Mechanisms of Cu/DMAP-cocatalyzed and DMAP-catalyzed C–N decarboxylative cross-coupling reactions

Abstract

A general and practical type of Cu/DMAP-cocatalyzed and DMAP-catalyzed C–N decarboxylative cross-coupling (DCC) reaction of carboxylic acids (or carboxylates) and azidoformates, which can construct both C(sp²)–N and C(sp³)–N bonds, was thoroughly investigated by density functional theory calculations. The Curtius rearrangements, i.e., extrusions of N₂, were found to be the rate-limiting steps for both Cu/DMAP-cocatalyzed and DMAP-catalyzed C–N DCC reactions. DMAP can facilitate the N₃⁻ transfer from the initial azidoformate species to the ensuing generated acyl azide intermediates. Then, the acyl azide intermediates undergo the Curtius rearrangements, overcoming a relatively low energy barrier. If DMAP was absent, the Curtius rearrangements would have to occur on the initial azidoformate species with energy barriers over 40.0 kcal mol⁻¹, which is not feasible at room temperature. Cu catalysts can further slightly facilitate the C–N DCC reactions.

---

Introduction

Decarboxylative cross-coupling (DCC) reactions have matured to become one of the most powerful approaches for the selective construction of C–C (carbon–carbon) and C–X (carbon–heteroatom) bonds, which is a vital tool for the synthesis of natural products, drugs, molecular devices, etc.¹ Meanwhile, conforming to the definition of Green Chemistry, DCC utilizes the readily available and environmentally benign carboxylate salts as nucleophile feedstock, rather than the expensive organometallic reagents which usually need to be pre-synthesized, and moreover, only non-toxic CO₂ is formed as a byproduct during the DCC reactions.^1,2

C–N bonds ubiquitously exist in natural products, biomolecules, pharmaceuticals and so on, and the efficient methods for the construction of C–N bonds have attracted a lot of interest.³ An array of methodologies to forge C(sp²)–N bonds, such as the Buchwald–Hartwig reaction,⁴ Ullmann coupling⁵ and Chan–Lam amination,⁶ have been reported, whereas the construction of C(sp³)–N bonds usually hinges on classical methods, including nucleophilic substitution and Mitsunobu alkylation, where the substrate species are quite limited.⁷ Unfortunately, the methods generalized for the formation of both C(sp²)–N and C(sp³)–N bonds are very scarce.

Recently, Lu and coworkers developed a Cu/DMAP co-catalyzed (DMAP, N,N-dimethylaminopyridine) C–N DCC reaction realizing the construction of both C(sp²)–N and C(sp³)–N bonds, in which naturally abundant carboxylic acids and easily synthesized azidoformates were utilized as carbon and nitrogen sources, respectively, as depicted in Scheme 1a.⁸ On the other hand, Hu's group demonstrated that an almost same C–N DCC reaction between carboxylates (carboxylic acid plus inorganic base) and azidoformates can occur under the catalysis of only DMAP (without Cu), as shown in Scheme 1b.⁹ The roles of Cu and DMAP catalysts in these C–N DCC reactions remain a mystery, and the mechanisms of the two types of C–N DCC reactions are also unknown, although envisioned mechanisms were proposed in the original experimental works.^8,9 In the present work, the mechanisms of the Cu/DMAP-cocatalyzed and DMAP-catalyzed C–N DCC reactions were investigated in depth by density functional theory (DFT) methods. Furthermore, to reveal the roles of Cu and DMAP in the C–N DCC reactions, a hypothetical DCC reaction without catalysts was also investigated, as shown in Scheme 1c. Different from the cases of Cu/DMAP-cocatalyzed and DMAP-catalyzed ones, the carbonyl motif (colored in black) of the product of the hypothetically catalyst-free C–N DCC reaction originates from the nitrogen source rather than the carbon source (colored in blue, see Scheme 1c). The mechanisms of Cu/DMAP-cocatalyzed, DMAP-catalyzed and catalyst-free C–N DCC reactions are depicted in Scheme 1a–c respectively. Noteworthily, the mechanisms proposed in the present work are quite different from the originally envisioned ones (see Scheme S1a and S1b in the ESI†). With respect to the Cu/DMAP-cocatalyzed C–N DCC reaction, the originally envisioned anhydride intermediate M1-SII (shown in Scheme S1a in the ESI†) is demonstrated not to be generated in the process of the reaction, and in its place, the ester intermediate M1-II and the complex intermediate M1-III consisting of acyl azide and carbonate are formed, as portrayed in Scheme 1a. With respect to the DMAP-catalyzed one, after the N₃⁻ attacks the anhydride intermediate, CO₂ is not extruded (M2-SII to M2-SIII, see Scheme S1b in the ESI†); instead, a carbonate intermediate M2-III₂ is formed, and the extrusion of CO₂ (M2-IV to prod⁻) follows the extrusion of N₂, i.e., the Curtius rearrangement, as shown in Scheme 1b. The mechanisms of the Cu/DMAP-cocatalyzed, DMAP-catalyzed, and hypothetically catalyst-free C–N DCC reactions are termed M1, M2, and M3, respectively.

Scheme 1 Mechanisms of (a) Cu/DMAP-cocatalyzed, (b) DMAP-catalyzed, and (c) hypothetically catalyst-free C–N DCC reactions.

Computational methods

All the calculations were conducted using the Gaussian 09 program package.¹⁰ Geometry optimizations were carried out, using the B3LYP functional¹¹ with the 6-31G(d,p) basis set for H, C, N, O and Cl, and the LanL2DZ basis set¹² with the corresponding effective core potential (ECP) for Cu (termed B3LYP/6-31G(d,p)∼Lan), and the self-consistent reaction field (SCRF) method with the SMD implicit solvation model¹³ and acetonitrile, the solvent primarily utilized in experiments, were utilized during calculations (denoted as level1). The same or similar methods were successfully used to investigate Cu-catalyzed reactions.¹⁴ Geometry structures are optimized without any constraints. Vibrational frequency analysis was performed at level1 and at 298 K to confirm each stationary point as a minimum or saddle point and to obtain the Gibbs free energies. Transition states were verified to be connected to the right reactants and products by internal reaction coordinate (IRC) calculations.¹⁵ Natural bond orbital (NBO)¹⁶ analysis was implemented at level1 of theory as the geometry optimization to determine the natural charge. Single-point calculations were also performed utilizing the B3LYP method and the 6-311+G(d,p) basis set for H, C, N, O and C, the SDD basis set with the corresponding ECP for Cu, and the same implicit solvation model was used for geometry optimization (termed B3LYP/6-311+G(d,p)∼SDD) on the structures optimized at level1 (denoted as level2). Energies and Gibbs free energies were calculated at level2 and level1, respectively, in the present work, unless otherwise stated. With regard to the carbonyl addition steps, single-point calculations with counterpoise corrections¹⁷ were also performed with B3LYP/6-311+G(d,p)∼SDD in the GAS phase to evaluate the basis set superposition errors (BSSE) of the H-bonded framework (denoted as level3); note that counterpoise corrections cannot be used with the implicit solvation model. The rate-limiting steps were also investigated with the wB97XD functional¹⁸ to evaluate the applicability of the B3LYP functional for the present work, as the B3LYP functional cannot properly estimate the weak dispersion interaction. The 3D diagrams of the structures were generated using the CYLview program.¹⁹

