Qingjun Zhang,Pengyuan Shi,Xigang Yuan,2,Youguang Ma,2,Aiwu Zeng,2,*

1 State Key Laboratory of Chemical Engineering,School of Chemical Engineering and Technology,Tianjin University,Tianjin 300350,China

2 Chemical Engineering Research Center,Collaborative Innovative Center of Chemical Science and Engineering (Tianjin),Tianjin 300350,China

Keywords:Inert C-H bond Carboxylation Solvent-free medium Base effect Density functional theory (DFT) study

ABSTRACT A feasible synthesis route is devised for realizing direct carboxylation of thiophene and CO2 in a relatively mild solvent-free carboxylate-assisted carbonate (semi) molten medium.The effects of reaction factors on product yield are investigated,and the phase behavior analysis of the reaction medium is detected through the thermal characterization techniques.Product yield varies with the alternative carboxylate co-salts,which is attributed to the difference in deprotonation capacity caused by the base effect within the system.Besides,the detailed mechanism of this carbonate-promoted carboxylation reaction is studied,including two consecutive steps of the formation of carbanion through breaking the C-H bond(s)via the carbonate and the nucleophile attacking the weak electrophile CO2 to form C-C bond(s).The activation energy barrier in C-H activation step is higher than the following CO2 insertion step whether for the formation of the mono-and/or di-carboxylate,which is in good agreement with that of kinetic isotope effect (KIE) experiments,indicating that the C-H deprotonation is slow and the forming presumed carbanion reacts rapidly with CO2.Both the activation energy barriers in deprotonation steps are the minimal for the cesium cluster system since there have the weak the cesium Cs-heteroatom S (thiophene)and Cs-the broken proton interactions compared to the K2CO3 system,which is likely to enhance the acidity of C-H bond,lowering the C-H activation barrier.Besides,these mechanistic insights are further assessed by investigating base and C-H substrate effects via replacing Cs2CO3 with K2CO3 and furoate(1a) with thiophene monocarboxylate (1b) or benzoate (1c).

1.Introduction

The conversion of CO2into the high-valued carboxylic acids is an economic and environmental route since the CO2is regarded as the abundant,non-toxic and renewable C1 source [1,2].Direct carboxylation of aromatic arenes with CO2has attracted some attentions owing to its high atomic economy and potential commercial value,while there are some challenges in realizing the C-H carboxylation under the moderate conditions due to the kinetically and thermodynamically stable CO2and less reactive aromatic compounds [3,4].Specific limiting reagents are taken to enhance their reactivities,such as Lewis acid[5-12],carbene metal complexes [13,14] or alkali-metaltert-butoxides [15-19],but the majority of these metal additives are expensive,non-renewable or unstable under the air environment.Therefore,the more environment-friendly and effective synthesis route should be developed.

The strong base carbonate can achieve the carboxylation of aromatic compounds through cleaving C-H bond(s) in the solvent or solvent-free environment [20-23].However,this carboxylation strategy is ineffective once the pKavalue of aromatic substrate exceeds 27 [20].Aromatic substrate with the weaker acidic C-H bond(s),such as cesium furan-2-carboxylate and benzoate,can achieve the carboxylation through abstracting proton and then reacting with CO2in the molten salt-CO2interface [21,24].Apart from these solid reaction mixtures,the more challenging liquidphase benzene(pKa＞40)[25]could also be slightly converted into carboxylic acids with the poor selectivity,but the additional cesium isobutyrate is indispensable to generate the necessary molten phase [21].However,there has extremely high reaction temperature (340-380 °C) and pressure (～7 × 106Pa) in this system and the combined yield of carboxylic acids is relatively poor owing to the chemical inertness and low solubility of benzene in the salt medium.

Among these aromatic substrates,achieving the direct carboxylation of thiophene (pKa≈32.5) is also an urgent problem since there is a large amount of coking thiophene in the traditional refining process in China,while it has not been reasonably used.The vast majority of thiophene in the production process is simply removed or destroyed as the sulfur impurities through the pickling or hydrogenation strategies,resulting in the inefficient use of this resource and the significantly environmental pollution.Besides,thiophene is similar in nature to benzene and can be substituted for benzene in many applications.Taking the generation of 2-thiophenecarboxylic acid as an elucidating example,there are three mainly conventional synthesis paths including the formylation [26],acetylation [27] and Grignard reagent [28] methods.These traditional synthesis pathways have some disadvantages,such as multiple reaction steps leading to the lower atomic efficiency,the application of toxic reagents and a large amount of waste discharges resulting in the environment pollution.As a result,a new feasible approach,direct carboxylation of thiophene with CO2,should be necessarily designed from the dual perspectives of green synthesis and atomic efficiency,taking the comprehensive consideration of the progress of the carbon fixation in recent years and the successful application of some aromatic arenes [5],such as toluene and phenol,carboxylation reactions.

The reaction efficiency is impacted on the carboxylates for the hydrogenation of CO2into multicarbon carboxylates[29],however,the various carboxylates are not assessed in the carboxylation of substrate benzene[21].Therefore,the effective carboxylate should be determined in this carboxylation pathway,which might affect the rate of proton abstraction.The simple cesium carboxylate(formate,acetate or oxalate) could be prioritized in this base-induced experiment because of its simple chemical structure and inexpensive price.Besides these carboxylates,the cesium pivalate should also be attempted because this pivalate anion is capable to activate inert C-H bond in the proton abstraction progress [30-32].

Fig.1.Reaction mechanism for the carbonate-promoted C-H carboxylation.

In this paper,we develop a feasible route to achieve the direct carboxylation of thiophene in a relatively mild solvent-free carboxylate-assisted cesium carbonate (semi) molten state.The effects of various reaction factors are explored.The combination of cesium carbonate and carboxylate co-salt exhibits the synergetic effect in the carboxylation reaction,and the cesium pivalateassisted Cs2CO3system affords the better product yield among these co-salts.The increasing use of carbonate-promoted synthesis of aromatic carboxylateviadirect C-H carboxylation necessitates a deeper understanding of the mechanism and regulating variables of the direct carboxylation reaction processes.Understanding the intricacies of these reactions should allow us to enhance yield and product selectivity while also expanding the substrate range.The Kanan group[21]previously proposed a multistep mechanism(Fig.1) for the carbonate-promoted C-H carboxylation that included the following steps: (i) the cleavage C-H bond of aromatic compound through the base carbonate leading to the formation of carbanion and bicarbonate;(ii) the presumed carbanion nucleophile attacking the weak electrophile CO2to form the C-C bond(s);(iii) the thermal decomposition of bicarbonate to form the carbonate.Nevertheless,there are still some crucial issues that have not been addressed in this proposed mechanism,as follows:(a) which step in the reaction process is the rate-and selectivitydetermining step?(b)what is the factor controlling the regioselectivity?(c)the deprotonation ability of the base affects the reaction result in this acid-base chemistry approach,so is there the base effect in this reaction? (d) what is the role of the counter-cation in the CO2insertion step? (e) what’s the variations in the interatomic interactions as the reaction proceeds? Therefore,the detailed mechanism of this carbonate-promoted carboxylation reaction is studied,and the corresponding distortion/interactionactivation strain model analysis is employed to elucidate the factor controlling the regioselectivity and the origins of the height of activation energy barrier for the bare and carboxylate-assisted Cs2CO3systems.

