Electrophilic substitution reactions of thiophene and thieno[2,3-b]thiophene heterocycles: a DFT study

Introduction

Electrophilic aromatic substitution (S_EAr) reactions (Scheme 1) occur when an aromatic ring reacts with an electrophile to generate a product in which one hydrogen atom is replaced by the electrophile [1, 2]. The S_EAr reaction often occurs in the presence of a catalyst [1, 3–7]. Aromatic compounds produced by S_EAr reactions are employed in agriculture, pharmacy and industry [6, 8–12].

Scheme 1:

S_EAr reaction mechanism (a) involving a benzene ring and a halogen; (b) at the α- and β-carbon atoms of a five-membered heterocycle and a halogen; (c) at the α- and β-carbon atoms of a 5:5 [2,3-b]-bicyclic heterocycle and a halogen. X and Y are heteroatoms.

The initial step of the S_EAr reactions of benzene involves an interaction between the benzene ring and the electrophile (E⁺) to form a π-complex [π-complex 1 in Scheme 1(a)] [1, 13]. The π-complex 1 reacts to generate a cationic intermediate, also known as the Wheland intermediate, which is an arenium ion; this is the rate-determining step. This cationic intermediate is called a σ-complex because of the formation of a new σ bond with benzene. The σ-complex is stabilised by resonance. Since the benzene ring loses its aromaticity, the σ-complex is less stable than the benzene ring [1]. Deprotonation takes place in the last step in order to restore the aromaticity of the benzene ring [14]. This happens through an attack of the counteranion of the E⁺ electrophile on the hydrogen atom of the ring [Scheme 1(a)] [1, 14].

The S_EAr reactions involving benzene derivatives are well-established. They are influenced by the electron-donating and electron-withdrawing groups attached to the aromatic ring [15]. The electron-donating and electron-withdrawing groups enhance and inhibit the S_EAr reaction by activating and deactivating the aromatic ring, respectively [2].

Other aromatic rings, such as five-membered, benzo-fused and 5:5 bicyclic heterocycles, also undergo S_EAr reactions. These heterocyclic compounds are important moieties and precursors which are useful mainly in drug discovery and material science [16], [17], [18], [19], [20], [21]. Five-membered heterocycles are described as π excessive [22]. The lone pair of electrons is delocalised; this completes the sextet of electrons which is required for aromaticity [22].

One of the simplest aromatic heterocycles has a five-membered ring with one heteroatom. The two carbon atoms where electrophilic substitution can take place are the α- and β-carbon atoms with respect to the heteroatom [Scheme 1(b)]. There are reports of high reactivities of the α-carbon atom of five-membered heterocycles, for example thiophene, pyrrole and furan rings [10, 17]. The synthesis of β-substituted five-membered heterocycle derivatives requires increasing the electron-withdrawing ability of a substituent at the α-carbon atom through the use of Lewis acid and protic acids [10, 17]. Belen’kii reported that the favourable site of electrophilic attack may be justified in terms of resonance stabilisation of the Lewis structure of the intermediate by the delocalisation of the positive charge [17]. There is greater resonance stabilisation of the σ-complex when substitution occurs at the α-carbon atom, as illustrated in Scheme 1(b). The trends in the S_EAr reactions of benzo-fused five-membered rings, such as indole, are predicted in a similar manner using resonance structures; this is shown in Scheme S1 in the Supplementary Information (SI) document.

Determining the favourable site for S_EAr reaction in 5:5 bicyclic systems using resonance structures is intriguing and complex. Scheme 1(c) shows that the stabilisation of the positive charge involves the disruption of the aromaticity of the adjacent five-membered ring for the α-attack but not for the β-attack. Hence, an electrophilic attack at the β-carbon atom is predicted through the resonance structures. However, experiments performed with 5:5 bicyclic heterocycles reported S_EAr reactions at the α- and α′-carbon atoms, instead of at the β-carbon atom. Bugge et al. carried out the bromination of thieno[2,3-b]thiophene and thieno[3,2-b]thiophene by using N-bromosuccinimide in glacial acetic acid and obtained the α-product [23]. N-Chlorosuccinimide (NCS) was reacted with thieno[2,3-b]thiophene in acetic acid to form the α-chlorothieno[2,3-b]thiophene and α,α′-dichlorothieno[2,3-b]thiophene products [24].

Such experimental results imply that resonance structures cannot be employed to determine the site which is the most prone to S_EAr reactions in the case of 5:5 bicyclic heterocycles.