Results and discussion

The reactions yielding (dehydrogenated) 2,2,2-trichloroethyl phenylcarbamate (construction of the C(aryl)–N bond) and (dehydrogenated) phenyl methylcarbamate (construction of the C(sp³)–N bond) were selected as the two model reactions and designated as RA and RB (as well as RA′ and RB′) to investigate the mechanisms of the C(sp³)–N and C(aryl)–N DCC reactions (Scheme 2a),^8,9 because analogous C(sp³)–C and C(aryl)–C DCC reactions do not always follow the same mechanism. For instance, some C(aryl)–C DCC reactions occur in an ortho-coupling fashion, that is, carboxyl as a directing group for C–H activation,²⁰ as shown in Scheme 2b, and this ortho-coupling fashion is impossible for C(sp³)–C DCC reactions. The ortho-coupling pattern was also considered for C(aryl)–N DCC reactions in the present work. The products of the two model reactions, namely, (dehydrogenated) 2,2,2-trichloroethyl phenylcarbamate and (dehydrogenated) phenyl methylcarbamate, are designated as prod_RA (or prod_RA′⁻) and prod_RB (or prod_RB′⁻), respectively. The two kinds of reactant species, i.e., carboxylic acids (or carboxylates) as carbon sources and azidoformates as nitrogen sources, are labeled as reac1 (or reac1⁻) and reac2, respectively, and the two types of catalysts, namely, DMAP and Cu species, are termed cat1 and cat2, respectively.

Scheme 2 (a) Model reactions for C–N DCC reactions and (b) C(aryl)–C DCC reactions with a C–H activation ortho-coupling pattern.

Mechanisms of Cu/DMAP-cocatalyzed C–N DCC reactions

The mechanisms of Cu/DMAP-cocatalyzed C(aryl)–N and C(sp³)–N DCC reactions were investigated utilizing the model reactions A and B (Scheme 2a), respectively. The results indicate that different from some cases of C(aryl)–C DCC reactions following a C–H active ortho-coupling pattern (Scheme 2b),²⁰ both Cu/DMAP-cocatalyzed C(sp³)–N and C(aryl)–N reactions follow the same mechanism consisting of two catalytic cycles, namely, DMAP and Cu catalysis cycles, as shown in Scheme 1a.

Ligand substitution between the precatalyst species Cu(OAc)₂ and DMAP precedes the two catalytic cycles,^8,21 as illustrated in the inset of Scheme 1a, leading to the formation of the active catalyst species Cu(DMAP)²⁺, which is more reactive than Cu(OAc)₂ (vide infra). Note that Cu(OAc)₂ is the experimentally utilized Cu catalyst species.⁸

The energy profile for the Cu/DMAP-cocatalyzed C(aryl)–N DCC reactions is illustrated in Fig. 1. Initially, the N(pyridine) of DMAP (cat1) attacks the C(carbonyl) of the 2,2,2-trichloroethoxycarbonyl azide N₃COOTCE (reac2_RA; –TCE, trichloroethyl, –CH₂CCl₃), forming a ylide intermediate M1_RA-2, in which positive and negative charges are mainly accumulated on the N(pyridine) and O(carbonyl), respectively (see Fig. S1 and Table S1 in the ESI†). Then, the N₃⁻ motif is detached from M1_RA-2, generating the carbamate intermediate (plus the N₃⁻ anion) M1_RA-3. Subsequently, the other reactant benzoic acid PhCOOH (reac1_RA) participates in the reaction, and combines with M1_RA-3, furnishing a complex intermediate M1_RA-4. The O–H motif of PhCOOH (M1_RA-4₁) adds to the C=O double bond of M1_RA-4₂, forming the ester intermediate (plus the N₃⁻ anion) M1_RA-5, and analogous to the previous works, extra H₂O molecules can work as a “water bridge” to facilitate this carbonyl addition step²² (see Fig. 2 and S2–S4, Table S2 in the ESI†). The detached N₃⁻ anion then attacks the C(carbonyl) of M1_RA-5₂, accompanied by the heterolysis of the C(carbonyl)–O(carboxyl) and C(ester)–N(pyridine) bonds, leading to the complex intermediate M1_RA-6 consisting of the benzoyl azide M1_RA-6₁ and trichloroethyl hydrogen carbonate M1_RA-6₂ (plus DMAP). Note that compared with the initial reactant species, the N₃⁻ motif is transferred from the initial azidoformate species (reac2_RA) to the formed acyl azide species (M1_RA-6₁), which facilitates the following rate-limiting step, i.e., extrusion of N₂ (M1_RA-8 to M1_RA-9, vide infra). Subsequently, the dissociated DMAP abstracts the H(hydroxyl) of M1_RA-6₂, forming the carbonate anion (M1_RA-7₂) plus the benzoyl azide (M1_RA-7₁) and the DMAPH⁺ (the DMAP catalysis cycle is complete). One role of DMAP is to assist the N₃⁻ transfer.

Fig. 1 (Free, in parentheses) Energy profile for the Cu/DMAP-cocatalyzed C(aryl)–N DCC reactions. Top, DMAP catalysis cycle; bottom, Cu catalysis cycle. Grey line, supposed Cu(OAc)⁺ instead of Cu(DMAP)²⁺ as the catalyst.

Fig. 2 Optimized structures of M1_RA-TS3 (two extra H₂O molecules labelled with a cycle act as a “water bridge” and can facilitate the hydrogen transfer, and transition vectors are presented by green arrows).

After that, the Cu catalysis cycle starts. The active catalyst Cu(DMAP)²⁺ (cat2) takes part in the reaction, and M1_RA-7₁ and M1_RA-7₂ coordinate to the Cu(II) center of cat2, forming a tetracoordinated intermediate M1_RA-8. Then, the Curtius rearrangement, i.e., extrusion of N₂, occurs on M1_RA-8, during which the phenyl motif (Ph) of M1_RA-8 migrates from the C(carbonyl) to N(azide), and N₂ secedes from M1_RA-8, giving rise to another tetracoordinated intermediate M1_RA-9. Different from the cases of C–C DCC reactions in which the decarboxylation steps are usually the rate-limiting steps,²³ the Curtius rearrangement step is the rate-limiting step of the whole C(aryl)–N DCC reaction though, and the energy barrier is 27.6 kcal mol⁻¹, which is 10.9 kcal mol⁻¹ higher than that (16.7 kcal mol⁻¹) of the ensuing decarboxylation step (M1_RA-9 to M1_RA-10); this step was also investigated using the wB97XD functional, and the energy barrier is only 2.1 kcal mol⁻¹ higher than that of B3LYP, indicating the validity of the B3LYP functional for the present work. CO₂ is subsequently extruded from M1_RA-9, affording a tricoordinated intermediate M1_RA-10. The trichloroethoxyl ligand (–OTCE) of M1_RA-10 is grafted from the Cu(II) center to C(isocyanate), forming a dicoordinated intermediate M1_RA-11. The dissociative DMAPH⁺ donates H⁺ to N(isocyanate) of M1_RA-11, leading to the product precursor M1_RA-12. Ultimately, the catalyst Cu(DMAP)²⁺ is detached from M1_RA-12, and the final coupled product 2,2,2-trichloroethyl phenylcarbamate (prod_RA) is formed.

