Dehydrogenation of 3-unsubstituted indolizines on platinum on carbon. A facile synthesis of biindolizines

Article abstract of DOI:10.1055/s-1998-2202

Biindolizines were synthesised by oxidative dimerisation on platinum on carbon without byproducts in almost quantitative yields.

Full text of DOI:10.1055/s-1998-2202

SYNTHESIS
Short Papers
1596
Dehydrogenation of 3-Unsubstituted Indolizines on Platinum on Carbon. A Facile
Synthesis of Biindolizines
Helmut Sonnenschein,*1 Hendrik Kosslick, Franz Tittelbach
Institut für Angewandte Chemie Berlin-Adlershof, Rudower Chaussee 5, D-12484 Berlin, Germany
Received 25 February 1998; revised 6 April 1998
Dedicated to Professor Ernst Schmitz on the occasion of his 70th birthday
Abstract. Biindolizines were synthesised by oxidative dimerisa-
tion on platinum on carbon without byproducts in almost quantita-
tive yields.
Key words: indolizines, dimerisation, dehydrogenation on plati-
num
3,3'-Biindolizines are reversible two-step redox systems
of theoretical and practical interest.2, 3 These connected
heteroaromatics as axially chiral compounds4 are useful
tools e.g. in syntheses of cyclophanes5 or as complex
ligands.6 Several methods for oxidative dimerisation of
indolizines unsubstituted at the 3-position have been de-
scribed in the literature. However, all the known methods
have some disadvantages: the coupling of indolizines with
potassium ferricyanide is accompanied by a byproduct.2
The dimerisation of indolizines with Fe3+ yielded biin-
dolizines only in some cases.7 For example the oxidation
Scheme 1
of the 2-phenylindolizine (1a) unsubstituted in 1-position
did not take place.
chlorobenzene on Pd/C as well as on Pt/C a byproduct 3a
was isolated. Since the new compound 3a shows the same
molecular ion in MS as 2a, we expected an unusual dimer-
isation involving both the 1- and 3-positions of indolizine.
The coupling by dehydrogenation on palladium on car-
bon (Pd/C) under refluxing conditions also allowed the
3,3-dimerisation of indolizines with higher oxidation po-
tentials such as indolizines unsubstituted at the 1-posi-
tion.8 For this reason we usually preferred this reaction.
But we often obtained, in accordance with literature,
yields lower than 40 % by use of p-xylene as solvent.
Therefore, we were interested in a more detailed investi-
gation of the oxidative coupling reaction depending on
the catalyst and/or the presence of oxygen during the re-
action.
Our investigations with Pd/C showed a significant in-
crease in yields by the use of higher amounts of catalyst or
by bubbling air through the heated mixture.9 The applica-
Scheme 2
tion of platinum (Pt/C, 10% Pt) instead of palladium on The nonsymmetric structure of a 1,3'-dimer was proven
carbon raises the yield of 2,2'-diphenyl-3,3'-biindolizine by 1H NMR. The spectrum shows two different singlets of
(2a). The dimer was obtained in 70% yield on bubbling air the isolated protons in the five-membered fused ring sys-
through the refluxing mixture in chlorobenzene. In con- tems. Only one of them is located near to the signal of the
trast to the reaction in the presence of Pd/C only half the hydrogen atom in the 1-position of the symmetric product
amount (0.1 instead of 0.2 equiv.) of catalyst is necessary 2a at δ = 6.91. The second singlet at δ = 8.18 we assigned
to obtain a comparable result.
to the proton at the 3'-carbon adjacent to the nitrogen. Ad-
ditional assignments of the 1H-signals were carried out by
H,H COSY, HSQC, and HMQC experiments.
The dimerisation of 2-phenylindolizine under the drastic
conditions of 132°C (boiling chlorobenzene) requires a
purification of the desired dimer by column chromatogra- The participation of the less reactive 1-position in the re-
phy. The time course of the experiments shows high con- action indicates drastic reaction conditions. Surprisingly,
version rates only during the starting phase. In all men- the dimerisation on platinum on carbon at room tempera-
tioned cases drastic decreases of dimerisation rates are ob- ture (Figure, – –) took place as rapidly as in the starting
A
served after about 25% conversion. As one reason of this phase in refluxing chlorobenzene (Figure, – –) and led
a
limited increase of yields we assumed side reaction. In to much higher yields of the dimer than under reflux con-
fact by reaction of 2-phenylindolizine (1a) in boiling ditions.

