Influence of Steric Effect on the Pseudo-Multicomponent Synthesis of N-Aroylmethyl-4-Arylimidazoles

A pseudo-three-component synthesis of N-aroylmethylimidazoles 3 with three new C–N bonds formed regioselectively under microwave conditions was developed. Products were obtained by reacting two equivalents of aroylmethyl bromide (ArCOCH2Br, 1) with the appropriate amidine salt (RCN2H3.HX, 2) and with K2CO3 as a base in acetonitrile. The bicomponent reaction also occurred, giving the expected 4(5)-aryl-1H-imidazoles 4. Notably, the ratio of products 3 and 4 is governed by steric factors of the amidine 2 (i.e., R = H, CH3, Ph). Therefore, a computational study was carried out to understand the reaction course regarding product ratio (3/4), regioselectivity, and the steric effects of the amidine substituent group.

It is important to note that few studies have been reported to obtain imidazoles using aroylmethyl bromides 1 and amidines 2 [21,22] (Scheme 1a), specifically N-substituted derivatives that we even obtained by a pseudo-multicomponent reaction (pseudo-MCR) [24,26] (Scheme 1b). Likewise, there are few reports on imidazole derivatives synthesis based on pseudo-MCRs and MAOS. In a pseudo-MCR two or more components are identical, which could be a limitation regarding the scope and functional flexibility. However, these reactions have the advantage of being efficient in obtaining products with molecular complexity. In particular, we developed a pseudo-three-component synthesis of N-aroylmethylimidazoles 3a-d under microwave (MW) conditions in good yields by reacting two equivalents of 1 with acetamidine hydrochloride (2a) [24][25][26]. This reaction also allowed us to obtain the expected 1H-imidazoles 4 as minor products (Scheme 1b). Possibly, the better solubility in organic solvents and the major steric effect in the cyclization step from amidines 2b [21] and 2d [22] (Scheme 1a) vs. 2a [24][25][26] (Scheme 1b) direct the reaction course. From these interesting findings and in search of testing our hypothesis of the steric effect, we decided to carry out the reaction using benzamidine hydrochloride (2b) and formamidinium acetate (2c) to ensure a good size difference of the substituent at position 2 of the imidazole ring (Scheme 1c). In addition, we performed a detailed computational study to understand the reaction course regarding the obtained results in the regioselective synthesis of N-aroylmethyl-4-arylimidazoles 3a-l and the expected 4(5)-aryl-1H-imidazoles 4. In the same way, this theoretical study would validate the proposed mechanism in our previous work. Ultimately, the new compounds are imidazole antifungal analogs [24][25][26][27], thus possessing significant biological potential. For example, by simple carbonyl group reduction, the aroylmethyl group of 3e-l can be converted to the privileged 2-aryl-2hydroxyethyl moiety present in various antifungal agents [24,26,27].
Results obtained for the 2-phenylimidazoles 3e-h and 4f-g synthesis agree with the expected steric and solubility effects since the 2-phenyl-1H-imidazoles are favored with 2b while N-substituted products are preferred using acetamidine 2a. Possibly, the better solubility in acetonitrile of 2b with regard to 2a (seven vs. two carbon atoms) favors both its interaction with the substrate 1a-b and the equimolar reaction, which would not happen with 2a. Moreover, the steric effect in the cyclization step from amidines 2a or 2b (CH 3 vs. Ph [21], see Scheme 1) also favors 1H-imidazoles 4f-g over products 3e-h. Compounds 4e-h have been reported in the literature through other synthetic ways; thus, we considered it sufficient to isolate only products 4f and 4g to extrapolate and generalize the proportions of products. This was developed based on our previous work, where we managed to isolate 2-methyl-NH-imidazoles in approximate yields (around 20%). In general, the elution of NHimidazoles by column chromatography is tedious because they tend to become significantly adsorbed on the silica gel; the process may take an entire day. However, the pseudo-MCR is more important because it allows access to novel N-substituted imidazoles.