In order to gain further insights into the preference of either α- or β-substitution in 5:5 bicyclic heterocycles, we performed computational mechanistic studies using the density functional theory (DFT) method. We investigated the mechanisms of the S_EAr reactions of a five-membered ring and a 5:5 bicyclic heterocycle through an analysis of the thermodynamic and kinetic parameters in the gas phase and in solution. NCS was used as the electrophile to determine the most preferred site in thiophene 1 and thieno[2,3-b]thiophene 2 for S_EAr reaction. NCS was selected as it was reported that N-halosuccinimide compounds are an effective source of electrophile for the direct halogenation of activated and deactivated aromatic systems [25], [26], [27], [28], [29]. Acetic acid was used as the solvent to mimic the experimental conditions used in previous reports []. The reactions investigated in this manuscript are shown in Scheme 2.

Scheme 2:

S_EAr reactions studied involving (a) thiophene 1 and (b) thieno[2,3-b]thiophene 2; (c) Geometry of NCS.

Methodology

The chlorination reactions of 1 and 2 were investigated with NCS using the B3LYP/6-311G(d) method. The B3LYP [33], [34], [35] functional was employed in several recent reports to study properties and mechanisms of heterocycles [36], [37], [38], [39], [40]. Full geometry optimisations were performed in the gas phase using the Gaussian 16 package [41] available in SEAGrid [42], [43], [44], [45]. The transition state (TS) was identified through vibrational frequency [46] and intrinsic reaction coordinate (IRC) [47] analyses. Ground state structures were determined through the absence of imaginary frequency. The relative electronic energies (ΔE) reported include the zero-point energy (ZPE) correction. The B3LYP/6-311G(d) method was employed based on the polarisable continuum model (PCM) [48] to perform bulk solvation with acetic acid. The GaussView 6.0 [49], Chemcraft 1.8 [50] and CYLview [51] software were used for the visualisation of molecules and reactions.

The computations were repeated by considering empirical dispersion (B3LYP-D3) [52] and by using the M06-2X [53] functional due to hydrogen bonding interactions. The trends observed were in agreement with the B3LYP/6-311G(d) method. Hence, only results obtained with the B3LYP/6-311G(d) method are reported in this paper; the results obtained with the B3LYP-D3/6-311G(d) and M06-2X/6-311G(d) methods are reported in the Supplementary Information.

Results and discussion

S_EAr reactions of thiophene

The S_EAr reaction of 1 may occur at the α- and β-carbon atoms. Unlike S_EAr of benzene [1, 13], we found that no cationic intermediate was formed in the reaction of 1. Two pathways were located for the α-S_EAr reaction, namely pathways A and B. Only one pathway was located for the β-S_EAr reaction (pathway C). The α- and β-S_EAr reaction pathways are initiated with a reactant complex (RC) which is formed by the interaction of 1 with NCS. This leads to a TS which forms a product complex (PC). The PC separates into the products (P). Figure 1 shows the potential energy surfaces (PESs) and the geometry of the TSs obtained for the α- and β-S_EAr reaction pathways of 1 with NCS. Table 1 lists the relative energy in terms of ΔE, ΔH and ΔG values of the RCs, TSs, PCs and Ps of the α- and β-S_EAr reaction pathways with respect to the separate reactants.

Fig. 1:

PESs of the 1 + NCS α- and β-S_EAr reactions. ΔE values are in kcal mol⁻¹. Bond distances are in Å. All values resulting from the solvent-phase computations are in parentheses.

Table 1:

ΔE, ΔH and ΔG, in kcal mol⁻¹, of the RCs, TSs, PCs and Ps of the α- and β-S_EAr reactions of 1 with NCS. The solvent-phase energy values are in parentheses.

	ΔE	ΔH	ΔG
Pathway A (α-SEAr reaction)
RC 1αA	−1.7 (−0.9)	−0.8 (+0.5)	+5.2 (+5.8)
TS 1αA	+58.5 (+57.2)	+59.0 (+58.1)	+69.5 (+67.3)
PC 1αA	−29.4 (−28.4)	−28.7 (−27.1)	−21.2 (−20.7)
P 1αA	−26.9 (−26.9)	−27.0 (−26.4)	−26.9 (−28.0)
Pathway B (α-SEAr reaction)
RC 1αB	−2.1 (−1.1)	−1.2 (+0.3)	+4.7 (+4.8)
TS 1αB	+58.2 (+56.8)	+58.6 (+57.8)	+69.3 (+67.0)
PC 1αB	−10.2 (−9.1)	−9.5 (−7.9)	−1.9 (−1.1)
P 1αB	−6.4 (−7.2)	−6.5 (−6.8)	−6.2 (−7.8)
Pathway C (β-SEAr reaction)
RC 1βC	−1.9 (−0.7)	−1.0 (+0.7)	+6.0 (+5.1)
TS 1βC	+65.5 (+63.8)	+66.0 (+64.9)	+76.2 (+73.6)
PC 1βC	−13.2 (−12.1)	−12.6 (−11.0)	−3.9 (−3.9)
P 1βC	−8.8 (−9.8)	−8.9 (−9.4)	−8.6 (−10.5)

