Efficient synthesis of aryldipyrromethanes in water and their application in the synthesis of corroles and dipyrromethenes

In this paper, we describe the efficient and selective synthesis of aryldipyrromethanes in aqueous medium by acid-catalyzed (HCl) condensations of aromatic aldehydes with 3 equivalents of pyrrole at room temperature. The precipitated aryldipyrromethanes can be isolated directly from the reaction mixture in an essentially pure state by simple filtration. Time control seems to be essential to avoid significant formation of the tripyrromethane analogue and the reaction time is strongly dependent on the nature of the aromatic aldehyde.


Introduction
Dipyrromethanes are widely being used as essential building blocks for the synthesis of a variety of functional porphyrins and (contracted and expanded) porphyrin analogues. 1Moreover, dipyrromethanes are the precursors of BODIPY dyes (4,4-difluoro-4-bora-3a,4a-diaza-s-indacene or boron dipyrromethene), which are currently receiving increasing attention due to their valuable properties, such as the relatively high absorption coefficients and fluorescence quantum yields, high (photo)chemical stability, and improved synthetic availability. 2Nowadays, meso-substituted dipyrromethanes are usually prepared via a one-flask method based on the acid-catalyzed (e.g.trifluoroacetic acid) condensation of an appropriate aldehyde with an excess of pyrrole (used as solvent) and flash chromatography is required in most cases to obtain the dipyrromethanes in high purity. 3Although in recent years significant progress has been made regarding the efficient preparation of these dipyrrolic building blocks (e.g. to eliminate the use of chromatography), 4 most synthetic approaches still require an excess of pyrrole to ensure an optimal yield of the dipyrromethane (over the higher oligocondensates, e.g.tripyrromethane).
Recently, the use of water as a cheap and non-toxic solvent for organic reactions is of particular interest due to the increasing environmental concerns.In the search for 'green chemistry' methods, two research groups have reported the synthesis of dipyrromethanes via acid-catalyzed condensations in aqueous medium. 5The first condensation reaction of unsubstituted pyrrole with carbonyl compounds (ketones and aldehydes) in water (at reflux temperature and catalyzed by HCl) has been described by Sobral et al. 5a Their one-step procedure afforded β-free dipyrromethanes in moderate to high yields by using only a 2/1 ratio of pyrrole over the carbonyl compound.Moreover, the required dipyrromethane could be isolated in essentially pure form by a simple filtration from the (cooled) reaction mixture.Precipitation of the dipyrromethane from the aqueous layer as it is formed, forces the reaction to completion and protects the product from further reactions.Kral et al. reinvestigated the condensation of various aromatic aldehydes with pyrrole (at room temperature) and they found that a mixture of dipyrromethanes (DPMs) and tripyrromethanes (TPMs) is usually obtained and the selectivity for both oligocondensates depends on the nature of the carbonyl compound and can be controlled by the molar ratio and concentrations of the starting products and the acid catalyst (HCl).5b A high initial concentration and a 6/1 ratio (pyrrole/aldehyde) lead to preferential formation of the DPM.Based on the work of Kral, Gryko et al. have developed an efficient method for the synthesis of meso-substituted A 3 -and trans-A 2 B-triarylcorroles in a H 2 O-MeOH mixture. 6The increased solubility of the aryldipyrromethanes on adding methanol to the reaction mixture, allowed the efficient preparation of the bilane precursor, which was then oxidized to the corrole macrocycle (with DDQ).
For the last decade our group has dedicated a considerable part of its research to the development of synthetic methods for a variety of (macrocyclic) oligopyrrolic compounds, e.g.porphyrins, 7 corroles, 8 and BODIPY derivatives, 9 for which aryldipyrromethanes are crucial building blocks.Due to the attraction of the novel approach toward aryldipyrromethanes in aqueous medium, 5 we decided to (re)investigate the condensation of several aromatic aldehydes with pyrrole in water and expand the range of possible aldehyde substrates.Especially those (pyrrole unsubstituted) aryldipyrromethanes that are useful for our own research purposes were pursued.