Once the 2-phenylimidazoles 3e-h and 4f-g were obtained, we carried out the reaction between two equivalents of 1a-d with one equivalent of formamidinium acetate (2c) under similar conditions. Although four products were formed, they did not correspond to the desired imidazoles but the aroylmethyl acetates 5a-d (Scheme 3a). These products were obtained by the reaction of the acetate group in 2c with 1a-d under these conditions. Thus, we increased the reaction temperature to 120 • C to favor the imidazoles' formation, possibly from 5a-d, with the formamidine in the reaction medium. Gratifyingly, the N-aroylmethylimidazoles 3i-l and 1H-imidazoles 4i-j were obtained in high yields and an estimated 7:2 ratio under these conditions (Scheme 3b). These results also agree with the steric effects; indeed, the N-substituted products are obtained in a higher proportion than when amidines 2a-b were used (3a-l/4a-l, H (3.5:1) vs. CH 3 (3:1) and Ph (1:1.25), Schemes 1b, 2 and 3b). Structures of all products obtained were characterized by 1 H and 13 C NMR studies and HRMS analysis (see the Experimental Section and Supporting Information for details). Regarding the imidazoles' synthesis 3 and 4 via the reaction of aroylmethyl bromide 1 with amidines 2, a plausible and general mechanism in Scheme 4 is depicted [24]. The N-substituted imidazoles 3a-l are formed by a regioselective pseudo-MCR of three steps (di-N-alkylation i/iib, cyclocondensation iii), while 1H-imidazoles 4 are obtained through a two-step process (N-alkylation i, cyclocondensation iia). It starts with a nucleophilic attack of the amidine 2 on substrate 1 (or the ketoester 5 in situ generated) to form the intermediate A, which can undergo a cyclocondensation reaction via the alcohol 4* to afford products 4 or a second N-alkylation with 1 (or 5) giving the cyclization intermediate B. This intermediate can then involve a cyclocondensation by attacking its nucleophilic nitrogen atom on a carbonyl group, with the later loss of a water molecule from 4*. It is important to emphasize that due to steric factors, the N-alkylation of A (intermolecular) or cyclocondensation of B (intramolecular) may be disadvantaged (Scheme 4). Despite the promising results in synthesizing imidazoles 3a-l and 4, we carried out some examples starting from benzamidine hydrochloride (2b) at 120 • C, as in the reaction with formamidine acetate (2c). In those reactions, microwave irradiation of reagents (substrates 3b/4-Cl and 3c/4-Br were tested) led to the formation of the expected imidazoles in poor yields, and the 3/4 product ratio turned out to be somewhat lower (Scheme 4 vs. Scheme 2). Possibly, the higher temperature promotes the reaction mixture deterioration; Scheme 4 shows photographs of the reaction mixtures at 100 • C and 120 • C, the one at 120 • C being a little darker due to its possible decomposition. These results were not observed using 2c because we believe the reaction occurs via ketoesters 5a-d as 1,2-biselectrophilic substrates; indeed, the obtention of these ketoesters led us to develop the reaction at 120 • C.
On only considering our experimental and literature research, it is possible to conclude that this reaction type is governed by the steric effect of the R group attached to the starting amidine. Studies have established that the smaller the R group in 2a-d, the more favored is the N-alkylation reaction of A (i.e., H > CH 3 > Ph > t-Bu) regarding the intramolecular cyclocondensation reaction; indeed, our new findings correspond with results reported in the literature [21,22,[24][25][26]. Remarkably, these reactions occur with high regioselectivity, and steric factors govern their course regarding yields and product ratio (imidazoles 3 vs. 4) (Scheme 5). It is important to mention that the formation of 3 from NH-imidazoles 4 was discarded because the N-alkylation of 4 under the established conditions did not take place; thus, a stronger base such as NaH in acetone must be used [24].
Intrigued by the observed reactivity, we carried out a computational DFT study to unveil its molecular determinants. The computational results are summarized in Scheme 6, Figure 2, and Table 1 (see the Supporting Information for more computational details). [a] Compounds type P 1 are 1H-imidazoles 4 and type P 2 N-substituted imidazoles 3.
[b] Ratio and approximate total yield of the mixture. Products type P 2 were only obtained by us under MW [24][25][26].

Scheme 6.
