Investigations of a novel process to the framework of benzo[c]cinnoline

Article abstract of DOI:10.1021/jo049102o

A novel synthetic process leading to the framework of benzo[c]cinnoline has been discovered and investigated. The process is composed of two separate reactions, the first of which is a partial reduction of the nitro groups of the 2,2′-dinitrobiphenyl, a process that we believe proceeds via a SET mechanism to yield the hydroxyamino and nitroso groups. In the following step the cyclization takes place under formation of the -N=N- bond. We believe that this process take place via a radical mechanism through the nitroso radical anion. The novel process affords either benzo[c]cinnoline or benzo[c]cinnoline N-oxide, both in high yields, 93% and 91%, respectively. To obtain benzo[c]cinnoline, the reaction is conducted with an alcohol as solvent and an alkoxide as the base, while for benzo[c]cinnoline N-oxide, water is used as solvent with sodium hydroxide as the base. To establish the latter procedure, statistical experimental design and multivariate modeling were utilized to reveal the response surface for the reaction and to determine the optimal conditions for the reaction. A proposal for the complex reaction mechanism is given. During the corroboration of the mechanism, a new deoxygenation reaction for converting benzo[c]cinnoline N-oxide into benzo[c]cinnoline was discovered. The reaction is conducted by treating the N-oxide with sodium ethoxide at elevated temperature to achieve near-quantitative conversion into benzo[c]cinnoline in a yield of 96%.

Full text of DOI:10.1021/jo049102o

In vestiga tion s of a Novel P r ocess to th e F r a m ew or k of
Ben zo[c]cin n olin e
Hans-Rene´ Bjørsvik,* Raquel Rodr´ıguez Gonza´lez, and Lucia Liguori
Department of Chemistry, University of Bergen, Alle´gaten 41, N-5007 Bergen, Norway
hans.bjorsvik@kj.uib.no
Received May 28, 2004
A novel synthetic process leading to the framework of benzo[c]cinnoline has been discovered and
investigated. The process is composed of two separate reactions, the first of which is a partial
reduction of the nitro groups of the 2,2′-dinitrobiphenyl, a process that we believe proceeds via a
SET mechanism to yield the hydroxyamino and nitroso groups. In the following step the cyclization
takes place under formation of the -NdN- bond. We believe that this process take place via a
radical mechanism through the nitroso radical anion. The novel process affords either benzo[c]-
cinnoline or benzo[c]cinnoline N-oxide, both in high yields, 93% and 91%, respectively. To obtain
benzo[c]cinnoline, the reaction is conducted with an alcohol as solvent and an alkoxide as the base,
while for benzo[c]cinnoline N-oxide, water is used as solvent with sodium hydroxide as the base.
To establish the latter procedure, statistical experimental design and multivariate modeling were
utilized to reveal the response surface for the reaction and to determine the optimal conditions for
the reaction. A proposal for the complex reaction mechanism is given. During the corroboration of
the mechanism, a new deoxygenation reaction for converting benzo[c]cinnoline N-oxide into benzo-
[c]cinnoline was discovered. The reaction is conducted by treating the N-oxide with sodium ethoxide
at elevated temperature to achieve near-quantitative conversion into benzo[c]cinnoline in a yield
of 96%.
SCHEME 1
In tr od u ction
We have recently reported a new catalytic oxidation
process that makes use of nitroarenes as catalysts in
aerobic oxidation.1,2 This catalytic process was derived
on the basis of a process where we used the nitroarenes
as stoichiometric oxidants.3 These two processes are both
suitable for the oxidation of aromatic compounds contain-
ing oxygen-functionalized benzylic carbons. Investigation
of the process revealed the involvement of single-electron-
transfer (SET) processes. Moreover, during this study a
reaction manifold was found that we initially believed
to be appropriate for the synthesis of diazenes but found
in practice that which it could not be used in this way, it
was suitable for the synthesis of N-heteroarenes contain-
ing the C-NdN-C moiety. The present report will reveal
our finding utilizing the reaction manifold for a process
leading to the framework of benzo[c]cinnoline which
provides an interesting class of compounds due to a
variety of biological activities, for example, topoisomerase
I inhibitors4 and compounds with fungicidal effects.5
Meth od s a n d Resu lts
Several years ago, Russell and co-workers6-8 demon-
strated by means of ESR spectroscopy that nitrosoben-
zene 1 and hydroxyaminobenzene 2 mixed in basic
medium react spontaneously to afford two entities of
nitrosobenzene anion radicals 3, pathway (a) of Scheme
1. When the reaction was performed with oxygen present,
the nitrobenzene radical anion 5 was observed,8 as in
pathway (b). On the other hand, if the base treatment of
1 and 2 was performed without oxygen present, N,N′-
* To whom correspondence should be addressed. Phone: +47 55 58
34 52. Fax: +47 55 58 34 52.