Results and Discussion
We have observed that several aryldipyrromethanes 2a-i could easily be obtained in high yields (69-97%) from the acid-catalyzed (HCl) condensation of the corresponding aldehyde precursors 1a-i and 3 equivalents of pyrrole in water (Table 1, Scheme 1a).In the same way, bisdipyrromethane 2j was obtained from terephthaldehyde (1j) and 6 equivalents of pyrrole (in 80% yield).The reactions were carried out in a 0.18 M aqueous HCl medium and 3 equivalents of pyrrole (compared to the 6-fold molar excess used by Kral et al. 5b ) were added, followed by the addition of 1 equivalent of the appropriate aromatic aldehyde 1a-i (~0.15 M).The reaction mixture was stirred at room temperature and the reaction progress was followed by both TLC and mass spectrometry (until complete disappearance of the aldehyde was observed).After the indicated time (Table 1), the precipitated (semi-) solid product, which often sticks to the walls of the flask and the stirring bar and might hamper the stirring, was filtered off and washed with water and petroleum ether to afford DPMs 2a-i in high yields.The reaction appears to be quite general.Only in the case of 4hydroxybenzaldehyde (1k), no aryldipyrromethane was obtained, probably due to interference of the electron rich phenol moiety.The amount of TPM (and other side products) was usually negligible (as observed by 1 H NMR, MS and TLC) and DPMs 2a-i were obtained in good purity (i.e.above 95%, as analyzed by 1 H NMR), although the DPMs (which are essentially white) often showed a grey or brown color.This result is similar to the observations made by Sobral et al. (at reflux temperature), 5a but is in contrast to the results obtained by Kral et al., 5b where significant amounts of TPM were detected.It has to be noted however that the latter authors did not isolate the aryldipyrromethanes by filtration, but through extraction (thereby losing the benefit of selective DPM precipitation), while the reaction was generally run for 17 h.Time control seems to be crucial to obtain the required DPMs in high purity.Longer reaction times lead to a higher content of TPM impurity.Analysis ( 1 H NMR, TLC) of the initially formed precipitate showed a very high purity of the DPM, while the precipitate that is formed after longer reaction times showed a gradual increase in TPM content.Hence, it might be beneficial to filter the reaction mixture at an early stage of the reaction if a batch of DPM with superior purity is required, but the subsequent batches (which can be obtained from the filtrate after a longer reaction time) still show only a minimal content of TPM and can generally be applied as bipyrrolic building blocks.From Table 1 it can be observed that for the less reactive aldehydes, e.g. the electron rich p-anisaldehyde (2e), and the sterically hindered aldehydes 2c and 2i, longer reaction times (8-12 h) are required to obtain the DPMs in an optimum yield, while for very reactive aldehydes, e.g.p-nitrobenzaldehyde (2d), a much shorter reaction time (only 2 h) was required.
Although these aryldipyrromethanes 2a-j have been prepared before by other authors using 'classical' conditions, 3,4 the presented mild and efficient green method has the advantage that a large excess of pyrrole is no longer required.Moreover, the dipyrromethanes can be isolated easily from the reaction medium by simple filtration in an essentially pure state.