2.Materials and Methods

2.1.Materials

Thiophene (99%),cesium or potassium carbonate (99.9%),cesium or potassium pivalate (＞97%),deuterium oxide (99%),potassium isobutyrate (＞98%),and cesium or potassium acetate(99.9%)were obtained from Aladdin;carbon dioxide(99.9%(mass))was purchased from Tianjin Liufang Industrial Gas Distribution Co.,Ltd;cesium oxalate (98% (mass)) was purchased from Jinjinle Chemical Co.,Ltd;cesium formate (98% (mass)) was supplied by Energy Chemical;anhydrous sodium tartrate (99.0% (mass)) was purchased from Jiaxing Sicheng Chemical Co.,Ltd;3-(trimethylsilyl)-1-propane sulfonic acid sodium salt (DSS,97%(mass)),potassium pivalate (＞98% (mass)) were purchased from Macklin.Thiophene was dried over 4A sieves before using,and other purchased reagents were used without further treatment.

2.2.Experimental procedure

The specifically calculated mixed salts (1-3 mmol) and thiophene (2-10 mmol) were charged successively into 50 ml reactor(316 L) equipped with quartz tube.The sealed reactor was backfilled through CO2three times,and reached the specific pressure through filling CO2gas reagent.This mixture was heated to the certain reaction temperature (180-340 °C) through using intelligent heating mantle and maintained the setting temperature for certain hours (1-5 h).Subsequently,this reactor was depressurized carefully after being cooled to ambient temperature.The unreacted thiophene was removed from the autoclave,and the remaining product mixtures were dissolved with 1 ml D2O,and then filtered through MCE syringe filter(2.2×10-7m)to remove the insoluble materials.The corresponding carboxylate yield was calculated by integrating the1H NMR peaks taking the anhydrous sodium tartrate as internal standard.The reacting carboxylate product was also verified through13C NMR with the added DSS as internal standard.

2.3.Characterization techniques

Products were identified and the corresponding yields were analyzed through the1H NMR and13C NMR.1H NMR and13C NMR signals were,respectively,recorded on the VARIAN INOVA 500 MHz and AVANCE III 400 MHz spectrometers.1H NMR chemical shifts were assigned relative to the residual proton signal(D2O,δ=4.79)and13C NMR chemical shifts were referenced with added DSS(δ=0).The relaxing delay and number of scans were assigned as 8 s and 64 in the1H NMR spectra.

Thermal stability for these mixed cesium salts were evaluated through using the Simultaneous Thermal Analyzer (NETZSCH STA 449F5).The sample was dried under vacuum for 24 hours to remove the absorbed water and grinded thoroughly to ensure the particle uniformity.The sample(1-10 mg)was placed on the Al2O3inner crucible and its mass change was recorded underN2atmosphere when the temperature was increased from ambient temperature to 600 °C at the heating rate of 10 °C·min-1.

The thermal event of cesium salt mixtures could be recorded through TA Instruments Q2000 Differential Scanning Calorimetry(DSC).The sample was also dried under vacuum for 24 h.The sample (5-10 mg) was weighed on the Tzero aluminum pan and then sealed hermetically before the DSC measurement.The analysis was carried out under flowingN2atmosphere (50 ml·min-1) when the temperature was increased from the ambient temperature to 280°C (or 300 °C) and then decreased to ambient temperature at the constant rate of 5 °C·min-1.

The phase transition of cesium salt mixtures could also be collected under vacuum throughin-situhigh temperature X-ray diffraction measurement (Bruker D8 advance) equipped with the reaction chamber using the Cu radiation (Kα=0.154 nm).The sample was heated from 100 °C to 280 °C with the ramp rate of 10 °C·min-1,and each diffraction data was collected in the 20 °C steps.Each XRD pattern was scanned over the 2θ range from 5° to 50° with the step size of 0.02° and the scan speed of 0.083 (°)·s-1.

2.4.Computational details

All the density functional theory (DFT) calculations were performedviathe quantum chemistry package of Gaussian 09 (Rev.D01)suites[33].Geometry optimizations for all referred structures of reactants or reactant complexes,intermediates and transition states and products or product complexes were carried out at the B3LYP-D3(BJ) theory level with the mixed basis set including the Stuttgart effective core potential SDD [34] for Cs atom and tripleζ split-valence 6-311G** for all other atoms.The combination of this theoretical method and the mixed basis set was named BS1.Frequency calculations were made to verify the nature of each structure taking the same theory level and basis set as the geometry structure optimization.The stationary points with 0 or 1 imaginary frequency,respectively,corresponded to the minimum structure and the transition state(TS)structure.And the characteristic of each TS structure was evaluated by the intrinsic reaction coordinate(IRC)calculations,which can ensure that the transition state was connected to the correct reactants and intermediates.The electronic single-point energies were computed by the M06-2X theory level with the def2-TZVPP basis set for the structures obtained by the BS1.

3.Results and Discussions

3.1.Carboxylation reaction in the cesium carbonate/carboxylate mixed salts

Direct thiophene carboxylation reaction was investigated in the solvent-free mixed cesium carbonate and carboxylate salts.The feasibility of this carboxylation path was initially explored at the small-scale of 2 mmol substrate with 3 mmol salt mixtures comprising 40% (mol) Cs2CO3and 60% (mol) cesium carboxylate(Table 1).Taking the cesium acetate (CsOAc) as an example,there was only the thiophene-2-carboxylate product at the reaction temperature of 200 °C (Entry 1),and the thiophene-2,5-dicarboxylate product was obtained at the reaction temperature over 220 °C(Entry 2).The corresponding carboxylate yield was gradually improved as the temperature increases(Entries 1-4),and the maximum carboxylate yield was reached at the temperature of 300 °C(Entry 4).The decreased carboxylate yield was observed when the temperature exceeds 300°C(Entry 5),probably due to the thermal decomposition of carboxylate products and mixed salts.To probe the possible causes,the thermal stability analyses of carboxylate products and cesium salt mixtures were conducted through the thermogravimetric analysis (TG).The thermal decomposition for the monocarboxylate occurred rapidly at approximately 340 °C,while it was not observed for the dicarboxylate within the reaction temperature range (Fig.S1 of Supplementary Material).Meanwhile,the minor decomposition of cesium salt mixtures was detected at 200-400 °C (Fig.2(a)),which also exerted a relatively smaller impact on the product yield.Apart from the combined product yield,the selectivity of carboxylate product was also influenced by the reaction temperature (Entries 1-5).The selectivity of thiophene-2-carboxylate was higher at the low temperature range,while the dicarboxylate constitutes approximately 80% of carboxylate products at the medium temperature range.The correspondingly detailed temperature effects were shown in Fig.3(a).Note that no carboxylate products were detected when the CO2is switched to nitrogen with the remaining conditions unchanged(Entry 6),which fully indicated that the C-C bond formation in the product carboxylate was traceable to the CO2.Besides,in comparison with the accelerated reactivity with the aid of cesium acetate,other simple carboxylates of formate (Entry 7) and oxalate(Entry 8) were also evaluated in this carboxylation system.As observed in Table 1,the enhancement effect for this carboxylation reaction varied with the carboxylate co-salt.The cesium acetate was the most effective among these simple carboxylate co-salts(formate/acetate/oxalate) since its basicity was the strongest among these three.