(1) Bjørsvik, H.-R.; Liguori, L.; Merinero, J . A. V. Tetrahedron Lett.
2002, 43, 4985-4987.
(2) Bjørsvik, H.-R.; Liguori, L.; Merinero, J . A. V. J . Org. Chem.
2002, 67, 7493-7500.
(5) Weaver, L. J . (Monsanto Chemical Co.). US Patent 3012909,
1959.
(6) Russell, G. A.; Geels, E. J . J . Am. Chem. Soc. 1965, 87, 122.
(7) Geels, E. J .; Konaka, R.; Russell, G. A. Chem. Commun. 1965,
13.
(8) Russell, G. A.; Geels, E. J .; Smentowski, F. J .; Chang, K.-Y.;
Reynolds, J .; Kaupp, G. J . Am. Chem. Soc. 1967, 89, 3821.
(3) Bjørsvik, H. R.; Liguori, L.; Minisci, F. Org. Process. Devel. Res.
2001, 5, 136
(4) Singh, S. K.; Yu, Y.; Ruchelman, A.; Sim, S.-P.; Liu, A.; Leroy,
F.; LaVoie, E. J . Abstract of Papers; 222nd ACS National Meeting,
Chicago, August 26-30, 2001; American Chemical Society: Washing-
ton, DC, 2001.
10.1021/jo049102o CCC: $27.50 © 2004 American Chemical Society
Published on Web 10/07/2004
7720
J . Org. Chem. 2004, 69, 7720-7727

Process to the Framework of Benzo[c]cinnoline
SCHEME 2
diphenyldiazene N-oxide 4 was obtained via a coupling
of the nitroso radical anion 3, pathway (c). We have from
our laboratory recently disclosed a new catalytic aerobic
oxidation1,2 process based on these early discoveries and
our more recently disclosed stoichiometric alkaline ni-
troarene oxidation method.3 Our recently unveiled aero-
bic oxidation process used highly electron-deficient ni-
troarenes as catalysts in our endeavors to oxidize various
aromatics with oxygen-functionalized benzylic carbon to
obtain the corresponding carboxylic acids. Here, we will
present our findings using oxygen free reaction environ-
ments, conditions that should favor the creation of the
-NdN- bond, path (c) of Scheme 1, which also consti-
tutes the complementary mechanism for our stoichio-
metric oxidation process.3
The C-NdN-C bond formation step is carried out by
reacting the 2,2′-dinitrobiphenyl 6 with acetophenone 7
in basic solution. Under such reaction conditions, 2,2′-
dinitrobiphenyl 6 operates as the oxidant for the ac-
etopheneone in a way similar to that as recently de-
scribed by us.2 Scheme 2 shows a proposal for this process
creation of the C-NdN-C moiety in the current context.
The nitrogroups can be reduced stepwise to the amino
group, a process that takes place over three steps, -NO2
f -NdO f -NHOH f -NH2. Under the reductive
conditions for the nitro groups of 2,2′-dinitrobiphenyl 6,
the partially reduced product N-(2′-nitrosobiphenyl-2-yl)-
hydroxylamine 8 will be formed during the course of the
redox process. The nitroso and hydroxylamine groups of
this intermediate react spontaneously8 to form two enti-
ties of nitroso radical anion (similar to pathway (a) of
Scheme 1). However, in the present case, the process
proceeds via an intramolecular SET mechanism to pro-
vide 2,2′-dinitrosoyl radical anion-biphenyl 9, pathway
(b) of Scheme 2. Under suitable reaction conditions, the
2,2′-dinitrosoyl radical anion-biphenyl 9 cyclizes to afford
benzo[c]cinnoline 10 or the less reduced compound benzo-
[c]cinnoline N-oxide 11, pathway (c) f (d) of Scheme 2:
this represents an complementary pathway to our aerobic
oxidation process.1,2
In our endeavor to accomplish the synthetic process
as outlined in Scheme 2, investigations were performed
along two different lines, namely (i) by using conditions
such as revealed for the stoichiometric alkaline ni-
troarene oxidation3 that is performed in aqueous sodium
hydroxide (40%) and (ii) by using an organic solvent with
an alkoxide as base. We believed that both of these redox
processes involve a “cage re-bonding” step. Such reaction
mechanisms are usually very dependent on the solvent
properties, especially the solvent viscosity, and thereby
the reaction temperature. A variety of solvents and bases
were thus investigated. The results of this screening are
portrayed in Scheme 3. It is clearly evident from these
data that the success of the process depends on both the
type of reaction medium and the type of base that is
utilized. Interestingly, some of the solvent and base
combinations afford both the benzo[c]cinnoline and the
benzo[c]cinnoline N-oxide (entries 1, 3, 4, 6, and 7). When
the solvent/base combinations EtOH/EtONa (entry 5),
i-PrOH/i-PrONa (entry 8), or PrOH/PrONa (entry 9) are
used, benzo[c]cinnoline 10 is obtained with excellent yield
to the exclusion of the N-oxide.