The reaction path following the originally envisioned mechanism (Scheme S1a in the ESI†) is not feasible, as depicted in Fig. S5 and S6 in the ESI.† Meanwhile, the C(sp³)–N DCC reaction was also investigated utilizing the model reaction B, and the results are presented in the ESI (see Fig. S7 and S8†).

Supposing Cu(OAc)⁺ instead of Cu(DMAP)²⁺ as the active catalyst

Moreover, the situation supposing Cu(OAc)⁺ rather than Cu(DMAP)²⁺ as the reactive catalyst is also considered (grey line of Fig. 1); noteworthily, due to the large steric congestion, the experimentally loaded Cu(OAc)₂ is very difficult to directly work as the active catalyst. In general, Cu(OAc)⁺ is significantly less active than Cu(DMAP)²⁺. The tetracoordinated intermediate M1_RA-8 formed from Cu(OAc)⁺ is 11.7 kcal mol⁻¹ less stable than that from Cu(DMAP)²⁺, and the underlying reason is twofold. On the one hand, the charge of the Cu(II) center of Cu(DMAP)²⁺ is more positive than that of Cu(OAc)⁺ as shown in Table S3,† and thus, compared with Cu(OAc)⁺, the attractive electrostatic interactions between the Cu(II) center of Cu(DMAP)²⁺ and O atoms of M1_RA-7₁ and M1_RA-7₂ are stronger. On the other hand, compared with Cu(OAc)⁺, the molecular orbital energy levels of Cu(DMAP)²⁺ are closer to those of M1_RA-7₁ and M1_RA-7₂ (summarized in Tables S3, S4 and S5 in the ESI†), resulting in that M1_RA-8 formed with Cu(DMAP)²⁺ has lower-energy bonding orbitals (see Fig. S9 in the ESI†).

The energy barrier (M1_RA-TS7 relative to M1_RA-9) for the extrusion of CO₂ step catalyzed by Cu(OAc)⁺ is 22.7 kcal mol⁻¹, which is 6.0 kcal mol⁻¹ higher than that (16.7 kcal mol⁻¹) of the same step catalyzed by Cu(DMAP)²⁺. Although the extrusion of CO₂ step is not the rate-limiting step, the energy barrier difference of 6.0 kcal mol⁻¹ still indicates that Cu(DMAP)²⁺ is more active than Cu(OAc)⁺. Houk's distortion-interaction energy analysis was utilized to analyze the energy barrier of the extrusion of CO₂ step,²⁴ and the structures of M1_RA-9 and M1_RA-TS7 were divided into five fragments, namely, CO₂, CuL (L = OAc⁻ or DMAP), DMAPH⁺, OTCE⁻, and PhNCO, as shown in Fig. S9 in the ESI.† The distortion-interaction analysis results are presented in Table 1, and the absolute distortion energy difference of the CuL fragment |ΔΔE_dist(CuL)| is 17.4 kcal mol⁻¹, which is significantly larger than those of the other four fragments (<1 kcal mol⁻¹), and the absolute interaction energy difference |ΔΔE_int| is significant (11.9 kcal mol⁻¹), indicating that the energy barrier difference |ΔΔE^‡| of 6.0 kcal mol⁻¹ mainly stems from the distortion energy of the CuL fragment and its interaction. Comparing the structure of M1_RA-9 and that of M1_RA-TS7, as for the DMAP/Cu(OAc)⁺-cocatalyzed case, the OAc⁻ ligand bidentately (η²-O) binds to the Cu(II) center in M1_RA-9 but monodentately (η¹-O) binds to Cu(II) in M1_RA-TS7, while with regard to the DMAP/Cu(DMAP)²⁺-cocatalyzed case, there is no significant structural variance of DMAP ligands in M1_RA-9 and M1_RA-TS7, as depicted in Fig. S10 in the ESI.† Therefore, Cu(DMAP)²⁺ is more reactive than Cu(OAc)⁺, and the other role of DMAP is to act as a coordinated ligand for the Cu catalyst species.

Table 1 Distortion (ΔE_dist) and interaction (ΔE_int) energy and energy barrier (E^‡) of the extrusion of CO₂ step (in kcal mol⁻¹)

	Cu(DMAP)^2+	Cu(OAc)^+	|ΔΔE|^a
a |ΔΔE| = |ΔE(Cu(DMAP)^2+) − ΔE(Cu(OAc)^+)|. b L = DMAP or OAc^−.
ΔEdist(CO2)	−53.8	−54.0	0.2
ΔEdist(CuL)^b	0.1	17.5	17.4
ΔEdist(DMAPH^+)	−0.3	−0.4	0.1
ΔEdist(OTCE^−)	−2.8	−3.0	0.2
ΔEdist(PhNCO)	0.5	1.3	0.8
∑ΔEdist	−56.3	−38.5	17.8
ΔEint	73.0	61.1	11.9
E ^‡	16.7	22.7	6.0

---

C–H activation ortho-coupling pattern

On the basis of the reported mechanism of C(aryl)–C DCC reactions in the literature,²⁵ the C–H activation ortho-coupling pattern (Scheme 2b) was also investigated with the model reaction A (Scheme 2a), and the energy profile is depicted in Fig. 4 (the mechanism for the C–H activation ortho-coupling pattern is labeled as M1′).

First, two ionized reactants, i.e., benzoate ions PhCOO⁻, coordinate to the Cu(II) center of the active catalyst Cu(DMAP)²⁺ (cat2), and a tetracoordinated intermediate M1′_RA-1 is formed, akin to the case of the C(aryl)–C DCC reaction.²⁵ With respect to M1′_RA-1, one PhCOO⁻ ligand monodentately (η¹-O) binds to the Cu(II) center, while the other PhCOO⁻ bidentately (η²-O) binds, as shown in Fig. 3 and S10 in the ESI.†M1′_RA-1 undergoes a concerted metalation-deprotonation step through M1′_RA-TS1, giving a chelate intermediate M1′_RA-2; in particular, a proton H⁺(ortho) of the monodentate PhCOO⁻ ligand transfers to the O(carboxyl) of the bidentate ligand, and the formed PhCOOH is dissociated. Then N₃COOTCE (reac2_RA) joins the reaction, and the O(carbonyl) of reac2_RA coordinates to the Cu(II) center of M1′_RA-2, yielding M1′_RA-3. Upon crossing the M1′_RA-TS2, the Curtius rearrangement occurs, and in particular, N₂ is expelled from M1′_RA-3, accompanied by the migration of the –OTCE motif from the C(carbonyl) to N and the Cu–O(carbonyl) bond dissociation, generating M1′_RA-4. This step is the rate-limiting step, and the relative energy barrier (39.5 kcal mol⁻¹, M1′_RA-TS2 relative to M1′_RA-3) is 11.9 kcal mol⁻¹ higher than that (27.6 kcal mol⁻¹, M1_RA-TS6 relative to M1_RA-8, see Fig. 1) of the rate-limiting step of the main path, and the overall energy barrier (42.9 kcal mol⁻¹, M1′_RA-TS2 relative to M1′_RA-2) is even 25.3 kcal mol⁻¹ higher. Then, the Cu–C(phenyl) of M1′_RA-4₁ breaks, and viaM1′_RA-TS3, M1′_RA-5 is formed. The C(isocyanate) of M1′_RA-5₂ binds to the deprotonated C(ortho) of M1′_RA-5₁, producing M1′_RA-6 through M1′_RA-TS4. The O(ether)–N bond of M1′_RA-6 breaks, and at the same time, the phenyl-containing motif migrates from the C(carbonyl) to N, generating M1′_RA-7viaM1′_RA-TS5. Upon passing over M1′_RA-TS6, the O of M1′_RA-7₂ binds to the C(isocyanate) of M1′_RA-7₁, and simultaneously, two protons H⁺ bind to N and C(ipso), forming M1′_RA-8. Ultimately, CO₂ as well as the active catalyst Cu(DMAP)²⁺ is released from M1′_RA-8viaM1′_RA-TS7, and the final product (prod_RA) is formed.