Scheme 1
From the obtained dipyrromethanes 2a,b,e,f,g, the corresponding dipyrromethenes 3a,b,e,f,g were synthesized according to a general procedure (Table 2, Scheme 1b). 10 Considering the limited stability of the (α'-unsubstituted) dipyrromethane skeleton, the observed yields (50-78%) are quite high.The oxidation of aryldipyrromethanes 2a,b,g was conducted with p-chloranil (in dichloromethane), while for dipyrromethanes 2e,f the use of DDQ (2,3-dichloro-5,6-dicyano-1,4benzoquinone) was necessary to obtain the corresponding dipyrrins in good yields.Dipyrromethenes 3a and 3g have been prepared by other groups, 10 but the other dipyrrin derivatives 3b,e,f have, to the best of our knowledge, never been isolated (and characterized) before.Corroles are ring-contracted porphyrin analogues lacking one meso-carbon atom.1c Until 1999, corroles were considered rather rare chemicals, quite unstable and difficult to prepare, but pioneering studies by Gross and Paolesse have afforded novel pathways toward (more) stable mesotriaryl-A 3 -corroles. 11This increased synthetic accessibility has triggered several other groups to search for efficient synthetic procedures for a variety of triarylcorroles, in order to be able to explore their intriguing properties and potential applications (e.g. in catalysis) to a further extent. 12he synthesis of trans-A 2 B-corroles via the 2+1 (MacDonald type) condensation of aryldipyrromethanes with aromatic aldehydes has been investigated by several research groups (mainly by D. Gryko et al.) and has evolved to one of the most important pathways toward functional corroles. 6,13nterest in the synthesis of corroles in our laboratory was initially triggered by a rather unexpected result.As part of a project to prepare porphyrin derivatives out of 4,6dichloropyrimidine-5-carbaldehyde, 7c a 1+1 condensation of this aldehyde and 5-(2,6dichlorophenyl)dipyrromethane (2i), catalyzed by boron trifluoride etherate (BF 3 •OEt 2 ), was performed (under Lindsey conditions).Surprisingly, the only identifiable product formed (in 25% yield) after p-chloranil oxidation, was the trans-A 2 B-corrole, without formation of the corresponding A 2 B 2 -porphyrin.8a Under similar conditions, 5-(2,6-dichlorophenyl)dipyrromethane (2i) could also be condensed with 2,6-dichlorobenzaldehyde (1i) and pentafluorobenzaldehyde (1h) to obtain the corresponding triarylcorroles (in 22 and 18% yield, respectively).However, oxidation was carried out using the Lee method 13a (DDQ and NH 4 Cl in propionitrile).Reaction of the same DPM 2i with benzaldehyde (1a) or 4-nitrobenzaldehyde (1d) did not afford any corrole under these conditions.Some of these corroles have already been used as analytically active compounds in liquid membrane electrodes (ISEs) that are sensitive toward neutral nitrophenol isomers or salicylic acid and salicylate.8b,8d,14 In the present work, some of the obtained sterically hindered aryldipyrromethanes were used for the synthesis of both known (optimized procedures) and novel trans-A 2 B-corroles 4a-h (Scheme 2, Structure Block 1, Table 3).

Scheme 2
As a starting point, the previously prepared 5,15-bis(2,6-dichlorophenyl)-10-(pentafluorophenyl)corrole 4a was chosen and we have tried to improve its yield (18%) by varying some reaction parameters.8a When the DPM/aldehyde content was raised to 1.5/1 and the BF 3 •OEt 2 concentration was raised to 0.85 equivalents, the only identifiable product was the corresponding A 2 B 2 -porphyrin 5a (Table 3, entry 1).While it is well known that the porphyrin-forming reaction, as well as many other macrocyclization reactions, requires a rather low concentration of reactants to achieve a reasonable yield of product and a rather high concentration of acid is also needed, it appears from Gryko's work 13 that exactly the opposite, i.e. a high concentration of reactants and a low concentration of acid catalyst is beneficiary if corroles are to be obtained.Increasing the DPM/aldehyde ratio to 3.9/1, while also lowering the concentrations of the substrates and the acid catalyst (0.24 equiv.) at the same time, provided corrole 4a in 8% yield, with 11% of the A 2 B 2porphyrin analogue 5a (entry 2).Also in this case Lee's method of oxidation was used.13a Finally, keeping relatively low concentrations of the substrates, but increasing the DPM/aldehyde ratio to 5/1 (approaching the preferred DPM/aldehyde ratio of Brückner 13c ), with a relatively low concentration of BF 3 •OEt 2 (0.37 equiv.)and using Lee-type oxidation (3 equiv.p-chloranil), corrole 4a was obtained in 22% yield without detectable porphyrin formation (entry 3).Using similar conditions but a shorter reaction time (30 min) resulted in a sharp decrease of the corrole yield (11%, entry 4).