The proposed reaction mechanism given in Scheme 2
was assembled as a complementary pathway based on
previously disclosed studies.3-8 The mechanism describes
the reduction and cyclization process leading to benzo-
[c]cinnoline N-oxide 11. However, the last reduction step,
namely the conversion of benzo[c]cinnoline N-oxide 11
into benzo[c]cinnoline 10, could not be accounted for on
the basis of previous experiments. In an attempt to
J . Org. Chem, Vol. 69, No. 22, 2004 7721

Bjørsvik et al.
TABLE 1. Exp er im en ta l Design w ith Ad ja cen t
Mea su r ed Resp on ses
process to benzo[c]cinnoline 10 as described above,
Scheme 4. In essence, a quantitative conversion of the
N-oxide 11 to the deoxygenated compound 10 was
achieved. Additional experiments at three different reac-
tion times, namely at 30, 90, and 140 min, were per-
formed. The sample contained initially approximately
87% of 11 and 13% of 10. The final time-concentration
profiles for the two compounds 10 and 11 are shown on
the right-hand side of Scheme 4. This plot reveals that
the conversion is completed after a reaction time of 140
min. A vertical parallel displacement of the conversion
curve (that thus initiates in 100%) reveals that an
additional ∼40 min is required for conversion of a pure
sample of benzo[c]cinnoline N-oxide; see the dashed curve
in Scheme 4. An additional experiment performed using
sodium hydroxide as base and ethanol as solvent revealed
a complete conversion of with excellent yield (>96%).
The experiment reported in entry 7 of Scheme 3 using
sodium hydroxide as base and water as the reaction
medium attracted our attention as a potential method
for the synthesis of the benzo[c]cinnoline N-oxide 11. On
the basis of this experiment, an experimental design
suitable for response surface modeling was laid out; see
entries 1-17 of Table 1. Adjacent to the experimental
conditions (x1, ..., x4) are shown the responses: % non-
converted 2,2′-dinitrobiphenyl (y6), percent yields of ben-
zo[c]cinnoline (y10), and percent yields of benzo[c]cinnoline
N-oxide (y11), respectively.
experimental variablesa
responsesc
b
no.
x1
x2
x3
x4
y6
y10
y11
1
2
1.8
2.2
1.8
2.2
1.8
2.2
1.8
2.2
2.0
1.6
2.4
2.0
2.0
2.0
2.0
2.0
2.0
2.4
3.0
2.4
3.4
15.0
15.0
25.0
25.0
15.0
15.0
25.0
25.0
20.0
20.0
20.0
10.0
30.0
20.0
20.0
20.0
20.0
25.0
25.0
30.0
30.0
140
140
140
140
160
160
160
160
150
150
150
150
150
130
170
150
150
157
157
167
167
3.0
5.0
5.0
3.0
5.0
3.0
3.0
5.0
4.0
4.0
4.0
4.0
4.0
4.0
4.0
2.0
6.0
4.5
4.5
2.0
2.0
87.64
37.01
50.32
57.94
42.70
22.13
55.00
14.74
49.74
55.04
20.04
38.72
34.12
94.22
25.12
43.17
38.39
11.00
2.71
1.68
5.45
6.25
11.01
4.67
4.25
5.07
5.02
4.75
4.11
5.35
4.82
4.03
0.00
4.95
7.21
3.74
5.01
9.49
5.41
9.49
3.45
57.92
45.75
35.28
42.28
56.09
45.17
70.85
48.07
36.80
61.95
54.19
59.19
0.00
57.10
44.80
50.03
78.87
90.71
71.60
86.47
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
34.91
0.00
a
Procedure: 2,2’-dinitrobiphenyl 6 (2.06 mmol, 0.503 g) was
added to a solution of NaOH and acetophenone 7. Experimental
variables: xk (definition), levels [-2, -1, 0, +1, +2]/unit; x1,
(amount of acetophenone) [1.6, 1.8, 2.0, 2.2, 2.4]/mmol; x2, volume
of NaOH (40%) [10, 15, 20, 25, 30]/mL; x3, reaction temperature
[130, 140, 150, 160, 170]/°C; x4, reaction time [2, 3, 4, 5, 6]/h.
b
Generator for the 24-1 design: x4 ) x1 × x2 × x3 (used to decide
the settings for experimental variable x4 based on the settings of
the three other experimental variables x1-x3). c The measured
responses are as follows: y6, 2,2′-dinitrobenzene/recovered (%); y10
The experimental variables of the design setup viewed
in Table 1 were scaled according to eq 1 in order to
facilitate the estimation of the regression coefficients, the
â’s, of eq 2.
