Conventional and Microwave-Assisted Reaction of N-Hydroxymethylphthalimide with Arylamines : Synthesis of N-( Arylaminomethyl )-phthalimides

Uma síntese eficiente e fácil dos compostos: 2-fenillaminometil-isoindol-1,3-diona (5a), 2-[(2-Clorofenilamino)metil]-isoindol-1,3-diona (5b), 2-[(3-Clorofenilamino)methyl]-isoindol1,3-diona (5c), 2-[(4-Clorofenilamino)metil)-isoindol-1,3-diona (5d), 2-[(2-Flúorfenilamino)metil]-isoindol-1,3-diona (5e), 2-[(3-flúorfenilamino)metil]-isoindol-1,3-diona (5f), 2[(4-Flúorfenilamino)metil]-isoindol-1,3-diona (5g), 2-[(2-Nitrofenilamino)metil]-isoindol-1,3diona (5h), 2-[(3-Nitrofenilamino)metil]-isoindol-1,3-diona (5i), 2-[(4-Nitrofenilamino)metil]isoindol-1,3-diona (5j), 2-[1H-(1,2,4)Triazol-3-il-aminometil)-isoindol-1,3-diona (5k) e 2([1,2,4]-Triazol-4-il-aminometil)-isoindol-1,3-diona (5l) está descrita. A síntese foi realizada partindo-se da N-hidróximetilftalimida 3 e de arile [1,2,4-triazol-3e 4-il]-aminas 4a-l através de procedimentos convencional e mediado por microondas. A reação de 3 com 4l aconteceu rapidamente e com altos rendimentos. Está descrita uma comparação entre estes dois métodos. São propostos três prováveis mecanismos de formação das N-(arilaminometil)-ftalimidas (um em solução e os outros dois em condições de aceleração por microondas). As análises cristalográficas de 5d forneceram as informações apropiadas sobre a conformação da mesma. Cálculos de orbitais moleculares Ab initio de 5d empregando um conjunto de base 6-31G* foi realizado e os resultados concordaram com os dados de raio-X.


Introduction
Phthalimide derivatives have been gaining considerable interest since 1979, when Chapman et al. 1 tested the hypolipidemic activity of 23 N-substituted phthalimide derivatives.Later on, Hall and co-workers 2,3 reported the antihyperlipidemic activity of phthalimide analogs in rodents and also the same activity was found by the administration of ortho-(Nphthalimido)acetophenone in sprague dawley rats.In 2001, we reported hypolipidemic activity in α-Dmannopyranosides containing phthalimidomethyl function as aglycone. 4There are other interesting biological aspects of these compounds which have been reviewed in 2003. 5A recent paper cites the synthesis and anticonvulsant behavior of N-substituted phthalimides. 6Besides, certain phthalimide derivatives are synthetically important and can be transformed to other useful products. 7 literature survey disclosed that only two compounds, i.e., N-(piperidinomethyl)phthalimide 8,9 and N-(morpholinomethyl)-phthalimide 8 were prepared initially.Later on, N-(phenylaminomethyl)-phthalimide was synthesized by Weaver et al. 10 In 1954, Winstead and Heine 11 published a note about the synthesis of N-(arylaminomethyl)-phthalimides by treating Nhydroxyphthalimide with aniline in an aqueous alcoholic solution and obtained 70% yield of N-(anilinomethyl)phthalimide.These authors also refluxed a mixture of phthalimide 1, 37% aqueous formaldehyde 2 and an aromatic amine 4 in ethanol to obtain the same product 5.They suggested the following reaction sequence (scheme 1).
With this, it is obvious that these imides are potential candidates for biological evaluations and need further exploration.Therefore, the present contribution reports the conventional (solution as well as solid-phase) and microwave-accelerated syntheses of ten N-arylamino-methyl-5a-j and two [1,2,4-triazol-3-and 4-yl]phthalimides 5k,l (Scheme 2).

Results and Discussion
Synthesis N-Hydroxymethylphthalimide 3 and an appropriate arylamine 4a-j or triazolamines 4k,l in methanol were refluxed for 3h.Work-up gave crystalline compounds 5a-l where the yields ranged from 68-99% with the exception of N-(ortho-chlorophenylaminomethyl) and N-(ortho-nitrophenylamino-methyl) phthalimides which gave lower yields (49 and 64%, respectively).Compound 5d was acetylated and 1 H-NMR spectra of 9d supported the proposed structure of 5d.The m.ps., crystallizing solvent and yields of these compounds are compiled in Table 1.The literature records the preparation of 5a-d by the conventional procedure. 11

Infrared Spectroscopy
The infrared spectra showed an absorption between 3392-3453 cm -1 for a secondary amine in each case.These are generally somewhat sharp signals.There are also two strong absorptions between 1747-1775 cm -1 for the symmetric and asymmetric stretching vibrations of the imide carbonyl functions.