Structure Block 1
Contrary to the synthesis of corrole 4a, preparation of 5,15-bis(2,6-dichlorophenyl)-10-(4nitrophenyl)corrole 4b had not been accomplished before.8a Therefore, we decided to try out synthetic conditions similar to those described by Gryko and Jadach. 13d Reacting DPM 2i and 4nitrobenzaldehyde (1d) in a 3.5/1 ratio in dichloromethane with trifluoroacetic acid (TFA, 0.32 equiv.)as the catalyst, with subsequent oxidation by direct addition of a 2-fold molar excess (with respect to the aldehyde) of p-chloranil, afforded corrole 4b in 6% yield (entry 5).However, several side-products were formed and the most prominent of them was identified as the corresponding A 3corrole (5,10,15-tris(2,6-dichlorophenyl)corrole).Its formation can be explained by the well-known phenomenon of acid-induced 'scrambling' or 'redistribution' of the pyrrole rings of oligopyrrolic intermediates, which is caused by the acidic reagents that are often used to induce macrocyclization. 15To avoid or at least lessen the amount of scrambling, we tried to carry out further experiments at 0 °C.However, this did not suppress the unwanted side reaction and the A 3corrole could always be observed and isolated in 1-2% yield (with respect to the starting DPM 2i).Applying a lower concentration of TFA (entry 6) caused an increase in the yield of the desired corrole 4b, in agreement with Gryko's findings.13d Monitoring the reaction by TLC suggested that the condensation did not require 7 h and indeed, shortening the reaction time to 2 h gave a similar yield of corrole 4b (10%).BF 3 •OEt 2 has also been tested as the acid catalyst for this reaction (entry 7).The A 3 -corrole was not detected among the reaction products, but the formation of the desired corrole 4b was also severely impeded (only 4%, even after 3 days of reaction).Using trichloroacetic acid (TCA) as the acid catalyst provided higher yields of corrole 4b (entry 8).Reaction with TCA in a concentration as low as ~1/75 (0.014 equiv.)compared to the concentration of the starting aldehyde 1d afforded corrole 4b in 17% yield.The reaction is rather sluggish and optimal yields are achieved only overnight.In this case, traces of the corresponding A 2 B 2 -porphyrin were also observed as well as a substantial amount of a red oligopyrrolic material.
Finally, prompted by the interesting results that were previously obtained on the oxidative Nalkylation of porphyrins, 16 it was reasoned that the same type of reaction could also be carried out with corroles.For that purpose, the synthesis of a meso-hydroxyphenyl substituted corrole 4f was required (Structure Block 1).However, procedures toward such corroles are hardly documented in the literature. 13,17An attempt to synthesize corrole 4f from commercial 3,5-di-tert-butyl-4hydroxybenzaldehyde (1l) and DPM 2i, using the optimized TCA-conditions, afforded only traces of the desired corrole (entry 11).As indicated by TLC, most of the DPM remained unreacted under these conditions.Usage of BF 3 •OEt 2 in a higher concentration and Lee's oxidation method lead to significant formation of the A 2 B 2 -(5f) and A 3 B-porphyrin.No corrole was detected in these experiments.In one case, with lowered concentrations of the starting compounds, the porphyrins were obtained in rather high yields, even though the DPM/aldehyde ratio should have strongly favoured corrole formation (entry 12).This turned actually out to be a good method for preparation of trans-A 2 B 2 -porphyrin 5f.The failure to obtain large amounts of the desired corrole 4f was first attributed to an oxidation of the di-tert-butyl-4-hydroxyphenyl substituent.Therefore aldehyde 1l was methylated to obtain 3,5-di-tert-butyl-4-methoxybenzaldehyde (1m). 18With this aldehyde, synthesis of the corresponding corrole 4g was attempted under the usual TCA-conditions.Unfortunately, corrole 4g was again obtained in a negligible quantity.Most of the starting materials remained unreacted under these conditions.Applying TFA as the acid catalyst (in a higher concentration) changed the picture (entry 13).Good yields of corrole 4g (up to 27%) could be isolated upon column chromatographic separation from a small amount of A 2 B 2 -porphyrin 5g.Demethylation of 4g (with boron tribromide or diisobutylaluminium hydride) to afford hydroxyphenyl-substituted corrole 4f was not successful.However, application of the obtained TFA-catalyzed conditions to the synthesis of corrole 4f afforded the desired A 2 B-corrole in 30% yield (entry 14).
Synthesis of the related meso-hydroxyphenylcorrole 4h has also been attempted under exactly the same conditions (entry 15).After a tedious chromatographic procedure, A 2 B-corrole 4h, as apparent from 1 H NMR, could be isolated in a yield as high as 34%.However, the product seems to be rather unstable and could not be kept in pure form.The apparent yield after the next column chromatographic purification (carried out 2 days after the previous yield estimation) was only 10%.
From the performed condensation reactions toward trans-A 2 B-corroles it can be concluded that TCA (and BF 3 •OEt 2 for 4a) was the acid catalyst of choice, giving best yields, for the synthesis of corroles 4a,b,c,d, starting from quite reactive aldehydes possessing electron-withdrawing substituents (and sterically hindered 5-(2,6-dichlorophenyl)dipyrromethane (2i)).On the other hand, TFA was preferred for the preparation of corroles 4f,g,h from less reactive substrates possessing electron-donating substituents (-OR).Direct oxidation of the tetrapyrromethane formed (without its isolation), with adding an excess of p-chloranil into the crude reaction mixture, was the preferred method for the final oxidative ring closure.
Once a good and reliable method for the synthesis of corrole 4f was established, it was possible to obtain enough of this product to serve as a substrate for the oxidative N-alkylation of the type described in one of our previous papers. 16The reaction was performed under the general Nalkylation conditions (Scheme 3).Within 15 min of the addition of the alkylating agent (ethyl bromoacetate (6)), the major spot on TLC was still that of the starting corrole 4f, without any spot similar to that of the oxidized intermediate in the analogous reaction with porphyrins.Running the reaction overnight changed the TLC picture into two (main) spots: one weaker, less polar greenbrownish spot with a weak red fluorescence (under 366 nm excitation) and another stronger, somewhat more polar, blue-greenish spot with a strong red fluorescence.Upon isolation of the two products by column chromatography, it became rapidly clear that two isomeric N-alkylated nonoxidized corroles 4i and 4j, in about 1/2 ratio, had been obtained (Scheme 3).