Fig. 3 (Free, in parentheses) Energy profile for the Cu/DMAP-cocatalyzed C(aryl)–N DCC reaction with a C–H activation ortho-coupling pattern.

The energy barrier of the rate-limiting step of the C–H activation ortho-coupling pattern is much higher than that of the main path; from the kinetic theory point of view, the ortho-coupling pattern is significantly less favorable than the main path. Meanwhile, under acid or neutral conditions, the main form of benzoic acid should be PhCOOH instead of PhCOO⁻, so it should be difficult to productively form M1′_RA-1. Therefore, the C–H activation ortho-coupling pattern can be safely excluded.

Mechanism of the DMAP-catalyzed (metal-free) C–N DCC reaction

The energy profile of the DMAP-catalyzed (metal-free) C–N DCC reaction is shown in Fig. 4. The first two steps (M2_RA′-1 to M2_RA′-3) are identical to those (M1_RA-1 to M1_RA-3, see Fig. 1) of the Cu/DMAP-cocatalyzed one. Different from the case of the Cu/DMAP-cocatalyzed DCC reaction, an inorganic base is used in the experiment, so the real carbon source species is the carboxylate anion rather than carboxylic acid. The PhCOO⁻ (reac_RA′⁻) combines with M2_RA′-3, furnishing the complex intermediate M2_RA′-4. The O(carboxyl) of M2_RA′-4₁ attacks C(carbonyl) of M2_RA′-4₂, affording M2_RA′-5. In contrast to the carbonyl addition step (M1_RA-4 to M1_RA-5 of Fig. 1) of the Cu/DMAP-cocatalyzed DCC reaction, the energy barrier (M2_RA′-TS3 relative to M2_RA′-4) of the carbonyl addition step of the DMAP-catalyzed one is merely 5.5 kcal mol⁻¹, and there is no chance to introduce extra H₂O molecules to act as a “water bridge”.

Fig. 4 (Free, in parentheses) Energy profile for the DMAP-catalyzed C–N DCC reactions.

The distortion-interaction energy analysis²⁴ is used to decipher the huge difference between the energy barrier (M2_RA′-TS3 relative to M2_RA′-4) of the carbonyl addition step catalyzed by DMAP and that (M1_RA-TS3 relative to M1_RA-4, Fig. 1) cocatalyzed by Cu/DMAP, and the structures are divided into three fragments, i.e., PhCOO⁻/PhCOOH (F1), DMAP-Troc⁺ (F2; –Troc, 2,2,2-trichloroethoxycarbonyl, –COOCH₂CCl₃) and N₃⁻ (F3), and the results are summarized in Table 2. Note that to maintain geometrical similarity, the carbonyl addition step cocatalyzed by Cu/DMAP without extra H₂O molecules (see Fig. S2 and S3a in the ESI†) was analyzed. The absolute distortion energy discrepancy of the PhCOO⁻/PhCOOH fragment |ΔΔE_dist(F1)| is as large as 95.3 kcal mol⁻¹, which is much larger than |ΔΔE_dist(F2)| and |ΔΔE_dist(F3)|. The |ΔΔE_int| is 50.9 kcal mol⁻¹ and is not large enough to offset the distortion energy. Therefore, the underlying reason for the discrepancy in the energy barriers of the carbonyl steps should originate from the strong O–H bond of PhCOOH, which is further demonstrated by a potential energy scan for O–H bond elongation (see Fig. S12 in the ESI†).

Table 2 Distortion (ΔE_dist) and interaction (ΔE_int) energy analysis and energy barrier (E^‡) of the carbonyl addition step catalyzed by DMAP and Cu/DMAP (in kcal mol⁻¹)

	DMAP	Cu/DMAP^a	|ΔΔE|^b
a The carbonyl addition step cocatalyzed by Cu/DMAP without extra H2O molecules was analyzed. b |ΔΔE| = |ΔE(Cu/DMAP) − ΔE(DMAP)|. c F1, F2 and F3 are three fragments, namely PhCOO^−/PhCOOH, DMAP-Troc^+ and N3^−.
ΔEdist(F1)^c	1.4	95.3	93.9
ΔEdist(F2)^c	17.0	14.3	2.7
ΔEdist(F3)^c	0.0	0.0	0.0
∑ΔEdist	18.4	109.6	91.2
ΔEint	−12.9	−63.8	50.9
E ^‡	5.5	45.8	42.3

---

Subsequently, upon crossing M2_RA′-TS4, DMAP is detached from M2_RA′-5, furnishing the anhydride intermediate (plus N₃⁻) M2_RA′-6. The dissociative N₃⁻ adds to the C(benzoate) of M2_RA′-6₁, and the C(benzoate)–O(anhydride) bond breaks, giving M2_RA′-7, which corresponds to M1_RA-7 in the Cu/DMAP-cocatalyzed C–N DCC reaction (see Fig. 1). Then, the rate-limiting step, namely, the Curtius rearrangement takes place on M2_RA′-7₁, during which the phenyl group (Ph) of M2_RA′-7₁ migrates from C(carbonyl) to N, and at the same time, N₂ is detached from M2_RA′-7₁, and viaM2_RA′-TS6, M2_RA′-8 is yielded. The energy barrier of the Curtius rearrangement step is 30.1 kcal mol⁻¹, which is 2.5 kcal mol⁻¹ higher than that (27.6 kcal mol⁻¹, M1_RA-TS6 relative to M1_RA-8 of Fig. 1) of the Cu/DMAP-cocatalyzed C–N DCC reaction, indicating that the Cu catalyst can slightly decrease the energy barrier of the Curtius rearrangement. Upon passing over M2_RA′-TS7, the –OTCE group of M2_RA′-8₂ transfers to M2_RA′-8₁, and CO₂ is released, forming the final product (prod_RA′⁻). The reaction path via the originally envisioned mechanism (Scheme S1b in the ESI†) is not feasible, as illustrated in Fig. S13 in the ESI.†

Influences of neutral and alkaline conditions on Cu/DMAP-cocatalyzed and DMAP-catalyzed (metal-free) C–N DCC reactions

Note that the Cu/DMAP-cocatalyzed C–N DCC reaction uses carboxylic acid (without inorganic base) as the carbon source (neutral condition),⁸ whereas the DMAP-catalyzed one used carboxylate (carboxylic acid plus inorganic base) as the carbon source (alkaline condition) in experiments,⁹ as shown in Scheme 1a and b. Therefore, there is a question whether the Cu/DMAP-cocatalyzed C–N DCC reactions can proceed under alkaline conditions, and whether the DMAP-catalyzed ones can proceed under neutral conditions. Mechanisms of the supposed Cu/DMAP-cocatalyzed C–N DCC reactions using carboxylate as the carbon source (alkaline condition) and DMAP-catalyzed ones taking carboxylic acid (neutral condition) as the carbon source are also investigated with the model reactions A′ and A (see Scheme 2a), respectively.