H Nuclear Magnetic Resonance Spectroscopy
The 1 H NMR spectra of compounds 5a-j presented a triplet around 5.0 ppm for the NH proton suggesting the existence of a CH 2 group in its vicinity.The methylene protons presented a doublet in the same region in each case supporting the proposed structure.The other two compounds of this series, viz., 5k e 5l also agreed with the structure but their chemical shifts apparently exhibited some difference compared to 5a-j.For example, the NH proton of compound 5k of the triazole ring appeared at low field ( = 11.34 ppm) and the CH heterocyclic ring proton absorbed at 8.68 ppm, four aromatic protons of the phthalimide moiety gave a multiplet between 7.87-7.75ppm and the CH 2 protons appeared at = 4.77 ppm.Compound 5l gave a singlet at = 8.48 ppm representing H-3' and H-5' protons of the triazole portion.The aromatic protons of this compound presented a singlet at 7.87 ppm, CH 2 at 4.77 (2H, d, J = 4.8 Hz) and NH at 7.52 as a triplet (J = 4.8 Hz).

Structure Verification
Although the structures of 5a-l appear reasonable, we considered other possibilities as well.In fact, there are three other isomers, besides 5a-l, which need contemplation.These are: 6, 7 and 8 (Figure 1), which have the same molecular formula, but are constitutional isomers.
In order to get more insight for the structures of the synthesized compounds, we acetylated one of these compounds, i.e., 5d with acetic anhydride and pyridine at room temperature.The reaction occurs in 30 min.and only one product 9d was isolated (Figure 2).The infrared spectrum of this compound didn't show any absorption for the ester carbonyl group; hence structures 6 and 7 are eliminated.Regarding isomer 8, its acetylation under mild conditions is not possible because of the amide function.Besides, the N-acetylation of the amide function requires severe conditions; hence structure 8 is also excluded.Therefore, it appears appropriate to say that the structures proposed for 5a-l are the correct ones.