Scheme 3
Lack of porphyrin-type conjugation probably inhibits formation of the necessary intermediates on the oxidation pathway. 16The ability to introduce only one substituent per macrocycle molecule reflects the smaller size of the corrole cavity.The distribution of the substituent groups over the core nitrogen atoms (i.e. the ratio of the isomers) is in accordance with earlier published research on N-alkylation of corroles. 19It has been shown that substitution on N(22) of triarylcorroles results in a more crowded environment (as compared to the N(21)-substituted isomer), reflected in higher deformation of the corrole ring from planarity and of the meso-aryl groups from perpendicular orientation, rendering this isomer (4i in this case) less stable.

Conclusions
We have optimized a convenient and mild methodology for the synthesis of 5-aryldipyrromethanes in aqueous medium, needing little or no work-up (simple filtration from the reaction mixture).The synthetic route is based on the condensation of an aromatic aldehyde with a small excess of pyrrole (3 equiv.) in water as a solvent and catalyzed by HCl (0.18 M).This method appears to be quite general and by careful control of the reaction time, which is strongly dependent on the nature of the aldehyde, the dipyrromethanes can be obtained in a very selective way (essentially no tripyrromethane or other oligomeric side products are observed).This 'green' synthetic route toward dipyrromethanes in water, which we feel is undervalued at present, has been (and will be) an impulse for research toward oligopyrrolic macromolecules.The aryldipyrromethanes obtained by this efficient method were engaged in the synthesis of dipyrromethenes (toward BODIPYderivatives) and trans-A 2 B-corroles and the results obtained show essentially the same yields as for dipyrromethanes synthesized by standard literature methods.Several novel aryldipyrromethenes were synthesized in high yields.A number of trans-A 2 B-corroles have also been prepared by several methods.Some novel corroles were obtained and the yields of others, known from literature, have been improved.Trichloroacetic acid (in very low concentration) was the acid catalyst of choice, giving best yields, for the preparation of corroles from sterically hindered 5-(2,6dichlorophenyl)dipyrromethane and more reactive aldehydes possessing electron-withdrawing substituents (-C 6 F 5 , -C 6 H 4 NO 2 , -C 6 H 4 CO 2 Me), while the use of stronger trifluoroacetic acid (in higher concentration) was preferred for less reactive substrates with electron-donating substituents (-C 6 H 2 R 2 (OH), -C 6 H 2 R 2 (OCH 3 ), -C 6 H 4 OH).An attempt to perform oxidative N-alkylation of an appropriately substituted trans-A 2 B-corrole resulted in the formation of two isomeric N-alkylated non-oxidized corroles.