Table 1 Summary reaction results for thiophene carboxylation as the function of temperature and salts

Fig.2.The thermal behaviors of the mixed Cs2CO3/CsOAc salt.(a) TG and DTG curves;(b) DSC curve.

To further explore the thiophene C-H carboxylation progress,the effects of the reaction factors such as CO2pressurePCO2,the amountnsaltsof carbonate and co-salt,carbonate proportion,substrate amountnthio,and reaction time on the reaction results were assessed taking the cesium acetate among these carboxylate salts.Enumerated in Fig.3 showed the impacts of these reaction factors on the carboxylate yields and the selectivity.As obtained in Fig.3(b),the corresponding carboxylate yields varied with the CO2pressure.The carboxylation reaction was proceeded at 300°C with the total amount of carboxylate products as 59.50 μmol·g-1of Cs2CO3when the initial CO2pressure was 2×105Pa,and the total reaction pressure of highly volatile thiophene and CO2was 6.6×105Pa.The highest carboxylate yield of 152.96 μmol·g-1of Cs2CO3was obtained at introducing 8 × 105Pa CO2(the total pressure of 1.76 × 106Pa),and the further increasing CO2pressure did not result in an increase in carboxylate yield since the carbonate could be sequestered in the form of bicarbonate due to the large amount of CO2[21,34].There was also the minor impact on the selectivity of carboxylate product at different CO2pressure.As illustrated in Fig.3(c),product yield increased almost linearly as the amount of the carbonate and co-salt increased,and this observation strongly implied that this direct carboxylation reaction was stochiometric in nature.The carbonate proportion in this salt medium also affected the C-H carboxylation reactivity(Fig.3(d)).When the low-melting cesium acetate (Tm=194 °C) [35] was employed alone,the trace yield was obtained.With the carbonate percentage increasing to 20% (mol),the total carboxylate yield reached the maximum.Once the increase of carbonate quantity was continued,the corresponding carboxylate yield was decreased subsequently because the excess high-melting cesium carbonate (Tm=793 °C)[35] might impede the appearance of uniform molten eutectic phase.In addition,the distinct effect of carbonate proportion on product selectivity was also observed.The selectivity of dicarboxylate was enhanced consistently with improved carbonate proportion,probably because available monocarboxylate was readily converted into dicarboxylate under the stronger alkali carbonate component.Similarly,it was also carried out varying the amount of thiophene (2-10 mmol).The carboxylate yield was gradually increased from 135.95 to 571.22 μmol·g-1of Cs2CO3,as shown in Fig.3(e),which indicated that the strong ability of carbonate conversion was beneficial to generate carboxylates with the excessive thiophene amount.Note that enumerated in Fig.3(d) demonstrated that use of 20% (mol) carbonate given the better results than that of 40% (mol) case.Therefore,the corresponding reaction was carried out with 20%(mol)case using 10 mmol thiophene,and the result was also shown in Fig.3(e)and denoted by a red asterisk.We could find that product yield was indeed better than all the possibilities in 40% (mol) case,and the corresponding selectivity was also more towards the monocarboxylate side (41.54%vs23.37%).As demonstrated in Fig.3(f),the carboxylate yields increased from 3.75% to 6.45% when the reaction time was raised from 1 to 3 h,but dropped when it is over 3 h.With longer reaction times,more water was created in the reaction system,which not only promoted the dissociation of the reactants and slowed down the reaction,but also dissolved some Cs2CO3,lowering the target product yields [34].

To determine whether the reaction proceeded at the molten salt interface [21,24],we performed the thermal behavior analysis for the mixed Cs2CO3/CsOAc salts through differential scanning calorimetry (DSC) with one heating/cooling cycle,in which endothermic or exothermic peak(s) indicated the possible phase transition behavior such as melting or crystallization (Fig.2(b)).The endothermic behavior (approx.73 °C) was regarded as the moisture removal due to its hygroscopic property,which was in keeping with the slight weight losses before 200 °C (Fig.2(a)).Another endothermic event (approx.134 °C) could be attributed to the eutectic transition of the cesium salt mixtures,and the reversible exothermic event (approx.117 °C) could be identified as the crystallization behavior.Therefore,the carboxylation reaction was proceeding at the molten salt interface in 200-340 °C,and the appropriate increase of carboxylation temperature might lead to the relatively uniform molten eutectic phase,which was beneficial for overcoming the mass transfer limitation.

It is now well established that carboxylate salts,especially acetate and pivalate,can be used to promote the stoichiometric and catalytic C-H activation effects in the transition metal catalysis[36].As also discovered in Table 1,cesium acetate indeed gave a better result than other salts.Then,other carboxylate-assisted species was further explored in the thiophene C-H carboxylation.We envisioned that whether the combination of cesium pivalate(CsOPiv) and Cs2CO3could promote thiophene C-H carboxylation.Therefore,it should be attempted in our following experiments whether the combination of cesium pivalate (CsOPiv) and Cs2CO3could promote thiophene C-H carboxylation efficiently.

Fig.3.The effects of various factors on product yields and selectivity in the Cs2CO3/CsOAc system.(a) reaction condition:2 mmol substrate,40% carbonate,8 × 105 Pa CO2 initial pressure,3 mmol total cesium salts,2 h;(b) reaction condition: 2 mmol substrate,300 °C,40% (mol) carbonate,3 mmol cesium salts,2 h;(c) reaction condition:2 mmol substrate,300°C,40%(mol)carbonate,8×105 Pa CO2 initial pressure,2 h;(d)reaction condition:2 mmol substrate,300°C,8×105 Pa CO2 initial pressure,3 mmol cesium salts,2 h;(e)reaction condition:300°C,8×105 Pa CO2 initial pressure,40%(mol)carbonate,3 mmol cesium salts,2 h;*-reaction condition:300°C,8×105 Pa CO2 initial pressure,20%(mol)carbonate,3 mmol cesium salts,2 h;(f)reaction condition:2 mmol substrate,300°C,8×105 Pa CO2 initial pressure,40%(mol)carbonate,3 mmol cesium salts.