Mechanism of formation of N-(arylaminomethyl)phthalimides 5a-l from N-hydroxymethylphthalimide 3 and arylamines 4a-l by conventional method
According to Winstead and Heine, 11 phthalimide first reacts with formaldehyde to produce Nhydroxymethylphthalimide followed by its reaction with an arylamine to give N-arylaminomethylphthalimide.Alternatively, formaldehyde may first react with an amine to provide N-hydroxymethylamine which would  subsequently condense with phthalimide to give arylaminomethyl phthalimides.Besides this, the authors didn't elaborate about the reaction mechanism.Therefore, we are proposing a more probable mechanism starting from N-hydroxymethylphthalimide with an arylamine.
The mechanism which accommodates best in solution involves the formation of an intimate ion pair.The solventseparated ion par is still an ion pair i.e., the cation and anion are not independent of each other, and may also yield products, or undergo internal return to give the starting reagents 3 and 4a-l.The complete mechanism of formation of the product is described below and given in Scheme 3.
Initially, arylamine 4a-l abstracts the proton from N-Hydroxymethylphthalimide 3 and generates phthalimide anion 10, formaldehyde 11 and protonated amine 12a-l.In methanol all three components are in a single solvation shell, where methanol immediately associates with formaldehyde forming a hemiacetal 13.Anion 10 may abstract a proton from 12a-l yielding phthalimide 1 and arylamine 4a-l, or alternatively a proton may transfer from 1 to 13 to provide the protonated hemiacetal 14, and phthalimide anion 10.All species are in equilibrium and are closely fitted ion pairs.Then, the protonated hemiacetal can easily lose methanol to furnish a protonated formaldehyde molecule, which in turn reacts with 4 to give an iminium ion 16a-l.
Finally the phthalimide anion 10 attacks the carbon atom of the iminium ion providing the product 5a-l.
In order to test the possibility whether this reaction indeed involves general base catalysis, we added a few drops of triethylamine to the methanol solution containing N-Hydroxymethylphthalimide 3 and an arylamine 4d and refluxed the contents.TLC monitoring showed that the reaction got completed in one hour instead of the normal reflux time of three hours without this base.Thus, we have been able to reduce the reaction time.The detection of phthalimide in one experiment by 1 H NMR spectroscopy in conjunction with the rate enhancement by the addition of Et 3 N strongly supports that this is a general base catalysis reaction, where the rate determining step is the formation of the tight ion pair which in a few steps leads to the final product.
In order to get more perception about the mechanism of formation of the products 5a-l from N-hydroxymethylphthalimide 3 and an amine 4a-l, we tried to do the reaction in a NMR tube.In a typical experiment, we dissolved compound 3 (0.017 mmol) and 4-chloroaniline 4d (0.017 mmol) in DMSO-d 6 (1 mL) at room temperature and obtained the spectrum at 25, 35 and 45 °C, but did not find any significant change.At 55 °C, after 40 min, we observed a change in the spectrum.There was a weak  The literature shows such as hydrogen at ≈ 11.29 ppm in DMSO-d 6 . 12owever, we didn't observe any hydrogen due to formaldehyde.It appears that the aldehyde remains as a hemiacetal in methanol solution.Although it was not possible to see an aldehyde signal in the spectrum, the formation of phthalimide 1 in the spectrum supports the mechanism of formation of the final products from 3 and an arylamine 4. The following experiments were performed to have an idea about the formation of hemiacetal as well as phthalimide.
Initially, one mol equiv. of each N-hydroxymethylphthalimide 1 and p-chloroaniline 4d were dissolved on methanol-d 4 in an NMR tube and the 1 H NMR spectrum recorded immediately at 20 °C .This produced the aromatic signals of compounds 3 and 4 as well.Besides, the -CH 2 O-and -NH 2 -of compounds 3 and 4 appeared at 5.13 and 4.92 ppm, respectively.After 5 min, two new signals at 4.65 and 4.57 ppm became visible.The former and the latter signal are presumably due to the methylene hydrogens of hemiacetal (CD 3 OCH 2 OH) and N-p-chlorophenylaminomethylphthalimide 5d, respectively.Increasing the temperature to 30 °C caused the decrease of the signal at 5.13 ppm and modification of the aromatic hydrogen signals.Finally, at 40 °C, the signal at 5.13 ppm disappeared, but the signal at 4.58 ppm became stronger and the -CH 2 -signal of the hemiacetal was visible at 4.65 ppm.
The interesting part was the aromatic region, which showed a narrow multiplet at 7.8 ppm similar to the aromatic hydrogens of phthalimide.
Further confirmation of the hemiacetal and phthalimide formation from N-hydroximethylphthalimide came when this compound was dissolved in CD 3 OD and gradually heated to 45 °C and later to 58 °C in a NMR tube.The signal at 4.65 ppm as well as the aromatic signals were quite clear.The above variable temperature experiments in the NMR tube strongly supported that compound 3 is in equilibrium with formaldehyde and phthalimide.
The same mechanism holds for dry conditions (without any solvent) like the reaction of 3 and 4a-l by conventional heating or under microwave irradiation.In both cases, addition of triethylamine increased the reaction rate and completed the reaction.Here also a tight ion pair is formed followed by the proton transfers to formaldehyde.This way, all the reagents remain together and lead to the iminium ion and finally to the product.This can also be considered a general basecatalyzed reaction.
There is another important observation which requires attention.This is about dry heating the reagents 3 and 4 in an oil bath without microwave assistance.The reaction takes a little longer (8 min) compared to 5 min in a domestic microwave oven.In both cases, it is heating but in the latter the microwaves do accelerate the reaction.

X-ray Crystallographic Analysis
The bond distances in general are as expected.However, a quite fascinating phenomenon was observed; the N(2)"C(9) bond length is 1.427 Å, which is much shorter then the normal bond distance of simple N"C bond (~1.47 Å).This shortening can be attributed to the superposition of the lone pair of electrons of N(2) with the antibonding (σ*) orbital of C( 9).This is similar to the anomeric effect which is common in carbohydrates and also in 2-alkoxytetrahydropyran systems. 13he bond angle C(1)"N(1)"C(9) of 122.27 degrees clearly shows that N(1)"C( 9) bond is coplanar with the imide ring.Another interesting observation is that the angle C(10)"N(2)"C(9) has 122.67 degrees.This deviation is due to interaction of the electron pair at N(2) and σ* of C( 9  It is fascinating to observe that the chlorine atom of one molecule attracts the oxygen atom of the other molecule which results in holding two molecules together in the unit cell (Figure 5).
Another interesting feature is that there is hydrogen bonding between the amino hydrogen atom of one molecule with the carbonyl oxygen atom of the other molecule, thus bringing both molecules closer.The bond distance between H-2 and 0-2 is 2.55 Å and the N(2)-H(2)-O(2) angle is 126° (Figure 6).
We have carried out ab initio molecular orbital calculations of N-(p-chlorophenylaminomethyl)phthalimide 5d using 6-31G* (Figure 8) as the basis set.The purpose was to compare the calculated values of 5d with the crystallographic data.Although the calculations are performed for a single molecule in vacuum, the experimental values are close with the calculated ones.These are described below: The N(2)-C(9) bond distance is 1.427 Å and the experimental value is also the same, showing a decrease in the C-N bond length.As observed in the X-ray crystallography data, the calculated value is remarkably the same.This supports the Overlap of C(9) σ* orbital with N(2) nonbonded orbital.
The bond angle C(1)-N(1)-C(9) has been found as 122.46° which is extremely close to the experimental one (122.27Å).Also, C(10)-N(2)-C(9) gives the value 122.92 degrees; the value is again close to the crystal data.
The remarkably close values of the experimental and calculated data show the effectiveness of the 6-31G* basis set (Figure 7).