Experimental Section
General Procedures.The chemicals used for synthetic procedures were of reagent grade quality.They were obtained from commercial sources and used as received.NMR spectra were acquired on commercial instruments (Bruker Avance 300 MHz or Bruker AMX 400 MHz) and chemical shifts (δ) are reported in parts per million (ppm) referenced to tetramethylsilane (TMS) ( 1 H) or the carbon signal of deuterated solvents ( 13 C).Coupling constants (J) are given in Hz.Detailed 13 C NMR peak assignments were obtained by careful analysis of DEPT, HMQC and HMBC NMR spectra.Mass spectra were run using a HP5989A apparatus (CI and EI, 70 eV ionisation energy) with Apollo 300 data system, a Kratos MS50TC instrument for exact mass measurements (performed in the EI mode at a resolution of 10000) or a Micromass Quattro II apparatus (electrospray ionization (ESI), usual solvent mixture: CH 2 Cl 2 -MeOH + NH 4 OAc) with MASSLYNX data system.UV-Vis spectra were taken on a Perkin-Elmer Lambda 20 Spectrometer.Melting points (not corrected) were determined using a Reichert Thermovar apparatus.For column chromatography 70-230 mesh silica 60 (E.M. Merck) was used as the stationary phase.

General procedure for the synthesis of aryldipyrromethenes 3a,b,e,f,g
To 1 g of aryldipyrromethane 2a,b,g or 2e,f, respectively, dissolved in 50 mL of CH 2 Cl 2 , was added p-chloranil or DDQ (1 equiv.),respectively (Table 2), dissolved in 10 mL of CH 2 Cl 2 , and the reaction mixture was stirred at room temperature for the indicated time (30 min to 1 h, Table 2).The solvent was removed under reduced pressure and the residue was purified by column chromatography (silica) to afford aryldipyrromethenes 3a,b,g and 3e,f, respectively (yields in Table 2).5-Phenyl-4,6-dipyrrin (3a) 10a,c was characterized by its melting point, mass (EI) and NMR ( 1 H and 13 C) spectra and showed essentially the same values as those reported in literature.8a To 197 mg (0.67 mmol) of 5-(2,6-dichlorophenyl)dipyrromethane (2i) and 25 mg (0.13 mmol) of pentafluorobenzaldehyde (1h), stirred in 30 mL of CH 2 Cl 2 under an Ar atmosphere at room temperature for 15 min (the flask was wrapped in aluminium foil for light protection), was added 60 µL of a 10% solution of BF 3 •OEt 2 in CH 2 Cl 2 (0.048 mmol).After 1 h, the reaction was quenched with dilute aqueous NaOH and the reaction mixture was washed with water and dried over MgSO 4 .After filtration and evaporation to dryness, 70 mg (1.3 mmol) of NH 4 Cl was added to the crude residue and the mixture was dissolved in 100 mL of propionitrile.To this mixture, 100 mg (0.4 mmol) of p-chloranil was added and the reaction mixture was stirred overnight.The solvent was evaporated and 22 mg (22%) of corrole 4 was isolated from the mixture by column chromatography (silica, eluent CH 2 Cl 2 -petroleum ether, 1-1 to 1-4, with addition of 1% Et 3 N).Purple solid; Mp >300 °C; 1

5,15-Bis(2,6-dichlorophenyl)-10-(4-carboxyphenyl)corrole (4e).
To 257 mg (0.356 mmol) of corrole 4d, dissolved in 20 mL of EtOH, was added 350 µL of a 5 M aqueous solution of NaOH (1.75 mmol).The reaction mixture was stirred at room temperature and then heated at reflux temperature.As a spot of the starting material could still be observed on TLC after 1.5 h of reflux, another aliquot (350 µL, 1.75 mmol) of NaOH solution was added.After an additional h of reflux, the reaction mixture was cooled to room temperature, 25 mL of water was added and the whole mixture was poured directly into 15 mL of water containing 0.3 mL of conc.HCl (3.6 mmol).The precipitate formed was filtered through a Büchner funnel and rinsed with an abundant volume of water.Upon drying, 223 mg (88%) of corrole 4e was obtained.Purple solid; Mp >300 °C; 1 H NMR (DMSO-d 6 , 400 MHz) δ 13.17