,
benzo[c]cinnoline/yield (%); y11, benzo[c]cinnoline N-oxide/yield (%).
The % values are all calculated on the basis of the initial quantity
of 2,2′-dinitrobiphenyl.
SCHEME 3
1
2
zi - zi,L
+
× (zi,H - zi,L
)
{
}
}
xi )
, i ) 1, ..., 4 (1)
1
2
zi,H - zi,L
+
× (zi,H - zi,L
)
{
xi of eq 1 is the experimental variable i given in scaled
units, zi is the experimental variable i given in real units,
and zi,L and zi,H are the selected low (-1) and high (+1)
experimental values (real unit), respectively, of the
experimental variable i. The model matrix was created
by including the design matrix after a column of ones,
the two-variable interactions (xi × xj), and the quadratic
terms (xi2) of each of the four experimental variables xi,
i ) 1, ..., 4. The final model matrix in scaled values (17
lines × 15 columns) with their corresponding yield values
y11 of target molecule benzo[c]cinnoline N-oxide was then
submitted to multivariate modeling in terms of the
multiple linear regression method (MLR)9 and the partial
least squares regression (PLSR) method10 to provide an
empirical model as depicted in eq 2.
I
I
I
I
yˆ11 ) ∠+ ∠x +
∠x x + ∠x 2, i ) 1, ..., 4
∑
∑∑
∑
0
i
i
ij
i
j
i i
i)1
i<j
i)1
(2)
establish if an additional final step 11 f 10, path (e),
operated during the influence of the alkoxides, a sample
of benzo[c]cinnoline N-oxide 11 was treated with sodium
ethoxide in ethanol under similar conditions as for the
(9) See, for example: (a) Draper, N. R.; Smith, H. Applied Regression
Analysis, 3rd ed.; Wiley: New York, 1998. (b) Montgomery, D. C.; Peck,
E. A. Introduction to linear regression analysis; Wiley: New York, 1982.
(10) Malinowski, E. R. Factor analysis in chemistry, 3rd ed.; Wiley:
New York, 2002.
7722 J . Org. Chem., Vol. 69, No. 22, 2004

Process to the Framework of Benzo[c]cinnoline
SCHEME 4
The final empirical mathematical predictive model for
the response yield (%) of benzo[c]cinnoline N-oxide 11
(y11) is shown in eq 3.
lower right corner-high yield at short reaction time and
elevated temperature, the experimental settings of entry
20 of Table 1 was selected. The model predicts a yield of
approximately 90%.11 However, when the experiment was
carried out, a yield of approximately 72% was achieved.12
When the quantity of acetophenone is augmented to 3.4
mmol, keeping the other experimental variables at the
same levels as in experiment 20, a yield of approximately
86% is achieved. Again the tendencies depicted in the
contour map were successfully applied to predict a high
yielding synthetic procedure, even if the required settings
represent long-range extrapolation.
At t em p t s To Syn t h esize Dip h en yld ia zen es. The
two synthetic methods for the preparation of benzo[c]-
cinnoline 10 and benzo[c]cinnoline N-oxide 11, respec-
tively, were used in attempts to prepare diphenyldiazene
derivatives 12 and 13 of Scheme 5. The experiments were
yˆ11 ) 47.9635 + 8.3630x1 + 2.9557x2 + 11.6359x3 +
2
2
5.4544x4 - 6.6337x3 × x4 + 2.0657x2 - 4.9691x3
(3)
This model was developed with the PLSR method using
a ) 4 PLS components. The product statistics indicates
a fairly good model; R112 ) 0.909, Q112 ) 0.627, RMSEP11
) 5.489, and RSD11) 5.844.
The model of eq 3 was used to estimate response
surfaces projected as iso-contour maps, shown in Figure
1. A straightforward interpretation of this response
surface reveals that it is compulsory to apply a rather
high reaction temperature, concomitant with large quan-
tities of sodium hydroxide (large volume) and acetophe-
none, and a medium-long reaction time (4.5-5 h). Such
conditions are found in the sub iso-contour map located
in the upper right corner of Figure 1. However, during
the investigation of the reaction leading to benzo[c]-
cinnoline N-oxide 11, it was observed that high reaction
temperatures (>170 °C) were detrimental to the process,
most probably due to concurrent or subsequent degrada-
tion reactions. Reaction temperatures below <140 °C
resulted in very low conversion and yield. By means of
the iso-contour projections of the response surface (Figure
1) and these additional constraints, the experimental
conditions for some few optimizing experiments with the
goal of evaluating the predictive capacity of the developed
model and finally to determine the optimal conditions for
the process were selected and carried out.