SYNTHESIS
November 1998
1597
the Figure (– –) only the first 25 % of monoindolizine
c
1c reacts quickly in the starting phase of dimerisation.
Under anaerobic conditions a conversion of 50% was
reached only after about 23 hours instead of the few min-
utes needed in the open system. Obviously, without oxy-
gen inactivation of the catalyst is already observed after a
short time. In the following the dehydrogenation by the
platinum takes place at slow rates. In the presence of ox-
ygen the potential of oxidation of the catalyst is regener-
ated to a certain extent. Higher temperatures favour side
reactions and give rise to an earlier inactivation of the cat-
alyst in contrast to the reaction at room temperature.
Figure. Dimerisation of 2-phenylindolizine (1a) and 1-methyl-2-phe-
nylindolizine (1c) with Pt/C in chlorobenzene: left: (– –): 1a re-
Unless otherwise noted, materials were obtained from commercial
suppliers and were used without further purification. Indolizines 1a–
e were prepared according to literature.10 Pd, 10% on carbon (Ald-
rich) and platinum, 10% on carbon (Merck) were used for the exper-
iments. For column chromatography, silica gel Merck KG 60 was
used. Mps are uncorrected. NMR spectra were recorded on a Bruker
WP 200 SY spectrometer. Chemical shifts are expressed in ppm
downfield from internal TMS. Mass spectra were recorded with a
Fisons Instruments VG Auto Spec.
a
fluxing; (– –): 1a stirring at r.t.; right: (– –): 1c stirring in open
A
C
vessel; (– –): 1c stirring under argon.
c
Under the reaction conditions no byproduct could be de-
tected as shown by HPLC. To confirm the generality of
this reaction, we checked the conditions of dimerisation of
some indolizines with Pt/C. The results are summarised in
the Table. Obviously a relationship exists between the
oxidation potential of the indolizines and the rate of
dimerisation. With the indolizines tested, those with the
lowest oxidation potential 1c and 1d needed shorter reac-
tion times. The electron-poor compound 2-phenylindoliz-
ine (1a) needs the longest reaction time. 1-Ethoxycarbon-
yl-2-phenylindolizine (1e) did not react under the de-
scribed conditions (see Table).
Dimerisation; General Procedure:
The reactions were carried out in a open vessel. 1 mmol of the desired
indolizine in chlorobenzene (10 mL) was stirred at r.t. with Pt/C
(10%, 195 mg, 0.1 mmol). The reaction time is given in the Table.
The catalyst was filtered off and the solvent was evaporated under
reduced pressure, giving the pure biindolizine. The yields are sum-
marised in the Table.
HPLC check of reactions: The reactions were carried out as described
above at r.t. or with heating, in presence of air or by exclusion of
oxygen. Loss of solvent by evaporation during the reaction time was
replenished by weight. Samples of 0.05 mL were taken at intervals,
diluted with MeCN(10 mL) and measured using a LC10-HPLC-sys-
tem (Shimadzu). The chlorobenzene of the 0.05 mL samples was used
as internal standard.
The corresponding products 2a–d were obtained from the
pale yellow solution by evaporation without separation
steps. The slightly varying yields (between 91 and 99%)
may be due to different adsorption on the coal-supported
catalyst. Lower amounts of the catalyst led to smaller
yields. The more reactive 1,2-dimethylindolizine (1d) is
less sensitive to reduced amounts of Pt/C than the mono-
substituted 2-phenylindolizine (1a).
2,2'-Dimethyl-3,3'-biindolizine (2b):
According to the general procedure by stirring of a mixture of 2-
methylindolizine (1b) (131 mg, 1 mmol) and Pt/C (10%, 195 mg,
0.1 mmol) in chlorobenzene (10 mL).
To prove the influence of oxygen on the dimerisation
of the reactive 1-methyl-2-phenylindolizine (1c) the
dimerisation was carried out after careful flushing of the
flask, the catalyst and the solvents by argon. This con-
sequent exclusion of air during the reaction resulted in
a drastic decrease in the rate of reaction. As shown in
m/z (EI, high resolution) = 260.13126, C18H16N2.
13C NMR (DMSO-d6): δ = 12.0 (CH3), 100.2, 110.1, 111.4, 117.5,
118.4, 122.7, 125.9, 133.0.
1H NMR (DMSO-d6): δ = 2.07 (s, 6H, CH3), 6.43Ð6.48 (m, 2H) 6.48
(s, 2H, H-1/1'), 6.71Ð6.77 (m, 2H), 7.16 (dd, J= 0.9, 7.0 Hz, 2H), 7.44
(td, J= 1.1, 8.4 Hz, 2H).
Table: Synthesis of Biindolizines in Chlorobenzene–Pt/C at Room Temperature
monomer dimer
isolated
yield/time
[%]/[h]
mp
[˚C]
mp (Lit)
[˚C]
oxidation
potential
[mV]a
time for
50%
conversion
1a
1b
1c
1d
1e
2,2'-diphenyl-3,3'-biindolizine (2a)
2,2'-dimethyl-3,3'-biindolizine (2b)
1,1'-dimethyl-2,2'-diphenyl-3,3'-biindolizine (2c)
1,1',2,2'-tetramethyl-3,3'-biindolizine (2d)
94/120
91/25
99/20
93/20
0
166–168
76–78
242–244
146–148
–
169–1717
0.908
0.813
0.777
0.640
1.198
12 h
–
2.5 h
25 min
20 min
∞
244–2467
144–1482
–
a oxidation potential of the compounds 1a–e, measured by cyclic voltammetry