The energy profile of the Cu/DMAP-cocatalyzed C–N DCC reaction assuming carboxylate instead of carboxylic acid as the carbon source (alkaline conditions) is shown in Fig. S14 in the ESI.† The DMAP catalysis cycle (M1_RA′-1 to M1_RA′-7 of Fig. S14 in the ESI†) is identical to the early stage steps (M2_RA′-1 to M2_RA′-7 of Fig. 4) of the DMAP-catalyzed C–N DCC reaction, while the Cu catalysis cycle (M1_RA′-7 to M1_RA′-12 of Fig. S14 in the ESI†) is analogous to that (M1_RA-7 to M1_RA-13 of Fig. 1) of the main path of the Cu/DMAP-cocatalyzed one. The energy barrier of the rate-limiting step (M1_RA′-8 to M1_RA′-9 of Fig. S14 in the ESI†), namely, the Curtius rearrangement, is 27.2 kcal mol⁻¹, which is almost identical to that (27.6 kcal mol⁻¹, M1_RA-TS6 relative to M1_RA-8 of Fig. 1) of the main path of the Cu/DMAP-cocatalyzed one, indicating that carboxylate is almost equivalent to carboxylic acid as the carbon source for the Cu/DMAP-cocatalyzed C–N DCC reaction.

With regard to the DMAP-catalyzed C–N DCC reaction assuming carboxylic acid rather than carboxylate (neutral condition) as the carbon source, there are two possible routes. In one route, the carboxylate is dissociated from carboxylic acid, and hence, the real carbon source is still the dissociated carboxylate, and then the mechanism is the same as that (see Scheme 1b and Fig. 4) of the main path of the DMAP-catalyzed C–N DCC reaction. However, given the fact that the degrees of dissociation of carboxylic acid are very low,²⁶ this route should not be feasible. In the other route, the carboxylic acid acts as the real carbon source, and the energy profile is portrayed in Fig. S15 in the ESI.† The front steps (M2_RA-1 to M2_RA-7 of Fig. S15 in the ESI†) are identical to those (M1_RA-1 to M1_RA-7 of Fig. 1) of the main path of the Cu/DMAP-cocatalyzed C–N DCC reaction, but the late-stage steps (M2_RA-7 to M2_RA-10 of Fig. S15 in the ESI†) are different. The energy barrier (M2_RA-TS6 relative to M2_RA-7 of Fig. S15 in the ESI†) of the rate-limiting step, i.e. the Curtius rearrangement, is 30.8 kcal mol⁻¹, which is slightly higher than that (30.1 kcal mol⁻¹, M1_RA′-TS6 relative to M2_RA′-7 of Fig. 4) of the main path of the DMAP-catalyzed C–N DCC reaction, indicating that carboxylic acid is also nearly equivalent to carboxylate as the carbon source for the DMAP-catalyzed DCC reactions.

These results suggest that neutral and alkaline conditions are almost equally favorable, with alkaline conditions being a little superior for both the Cu/DMAP-cocatalyzed and DMAP-catalyzed C–N DCC reactions.

Mechanism of the supposed catalyst-free C–N DCC reactions

The supposed catalyst-free C–N DCC reaction (Scheme 1c) was also investigated in parallel with the model reaction A′ (Scheme 2a), aiming to further elucidate the roles of Cu and DMAP catalyst species, and the energy profile is depicted in Fig. 5.

Fig. 5 (Free, in parentheses) Energy profile for the supposed catalyst-free C–N DCC reactions.

In contrast to the cases of Cu/DMAP-cocatalyzed and DMAP-catalyzed C–N DCC reactions, due to the fact that DMAP is absent in this supposed reaction, the N₃⁻ cannot transfer from the initial azidoformate species to the ensuing formed acyl azide species, and therefore, the Curtius rearrangement has to occur on the initial azidoformate species reac2_RA′ instead of the acyl azide species (M1_RA-7₁ of Fig. 1, M1_RB-7₁ of Fig. S7 in the ESI,†M2_RA′-7₁ of Fig. 4 and M2_RB′-7₁ of Fig. S16 in the ESI†), and the alkoxyisocyanate intermediate M3_RA′-2 is formed, accompanied by the extrusion of N₂. The energy barrier is as high as 40.1 kcal mol⁻¹, which is about 10.0 kcal mol⁻¹ higher than those occurring on acyl azide species. The –OTCE group of M3_RA′-2 migrates from N to C(isocyanate). Then, the PhCOO⁻ (reac1_RA′⁻) takes part in the reaction, forming the dehydrogenated iminodicarbonate intermediate M3_RA′-3. Subsequently, the –NCOOTCE motif of M3_RA′-3 migrates from O(carboxyl) to the benzene ring, generating M3_RA′-4viaM3_RA′-TS3. This step is the rate-limiting step, and the energy barrier is as large as 44.5 kcal mol⁻¹, and it is impossible to overcome such a high energy barrier at room temperature. In M3_RA′-4, N binds to the C(ipso) and C(ortho) of the benzene ring. Upon crossing M3_RA′-TS4, CO₂ is detached, and the final product prod_RA′⁻ is formed.

Compared with the Cu/DMAP-cocatalyzed and DMAP-catalyzed ones, the energy barriers of the catalyst-free C–N DCC reactions are much higher, and such high energy barriers are very difficult to be overcome. Hence, the catalyst-free C–N DCC reaction cannot take place at room temperature.

Curtius rearrangement

The rate-limiting steps of both Cu/DMAP-cocatalyzed and DMAP-catalyzed C–N DCC reactions (as well as the supposed Cu/DMAP-cocatalyzed with C–H activation ortho-coupling pattern one) are the Curtius rearrangements, depicted in Fig. 6a. The N₃⁻ anion can feasibly transfer from the initial azidoformate species (reac2, see Scheme 1) to the ensuing acyl azide intermediates (M1_RA-6₁ of Fig. 1, M1_RB-6₁ of Fig. S7 in the ESI,†M2_RA′-7₁ of Fig. 4, and M2_RB′-7₁ of Fig. S16 in the ESI†) insofar as DMAP catalyst species was loaded. As thus, the Curtius rearrangements occur on the acyl azide intermediates, and the energy barriers of the Curtius rearrangements are about 30.0 kcal mol⁻¹ (30.1, 29.3, 27.6 and 28.9 kcal mol⁻¹), as presented in Table 3. In contrast, if DMAP was not added, the Curtius rearrangement would have to occur on the initial azidoformate species (M3_RA′-1 to M3_RA′-2 of Fig. 5 and M3_RB′-1 to M3_RB′-2 of Fig. S16 in the ESI†), and the energy barriers are about 40.0 kcal mol⁻¹ (40.1 and 37.6 kcal mol⁻¹), as listed in Table 3, which is about 10.0 higher than those on the acyl azide intermediates. Note that as for the supposed C–H activation ortho-coupling pattern, the Curtius rearrangement (M1′_RA-3 to M1′_RA-4 of Fig. 3) also takes place on the azidoformate species, though DMAP catalyst species exist, and the energy barrier is also as high as 39.5 kcal mol⁻¹.