Entries 18 and 19 of Table 1 show successfully how
the response surface iso-contour projections can operate
as a tool for predictive purposes. If the response surface
iso-contour projection is interpreted carefully, one can see
that by moving horizontally to the right on the outer
abscissa axis (variable x1) and keeping the other variables
constant, an increased yield is predicted. This observation
was used to decide the experimental conditions for the
trial reported in entry 19, which only represents a small
variation of trial 18, augmenting the quantity of ac-
etophenone from 2.4 to 3.0 mmol. These settings however,
embody a subresponse surface iso-contour projection that
is not included in Figure 1. The anticipated improvement
was realized when the experiment was carried out; see
entry 19. Similarly, an improvement was also achieved
when a reaction was conducted at a shorter reaction time
concomitant with an elevated reaction temperature.
Motivated by the predictions of the subcontour maps
SCHEME 5
carried out at two different temperatures, namely at 75
and 157 °C, and conducted for 2 and 5 h with 1,3-
dinitrobenzene 15 as the substrate. Furthermore, two
solvent and base systems, (i) water and sodium hydrox-
ide, and (ii) i-PrOH and i-PrONa, were used in the trials.
GC-MS analysis of isolated raw products were com-
pared with previous experimental results.13,14 These
analyses revealed that the reaction that was carried out
in water, with NaOH as base, at 75 °C contained the
expected product N,N′-bis(3-nitrophenyl)diazene N-oxide
12 (16%), minor amounts of bis(3-nitrophenyl)diazene 13
(3%), and the reduction product 3-nitrophenylamine 14
(25%). When the reaction was conducted at 75 °C with
(11) The prediction is performed by using the following conditions:
x1 ) 2.4 mmol, x2 ) 30 mL, x3 ) 167 °C, and x4 ) 2 h. This position is
found in the subplot in the upper-right corner of Figure 1. The lower
right corner of this subplot predicts 90% yield.
(12) The developed model predicts an “optimized” yield at the
settings of short reaction time and elevated reaction temperature. Most
probably this is due to a model error that becomes visible when “long-
range” extrapolation is performed. In fact, only one experiment has
previously been performed at such a short reaction time and included
in the training set (entry 16 of Table 1) for the predictive model eq 3.
(13) El’tsov A. V.; Kuznetsova N. A.; Frolov A. N. Zh. Org. Khim.
1971, 7 (4), 817-820.
(14) Khan M. S.; Lone A.; Kashmiri M. A. J . Nat. Sci. Math. 1990,
30 (2), 77-81.
J . Org. Chem, Vol. 69, No. 22, 2004 7723

Bjørsvik et al.
F IGURE 1. Multidimensional contour plot showing the predicted yield as a function of four experimental variables: quantity of
acetophenone [mmol] (x1), volume of NaOH (40%) solution [mL] (x2), the reaction temperature [°C] (x3), and the reaction time [h]
(x4). Multidimensional response surface plots in terms of contour plots describe variation in one or more responses that is given
by the contour lines and two or more experimental variables. In the present plot, the contour lines show the yield of target
molecule benzo[c]cinnoline N-oxide when the four experimental variables x1-x4 are varied. The experimental variables x3 and x4
are varied continuously in the ranges of 130-170 °C and 2-6 h, respectively. The two variables x1 and x2 are both varied on five
discrete experimental levels each that are [1.6, 1.8, 2.0, 2.2, 2.4 ]/mmol and [10, 15, 20, 25, 30]/mL, respectively.
2-propanol and sodium isopropoxide as solvent and base,
respectively, the only product observed in substantial
quantity was 3-nitrophenylamine 14 (27%). Minor quan-
tities (<3%) of the benzene-1,3-diamine were also de-
tected. The corresponding reactions performed at the
elevated temperature (157 °C) resulted in each case only
in degradation products and a low yield of the diazene
products; this supports our proposed mechanism involv-
ing a “cage re-bonding” step. Such a mechanism disfavors,
of course, the combination of two separate molecular
entities as in Scheme 5. Moreover, several other reactive
species (Scheme 2) that are present during the redox
process will compete for the 3-nitronitrosyl radical anion
species when liberated from the solvent cage.
Attem p t To Syn th esize Va r iou s Su bstitu ted Ben -
zo[c]cin n olin es. To validate the generality of our new
process to the framework of benzo[c]cinnoline, a series
of various substituted 2,2′-dinitrobiphenyls (23-30)
were subjected to the conditions of the optimized reac-
tion conditions. The series of 2,2′-dinitrobiphenyls
23-30 shown in Table 2 were prepared utilizing the
Ullmann coupling reaction15-17 with 2-nitrobromoben-
(15) Ullmann, F. Ann. 1904, 332, 38.