The carboxylation reaction was similarly carried out in the presence of 40% (mol) Cs2CO3/60% (mol) CsOPiv.The reaction effects were assessed at the temperature range of 240-320°C for 2 h after filling 8 × 105Pa CO2at the ambient temperature (Fig.4(a)).The onset temperature (260 °C) for carboxylation reaction was higher than the Cs2CO3/CsOAc medium.The optimal reaction temperature was 280°C with total carboxylate yield of 255.87 μmol·g-1of Cs2-CO3at the combined yield of 8.33%,which was higher than those in the Cs2CO3/CsOAc medium.This comparable carboxylate yield demonstrated that cesium pivalate certainly possessed the prominent roles in the C-H carboxylation of thiophene.The selectivity of thiophene-2-carboxylate (51.39%) was also larger than that(22.54%) in the Cs2CO3/CsOAc system,indicating that the secondary C-H carboxylation of monocarboxylate was weaker in the Cs2CO3/CsOPiv medium.The total carboxylate yield was then decreased as the reaction temperature continued to elevate,which could be similarly attributed to product decomposition since the Cs2CO3/CsOPiv medium were consistently steady in the reaction temperature regime (Fig.5(a)).We next confirmed that carbonate proportion also notably affected the carboxylate yield and the selectivity in the Cs2CO3/CsOPiv medium(Fig.4(b)).The trace yield of carboxylate products was obtained when the cesium pivalate was employed alone,although this carboxylate salt (Tm=344-348°C) was not melted at 280 °C.The maximum carboxylate yield is achieved at the 40% carbonate.As the carbonate proportion increases,the product gradually tilts towards the di-carboxylate.The effect of CO2pressure on the carboxylate yield was explored subsequently in the pressure range of 2 × 105Pa to 1.2 × 106Pa(Fig.4(c)).The carboxylate yield reached the 81.26 μmol·g-1of Cs2CO3at the combined yield of 2.65%with the total reaction pressure of approximately 6.8 × 105Pa when introducing 2 × 105Pa CO2at ambient temperature.Optimal carboxylate yield was obtained when the CO2pressure was 8 × 105Pa,and the further yield enhancement was not observed with the elevated pressure(1 × 106Pa -1.2 × 106Pa).Note that there was no significant effect of CO2pressure on product selectivity.Fig.4(d) illustrated the effects of thiophene amount on this carboxylation reactivity were different from that in the Cs2CO3/CsOAc medium.The carboxylate yield was increased gradually to 42.38% from 8.31% with attaining the maximum carboxylate products of 1301.92 μmol·g-1of Cs2CO3if the thiophene amount was elevated to 8 mmol from 2 mmol.This increased yield indicated that the carboxylation reactivity might not be released under relatively lower substrate amounts due to the strong-base property of alkali metal salts.However,the carboxylate yield was declined to 39.14% after continually increasing to 10 mmol,which still remained better than that in the Cs2CO3/CsOAc system.Besides,the carboxylation reaction could also be promoted with the increase of the quantities of cesium salt mixtures based on the inherent stochiometric characteristics of this reaction (Fig.4(e)),and the selectivity of monocarboxylate was stabilized as approximately 50% within this amount range.When the reaction time was raised from 1 to 3 h,the carboxylate yield increased from 164.56 to 296.35 μmol·g-1of Cs2CO3,while it dropped at the reaction time over 3 h(Fig.4(f)).

Fig.4.The effects of various factors on product yields and selectivity in the Cs2CO3/CsOPiv system.(a)reaction condition:2 mmol substrate,40%(mol)carbonate,8×105 Pa CO2 initial pressure,3 mmol total cesium salts and 2 h;(b) reaction condition: 2 mmol substrate,280 °C,8 × 105 Pa CO2 initial pressure,3 mmol cesium salts and 2 h;(c)reaction condition: 2 mmol substrate,280 °C,40% (mol) carbonate,3 mmol cesium salts and 2 h;(d) reaction condition: 280 °C,8 × 105 Pa CO2 initial pressure,40% (mol)carbonate,3 mmol cesium salts and 2 h;(e)reaction condition:2 mmol substrate,280°C,8×105 Pa CO2 initial pressure,40%(mol)carbonate and 2 h;(f)reaction condition:2 mmol substrate,280 °C,8 × 105 Pa CO2 initial pressure,40% (mol) carbonate,3 mmol cesium salts.

Similarly,to determine whether the reaction was going on at the molten salt interface,the phase behavior of 40% (mol) Cs2-CO3/60% (mol) CsOPiv medium was also measured through the DSC andin-situhigh temperature X-ray diffraction (in-situXRD).The small endothermic event (approx.117 °C and 146 °C),which was possibly caused by the melting behavior,was observed during the heating process in the DSC curve (Fig.S2 of Supplementary Material);however,the corresponding crystallization point seemed to disappear during the cooling process although the Cs2-CO3/CsOPiv system was consistently steady(Fig.5(a)).The cause of this anomaly was not clear,but it may be attributed to the insufficient cooling intensity.However,their trends for improving the cooling speed were similar to the previous one.Therefore,a more transparent thermal analysis method ofin-situXRD measurement was adopted to detect the phase transition for Cs2CO3/CsOPiv medium to precisely determine whether this reaction also occurred at the molten salt interface.The intensities of crystalline diffraction peaks were collected at the temperature of 100-280 °C (Fig.5(b)).The diffraction intensity of crystalline CsOPiv was weakened distinctly relative to the Cs2CO3peaks at 120-160 °C,and then was not detectable at 180 °C in the XRD pattern,although diffraction peaks of crystalline Cs2CO3remained in the measured temperature region.The disappearance of the CsOPiv diffraction peaks was assigned to the eutectic melting behavior and the characterization results revealed that this reaction was also proceeded in the presence of the molten eutectic phase and solid Cs2CO3phase.

Fig.5.The thermal behaviors of mixed Cs2CO3/CsOPiv salt.(a) TG and DTG curves;(b) in-situ XRD pattern.

In order to further determine the influence of the abovementioned possible base effect in this reaction,we performed the relevant experimental studies on the replacement of cesiumpotassium metal cations in the mixed salts,enumerated in Table 2,which mainly contained two aspects: one was to replace Cs2CO3with K2CO3when the auxiliary carboxylate co-salt remained invariable (Entry 1vsEntry 2);another was to directly replace the metal cation of carboxylate without changing the carbonate(Entry 1vsEntry 3 and Entry 4).The replacement of Cs2CO3with K2CO3leaded to reduce the yield from 8.33% to 1.08% under the same conditions as the pure cesium salts case,while it was declined to 4.29%at simply changing carboxylate(CsOPiv to KOPiv replacement at constant carbonate).