Fig. 6 (a) Curtius rearrangement and (b) the highest occupied molecular orbitals of M3_RA′-1 and M3_RB′-1.

Table 3 Energy barriers (E^‡, in kcal/mol) of the Curtius rearrangements, natural charges (q(₁N) and q(₁X) in e) on ₁N and ₁X (X = O or C), distance (d(₁N–₁X) in Å) and electrostatic force (F(₁N–₁X) × 10⁻⁹ in N) between ₁N and ₁X

	E ^‡	q(1N)	q(1X)	d(1N–1X)	F(1N–1X)
M3RA′-1	40.1	−0.395	−0.540	2.211	1.005
M3RB′-1	37.6	−0.398	−0.524	2.208	0.986
M1′RA-3	39.5	−0.396	−0.530	2.208	0.993
M2RA′-7	30.1	−0.392	−0.158	2.246	0.244
M2RB′-7	29.3	−0.406	−0.803	2.245	1.279
M1RA-8	27.6	−0.371	−0.151	2.402	0.224
M1RB-8	28.9	−0.381	−0.788	2.422	1.182

---

The large discrepancies between the energy barriers of the Curtius rearrangement occurring on the acyl azide species and those on azidoformates stem from the bond type. As shown in Fig. 6a, during the Curtius rearrangement, the ₁N–₂N bond breaks, and ₁N inserts into the C*–₁X bond. As for the azidoformates (X = O) M3_RA′-1 and M3_RB′-1 (see Fig. S16 in the ESI†) as well as M1′_RA-3, one of the lone pair p-orbitals of ₁O atoms overlaps with the π orbital of C*=O, such that a p–π conjugation on ₁O–C*=O is formed. This way, the C*–₁O bonds of azidoformates are of partial double bond characteristics, and are difficult to break, and the energy barriers are higher. As shown in Fig. 6b, the C*–₁O bond lengths of M3_RA′-1 and M3_RB′-1 are 1.343 and 1.346 Å, respectively, which are much shorter than the adjacent C_R–₁O bonds (1.431 and 1.407 Å), and also shorter than the C–O bond (1.415 Å) of the optimized benchmark molecule methoxymethane, as shown in Fig. S16 in the ESI,† indicative of the partial C*–₁O double bond, which is further corroborated by the π orbital characteristics of the highest occupied molecular orbital (HOMO) on C*–₁O, as shown in Fig. 6b. Meanwhile, during the Curtius rearrangement, the repulsive electrostatic forces between ₁N and ₁X must be overcome to form the ₁N–₁X bond. Cu catalyst species can slightly reduce the charges on ₁N and ₁X and the repulsive electrostatic forces between ₁N and ₁X. As shown in Table 3, the electrostatic forces for M1_RA-8 and M1_RB-8 are slightly smaller than those for M2_RA′-7 and M2_RB′-7, and this should be the reason for the fact that the energy barriers of the Curtius rearrangement in the Cu/DMAP-cocatalyzed C–N DCC reactions are marginally smaller than those of the DMAP-catalyzed ones.

Conclusions

The mechanisms of the Cu/DMAP-cocatalyzed and DMAP-catalyzed C–N DCC reactions using carboxylic acids (or carboxylates) and azidoformates as the carbon and nitrogen sources, respectively, have been investigated in depth by DFT methods. In contrast to some cases of C(aryl)–C DCC reactions conforming to a C–H activation ortho-coupling mechanism, both the C(aryl)–N and C(sp³)–N DCC ones follow the same mechanism. Extrusions of N₂, namely, the Curtius rearrangements, instead of extrusions of CO₂, which are usually the rate-limiting steps for C–C DCC reactions, are the rate-limiting steps for both the Cu/DMAP-cocatalyzed and DMAP-catalyzed C–N DCC reactions. The DMAP catalyst is a requisite for these two types of C–N DCC reactions and can assist the transfer of N₃⁻ from the initial azidoformate species to the following formed acyl azide intermediates, and the Curtius rearrangements occur on the acyl azide intermediates with a relatively low energy barrier. If the DMAP catalyst was not utilized, the initial azidoformate would have to undergo the Curtius rearrangement with a very high energy barrier, which is impossible to be overcome at room temperature. The Cu catalyst can further slightly reduce the energy barrier of Curtius rearrangements, and in addition to assist the N₃⁻ transfer, DMAP also acts as a ligand for the active Cu catalyst species. The calculated results indicate that alkaline conditions may be a little more favorable than neutral conditions for both Cu/DMAP-cocatalyzed and DMAP-catalyzed C–N DCC reactions.

Abbreviations

DCC	Decarboxylative cross-coupling
DMAP	N,N-Dimethylaminopyridine
M1	Mechanism of the C–N DCC reaction co-catalyzed by Cu/DMAP
M1′	Mechanism of the C–N DCC reaction with the C–H activation ortho-coupling pattern co-catalyzed by Cu/DMAP
M2	Mechanism of the C–N DCC reaction catalyzed by DMAP
M3	Mechanism of the non-catalytic C–N DCC reaction
–TCE	Trichloroethyl, –CH2CCl3
–Troc	2,2,2-Trichloroethoxycarbonyl, –COOCH2CCl3
Level1	SMD&B3LYP/6-31G(d,p)∼Lan
Level2	SMD&B3LYP/6-311+G(d,p)∼SDD//SMD&B3LYP/6-31G(d,p)∼Lan
Level3	counterpoise&gas&B3LYP/6-311+G(d,p)∼SDD//SMD&B3LYP/6-31G(d,p)∼Lan

Data availability

The data underlying this study are available in the published article and its ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Science Foundation of China (Grant No. 22363009), the Natural Science Foundation of Inner Mongolia Autonomous Region (Grant No. 2021BS02020 and 2022MS02012), the Program for Innovative Research Team in Universities of Inner Mongolia Autonomous Region (Grant No. NMGIRT2302), the Fundamental and Applied Basic Research Project for Hohhot (Grant No. 2024-GUI-JI-25), the Fundamental Research Funds for Inner Mongolia Normal University (Grant No. 2022JBQN086), the Program for Restructured Inner Mongolia Key Laboratory of Green Catalysis (Grant No. 2060404), and the Department of Education of Inner Mongolia Autonomous Region.