7724 J . Org. Chem., Vol. 69, No. 22, 2004

Process to the Framework of Benzo[c]cinnoline
TABLE 2. Va r iou s 2,2′-Din itr obip h en yl Der iva tives Su bm itted for th e Novel P r ocessa to th e Ben zo[c]cin n olin e
F r a m ew or k
a
2,2′-Din itr obip h en yl Der iva tives. To a solution of the 1-bromo-2-nitrobenzene in DMF was added Cu-Bronze (in two portions,
adding the second after 4 h of reaction time). The mixture was heated at 160 °C for 19 h with reflux, at atmospheric pressure. The
reaction mixture was then quenched with water. The solid crude product was isolated by filtration. The isolated solid was extracted with
MeOH and filtered to remove undissolved inorganics and tars. The solvent was removed under vacuum to obtain the 2,2′-dinitrobiphenyl
derivative. Ben zo[c]cin n olin e d er iva t ives. 2,2′-Dinitrobiphenyl and acetophenone were added to a basic solution of EtONa (from
dissolving Na0 in EtOH). The resulting solution was placed into a sealed tube under N2, heated at 160 °C for 5 h, quenched with H2O (40
b
mL), and extracted with EtOAc (3 × 40 mL). The major product is the benzo[c]cinnoline derivative. In this case, the redox process
proceeds intramolecularly by means of three different functional group transformations. c The reaction mixture contained several products,
the structures of which were not determined. A rough estimate for the yield of target molecule was performed by means of GC-MS.
zenes (15-22) as the starting materials. The 2,2′-dini-
trobiphenyls, were variably functionalized, but -NH2,
-OH, and -CN were excluded, since previous experi-
ments under comparable redox conditions revealed a
mismatch for those groups. Such groups easily form
radicals resulting only in degradation products and tars.
Likewise, the functional groups -COCH3, -CHO, and
-CH2OH are not compatible with the reaction conditions
because such groups can participate in the redox process
ultimately oxidizing them to the carboxylic acid groups.
However, in the series of compounds that were submitted
to the conditions of the novel process, a compound with
the acetyl group was included (entry 7, Table 2). This
compound (28) was especially interesting, as this molec-
ular entity could tentatively operate both as substrate
and reagent in the redox process; however, it failed to
react. The other experiments performed included the
followingfunctionalgroups: -CH3, -OCH3, -Ph, -COOH,
-COOCH3, -Cl, and -CF3. The conditions utilized and
results are summarized in Scheme 6 and Table 2. Except
(16) Ullmann, F.; Sponagel, P. Ber. 1905, 38, 2211.
(17) See, e.g.: Furniss, B. S.; Hannaford, A. J .; Rogers, V.; Smith,
P. W. G.; Tatchell, A. R. Vogel’s textbook of practical organic chemistry;
Longman Scientific and Technical: Burnt Mill, Harlow, 1978; pp 610-
611.
J . Org. Chem, Vol. 69, No. 22, 2004 7725

Bjørsvik et al.
SCHEME 6
for the compound carrying an acetyl group (28), it was
only the 2,2′-dinitrobiphenyl substituted with the -COOH
group (compound 26) that provided a low yield. The
corresponding methylester (27) of that compound and
2,2′-dinitrobiphenyl derivative 29 both provided accept-
able yields (∼50%). All of the other 2,2′-dinitrobiphenyl
derivatives provided excellent yields.
is <140 °C, the redox process did not operate; see, e.g.,
experiment entry 14. (ii) When elevated temperatures are
utilized, high conversions of 2,2′-dinitrobiphenyl 6 are
accomplished, but most probably this is due to successive
degradation reactions. Moreover, when the process is
carried out at reaction temperatures that exceed 170 °C,
the solvent cage that encloses the reactive species may
be weakened, which results in leakage of the reactive
species that can provide other reduction products or tars.
We believe that the disclosed process can find applica-
tion for synthetic purposes, even if the process requires
a slightly elevated temperature to operate satisfactory.
The elevated temperature is a minor drawback compared
to the advantages of the novel process: (i) the process
operates in the absence of any transition (or heavy)
metals in catalytic or stoichiometric quantities, (ii) no
toxic or hazardous reagents are required, (iii) the process
is operated with cheap reagents and solvents, (iv) the
novel process provides as high selectivity and yield as
the most of the other processes and methods previously
disclosed by other researchers,18-24 and (v) no harmful
side products are produced during the process. In all, the
disclosed process appears to be an attractive method for
benzo[c]cinnolins and may thus provide an alternative
for industrial applications.
Con clu sion s
A process leading to benzo[c]cinnoline 10 and benzo-
[c]cinnoline N-oxide 11 has been discovered and inves-
tigated. During the investigation of the process it was
discovered that the types of solvent and base were of
paramount importance for the course of the process.