Table 2 The impacts of the combination effects of carbonate and pivalate on the reaction results

3.2.Reaction mechanism of cesium pivalate-assisted Cs2CO3-promoted C-H carboxylation with CO2

In this reaction model,the full molecular system comprising both counter-cation and corresponding anion was taken into account into the calculations without the truncations or any symmetry constrains.This reaction mainly consisted of two consecutive steps including that the formation of carbanion represented as the organocesium was the result of base induced deprotonation and the nucleophile attacked the weak electrophile CO2to form the C-C bond(s)in the form of cesium carboxylate.The counter-cation Cs+was considered in this present study,and there were two main reasons: one was the inherent instability of the formed carbanion in the gas phase [37];another was that the metal-CO2interaction was required to polarize and activate CO2[38].The experimental results showed that the cesium pivalate-assisted Cs2CO3system had the best effect among a series of carboxylates.Therefore,this system was taken,in this paper,as an example to elaborate the reaction mechanism details of this reaction process.

Fig.6. Syn-and anti-attack modes of substrate(s) to the Cs cluster.

Fig.7.(a)Reaction energy profiles of the first carboxylation reaction.Energies are relative to the reactants of Reactant_1 and Cs cluster.(b)All reported structures and their important geometry parameters (distances are in Angstroms (10-10 m)).

The cesium cluster,shown in Fig.6,resembling boat-like shape forms with 38.8 kcal·mol-1exergonic(Fig.7)when cesium carbonate and cesium pivalate were combined together.The substrate can attack to this Cs cluster from either thesynorantidirection to the curved side of the intermediate Cs cluster on account of the asymmetric characteristics of the pivalate group (Fig.6).Calculations indicated that there proceeded the same reaction mechanism for these two alternative attack pathways,while the activated energy barrier in cleaving the C-H bond forsyn-attack route was lower than that of theanti-attack alternative (shown in Figs.S3 and S4 of Supplementary Material).As a result,we solely elaborated on the reaction mechanism details of the substratesyn-attack on the Cs cluster.

Fig.8.(a) Reaction energy profiles of second carboxylation reaction.Energies are relative to the reactants of Reactant_3 and Cs cluster.(b) All reported structures and their important geometry parameters (distances are in Angstroms (×10-10 m)).

The stable intermediate Reactant Complex_1 was formed with 9.7 kcal·mol-1exergonic when adding the boat-like Cs cluster into the system,as shown in Fig.7 (a).A careful investigation of this structure revealed that these failed in the hydrogen bond range[32]for the two pairing H-O atoms,that is,the distances between H1of thiophene and O1and O2of center Cs2CO3in cluster were 2.10 × 10-10m,and 2.59 × 10-10m,respectively,illustrated in Fig.7(b),which indicated that these two oxygen atoms can serve as the proton acceptors to generate the C4 carbanion.These interatomic interactions also leaded to the elongation of the C-H1bond.Besides,the natural population analysis(NPA)charge for these two oxygen atoms was,respectively,-0.971 a.u.and-0.950 a.u.(1 a.u.=1.6021892×10-19C).In principle,from the above two perspectives,the O1atom was the most promising proton acceptor.However,the deformation energy of the system was higher than that of the proton acceptor O2atom as the proton gradually approached.This inference was confirmed by the consequence of relaxed scanning of broken C-H1bond.Therefore,the atom O2in the Cs cluster acted as the proton acceptor,which was also demonstrated by the structure of transition state TS_1.In TS_1,the activated C-H bond was elongated from 1.09×10-10m to 1.53×10-10m,while it was shortened from 2.59 × 10-10m to 1.12 × 10-10m for the O2-H1bond.Meanwhile,the distance between the counter-cation Cs+and the carbanion was also gradually decreased.The activation barrier of this deprotonation step was 17.0 kcal·mol-1when taking the Reactant Complex_1 as the reference.As the activated C-H bond distance gradually increased,the less kinetically stable intermediate Int_1 was produced and then transformed to the more thermodynamically favorable intermediate Int_2′: there was 39.4 kcal·mol-1exergonic for the reaction Cs2CO3+CsOPiv+Reactant 1 →Int2′.Note that the bicarbonate was thermally unstable and it will undergo the thermal decomposition(175°C)into the carbonate.Therefore,in this reaction mechanism,the detailed decomposition steps were not considered,although the onset temperature of this carboxylation reaction was higher than the thermal decomposition temperature of bicarbonate.In order to more clearly judge the difference in the activation energy barrier between the key steps in this reaction process,we simply took the energy difference between the generated nucleophile and the bicarbonate as the starting point for the next step.The elimination of the cesium bicarbonate cluster from the intermediate Int_2′leaded to the formation of complex Reactant_2 with the metal-nucleophile bond.

The other key step for achieving this carboxylation reaction was the activation and polarization of chemical inertness CO2molecule.The detailed insights of CO2insertion step were shown in Fig.7(a).As the interaction of the carbanion and metal with the CO2leaded to the formation of the stable intermediate Reactant Complex_2 with 0.7 kcal·mol-1(1 kcal=4.186 kJ) exergonic.The bond angle O-C-O of CO2changed from 180° to 174.95° to expose the Ccenter of CO2as much as possible,and the metal-nucleophile bond was elongated to 3.01 × 10-10m and the corresponding details were shown in Fig.7 (b).Gradually attacking the electrophile CO2by the C4 nucleophile with the aid of the metal-CO2interaction,the transition state structure TS_2 was gained,wherein the distance between the C-center of CO2and nucleophile C of thiophene ring was shortened from 3.10 × 10-10m to 2.77 × 10-10m and the corresponding bond angle was bent to 168.96°.The activation barrier of this CO2insertion step was 0.8 kcal·mol-1when taking the Reactant Complex_2 as the benchmark.As the activated C-C bond distance gradually reduced,the more kinetically and thermodynamically favorable product Product_1 was formed:there was 31.4 kcal·mol-1exergonic for the reaction CO2+Reac tant_2 →Product_1.

The further mechanistic insights for the secondary C-H carboxylation of monocarboxylate were the similar as the aforementioned steps.The detailed reaction mechanisms were shown in Fig.8(a),and the correspondingly important geometry structure parameters of the reactant complexes,transition state and intermediates were presented in Fig.8(b).The activation barrier of this second deprotonation step was 19.6 kcal·mol-1when taking the Reactant Complex_3 as the reference,which was higher than that of the thiophene C-H activation step,while it was 1.0 kcal·mol-1for the following CO2insertion step relatively to the Reactant Complex_4.We can find that the energy barrier of C-H activation step was higher than that of the following CO2insertion step whether for the formation of the mono-and/or di-carboxylate,which indicated that the C-H deprotonation induced by the base was slow and the resulting carbon-centered nucleophile reacted quickly with CO2.In other words,the proton abstraction in carboxylation reaction was a rate-limiting step.Furthermore,the second C-H activation of monocarboxylate was a rate-determining step of these whole CsOPiv-assisted Cs2CO3-promoted C-H carboxylation reaction.