SYNTHESIS
Short Papers
1598
2,2'-Diphenyl-3,1'-biindolizine (3a):
(1) New address: H. Sonnenschein, Institut für Nichtklassische
Chemie an der Universität Leipzig, Permoserstr. 15, D-04303
Leipzig.
(2) Hünig, S.; Linhart, F. Liebigs Ann. Chem. 1976, 317.
(3) Cardellini, L.; Carloni, P.; Greci, L.; Tosi, G.; Andruzzi, R.;
Marrosu, G.; Trazzo, A. J. Chem. Soc., Perkin Trans. 2 1990,
2117.
(4) Leitner, M. B.; Kreher, T.; Sonnenschein, H.; Costisella, B.;
Springer, J. J. Chem. Soc., Perkin Trans. 2 1997, 377.
(5) Sonnenschein, H.; Kreher, T.; Gründemann, E.; Krüger, R. P.;
Kunath, A.; Zabel, V. J. Org. Chem. 1996, 61, 710;
Kreher, T.; Sonnenschein, H.; Costisella, B.; Schneider, M.
J. Chem. Soc., Perkin Trans. 1 1997, 3451.
(6) Köckritz, A.; Sonnenschein, H.; Bischoff, S.; Theil, F.; Gloede,
J. Phosphorus, Sulfur Silicon, in press.
(7) Andruzzi, R.; Cardellini, L.; Greci, L.; Stipa, P.; Poloni, M.;
Trazza, A. J. Chem. Soc., Perkin Trans. 1 1988, 3067.
(8) Kakehi, A.; Ito, S.; Hamaguchi, A.; Okano, T. Bull. Chem. Soc.
Jpn. 1981, 54, 2833.
The compound was obtained as byproduct of 2a by boiling a solution
of 2-phenylindolizine (1a) (191 mg, 1 mmol) in chlorobenzene
(10 mL) in presence of Pd/C (10%, 212 mg, 0.2 mmol) or Pt/C (10%,
195 mg, 0.1 mmol) for 20 h. Insoluble substances were removed by
filtration and the filtrate was concentrated at reduced pressure. The
residue was separated by column chromatography on silica gel (hep-
tane/toluene 10:3). 3a was isolated as an oil in yields of about 5%.
m/z (EI, high resolution) = 384.16320, C28H20N2.
NMR assignments by H,H COSY, HSQC, and HMQC experiments
and according to literature.11
13C NMR (DMSO-d6): δ = 98.2 (C-1), 99.1 (C-2'), 110.3 (C-6), 111.2
(C-6'), 112.2 (C-3'), 113.8 (C-2), 116.8 (C-7'), 117.4 (C-7), 118.6 (C-
8'), 118.9 (C-8), 122.5 (C-5), 126.5 (C-5'), 125.3, 125.9, 126.1, 126.8,
127.9, 128.3, 128.6, 128.7, 134.6, 135.9 (2 × phenyl, C-3, C-1');
132.1, 132.4 (8a, 8a').
1H NMR (DMSO-d6): δ = 6.33 (dt, J= 1.0, 6.8 Hz, 1H, H-6), 6.61Ð
6.65 (m, 2H, H-6', 8'), 6.66Ð6.71 (m, 2H, H-7, 7' ), 6.91 (s, 1H, H-1),
7.05Ð7.27 (m, 8H, phenyl), 7.31 (dd, J= 1.0, 6.8 Hz, 1H, H-5), 7.36Ð
7.40 (m, 2H, phenyl), 7.47 (d, J= 9.3 Hz, 1H, H-8), 8.18 (s, 1H, H-
3'), 8.34Ð8.38 (m, 1H, H-5').
(9) Kreher, T.; Sonnenschein, H.; Schmidt, L.; Hünig, S. Liebigs
Ann. Chem. 1994, 1173.
Sonnenschein, H. unpublished results.
(10) For review on the synthesis of indolizines see: Uchida, T.;
Matsumoto, K. Synthesis 1976, 209.
(11) Bode, M. L.; Kaye, T, J. Chem. Soc., Perkin Trans 1 1993,
1809.
This work was supported by the Deutsche Forschungsgemeinschaft
(KO 1639/2-1).