Likewise, the reaction temperature should carefully be
kept within the range 155-167 °C to achieve optimal
yields. The quantity of acetophenone, which functions as
a reductant, should, when using the process for the
preparation of benzo[c]cinnoline 10, be used in slightly
less than molar quantities. When synthesizing benzo[c]-
cinnoline N-oxide 11 a moderate molar excess of ac-
etophenone is required. When the optimized experimen-
tal conditions are used, excellent selectivity and yields
are achieved for either of the two products, namely 93%
(10) and 91% (11). The novel process to benzo[c]cinnoline
can successfully be utilized for the preparation of various
substituted benzo[c]cinnolines in most cases with me-
dium to high yields, Table 2. As for the previously
disclosed catalytic aerobic oxidation process,1,2 the novel
combined redox and cyclization process forming the
benzo[c]cinnoline framework is not compatible with
functional groups such as -CN, -NH2, and -OH, as such
groups easily forms free radicals under the redox condi-
tions and result in nonselective processes: these form
only degradation products or tars.
We believe that the reaction proceeds by a complex
reaction mechanism, some facets of which are outlined
in Scheme 3. Likewise, we think that the initial redox
process (path (b), Scheme 2) forming the radicals is
analogous to the mechanisms we have previously dis-
closed.1-3 Additionally, the work of Russell and co-
workers6-,8 described conclusive evidence for the single
electron-transfer reduction of nitroarenes in basic
media, and for the formation of C-centered radicals and
nitroarene radical anions by reactions of carbanions with
nitroarenes.
Exp er im en ta l Section
Ben zo[c]cin n olin e: Wor k u p w ith F la sh Ch r om a tog-
r a p h y. The reaction was carried out in a sealed tube reactor
of 50 mL capacity. The reaction tube was flushed with
nitrogen, and the reactants 2,2′-dinitrobiphenyl (Aldrich 97%)
(2.0 mmol, 0.488 g) and acetophenone (1.8 mmol, 0.212 mL)
were dissolved in ethanol (15 mL) and placed in the reaction
tube. Finally, sodium ethoxide (10 mmol, 0.230 g) was added.
The reaction mixture was heated at 160 °C for 5 h and
quenched with water (50 mL) followed by filtration on Celite.
The filtrate was extracted with ethyl acetate (3 × 50 mL). The
organic phase, which contained the target product, was dried
over anhydrous sodium sulfate and filtered and the solvent
removed under vacuum to obtain the crude product (564 mg).
This material was purified and separated by means of stan-
dard flash chromatography25 using a column of i.d. 3 cm. The
eluent was ethyl acetate and n-hexane in the ratio 1:4 (Rf (10)
) 0.334, Rf(11) ) 0.182). The crude material (564 mg) was
(18) Ettienne, A.; Izoret, G. Bull. Soc. Chim. Fr. 1964, 2897.
(19) Everett, J . L.; Ross, W. C. J . Chem. Soc. 1949, 1972.
(20) King, F. E.; King, T. J . J . Chem. Soc. 1945, 824.
(21) Ross, S. D.; Kahn, G. J .; Leach, W. A. J . Am. Chem. Soc. 1952,
74, 4122.
(22) Laskar, D. D.; Prajapti, D.; Sandhu, J . S. J . Chem. Soc., Perkin
Trans. 1 2000, 67-69.
(23) Pink, M.; Young, V. G. J . Org. Chem. 2000, 65, 6388.
(24) Wada, S.; Urano, M.; Suzuki, H. J . Org. Chem. 2002, 67, 8254-
8257.
Experiments conducted at (i) prolonged reaction times
at low temperatures or (ii) at short reaction times with
elevated reaction temperature did not provide high yields
as anticipated by the response surface. Explanations for
this may be as follows: (i) When the reaction temperature
(25) Still, W. C.; Khan, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923.
7726 J . Org. Chem., Vol. 69, No. 22, 2004

Process to the Framework of Benzo[c]cinnoline
loaded on the column. In total, 150 fractions of 10 mL were
collected. Fractions 49-123 were determined to contain pure
benzo[c]cinnoline (281 mg). The isolated sample of benzo[c]-
One hundred and twelve fractions of 10 mL each were
collected. Fraction numbers 54-112 were determined to
contain pure benzo[c]cinnoline N-oxide (305 mg). The spectral
data obtained for the isolated sample of benzo[c]cinnoline
N-oxide were compared with those of an authentic sample.26
N,N′-Bis(3-n itr op h en yl)d ia zen e N-Oxid e. A sealed tube
reactor was flushed with nitrogen, and aqueous NaOH (40%,
15 mL), 1,3-dinitrobenzene (0.672 g, 4 mmol), and acetophe-
none (0.281 mL, 2.4 mmol) were then added.
The reaction mixture was heated at 75 °C (oil bath) with
stirring for 2 h. The basic water phase was diluted with water
(50 mL), passed through a pad of Celite, and extracted with
ethyl acetate (3 × 70 mL). The organic phase was then dried
over anhydrous sodium sulfate, filtered, and evaporated under
vacuum. The isolated compound was analyzed via GC-MS,
and this showed the presence of N,N′-bis-(3-nitrophenyl)-
diazene N-oxide (16%), bis-(3-nitrophenyl)diazene (3%), and
3-nitrophenylamine (25%).