Aiming to precisely disclose what factor caused the difference in activation energy barrier for these two attack pathways in this C-H carboxylation reaction proceed,the distortion/interaction-ac tivation strain analysis model was employed [35-37,39-41].The essence of this model was to decompose the relative energy ΔE(ζ)of the potential energy surface into two contributions,including the activation strain energy ΔEstrain(ζ),which was the energy required to twist the structures of the referenced two fragments,and the interaction energy ΔEint(ζ) between the two warped fragments along the reaction coordinate ζ.Note that we focused on the analyses of these two energy changes in the entire reaction energy profile along the reaction coordinate to avoid the misleading insights observed in the single-point analysis only in the transition state structure.Fig.9 gived the results of activation strain analyses for the proton abstraction step of the alternative attacking pathways in the thiophene and monocarboxylate.As discovered in Fig.9(a) that these two possibilities had the similar interaction energy term along the reaction coordinate,while the higher activation energy barrier inanti-attack path was stemmed from the more destabilizing strain energy term when compared to thesyn-attack path.This indicated that the activation strain energy was the origin of the difference in energy barrier in two alternative attacking paths for thiophene substrate.Depicted in Fig.9(b) indicated that the less activation energy barrier insyn-attack proceed originated from the more stable interaction energy term for monocarboxylate substrate even with the little energy penalty in comparison with theanti-attack route.As found in Fig.9(c) that the differences in both two energy terms determined the height of activation energy barrier for the thiophene and monocarboxylate.Note that the difference in the activation strain energy term was greater than that in the interaction energy term.Hence,the reason why the energy barrier in the second proton removal step was higher than that in the first step was the combination effect of the less stabilizing factor and the stronger strain curve.

Fig.9.Distortion/Interaction-activation strain analyses for the alternative syn-and anti-attack modes of deprotonation step along the reaction coordinates:(a)thiophene;(b)thiophene-2-carboxylate;(c) thiophene vs thiophene-2-carboxylate.A dot designates a TS (1 kcal=4.186 kJ).

3.3.Kinetic isotope experiments

From the above DFT quantitative calculation results,it can be seen that the rate-determining step in the direct carboxylation of thiophene promoted by the carboxylate-carbonate system was the cleavage of the C-H bond of the aromatic substrate.There was a great challenge of observing the transition state structure in the reaction mechanism,while it was relatively easy to confirm whether a particular step in the reaction process such as C-H bond cleavage was the rate-determining step of the reaction process.Therefore,in order to determine the rationality and reliability of this quantitative calculation process,the kinetic isotope effect(KIE) was used to determine whether the C-H bond cleavage was really included in the rate-determining step of this reaction process [37].

Fig.10 gave the corresponding results of intermolecular competition experiments [42].Direct carboxylation experiments were performed with different ratios of protonated and deuterated substrates.The KIE value of the direct carboxylation of C-H promoted by K2CO3or Cs2CO3system was 2.1 or 1.7.The results of the KIE experiments were consistent with the reaction mechanism under the above DFT calculation,that is,the C-H deprotonation induced by the base was slow and the resulting carbon-centered nucleophile reacted rapidly with CO2.

3.4.Clarification of the origin of regioselectivity

Fig.10.KIE values for benzothiophene C-H carboxylation were measured via intermolecular competition experiments.

As indicated by the reaction results,the strategy of the cesium carbonate-promoted thiophene carboxylation reaction had a very specifically exclusive regioselectivity.Therefore,it was necessary to further determine what factors affect regioselectivity in this reaction,that is,the origins of this exclusive regioselectivity in this carboxylation path needed to be revealed.

Based on the above-mentioned mechanism insights,the most likely factor affecting the regioselectivity in this reaction can only be the proton abstraction step,while the CO2insertion step was not possible.Once the carbanion was generated,the following CO2insertion step will get the corresponding carboxylate.Therefore,in this reaction,the proton abstraction step was the regioselectivity determining step.Then we studied the effect of the activated H positions of thiophene and thiophene-2-carboxyalte on the C-H activation mechanisms,and the corresponding results were illustrated in Table 3.For thiophene,there was 5.82 kcal·mol-1less activation energy barrier in alpha-H cleavage step in comparison with that of beta-H removal path.The activated C-H bond length in alpha-H TS structure was shorter than that in beta-H TS structure,indicating that the beta-H transition state was later than the other.The energy barriers of the 3-H and 4-H elimination paths were,respectively,6.66 kcal·mol-1and 3.25 kcal·mol-1higher than that of the 5-H one.The activated C-H bond lengths in the former two TS structures were longer than that in the latter,which demonstrated that the 5-H transition state was earlier than the other two alternatives.

There were some factors that affect the energy required to break the C-H bond during the reaction.Among them,the key factor was the inherent reactivity of the C-H bond(s),which was also the root of the height of the activation energy barrier in the reaction process.So as to more clearly understand the difference in the activation energy barriers in the C-H bond cleavage step for different activated H-positions,we explored two possible factors comprising the acidity and the dissociation energy (BDE) of the C-H bond.As discovered in Table 3,the more positive of the NPA charge of H atom in reactant,the smaller the activation energy barrier.However,there was no correlation between the height of activation energy barrier and the BDE of C-H bond.For thiophene,the C-H activation barrier of alpha-H removal path was lower than that of beta-H abstraction route,while the BDE tendency was indeed the opposite,and it was also the same for the monocarboxylate.

Table 3 Important parameters for C-H activation step of different locations for thiophene and monocarboxylate (1 kcal=4.186 kJ)

Fig.11 gave the results of activation strain analyses for the proton abstraction step of different activated H-positions in the thiophene and monocarboxylate.As discovered from Fig.11(a),these two possibilities had the similar interaction energy term along the reaction coordinate,while the higher activation energy barrier in the beta-H removal path was stemmed from the more destabilizing strain energy term when compared to the alpha-H abstraction step,which will make the reaction more endothermic.Besides,this more destabilizing term resulted in shifting the transition state to a later stage and the product side.This indicated that the activation strain energy was the origin of the regioselectivity in the proton abstraction step for thiophene substrate.Depicted in Fig.11(b) illustrated that the cooperation of these two energies determined the regioselectivity in the C-H deprotonation step ofmonocarboxylate substrate.Herein,the largest difference in the interaction energy term for these two alternatives was about 15 kcal·mol-1,while it was 10 kcal·mol-1for the other term.Therefore,the less activation energy barrier in 5-H activation proceed originated from the more stable interaction energy term even with the little energy penalty in comparison with the 4-H monocarboxylate reaction step,which leaded to the transition state of 4-H pathway shifting to the later stage.As inspired from Fig.11(c)that we can determine the reason why the energy barrier in the second proton removal step was higher than that in the first step.We found that their results were exactly opposite to those shown in Fig.11(b),although the regioselectivity of both processes was determined by two energy terms simultaneously.Note that the difference (about 20 kcal·mol-1) in the activation strain energy term was greater than that (13 kcal·mol-1) in the interaction energy term.Hence,the formation of the later stage transition structure was caused by the combination effect of the less stabilizing factor and the stronger strain curve.

Fig.11.Distortion/Interaction-activation strain analyses for the deprotonation step along the reaction coordinates:(a)thiophene;(b)thiophene-2-carboxylate;(c)thiophene vs thiophene-2-carboxylate.A dot designates a TS (1 kcal=4.186 kJ).