References

1. (a) N. Rodriguez and L. Goossen, Decarboxylative Coupling Reactions: A Modern Strategy for C–C-bond Formation, Chem. Soc. Rev., 2011, 40, 5030–5048; (b) A. Varenikov, E. Shapiro and M. Gandelman, Decarboxylative Halogenation of Organic Compounds, Chem. Rev., 2021, 121, 412–484; (c) J. D. Weaver, A. R. Recio, A. J. Grenning and J. A. Tunge, Transition Metal-Catalyzed Decarboxylative Allylation and Benzylation Reactions, Chem. Rev., 2011, 111, 1846–1913; (d) Y. Wei, P. Hu, M. Zhang and W. Su, Metal-Catalyzed Decarboxylative C-H Functionalization, Chem. Rev., 2017, 117, 8864–8907.
2. (a) P. Anastas and N. Eghbali, Green Chemistry: Principles and Practice, Chem. Soc. Rev., 2010, 39, 301–312; (b) M. Simon and C. Li, Green Chemistry Oriented Organic Synthesis in Water, Chem. Soc. Rev., 2012, 41, 1415–1427; (c) T. Dalton, T. Faber and F. Glorius, C–H Activation: Toward Sustainability and Applications, ACS Cent. Sci., 2021, 7, 245–261.
3. (a) M. Taillefer and D. Ma, Amination and Formation of sp² C–N Bonds, Springer, 2013; (b) A. Ricci, Amino Group Chemistry: From Synthesis to the Life Science, Weinheim: Wiley-VCH, 2007; (c) H. G. Cutler and S. J. Cutler, Biologically Active Natural Products: Agrochemicals, CRC Press, 1999; (d) G. Evano, N. Blanchard and M. Toumi, Copper-Mediated Coupling Reactions and Their Applications in Natural Products and Designed Biomolecules Synthesis, Chem. Rev., 2008, 108, 3054–3131; (e) J. Bariwal and E. V. Van der Eycken, C–N bond Forming Cross-Coupling Reactions: An Overview, Chem. Soc. Rev., 2013, 42, 9283–9303; (f) M. Jouny, J. Lv, T. Cheng, B. H. Ko, J. J. Zhu, W. A. Goddard III and F. Jiao, Formation of Carbon–Nitrogen Bonds in Carbon Monoxide Electrolysis, Nat. Chem., 2019, 11, 846–851.
4. (a) A. S. Guram and S. L. Buchwald, Palladium-Catalyzed Aromatic Aminations with in situ Generated Aminostannanes, J. Am. Chem. Soc., 1994, 116, 7901–7902; (b) F. Paul and J. F. Hartwig, Palladium-Catalyzed Formation of Carbon-Nitrogen Bonds, Reaction Intermediates and Catalyst Improvements in the Hetero Cross-Coupling of Aryl Halides and Tin Amides, J. Am. Chem. Soc., 1994, 116, 5969–5970; (c) P. Ruiz-Castillo and S. L. Buchwald, Applications of Palladium-Catalyzed C–N Cross-Coupling Reactions, Chem. Rev., 2016, 116, 12564–12649; (d) R. Dorel, D. C. P. Grugel and A. M. Haydl, The Buchwald–Hartwig Amination After 25 Years, Angew. Chem., Int. Ed., 2019, 58, 17118–17129.
5. C. Sambiagio, S. P. Marsden, A. J. Blacker and P. C. McGowan, Copper Catalysed Ullmann Type Chemistry: From Mechanistic Aspects to Modern Development, Chem. Soc. Rev., 2014, 43, 3525–3550.
6. (a) P. Y. S. Lam, Chan–Lam Coupling Reaction: Copper-promoted C–Element Bond Oxidative Coupling Reaction with Boronic Acids, in Synthetic Methods in Drug Discovery, Royal Society of Chemistry, Cambridge, 2016, vol. 1, pp. 242–273; (b) M. J. West, J. W. B. Fyfe, J. C. Vantourout and A. J. B. Watson, Mechanistic Development and Recent Applications of the Chan–Lam Amination, Chem. Rev., 2019, 119, 12491–12523.
7. (a) R. N. Salvatore, C. H. Yoon and K. W. Jung, Synthesis of Secondary Amines, Tetrahedron, 2001, 57, 7785–7811; (b) K. C. K. Swamy, N. N. B. Kumar, E. Balaraman and K. V. P. P. Kumar, Mitsunobu and Related Reactions: Advances and Applications, Chem. Rev., 2009, 109, 2551–2651.
8. Y. Zhang, X. Ge, H. Lu and G. Li, Catalytic Decarboxylative C-N Formation to Generate Alkyl, Alkenyl, and Aryl Amines, Angew. Chem., Int. Ed., 2021, 60, 1845–1852.
9. J. Zhang, Y. Hou, Y. Tang, J. Xu, Z. Liu, G. Yang and X. Hu, Transition-metal-free decarboxylative ipso amination of aryl carboxylic acids, Org. Chem. Front., 2021, 8, 3434–3439.
10. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, et al., Gaussian 09, Gaussian, Inc., Wallingford, CT, 2009.
11. (a) A. D. Becke, Density–Functional Thermochemistry. III. The Role of Exact Exchange, J. Chem. Phys., 1993, 98, 5648–5652; (b) C. Lee, W. Yang and R. G. Parr, Development of the Colle-Salvetti Correlation-Energy Formula into a Functional of the Electron Density, Phys. Rev. B, 1988, 37, 785.
12. (a) P. J. Hay and W. R. Wadt, Ab initio Effective Core Potentials for Molecular Calculations. Potentials for the Transition Metal Atoms Sc to Hg, J. Chem. Phys., 1985, 82, 270–283; (b) W. R. Wadt and P. J. Hay, Ab initio Effective Core Potentials for Molecular Calculations. Potentials for Main Group Elements Na to Bi, J. Chem. Phys., 1985, 82, 284–298; (c) P. J. Hay and W. R. Wadt, Ab initio Effective Core Potentials for Molecular Calculations. Potentials for K to Au Including the Outermost Core Orbitals, J. Chem. Phys., 1985, 82, 299–310.
13. A. V. Marenich, C. J. Cramer and D. G. Truhlar, Universal Solvation Model Based on Solute Electron Density and on a Continuum Model of the Solvent Defined by the Bulk Dielectric Constant and Atomic Surface Tensions, J. Phys. Chem. B, 2009, 113, 6378–6396.
14. (a) P. Brandt, M. J. Sodergren, P. G. Andersson and P. Norrby, Mechanistic Studies of Copper-Catalyzed Alkene Aziridination, J. Am. Chem. Soc., 2000, 122, 8013–8020; (b) Y. Yamamoto, Theoretical Study of the Copper-Catalyzed Hydroarylation of (Trifluoromethyl)alkyne with Phenylboronic Acid, J. Org. Chem., 2018, 83, 12775–12783; (c) J. M. Fraile, J. I. Garcia, V. Martiez-Merino, J. A. Mayoral and L. Salvatella, Theoretical (DFT) Insights into the Mechanism of Copper-Catalyzed Cyclopropanation Reactions. Implications for Enantioselective Catalysis, J. Am. Chem. Soc., 2001, 123, 7616–7625; (d) M. Wang, T. Fan and Z. Lin, DFT Studies on Copper-Catalyzed Arylation of Aromatic C-H Bonds, Organometallics, 2012, 31, 560–569; (e) H. Zhao, L. Dang, T. B. Marder and Z. Lin, DFT Studies on the Mechanism of the Diboration of Aldehydes Catalyzed by Copper(I) Boryl Complexes, J. Am. Chem. Soc., 2008, 130, 5586–5594.