Ullm a n n Cou p lin g Rea ction . To a solution of the 2-ni-
trobromobenzene (2.0 mmol) in DMF (6.0 mL) was added Cu-
bronze in two portions (0.30 g at start of the reaction, then
another portion of 0.30 g after 4 h.). The reaction mixture was
heated at 160 °C for 19 h with reflux condenser at atmospheric
pressure. The final solution was cooled and then quenched in
water (60 mL). The solid was filtered off and extracted (3 ×
70 mL) with warm methanol (65 °C). The ethanol extracts were
combined and filtered, and the solvent was removed under
vacuum. The isolated crude contained the Ullmann coupling
product as the major product in yields of 19-97%.
Mu ltiva r ia te Ca lcu la tion s a n d Gr a p h ics. Calculations
and the graphical representations were performed by means
of in-house developed procedures for MATLAB version 6.1.27
The MATLAB procedures have previously been validated
by comparison of the calculated results with computational
results obtained from several commercial computer programs
for statistics and mathematical model building.
1
cinnoline was compared with H NMR spectra of an authentic
sample.24,26
Ben zo[c]cin n olin e: Wor k u p w ith Cr ysta lliza tion . Ac-
etophenone (3.6 mmol, 0.424 mL) and 2,2′-dinitrobiphenyl
(4.12 mmol, 1.070 g) were added to a basic solution of sodium
ethoxide, obtained from reacting Na-metal (20 mmol, 0.460 g)
with ethanol (30 mL). The resulting solution was placed in a
sealed tube reactor under nitrogen atmosphere and heated at
160 °C for 5 h. The reaction mixture was then quenched with
water (70 mL) followed by filtration on Celite and silica.
The filtrated aqueous solution was extracted with ethyl
acetate (3 × 70 mL). The organic phase, which contained the
target product, was dried over anhydrous sodium sulfate and
filtered, and the solvent was removed under vacuum to obtain
the crude product (786 mg). The sample was analyzed on GC
utilizing a correction factor to determine the quantity of benzo-
[c]cinnoline (91% yield, 90% purity). The sample from the flash
chromatography was dissolved in refluxing n-hexane (70 mL)
and filtrated. The resulting yellow solution was allowed to cool
on an ice bath for a period of 1 day to achieve yellow crystals
(462 mg) of benzo[c]cinnoline of a purity of practically 100%
measured on GC.
Ben zo[c]cin n olin e N-Oxid e. To a 40% aqueous solution
of NaOH (25 mL) were added acetophenone (0.281 mL, 2.4
mmol) and 2,2′-dinitrobiphenyl (0.503 g, 2.0 mmol). The
reaction was carried out at 157 °C (oil bath) with stirring for
4 h 30 min at atmospheric or slightly elevated pressure by
means of a sealed tube reactor. After a few minutes of reaction
time, the reaction mixture became dark orange, which during
the course of the reaction changed to dark brown.
The basic water phase was diluted by adding more water
(50-70 mL) and extracted with ethyl acetate (3 × 70 mL). The
organic and water phases were passed through a Celite layer
in order to remove tars.
The basic water phase was acidified using concentrated HCl
(37% solution) until pH 1-2 and then extracted with ethyl
acetate (3 × 70 mL). This acidic phase contained pure benzoic
acid.
Ack n ow led gm en t. Economic support from the Re-
search Council of Norway is gratefully acknowledged.
Professor George Francis is acknowledged for linguistic
assistance.
The organic phase contained benzo[c]cinnoline, benzo[c]-
cinnoline N-oxide, unreacted 2,2′-dinitrobiphenyl, and ac-
etophenone analyzed by GC-MS to give a preliminary con-
clusion that benzo[c]cinnoline (9%) and benzo[c]cinnoline
N-oxide (90% yield) were the two major products in the final
crude (530 mg). The isolated crude product was separated and
purified by means of flash chromatography. A column of i.d. 4
cm was used. The eluent was ethyl acetate and n-hexane in
the ratio 2:3 (Rf(10) ) 0.566, Rf(11) ) 0.440). The total quantity
of crude material that was loaded on the column was 864 mg.
Su p p or tin g In for m a tion Ava ila ble: General experimen-
tal information and copies of 1H NMR spectra for compounds
10, 11, 23-25, 28-33, and 38. This material is available free
of charge via the Internet at http://pubs.acs.org.
J O049102O
(27) (a) Using Matlab, Version 6, The MathWorks Inc., Natick, MA.
(b) Using Matlab Graphics Version 6, The MathWorks Inc., Natick,
MA.
(26) Kilic, E.; Tu¨zu¨n, C. Org. Prep. Proced. Int. 1990, 22, 485-493.
J . Org. Chem, Vol. 69, No. 22, 2004 7727