOAc); λmax(EtOH)/nm 223, 320 and 331 (ε/dm3 molϪ1 cmϪ1
63700, 7850 and 8670); νmax(KBr)/cmϪ1 1606, 1555; δH 0.93
[6 H, d, J 6.8, 2 × CH(CH3)2], 1.05 [6 H, d, J 6.6, 2 ×
CH(CH3)2], 2.98 [2 H, m, 2 × CH(CH3)2], 7.61 (2 H, dd, J 7.1
and 1.4, 7- and 7Ј-H), 7.67 (2 H, d, J 5.5, 4- and 4Ј-H), 7.71
§ CCDC reference number 207/374. See http://www.rsc.org/suppdata/
p1/1999/3677 for crystallographic files in .cif format.
3682
J. Chem. Soc., Perkin Trans. 1, 1999, 3677–3683

View Article Online
8 (a) K. Hirao, R. Tsuchiya, Y. Yano and H. Tsue, Heterocycles, 1996,
data of 1b and 1c were measured on an Enraf-Nonius CAD-4
diffractometer with graphite-monochromated Mo-Kα radi-
ation (λ = 0.710 73 Å). The structure of 1b was solved by direct
methods using MULTAN7823 and refined by block-diagonal
least-squares based on |Fo|, and that of 1c was solved by direct
methods using SAPI9126 and refined by full-matrix least-
squares. All calculations on 1b and 1c were performed using
UNICS-III program system27 and TeXsan,28 respectively.
Hydrogen atoms were located from a difference Fourier syn-
thesis. Anisotropic temperature factors for all non-hydrogen
atoms and isotropic temperature factors for hydrogen atoms
were applied. ORTEP plots of 1b and 1c are shown in Fig. 2.
42, 415; (b) H. Tsue, H. Fujinami, T. Itakura, R. Tsuchiya,
K. Kobayashi, H. Takahashi and K. Hirao, Chem. Lett., 1999, 17.
9 G. Chelucci, M. A. Cabras, A. Saba and A. Sechi, Tetrahedron:
Asymmetry, 1996, 7, 1027.
10 H. Tsue, H. Fujinami, T. Itakura and K. Hirao, unpublished results.
11 J. B. Hendrickson and C. Rodrgíuez, J. Org. Chem., 1983, 48, 3344.
12 M. M. Robison and B. L. Robison, J. Org. Chem., 1956, 21, 1337.
13 M. Iyoda, H. Otsuka, K. Sato, N. Nisato and M. Oda, Bull. Chem.
Soc. Jpn., 1990, 63, 80.
14 A. J. Clarke, S. McNamara and O. Meth-Cohn, Tetrahedron Lett.,
1974, 2373.
15 L.-X. Dai, Z.-H. Zhou, Y.-Z. Zhang, C.-Z. Ni, Z.-M. Zhang and
Y.-F. Zhou, J. Chem. Soc., Chem. Commun., 1987, 1760.
16 N. Harada and K. Nakanishi, Acc. Chem. Res., 1972, 5, 257.
17 (a) S. F. Mason, R. H. Seal and D. R. Roberts, Tetrahedron, 1974,
30, 1671; (b) I. Hanazaki and H. Akimoto, J. Am. Chem. Soc., 1972,
94, 4102.
18 M. J. S. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Stewart,
J. Am. Chem. Soc., 1985, 107, 3902.
19 J. J. P. Stewart, J. Comput. Chem., 1989, 10, 209.
20 M. Nakamura, M. Oki, H. Nakanishi and O. Yamamoto, Bull.
Chem. Soc. Jpn., 1974, 47, 2415.
21 K. Fuji, M. Sakurai, N. Tohkai, A. Kuroda, T. Kawabata, Y.
Fukazawa, T. Kinoshita and T. Toda, Chem. Commun., 1996, 1609.
22 M. Kranz, T. Clark and P. von R. Schleyer, J. Org. Chem., 1993, 58,
3317.
23 A program for the Automatic Solution of Crystal Structures from
X-ray Diffraction Data, P. Main, S. E. Hull, L. Lessinger,
G. Germain, J.-P. Declercq and M. M. Woolfson, University of
York, England, and Louvain, Bergium, 1978.
Acknowledgements
This work was supported by Grant-in-Aids (No. 10877347 to
K. H. and No. 09740456 to H. T.) from the Ministry of
Education, Science, Sports and Culture, Japan.
References
1 E. L. Eliel, S. H. Wilen and L. N. Mander, Stereochemistry of
Organic Compounds, Wiley, New York, 1994, ch. 14.5.
2 (a) M. Oki, The Chemistry of Rotational Isomers, Springer-Verlag,
Berlin, 1993, ch. 3; (b) A. S. Cooke and M. M. Harris, J. Chem.
Soc., 1963, 2365; (c) Y. Badar, A. S. Cooke and M. M. Harris,
J. Chem. Soc., 1965, 1412; (d) A. S. Cooke and M. M. Harris,
J. Chem. Soc. C, 1967, 988.
3 (a) R. B. Kress, E. N. Duesler, M. C. Etter, I. C. Paul and D. Y.
Curtin, J. Am. Chem. Soc., 1980, 102, 7709; (b) A. K. Colter and
L. M. Clemens, J. Phys. Chem., 1964, 68, 651.
4 K. Tani, T. Yamagata, S. Otsuka, S. Akutagawa, H. Kumobayashi,
T. Taketomi, H. Takaya, A. Miyashita and R. Noyori, in
Asymmetric Reactions and Processes in Chemistry, ACS Symposium
Series 185, ed. E. L. Eliel and S. Otsuka, American Chemical
Society, Washington, D.C., 1982, p. 187.
5 T. Hayashi, K. Tomioka and O. Yonemitsu, Asymmetric Synthesis,
Kodansha, Tokyo, 1998.
24 M. T. Ashby, G. N. Govindan and A. K. Grafton, J. Am. Chem.
Soc., 1994, 116, 4801.
25 G. L. Jones, J. Chem. Soc., 1960, 1918.
26 Structure Analysis Program with Intelligent Control, F. Hai-Fu,
Rigaku Corporation, Tokyo, Japan, 1991.
27 T. Sakurai and K. Kobayashi, Rikagaku Kenkyusho Hokoku, 1979,
55, 69.
28 TeXsan: Crystal Structure Analysis Package, Molecular Structure
Corporation, The Woodlands, Texas, USA, 1985 and 1992.
6 M. Crawford and I. F. B. Smyth, J. Chem. Soc., 1954, 3464.
7 M. Fujii and A. Honda, (a) J. Heterocycl. Chem., 1992, 29, 931;
(b) Chem. Express, 1992, 7, 329.
Paper 9/05161I
J. Chem. Soc., Perkin Trans. 1, 1999, 3677–3683
3683