Schematic representation of the structure of critical points determined on path 1, yielding 1H-imidazoles (4, type P 1 ) and path 2, leading to N-substituted imidazoles (3, type P 2 ). We considered a model system including the phenacyl bromide (1') and the anion I' 1 (A in Scheme 4). We determined that the initial reactive species is I' 1 , deriving from the deprotonation of intermediate I 1 using the CO 3 2− present in the reaction medium. This anion has a negative and nucleophilic nitrogen atom (N3). The deprotonation at sp 2 nitrogen atom (N1) was ruled out on energy grounds (Scheme S1). The key step of the mechanism is the competence between the intramolecular cyclocondensation reaction (path 1: I' 1 → T S 1 → I 3 , route iia in Scheme 5) and the N-alkylation reaction (path 2: I' 1 → T S 2 → I 2 , route iib in Scheme 4), which eventually lead to 4-aryl-1H-imidazoles P 1 or N-substituted 4-aryl-1H-imidazoles P 2 (compounds 4 or 3 in Scheme 4), respectively (Scheme 6 and Figure 2).
We performed calculations of the transition states T S 1 and Ts2 with amidines, featuring different R groups to investigate steric effects. The results, reported in Table 1 together with the experimental ratio (%P 1 /%P 2 or 4 vs. 3), clearly indicate that the more cumbersome R, the more favored is path1, in agreement with experimental data. Analyzing the transition states geometries, a notable change in the N1-C2-N3-C8 dihedral (137 • to 105 • ) passing from H to t-Bu can be noted. These variations reflect the growing unfavorable steric interaction of the R group with the substrate 1', as the size of this substituent group increases (t-Bu > Ph > CH 3 > H in size).
In addition, we performed calculations of the transition states T S 1 and Ts2 with amidines featuring different R groups to investigate steric effects. The results, reported in Table 1 together with the experimental ratio (%P 1 /%P 2 or 4 vs. 3), clearly indicate that the more cumbersome R, the more favored is path1, in agreement with experimental data. Analyzing the transition state geometries, a notable change in the N1-C2-N3-C8 dihedral (137 • to 105 • ) passing from H to t-Bu can be noted. These variations reflect the growing unfavorable steric interaction of the R group with the substrate 1', as the size of this substituent group increases (t-Bu > Ph > CH 3 > H in size).
As expected, with lower P 1 /P 2 (or imidazoles 4 vs. 3) ratios (Entries 1 and 2), a small ∆∆G (T S 1−T S 2) was found. In particular, with R = CH 3 , the calculations predicted a very slight preference for P 1 (instead of the experimentally observed slight preference for P 2 ); we can consider this small discrepancy as an effect of the intrinsic DFT error. The different nature and solvent interaction of the two tested amidine salts (2a/chloride/CH 3 vs. 2c/acetate/H) may also cause this slight divergence. Importantly, a clear pattern is evidenced (the more cumbersome R, the more favored the P 1 formation), which facilitates the reaction control to obtain the desired compounds, 1H-imidazoles or N-substituted imidazoles (Table 1 and Figure 2b).
Concerning the intermediates following T S 1 and Ts2, we found that I 3 was higher in energy than I 2 for every combination of reactants considered, probably due to the constrained ring conformation and the negative charge on the oxygen atom. In the following steps, which are assumed not to be rate-limiting, I 3 is stabilized by protonation of O6 and the aromaticity engendered through a water molecule elimination. On the other hand, I 2 converts to I 4 through an intramolecular cyclization. Consequently, P 2 is obtained by the same final route proposed for the intermediate of cyclization I 3 , a cyclocondensation reaction with the loss of a water molecule (Scheme 6). Notably, these theoretical analyses allowed us to validate the proposed mechanism to obtain imidazoles 3 and 4 (Scheme 5).
An alternative pathway (path 3) to obtain the N-substituted imidazoles 3 (1,4-di-or 1,2,4-tri-substituted compounds, type P 2 ) involves the interaction of the nitrogen atom of the anion A (type I' 1 ) with the carbonyl group of the α-bromoketone (I' 1 → T S 3 → I 5 ), was also examined. We found that only with formamidine (R=H), is the initial transition state along this reaction channel, T S 3, competitive with T S 1 and T S 2. However, after analyzing the next steps involving second deprotonation and an intramolecular nucleophilic substitution, we could rule out path 3 on energy grounds (Scheme S2a). On the other hand, we also considered a pathway that would lead to the 1,5-di-or 1,2,5-tri-substituted regioisomers, type P 3 (Scheme S2b). Likewise, in this case, the calculated energies were higher than those of path 1, and consequently, this pathway is not competitive (Table S1, Supplementary Materials).

