Regiospecific synthesis of 5,7-disubstituted quinoxalino[2,3-b ]phenazines

Hydrogenation of the readily prepared dinitrobenzenediamines 7 followed by air oxidation affords the green colored 5,7-disubstituted-5 H ,12 H -quinoxalino[2,3-b ]phenazines 3 in good yields. Mechanistic rationale, compound characterisation and full experimental details are provided.


Introduction
Interest in the heterocyclic system quinoxalino [2,3-b]phenazine (Flourindine) has reemerged.Recently theoretical and experimental studies on the 5,7-diphenyl-5H,12H-quinoxalino [2,3b]phenazine 3 established it has a singlet ground state resulting in a zwitterionic structure. 1A 5,7-bisoctadecyl derivative has shown interesting high temperature liquid crystalline properties. 2,3These studies were made possible by an improved high yielding and regiospecific synthesis of the 5,7-disubstitued isomers.We now wish to report the full synthetic details.The parent system quinoxalino[2,3-b]phenazine 1 is not known.The only dihydroquinoxalino [2,3-b]phenazine is the 5H,14H-dihydro derivative 2 thought to be in equilibrium with the 5H,12H-dihydro isomer. 4However, 5,7-diphenyl-5H,12H-quinoxalino [2,3-b]-phenazine (diphenylisofluorindine, 5,7-DPQP, 3a) exists. 5The preparation of 5,7-DPQP from the treatment of 3-imino-N,5-diphenyl-3H,5H-2-phenazinamine (3-anilinoaposafranine) with N-phenyl-1,2benzenediamine 5 and two equivalents of mineral acid in refluxing benzoic acid was reported over a 100 years ago, however, at that time its electronic structure was not understood. 5The product which did not melt (up to 260 o C) was identified by microanalysis and by comparison of its physical appearance and color in solution with its more commonly known isomer 5,12-diphenyl-5H,12H-quinoxalino [2,3-b]phenazine (diphenylfluorindine, 5,12-DPQP, 4). 6Both isomers dissolve in acid to give a blue solution with a red fluorescence but only the free base of 5,12-DPQP was observed to fluoresce strongly to the naked eye whilst that of the 5,7-DPQP did not.Both compounds crystallize to give blue-green crystals with a metallic luster.Various preparations of 5,12-DPQP are reported. 6In particular the treatment of 3-anilinoaposafranine with N-phenyl-1,2-benzenediamine 5 and one equivalent of mineral acid to give the isomer 5,12-DPQP 6 suggested to us that the formation of a mixture of both isomers was likely via this route and would therefore require separation.Furthermore the synthesis of 3-anilinoaposafranine was derived from the oxidative coupling of two equivalents of N-phenyl-1,2-benzenediamine 5 which gives a mixture of two isomeric products that again require careful separation.

Synthesis
We proposed and successfully carried out a rational synthesis that affords 5,7-DPQP unambiguously and in good yield (Scheme 1).1,5-Difluoro-2,4-dinitrobenzene 6 reacts with N-substituted-1,2-benzenediamines 5 to give dinitrobenzenediamines 7 in good yields.Hydrogenation of compounds 7 gave the benzenetetraamines 8, which on simple heating in ethanol in the presence of air gave the free base 3.The benzenetetraamines 8 were very susceptible to oxidation and their isolation and characterization was only carried out with one example (c.f.compound 8a, Experimental section).Treatment of 8a with ethanol and hydrochloric acid gave the hydrochloride salt of 5,7-DPQP which could be liberated with aqueous hydroxide.The N-aryl-1,2-benzenediamines 5 were prepared from 1-fluoro-2-nitrobenzene and anilines in the presence of potassium fluoride, 8 followed by hydrogenation.The N-alkyl derivatives were prepared from the action of the more nucleophilic alkylamines on 1-fluoro-2-nitrobenzene in refluxing ethanol, followed by hydrogenation.Unsymmetrical quinoxalino [2,3-b]phenazine 3e was prepared from the selective displacement of one fluoride from 1,5-difluoro-2,4-dinitrobenzene 6 which was achieved under mild conditions to give 5-fluoro-2,4-dinitrobenzamine 9 in good yield (Scheme 2).
Nearly quantitative yields (c.f.Method 1, compound 7a, Experimental section) were obtained for the preparation of compounds 7 with the use of 4 equivalents of benzenediamine 5.The cost, however, of preparing more complex diamines 5 prevented the repeated use of 4 fold excesses and despite lower yields the use of 2 equivalents of diamine 5 followed by 2 equivalents of Hünig's base was preferred (c.f.Method 2, compound 7a, Experimental section).The overall synthesis, analogous to that used for the preparation of 5H,14H-quinoxalino [2,3-b]phenazine from 1,5-dichloro-2,4-dinitrobenzene and excess 1,2-benzenediamine, 9 allows the preparation of a variety of 5,7-disubstituted quinoxalinophenazines (Table 1).