3.5.Variations of the interatomic interactions during the reactions

For the purpose of deeply understanding the characteristics of this carbonate-mediated C-H carboxylation reaction,we visually analyzed the variations in the interatomic interactions as the reaction proceeds taking the interaction region indicator (IRI) analysis method [43] performed by the Multiwfn program [44].We took the thiophene C-H deprotonation step and the following CO2insertion step as examples to briefly illustrate the changes in interatomic interactions in the reaction process.The standard IRI isosurface colorbar and their corresponding chemical explanations were elucidated in Fig.12.

Fig.12.The standard IRI color-bar and the corresponding chemical explanations of sign(λ2)ρ on IRI isosurface.

Fig.13.Variation of isosurface map of IRI=1.0 during the proton abstraction step of thiophene and Cs2CO3.

Fig.13 gave the IRI isosurface maps with some illustrative points along the IRC route of the cesium carbonate induced thiophene deprotonation.The reaction complex was represented by the first point from left to right.According to the color distribution of the IRI isosurface,we can confirm that there was hydrogen bond formed by the H atom in thiophene and the O atom in Cs2CO3,vdW interaction,and steric effect in the thiophene ring apart from the covalent interactions of chemical bonds and ionic bonds of cation Cs+and theIRI isosurface maps showed that as the reaction progressed,the related H-bond noncovalent interaction transitioned to a normally chemical O-H bond,thus the IRI isosurface maps were becoming increasingly blue in the appropriately enlarged regions,whereas the original C-H bond progressively broke,lowering the electron density in the correspondingly bonding region leading to the transition of the strong covalent interaction to the weak attraction interaction between the carbanion and the removed proton.Besides,the electrostatic attractive interaction regions between the counter-cation Cs+and the carbanion were gradually reinforced together with the more vdW interaction regions.

Fig.14 visually showed the interaction changes in the reaction process of the forming carbanion attacking on CO2.The first point from the right to left was the reactant complex,wherein there was vdW interaction between the CO2and metal-nucleophile complex.As the reaction went on,the C-C bond that will be formed were gradually strengthened and then elongated along the bonding direction,resulting in the accumulation of the electron density in the bonding regions,which were mainly manifested in that the color of the IRI isosurface map became more and more blue in the corresponding regions.Besides,the structure of the thiophenecarboxylic acid anion was constantly adjusted to finally form a stable product.The weak attraction interactions and vdW interactions between the counter-cation Cs+and the host were gradually weakened,while the electrostatic attraction interactions between the cation Cs+and the formed carboxylate anion became stronger and stronger since the color of the IRI isosurfaces in this corresponding gradually increased region became more and more blue.Meanwhile,the corresponding steric effect also appeared in the regions between the ring and carboxylate anion,as a result,the matching isosurface gradually changed from green to orange.

3.6.Proof of the concept

Aiming to validate the previously stated mechanism details of the stoichiometric reaction of the direct C-H carboxylation reaction process,mainly for the cleavage of the C-H bond,there were also two other items being investigated from the experimental and computational perspectives,respectively,including the substitution of Cs and K in the mixed salt (containing K2CO3system with auxiliary carboxylate invariable and potassium pivalate KOPiv system with carbonate invariable)and the substitution of the reactant substrate of furoate(instead of thiophene monocarboxylate,i.e.,S atom was replaced by the O atom) and benzoate (instead of five-membered heterocyclic thiophene monocarboxylate,i.e.,the five-membered heterocyclic ring replaced by six-membered ring).

Fig.14.Variation of isosurface map of IRI=1.0 during the CO2 insertion step for formation product_1.

3.6.1.Impact of the combination effect of carbonate and pivalate on reaction outcome

As observed from the aforementioned experimental results that the replacement of Cs2CO3with K2CO3leaded to reduce the yield of reaction process from 8.33% to 1.08% under the condition that the auxiliary carboxylate (CsOPiv) remained unchanged.Furthermore,simply changing the carboxylate (CsOPiv to KOPiv replacement at constant carbonate)decreased the reaction yield to 4.29%.We performed a parallel calculation investigation on the direct carboxylation reaction of thiophene and CO2in the presence of K2CO3or potassium pivalate in order to provide a more in-depth understanding of the reaction mechanism of deprotonation driven by the base effect.

Fig.15.(a) Reaction energy profiles of the carboxylation reaction of K2CO3 and CsOPiv system.Energies are relative to the reactants of Reactant_3 and Cs cluster.(b) All reported structures and their important geometry parameters (distances are in Angstroms (10-10 m)).

Fig.16.(a)Reaction energy profiles of the carboxylation reaction K2CO3 and CsOPiv system.Energies are relative to the reactants of Reactant_3 and Cs cluster.(b)All reported structures and their important geometry parameters (distances are in Angstroms (10-10 m)).

The specified mechanism details of cesium pivalate-assisted K2CO3-promoted C-H carboxylation reaction (named CsOPiv-K2CO3system) were illustrated in Figs.15 and 16.The structures of the intermediates and transition states were the similar as the Cs2CO3alternative.As calculations demonstrated that the activation energy barrier in the C-H cleavage step was,respectively,20.5 kcal·mol-1and 40.9 kcal·mol-1for the thiophene and monocarboxylate,which were greater than 17.0 kcal·mol-1and 19.6 kcal·mol-1,respectively,for the corresponding Cs2CO3-mediated C-H carboxylation reaction.The correspondingly detailed reason for the difference in the calculated energy barrier in deprotonation step of Cs2CO3being replaced by K2CO3was shown in the activation strain analysis below.Note that there were also some slight differences in the CO2insertion step from the Cs2-CO3base case to the K2CO3base case.In the transition state structure,both the bond angle of O-C-O in CO2and the bond length of nucleophile-carbon in the K2CO3system were smaller than that in the Cs2CO3system whether for the thiophene and/or monocarboxylate,which was mainly due to the difference in the ion radius of different counter-cations,and its corresponding energy barrier in CO2activation step was higher.Besides,the corresponding reaction energy profiles of simply changing the carboxylate (named KOPiv-Cs2CO3system)were shown in Figs.S5 and S6 of Supporting Information owing to their similar structures of intermediates and transition states as the CsOPiv-Cs2CO3system.As calculations demonstrated that the activation energy barrier in the C-H cleavage step was,respectively,18.2 kcal·mol-1and 21.6 kcal·mol-1for the thiophene and monocarboxylate,which were slightly greater than those,17.0 kcal·mol-1and 19.6 kcal·mol-1,reported for the corresponding CsOPiv-assisted Cs2CO3-mediated C-H carboxylation reaction,while they were less than those found in CsOPiv-assisted K2CO3system.The correspondingly detailed reason for the difference in the calculated energy barrier in deprotonation step of simply changing the carboxylate was also shown in the following activation strain analysis.As observed that the computations predicted the trend in the reactivity of the stoichiometric reaction of thiophene with CO2was the CsOPiv-K2CO3＜KOPiv-Cs2CO3＜CsOPiv-Cs2CO3,and the corresponding experimental results (Table 2) elucidated that the reaction yield was decreased when either carbonate or pivalate was replaced,which was in line with the trend predicted by the calculations.To put it differently,the base effect was slightly stronger in cesium system than in potassium case.