General
Synthesis and melting points. Reagents were purchased from commercial sources and used without further purification. All starting materials were weighed and handled in the presence of air at room temperature. Reactions were monitored by thin-layer chromatography (TLC) visualized by a UV lamp (254 nm or 365 nm). Flash chromatography was performed on silica gel (230-400 mesh). MW-assisted reactions were performed in a CEM Discover SP-focused microwave (ν = 2.45 GHz) reactor bearing a built-in pressure measurement sensor and a vertically focused IR temperature sensor. Sealed reaction vessels (10 mL, max pressure = 300 psi) containing a Teflon-coated stir bar (obtained from CEM) were used for these reactions. The temperature, power, and time settings were controlled and used for all reactions. Melting points were conducted using a capillary melting point apparatus and are uncorrected.

General Procedures
Synthesis of 4-aryl-2-phenylimidazoles 3e-h and 4f-g. A mixture of benzamidine hydrochloride (2b, 82 mg, 0.52 mmol), potassium carbonate anhydrous (K 2 CO 3 , 141 mg, 1.02 mmol), and the appropriate aroylmethyl bromide 1a-d (1.00 mmol) in acetonitrile anhydrous (1.0 mL) was subjected to microwave irradiation at 100 • C (150 W, monitored by an IR temperature sensor) and maintained at this temperature for 40 min in a sealed tube containing a Teflon-coated magnetic stirring bar. The resulting reaction mixture was cooled to 50 • C by airflow and was neutralized with dilute hydrochloric acid (HCl, 10%) and then partitioned between water (5.0 mL) and ethyl acetate (2 × 10.0 mL). The organic layer was washed with brine (2 × 5.0 mL) and dried over anhydrous magnesium sulfate (MgSO 4 ). Then, the solvent was removed under reduced pressure, and the residue purified by flash chromatography on silica gel (eluent: first CH 2 Cl 2 and then CH 2 Cl 2 /CH 3 OH 20:1 v/v).
Synthesis of 4-arylimidazoles 3i-l and 4i-j, and aroylmethyl acetates 5a-d. A mixture of formamidinium acetate (2c, 54 mg, 0.52 mmol), K 2 CO 3 anhydrous (141 mg, 1.02 mmol), and the appropriate substrate 1a-d (1.00 mmol) in acetonitrile anhydrous (1.0 mL) was subjected to microwave irradiation at 120 • C (150 W, monitored by an IR temperature sensor) and maintained at this temperature for 40 min in a sealed tube containing a Tefloncoated magnetic stirring bar. The reaction treatment and purification of the products were carried out in the same way as in the previous reaction. The first fraction eluted by flash chromatography contained N-aroylmethylimidazoles 3i-l (major product, 68-72%), while the second fraction contained compounds 4i-j (minor product, 17-22%). Aroylmethyl acetates 5a-d were obtained when this reaction was carried out at 100 • C under similar conditions; however, the residue obtained was purified by flash chromatography (eluent: CH 2 Cl 2 ) to afford the pure products 5a-d in high yields (82-87%).

Conclusions
To sum up, we synthesized a novel family of N-aroylmethyl-4-arylimidazoles 3e-l via a pseudo-three-component reaction that proceeds with the formation of three new C-N bonds in a regioselective manner. Products were obtained via the reaction of aroylmethyl bromides with amidine salts in a 2:1 ratio; however, the bicomponent reaction (1:1) that allows 1H-imidazoles 4 to be formed also occurred. This protocol provides high yields (up to 94% for the mixture of 3 and 4) using K 2 CO 3 as a base in acetonitrile under microwave irradiation conditions. Remarkably, the proportion of imidazoles obtained, 3 and 4, is governed by steric factors of the R group attached to the starting amidine, which together with the proposed reaction mechanism, was validated by DFT calculations; a clear pattern was evidenced: the smaller the R group, the more favored is the formation of the Nsubstituted products 3 (type P 2 ). In general, the computational and experimental studies were positively complemented to better explain and understand our interesting findings.
Supplementary Materials: Supporting information for this article (i.e., copies of NMR spectra, HRMS analysis, and computational details) can be downloaded. Scheme S1: Two possibilities deprotonation routes of I 1 using K 2 CO 3 ; Scheme S2: Possible pathways leading to N-substituted imidazoles P 2 and P 3 ; Table S1: Energies of key transition states in the discussed mechanism of reaction; Figure