I
Mechanistic rationale.The mechanism for the cyclization closely follows that proposed for the synthesis of dihydroquinoxalino[2,3-b]phenazine. 9A dilute solution of tetraamine 8 in DCM (λ max 241 nm), at ca. 20 C, becomes brown in color (over 24 h) and absorption spectroscopy shows the formation of two new strong absorptions at 283 and 472 nm; the absorption at 241 nm is no longer visible.Over a period of 7 days the intensity of the absorption at 472 nm decreases until the spectrum resembles that of 5,7-DPQP; the solution's color changing from brown to green.This suggests that the first cyclization to give presumably phenazine 14 is more rapid than the second cyclization to give 5,7-DPQP 3a.A probable mechanism is described in Scheme 3. Air oxidation of benzenetetraamine 8a gives species 12 which can cyclize via nucleophilic attack of the diphenylamine on the NH imine to give 13 and ultimately quinoxalinophenazine 3a.

Conclusions
We have developed a regiospecific and high yielding synthesis for 5,7-disubstituted quinoxalino [2,3-b]phenazines.The synthetic route makes these unusual zwitterions readily available for further study.

Experimental Section
General Procedures.Reactions and column eluents were monitored by TLC using plasticbacked thin layer chromatography plates (Kodak) viewed under UV light at 254 and 350 nm.Dry flash chromatography on Bodman flash silica 32-63 was used for separations.UV/vis spectra were measured on HP 8453 UV-visible spectrometer.IR spectra were measured on a Mattson Infinity Series FTIR spectrometer. 1 H and 13 C NMR spectra were measured on Brucker AMX500, AMX400 and AC200 machines.Mass spectra were recorded on VG ZAB-SE or Autospec "Q" machines.Microanalyses were carried out by Desert Analytics, Inc.

N',N'''''-Bis[2-(N-phenylbenzamino)]-1,2,4,5-benzenetetraamine (8a).
To a stirred suspension of 1,5-bis[N-(N'-phenyl-1,2-benzenediamino)]-2,4-dinitrobenzene 7a (100mg, 0.188mmol) in EtOH (20mL) at ca. 20 o C, under argon, (10%) palladium on carbon (100mg) was added in one portion.The reaction mixture was evacuated (to 25mmHg) and flushed with argon 3 times then the mixture was evacuated (to 25mmHg) and flushed with hydrogen 3 times.The reaction mixture was then left to stir under an atmosphere of hydrogen.The color of the mixture became dark red and after 1 h the color disappeared and a cream colored precipitate was observed.The suspension was diluted with sufficient dichloromethane to dissolve the precipitate, which was filtered through a celite pad and all volatiles were removed to afford brown oil.This was diluted with cold EtOH and triturated to afford a crude specimen of the title compound 8a (80mg, 90%) as brown needles.A sample of the crude product was further purified by dry flash chromatography on silica (Et The color of the reaction mixture rapidly became brown then green and a green crystalline precipitate was formed within 20 min.TLC monitoring over a period of 7 h showed consumption of the starting material had ceased and so the precipitate was collected by filtration and dried to give 101.8mg of a green crystalline material.The filtrate was heated to reflux and rapidly precipitated a further 99.2mg which was removed by filtration.The process was repeated collecting a further four fractions (64.1, 38.0, 49.1 and 16.8mg) until it became impractical to recover more precipitate.No further purification was required, the total recovery of compound 3a was 369mg, 84% yield; the sample was identical to that described above.Method 3. To a stirred suspension of 1,5-bis[N-(N'-phenyl-1,2-benzenediamino)]-2,4dinitrobenzene 7a (100mg, 0.188mmol) in EtOH (20mL) at ca. 20 C, under argon, (10%) palladium on carbon (100mg) was added in one portion.The reaction mixture was evacuated (to 25mmHg) and flushed with argon 3 times, then evacuated (to 25mmHg) and flushed with hydrogen 3 times.The reaction mixture was then left to stir under a hydrogen atmosphere.The color of the mixture became dark red and after 3 h this red color disappeared and a cream colored precipitate was observed.The mixture was heated gently to dissolve the precipitated amine and then hot-filtered through a short pad of celite to remove the palladium catalyst.The dark ethanolic solution of the amine was heated exposed to atmospheric oxygen until a green precipitate was formed.This was filtered off and the filtrate was taken to reflux until more precipitate was formed.The precipitate was removed and the process repeated until there was no further precipitate.Combining the precipitated material gave the title compound 3a (70mg, 85%) identical to an authentic sample.
bRecrystallised from ethanol and dried overnight under vacuum (30mmHg) at 60 oC.
at ca.20 C, under an atmosphere of air, hydrochloric acid (36%, 10mL) was added in one portion.The color of the reaction mixture became lilac, then blue.The reaction mixture was heated under reflux for 2 h then allowed to cool to ca. 20 C. A green-blue precipitate was observed and assumed to be the hydrochloride salt of compound 3a.