The gas-phase pathway A of the α-S_EAr reaction starts with RC _1αA in which the oxygen atom of NCS interacts with the α-hydrogen atom of the thiophene ring at a distance of 2.407 Å (Figure S1 in Supplementary Information). This O‧‧‧H interaction is not stabilising enough such that the ΔE of RC _1αA (−1.7 kcal mol⁻¹; Table 1) is close to that of the separate reactants. RC _1αA leads to TS _1αA with a ΔE of +58.5 kcal mol⁻¹. The geometry of TS _1αA comprises of a four-atom ring which involves the carbon, chlorine, nitrogen and hydrogen atoms with bond distances as shown in Fig. 1(a). TS _1αA generates PC _1αA (ΔE = −29.4 kcal mol⁻¹) in which the chlorine atom from NCS substitutes the α-hydrogen atom of 1. This hydrogen atom bonds to the nitrogen atom of NCS. Moreover, an interaction between the β-hydrogen and the oxygen atoms occurs at a distance of 2.457 Å in PC _1αA (Figure S1).

RC _1αB in the pathway B of the gas-phase α-S_EAr reaction is also formed through an interaction of the β-hydrogen atom of 1 and the oxygen atom of NCS. The O‧‧‧H distance is 2.461 Å (Figure S1) and the ΔE is −2.1 kcal mol⁻¹ (Table 1). RC _1αB overcomes an activation barrier leading to TS _1αB in which a ring is formed as depicted in Fig. 1(b). Compared to TS _1αA , in TS _1αB , the oxygen atom of NCS interacts with the α-hydrogen atom of 1 at a distance of 2.010 Å. TS _1αB has a ΔE value of +58.2 kcal mol⁻¹ (Table 1), which is comparable to that of TS _1αA (+58.5 kcal mol⁻¹; Table 1) and thus, these two pathways are in competition. TS _1αB leads to PC _1αB where an interaction occurs between the nitrogen and β-hydrogen atoms at 2.709 Å. The relative energy of PC _1αB is −10.2 kcal mol⁻¹, which indicates that the formation of the O–H bond is less favoured than the N–H.

In pathway C of the β-S_EAr reaction in the gas phase, the oxygen atom of NCS interacts with the β-hydrogen atom of 1 in RC _1βC (Figure S1). The ΔE of RC _1βC is −1.9 kcal mol⁻¹ (Table 1). Contrary to the pathway A, the pathway C of the β-S_EAr reaction features a TS _1βC with a five-atom ring involving the carbon, chlorine, nitrogen, oxygen and hydrogen atoms [Fig. 1(b)]. A shorter distance (2.060 Å) occurs between the oxygen and hydrogen atoms of TS _1βC , compared to the nitrogen and hydrogen atoms in TS _1αA (2.356 Å; Fig. 1). However, the O‧‧‧H distance in TS _1βC is longer than in TS _1αB (2.010 Å), indicating that the O‧‧‧H interaction in TS _1βC is stronger than in TS _1αA and weaker than in TS _1αB . The ΔE value of TS _1βC is +65.5 kcal mol⁻¹, which is higher than that of the TSs of the α-S_EAr reaction. Hence, the α-S_EAr reaction is kinetically more favoured than the β-S_EAr reaction. TS _1βC leads to PC _1βC (−13.2 kcal mol⁻¹) in which the chlorine atom replaces the β-hydrogen atom of 1. The hydrogen atom bonds to the oxygen atom of NCS as observed in the optimised structure of PC _1αB . This is different from the PC _1αA where the N–H bond is formed. The nitrogen atom in PC _1βC forms a double bond with the neighbouring C atom and interacts with the α′-hydrogen atom (Figure S1).