Fig.17.Distortion/Interaction-activation strain analyses of substrate deprotonation step along the reaction coordinates for alternative combination of carbonate and pivalate: (a) thiophene;(b) thiophene-2-carboxylate.

Fig.18.The interaction analyses between the substrate and the mixed Cs2CO3-CsOPiv (left) K2CO3-CsOPiv (right) system: (a) reactant complex_3;(b) transition state structure TS_3；(c) intermediate Int_3.

In order to clearly reveal the origin of the difference in activation energy barrier for these two alternative substitutions in this C-H carboxylation reaction proceed,we also performed the distortion/interaction-activation strain analysis in the entire reaction energy profile along the reaction coordinate for the proton abstraction step in thiophene and monocarboxylate.It can be found from Fig.17 (a) that the potential barrier of the system where Cs2CO3was replaced by K2CO3was the highest among these three alternatives,which was caused by the combined effect of the weaker interaction energy and the stronger activation strain energy.Note that there was only a slight difference in distortion energy with the similar interaction energy term when simply changing the carboxylate.Depicted in Fig.17(b) indicated that the interaction energy term determined the height of activation energy barrier of the C-H activation process in monocarboxylate substrate.The most obvious support for this judgement was the case for substituting carbonate.These two possibilities had the similar distortion energy term along the reaction coordinate,while the higher activation energy barrier in CsOPiv-K2CO3system was stemmed from the less stabilizing interaction energy term when compared to the CsOPiv-Cs2CO3base case,because the interaction between the counter-cation K+and the heteroatom S on the thiophene ring and broken proton was very weak,respectively,compared with that of the counter-cation Cs+,resulting in a higher reaction energy barrier,and the corresponding interaction analyses IRI results for each intermediates were shown in Fig.18 for these two carbonate system.When simply replacing the carboxylate,the variation trends of the two energy terms were similar.Therefore,the less activation energy barrier in Cs2CO3system originated from the more stable interaction energy term in comparison with the K2CO3case.

Fig.19.Distortion/Interaction analysis for the direct carboxylation transition states of alternative substrates (1 kcal=4.186 kJ).

3.6.2.Effect of C-H substrate

As was well-known that the pKavalue was a very important indicator of whether the carbonate-mediated direct C-H carboxylation can be achieved.Another point that required further investigation was whether this linked pKawas the only factor that influenced the carbonate-promoted direct C-H carboxylation.Besides,according to the above results,the energy barrier of C-H activation step was higher than that of the following CO2insertion step whether for the formation of the mono-and/or dicarboxylate.In other words,the proton abstraction in carboxylation reaction was a rate-limiting step.Furthermore,the second C-H activation was a rate-determining step of these whole cesium carbonate-promoted C-H carboxylation reaction.Therefore,we took the proton abstraction of monocarboxylate,such as furoate(1a),thiophene monocarboxylate (1b) and benzoate (1c),as the research point to further elucidate the above question in terms of the experimental and computational insights.

Experimental results indicated that there were the harsh conditions to achieve the quantitative effect for realizing the C-H carboxylation for the other two substrates with respect to the substrate 1a,as shown in Fig.19,which meant that the order of the reactivity of the substrate was the 1a ＞1b ＞1c.The πelectron density in these aromatic carboxylates gradually decreased as observed by the close examination of the substrates,with the order being 1a ＞1b ＞1c.As a result,this electronic property could have a significant impact on the C-H carboxylation process induced by cesium carbonate.The corresponding calculation results also illustrated this point,that is,there were,respectively,the activation energy barrier in the cleavage of the C-H bond increased by 3.6 kcal·mol-1and 5.9 kcal·mol-1when the substrate changed from 1a to 1b or 1c,which also provided the evidence of the reactivity of the substrates.Therefore,from the perspective of experiment and calculation,it was proved that the substrate reactivity was consistent with the order of π electron density in their structure,which was contrary to the order of substrate reactivity given by the acidity coefficient (pKa(1a)≈35,pKa(1b)≈32.5 and pKa(1c)≈40).

In order to further disclose what factors may have caused the difference in activation energy barrier and reactivity of different substrates in this C-H carboxylation reaction proceed,the distortion/interaction-activation strain analysis model was employed.Depicted in Fig.19 indicated that the activation strain energy in the system mainly came from the substrate,while the small portion was from the carbonate itself during the process of the reactants in the free state structures shifting to their transition state structures.The total distortion energy of carbonate and substrate,destabilizing factor,was respectively,4.9 kcal·mol-1and 4.2 kcal·mol-1smaller for substrate 1a and 1b,compared with the substrate 1c.And the interaction energy term,the stabilizing factor,between the two twisted fragments of carbonate and substrate was decreased by 7.5 kcal·mol-1and 12.4 kcal·mol-1for substrate 1b and 1c even with the certain degree of energy penalty,respectively,with regards to the substrate 1a.Hence,the larger stabilizing interaction term leaded to the lower the reaction energy barrier in the case of low energy penalty.In other words,compared with substrate 1a,the higher energy barrier in proton abstraction step may be due to the weak interaction between the substrate and carbonate.Therefore,the π-electron density in these aromatic carboxylates was responsible for this phenomenon,which indicated that the lower height of activation energy barrier in the direct C-H carboxylation system was the result of the aromatic carboxylate containing the richer π-electron density.

4.Conclusions

Direct carboxylation of thiophene with CO2was carried out in the presence of mixed cesium carbonate and carboxylate salts in terms of the experiment and DFT calculations.The carboxylate salt could create the synergistic effect on the carbonate-promoted carboxylation and the reaction effect varied with the carboxylates owing to their different deprotonation abilities.As concluded from the phase behavior analysis that the reaction was proceeding at the(semi) molten salt interface through the thermal characterization techniques of DSC andin-situXRD.The pivalate-assisted Cs2CO3-promoted C-H carboxylation reaction had the best reaction effect since both the activation energy barriers in the proton abstraction steps were the minimal because of the weak Cs-heteroatom(thiophene) and Cs-H (removed) interactions,which were likely to enhance the acidity of C-H bond of coordinated monocarboxylate,lowering the C-H activation barrier.The energy barrier of C-H activation step was higher than that of the following CO2insertion step whether for the formation of the mono-and/or dicarboxylate,which indicated that the C-H deprotonation induced by the base was slow and the resulting carbon-centered nucleophile reacted quickly with CO2,which was consistent with the KIE’s results.Besides,these corresponding mechanistic insights were validated by further investigating the base and C-H substrates effects.Note that the substrate reactivity was consistent with the order of π electron density in their structure,which was contrary to the order of substrate reactivity given by the acidity coefficient.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.