15. (a) K. Fukui, The Path of Chemical-Reactions - The IRC Approach, Acc. Chem. Res., 1981, 14, 363–368; (b) H. P. Hratchian and H. B. Schlegel, Theory and Applications of Computational Chemistry: The First 40 Years, ed. C. E. Dykstra, G. Frenking, K. S. Kim and G. Scueria, Elsevier, Amsterdam, 2005, pp. 195–249; (c) H. P. Hratchian, Using Hessian Updating to Increase the Efficiency of a Hessian Based Predictor-Corrector Reaction Path Following Method, J. Chem. Theory Comput., 2005, 61–69.
16. (a) J. P. Foster and F. Weinhold, Natural Hybrid Oribtals, J. Am. Chem. Soc., 1980, 102, 7211–7218; (b) A. E. Reed, R. B. Weinstock and F. Weinhold, Natural Population Analysis, J. Chem. Phys., 1985, 83, 735–746.
17. (a) S. F. Boys and F. Bernardi, Calculation of Small Molecular Interactions by Differences of Separate Total Energies – Some Procedures with Reduced Errors, Mol. Phys., 1970, 19, 553–566; (b) S. Simon, M. Duran and J. J. Dannenberg, How does Basis Set Superposition Error Change the Potential Surfaces for Hydrogen Bonded Dimers?, J. Chem. Phys., 1996, 105, 11024–11031.
18. (a) J. D. Chai and M. Head-Gordon, Systematic Optimization of Long-Range Corrected Hybrid Density Functionals, J. Chem. Phys., 2008, 128, 084106; (b) J. D. Chai and M. Head-Gordon, Long-Range Corrected Hybrid Density Functionals with Damped Atom-Atom Dispersion Corrections, Phys. Chem. Chem. Phys., 2008, 10, 6615–6620.
19. C. Y. Legault, CYLVeiw20, Universite de Sherbrooke, Quebec, Montreal, Canada, 2020, https://www.cylview.org.
20. (a) N. Y. P. Kumar, A. Bechtoldt, K. Raghuvanshi and L. Ackermann, Ruthenium(II)-Catalyzed Decarboxylative C-H Activation: Versatile Routes to meta-Alkenylated Arenes, Angew. Chem., Int. Ed., 2016, 55, 6929–6932; (b) J. Zhang, R. Shrestha, J. F. Hartwig and P. Zhao, A Decarboxylative Approach for Regioselective Hydroarylation of Alkynes, Nat. Chem., 2016, 8, 1144–1151; (c) M. Simonetti and I. Larrosa, Good Things Come in Three, Nat. Chem., 2016, 8, 1086–1088; (d) L. Huang, A. Biafora, G. Zhang, V. Bragoni and L. J. Gooßen, Regioselective C-H Hydroarylation of Internal Alkynes with Arenecarboxylates: Carboxylates as Deciduous Directing Groups, Angew. Chem., Int. Ed., 2016, 55, 6933–6937.
21. (a) M. Shahid, A. A. Tahir, M. Hamid, M. Mazhar, M. Zeller, K. C. Molloy and A. D. Hunter, Copper(II) Oligomeric Derivatives for Deposition of Copper Thin Films, Eur. J. Inorg. Chem., 2009, 1043–1050; (b) M. Guan, C. Wang, J. Zhang and Y. Zhao, Practical Organic Solvent-Free Cu(OAc)₂/D MAP/TEMPO-Catalyzed Aldehydes and Imines Formation from Alcohols under Air Atmosphere, RSC Adv., 2014, 4, 48777–48782; (c) C. I. Someya, S. Inoue, E. Irran and S. Enthaler, Exploring the Coordination Chemistry of O,N,O′-Ligands Modified by 2-Thienyl-Substituents to Nickel, Inorg. Chem. Commun., 2014, 44, 114–118.
22. (a) Z. Chen, Z. Tian, K. Kallio, A. L. Oleson, A. Ji, D. Borchardt, D. Jiang, S. J. Remington and H. Ai, The N–B Interaction through a Water Bridge: Understanding the Chemoselectivity of a Fluorescent Protein Based Probe for Peroxynitrite, J. Am. Chem. Soc., 2016, 138, 4900–4907; (b) E. G. Schnitzler, C. Badran and W. Jager, Contrasting Effects of Water on the Barriers to Decarboxylation of Two Oxalic Acid Monohydrates: A Combined Rotational Spectroscopic and Ab Initio Study, J. Phys. Chem. Lett., 2016, 7, 1143–1147; (c) H. Chen, J. Chang and H. Chen, A Computational Study on the Decomposition of Formic Acid Catalyzed by (H₂O)_x, x = 0–3: Comparison of the Gas-Phase and Aqueous-Phase Results, J. Phys. Chem. A, 2008, 112, 8093–8099; (d) R. Zhao, X. Zhao and X. Gao, Molecular-Level Insights into Intrinsic Peroxidase-Like Activity of Nanocarbon Oxides, Chem. – Eur. J., 2015, 21, 960–964; (e) R. Zhao and X. Zhao, Quantum chemical insights into regioselective hydrolysis of C₆₀F₃₆, RSC Adv., 2014, 4, 62465–62471.
23. (a) S. Zhang, Y. Fu, R. Shang, Q. Guo and L. Liu, Theoretical Analysis of Factors Controlling Pd-Catalyzed Decarboxylative Coupling of Carboxylic Acids with Olefins, J. Am. Chem. Soc., 2010, 132, 638–646; (b) D. Tanaka, S. P. Romeril and A. G. Myers, On the Mechanism of the Palladium(II)-Catalyzed Decarboxylative Olefination of Arene Carboxylic Acids. Crystallographic Characterization of Non-Phosphine Palladium(II) Intermediates and Observation of Their Stepwise Transformation in Heck-like Processes, J. Am. Chem. Soc., 2005, 127, 10323–10333; (c) Y. Jiang, Y. Fu and L. Liu, Mechanism of Palladium-Catalyzed Decarboxylative Cross-Coupling between Cyanoacetate Salts and Aryl Halides, Sci. China: Chem., 2012, 55, 2057–2062.
24. (a) D. H. Ess and K. N. Houk, Distortion/Interaction Energy Control of 1,3-Dipolar Cycloaddition Reactivity, J. Am. Chem. Soc., 2007, 129, 10646–10647; (b) C. Y. Legault, Y. Garcia, C. A. Merlic and K. N. Houk, Origin of Regioselectivity in Palladium-Catalyzed Cross-Coupling Reactions of Polyhalogenated Heterocycles, J. Am. Chem. Soc., 2007, 129, 12664–12665.
25. J. Yu, S. Zhang and X. Hong, Mechanisms and Origins of Chemo- and Regioselectivities of Ru(II)-Catalyzed Decarboxylative C-H Alkenylation of Aryl Carboxylic Acids with Alkynes: A Computational Study, J. Am. Chem. Soc., 2017, 139, 7224–7243.
26. LibreTexts Chemistry, https://chem.libretexts.org/Ancillary_Materials/Reference/Reference_Tables/Equilibrium_Constants/E1%3A_Acid_Dissociation_Constants_at_25C.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4qo02352h

‡ These authors contributed equally to this work and share the first authorship.

---