The pathways of the two S_EAr reactions end with different products, P _1αA , P _1αB and P _1βC . The ΔH of P _1αA (−27.0 kcal mol⁻¹) is lower than that of P _1αB (−6.5 kcal mol⁻¹) and P _1βC (−8.9 kcal mol⁻¹); the pathway A is more exothermic than the pathways B and C. P _1αA has a relative Gibbs free energy of −26.9 kcal mol⁻¹, which is lower than that of P _1αB (−6.2 kcal mol⁻¹) and P _1βC (−8.6 kcal mol⁻¹). This difference shows that P _1αA is thermodynamically more stable than P _1αB and P _1βC . Thus, the pathway A of the α-S_EAr reaction is both kinetically and thermodynamically more favoured than the β-S_EAr reaction. Hence, electrophilic attack occurs at the α-position in 1 through pathway A. This observation is in line with the deduction which was reached through an analysis of the resonance structures in Scheme 1(b).

When the reactions were studied in acetic acid, the two pathways A and B were obtained for the α-S_EAr reaction and the pathway C was obtained for the β-S_EAr reaction, similar as in the gas phase. The geometries of the RCs and PCs are shown in Figure S2 in the Supplementary Information. On comparing the energy values of the solvent-phase reaction (Table 1), it was observed that the energy values are comparable.

S_EAr reactions of thieno[2,3-b]thiophene

The S_EAr reaction may occur at the α- and β-carbon atoms of 2. Two pathways (A and B) and one pathway (C) were located for the α- and β-S_EAr reactions, respectively. Figure 2 depicts the PESs of the α- and β-S_EAr reactions of 2 with NCS and the geometries of TS _2αA , TS _2αB and TS _2βC . Table 2 lists the relative energy in terms of ΔE, ΔH and ΔG of the RCs, TSs, PCs and Ps of the α- and β-S_EAr reactions, in kcal mol⁻¹, with respect to the separate reactants.

Fig. 2:

PESs of the 2 + NCS α- and β-S_EAr reactions. ΔE values are in kcal mol⁻¹. Bond distances are in Å. All values resulting from the solvent-phase computations are in parentheses.

Table 2:

ΔE, ΔH and ΔG, in kcal mol⁻¹, of the RCs, TSs, PCs and Ps of the α- and β-S_EAr reactions of 2 with NCS. The solvent-phase energy values are in parentheses.

	ΔE	ΔH	ΔG
Pathway A (α-SEAr reaction)
RC 2αA	−2.1 (+1.5)	−1.1 (+2.9)	+4.6 (+8.3)
TS 2αA	+54.0 (+54.2)	+54.6 (+55.3)	+65.2 (+64.2)
PC 2αA	−30.2 (−26.2)	−29.3 (−24.8)	−21.4 (−19.4)
P 2αA	−27.1 (−28.1)	−27.1 (−27.6)	−27.0 (−29.1)
Pathway B (α-SEAr reaction)
RC 2αB	−2.4	−1.5	+4.4
TS 2αB	+53.9	+54.4	+65.0
PC 2αB	−10.9	−10.2	−1.4
P 2αB	−6.6	−6.6	−6.3
Pathway C (β-SEAr reaction)
RC 2β	−2.4 (+1.1)	−1.4 (+2.5)	+4.5 (+8.4)
TS 2β	+61.0 (+60.6)	+61.6 (+61.8)	+71.5 (+70.8)
PC 2β	−13.1 (−9.7)	−12.3 (−8.4)	−3.6 (−1.3)
P 2β	−8.7 (−11.0)	−8.7 (−10.5)	−8.4 (−11.5)

In RC _2αA of the pathway A of the gas-phase α-S_EAr reaction, the oxygen atom of the NCS molecule interacts with the β- and β′-hydrogen atoms of 2 (Figure S3 in Supplementary Information) at distances of 2.673 and 2.678 Å, respectively. These O‧‧‧H interactions stabilise the RC _2αA with a relative electronic energy of −2.1 kcal mol⁻¹ (Table 2). The pathway A of the α-S_EAr reaction features TS _2αA (ΔE = +54.0 kcal mol⁻¹) in which the NCS electrophile interacts with 2 in a similar manner as with 1. The C–Cl–N–H ring is formed with an N‧‧‧H distance of 2.352 Å and a C‧‧‧Cl distance of 2.502 Å. The nitrogen atom completely abstracts the α-hydrogen atom in PC _2α . The oxygen atom interacts with the β- and β′-hydrogen atoms at distances of 2.795 and 2.779 Å, respectively. PC _2αA has a ΔE of −30.2 kcal mol⁻¹.