Conclusions
We have accomplished the synthesis of twelve Narylamino-methylphthalimides 5a-l employing conventional, dry heating and microwave-accelerated procedures.The last method was the speediest one causing a drastic reduction of the reaction time.Tree reasonable mechanisms of formation of the final products have been proposed.The X-ray data analyses in conjunction with the molecular orbital calculations gave the correct conformation of the molecule particularly for compound 5d.

General
All compounds were checked for their structures by infrared (IR), UV, and 1 H NMR spectroscopy.Melting points were determined on a Digital Electrothermal serie IA 9100 melting point apparatus and are uncorrected.The microanalyses were performed in a Carlo Erba Mod.EA1110 equipment.UV spectra were recorded with a U-3200 Hitachi spectrometer.IR spectra were measured with a Bruker model IF S66 FTIR spectrometer using potassium bromide discs.NMR spectra were recorded in CDCl 3 (for compounds 5a-j) or DMSO-d 6 (for compounds 5k,l) using tetramethylsilane (TMS) as an internal standard, on Varian Unity Plus 300 MHz spectrometer.Assignments of 2-[(4-Fluorophenylamino)methyl]-isoindole-1,3-dione and 2-[(3-Fluorophenylamino)methyl]-isoindole-1,3-dione were made using 1 H homonuclear irradiation and gHMBC (gradient Heteronuclear Multi Bonds Correlation) NMR spectra.The reactions in microwave were realized in Oven Microwave Sanyo EM-3500B, 220V/1350W/2450MHz.Since we do not have any specific microwave oven, we did some experiments and located the correct heating place in the oven.Each experiment was repeated at least three times in order to obtain reproducible results.Therefore, we feel confident that these experiments can be repeated using any microwave oven.

General procedure for the synthesis of compounds 5a-l by heating
Compounds 5a-l were obtained by refluxing a mixture of N-hydroxymethylphthalimide 3 (0.6 mmol) and an appropriate aromatic amine 4a-l (0.6 mmol) in MeOH (3.0 mL) for 3 hours under N 2 atmosphere.The progress of the reaction was accompanied by thin-layer chromatography.After solvent evaporation, the solid material was crystallized and recrystallized from an appropriate solvent, except the products 5k,1 which were filtered through a short column containing silica gel.This procedure removed the impurities and facilitated crystallization.Table 1 contains the crystallizing solvent, yields and melting points of compounds 5a-l.
) discussed above.Only such an occurrence can be compatible with this value.The torsion angles C(1)"N(1)"C(9)"N(2) and C(8)"N(1)"C(9)"N(2) are 69.7 and "114.42°clearly indicate that C(9)"N(2) bond is ~70° out of the phthalimide plane.This also supports the overlap of N(2) electronpair and C(9)"N(2)"C(10)"C(11) is 14.9°, which is necessary to stabilize the conformation.An Ortep diagram is given in Figure 3.The unit cell diagram (Figure 4) contains two molecules of 5d, where the phenyl ring of one molecule is arranged in a parallel fashion with the phenyl ring of the other molecule in the crystal packing; however the chlorine atoms of the two molecules are set in opposite direction.

Figure 4 .
Figure 4. Unit cell diagram showing two molecules of 5d.

Figure 5 .
Figure 5. Unit cell structure showing the interaction between the chlorine atom of one molecule and the oxygen atom of the other molecule of phthalimide.

Table 1 .
Solvent of crystallization, reaction time, yields and melting points of compounds 5a-l are given.The results of conventional heating as well as microwave irradiation are also provided Figure 1.Possible isomeric structure of compounds 5a-l.