Contrary to RC _2αA , the oxygen atom of the NCS moiety in RC _2αB (ΔE = −2.4 kcal mol⁻¹; Table 2) interacts with the α-hydrogen atom of 2 at a distance of 2.505 Å (Figure S3) in pathway B in the gas phase. RC _2αB leads to TS _2αB in which the oxygen atom is also involved in the ring in the geometry of TS _2αB . TS _2αB has a ΔE of +53.9 kcal mol⁻¹, which is comparable to TS _2αA . TS _2αB leads to PC _2αB in which an O–H bond is formed and the nitrogen atom interacts with the β- and β′-hydrogen atoms at 2.905 and 2.961 Å, respectively. This PC has a higher ΔE (−10.9 kcal mol⁻¹) than PC _2αA .

The pathway C of the gas-phase β-S_EAr reaction starts with RC _2βC (−2.4 kcal mol⁻¹) in which an α′-H‧‧‧O interaction occurs at a distance of 2.467 Å (Figure S2 in Supplementary Information). RC _2βC overcomes an activation energy of +61.0 kcal mol⁻¹ to reach TS _2βC which is structurally comparable to TS _1βC . This activation energy is higher than that of the α-S_EAr reaction; the α-S_EAr reaction is kinetically more favoured than the β-S_EAr reaction. The oxygen atom interacts with the β-hydrogen atom at a distance of 2.116 Å in TS _2βC . The oxygen atom abstracts the β-hydrogen atom in PC _2βC . The nitrogen atom forms a double bond with the neighbouring carbon atom and interacts with the β′-hydrogen atom at a distance of 2.394 Å. PC _2βC (−13.1 kcal mol⁻¹) is higher in ΔE than PC _2αA (−30.2 kcal mol⁻¹) and lower in ΔE than PC _2αB (−10.9 kcal mol⁻¹).

Table 2 shows that the ΔH of P _2αA is lower than that of P _2αB and P _2βC ; therefore, the pathway A of the α-S_EAr reaction is more exothermic than the pathways B and C. The pathway A for the α-S_EAr reaction of 2 is also thermodynamically more favoured owing to the lower ΔG of P _2αA . The electrophilic attack is expected to occur at the α-carbon atom with NCS as the α-S_EAr reaction is both kinetically and thermodynamically more preferred than the attack at the β-carbon atom. The same trend was observed for the reaction of 1 with NCS. However, contrary to the reaction with 1, the preference for the α-S_EAr reaction does not correlate with the deduction which was reached through the resonance structures [Scheme 1(c)]. This highlights the importance of mechanistic studies to predict reaction outcomes.

On considering the effect of acetic acid as solvent, the pathway B of the α-S_EAr reaction was not located. The geometries of the RCs and PCs for pathways A and C are shown in Figure S4 in the Supplementary Information. Similar to the reactions of 1, the energy values computed in acetic acid are close to those computed in the gas phase. The trends observed also remain unchanged, that is, the reactants 2 and NCS prefer the α-S_EAr reaction as shown by the lower activation barrier of the pathway A (+54.2 kcal mol⁻¹) compared to the pathway C (+60.6 kcal mol⁻¹).

Conclusions

The S_EAr reactions of thiophene 1 and thieno[2,3-b]thiophene 2 with NCS were investigated to determine whether electrophilic attack occurs at the α- or the β-positions. The investigations were performed in the gas phase and in acetic acid. The Rs, RCs, TSs, PCs and Ps along the PESs of both S_EAr reaction mechanisms were optimised using the B3LYP/6-311G(d), B3LYP-D3/6-311G(d) and M06-2X/6-311G(d) methods. The trends observed remain unchanged with the different methods; hence, only the B3LYP/6-311G(d) results are reported. An analysis of the resonance structures shows that electrophilic attack is expected to occur at the α-carbon atom in 1 and at the β-carbon atom in 2. However, on comparing the energetic parameters, the pathway A of the α-S_EAr reaction was found to be kinetically and thermodynamically more favoured for both 1 and 2 in the gas phase and in acetic acid. Hence, chlorination with NCS will occur at the α-carbon atom for both 1 and 2, despite what is predicted by the resonance structures. This demonstrates how mechanistic studies help in predicting reaction outcomes. Further mechanistic studies involving the presence of substituents and different electrophiles would aid in deciphering general rules about S_EAr reactions of 5:5 bicyclic heterocycles. In the future, we will be applying the methodology used in this paper to examine other 5:5 fused heterocycles: with other heteroatoms, with heteroatoms in other positions, and with different heteroatoms in the two rings. A further extension will deal with 5:5 systems with more than two heteroatoms.