Selective nucleophilic α-C alkylation of phenols with alcohols via Ti=Cα intermediate on anatase TiO2 surface

C−C bond forming reaction by alkylation of aryl rings is a main pillar of chemistry in the production of broad portfolios of chemical products. The dominant mechanism proceeds via electrophilic substitution of secondary and tertiary carbocations over acid catalysts, forming multiple aryl alkylation products non-selectively through all secondary and tertiary carbons in the alkyl chains but producing little α-C alkylation products because primary carbocations are poorly stable. Herein, we report that anatase TiO2 (TiO2-A) catalyzes nucleophilic α-C alkylation of phenols with alcohols in high selectivity to simply linear alkylphenols. Experimental and computational studies reveal the formation of Ti=C− bond with the α-carbon of the alkyl group at oxygen vacancies of the TiO2-A surface. The subsequent α-C alkylation by selective substitution of phenol ortho-C−H bond is verified by deuterium exchanged substrate and DFT calculations.


Response to Reviewers
Response to Reviewer #1 Comment (1): The authors should give the calculation methods for the product yield and the carbon balance.
Response: Thank you for your comments. We have added the calculation methods and the marked sentences in the Methods section of the manuscript as "The conversion of phenol, selectivity of phenolic compounds and product yield are calculated by Eq. 1, Eq. 2 and Eq. 3, respectively. In these equations, ninitial phenol and nfinal phenol are the molar amount of phenol before and after reaction. nproduct i is the molar amount of aromatic product i in the reaction mixture.
We also studied the molar mass balance of alcohols. The conversion of alcohol and product yield are calculated by Eq. 4 and Eq. 5. In these equations, ninitial alcohol and nfinal alcohol are the molar amount of alcohol before and after reaction. nproduct i is the molar amount of mono alkyl group in product i after the reaction."

Comment (2):
The boiling points of polyalkylation products are generally very high.
Is GC-MS accurate for quantitative detection.
Response: Thank you for your comments. The polyalkylation products generated by solid acid catalysts were identified by GC-MS, and calculated by the molar balance of phenol in our previous version of manuscript. In the revised manuscript, we recalculate the yield of polyalkylation products by 1 H NMR spectra with an internal standard method. By combining results of GC-FID and NMR methods, all alkylation products are accurately quantified now. We have updated the results in Fig. 1 and added the marked sentence in the manuscript as "Several isopropyl polyalkylation products are identified by GC-MS and calculated by NMR spectra with the internal standard method."

Comment (3):
The feeding ratio between phenol and 1-propanol is 1:4, why? Does the ratio significantly influence the catalytic performance?
Response: Thank you for your comments. As shown in Fig. S1C in the revised SI, we examined the feeding ratio between phenol and 1-propanol under standard reaction conditions. With the feeding ratio varying from 1:1 to 1:4, product selectivity had no obvious changes, and conversions of phenol gradually increased. This is due to the relatively high concentration of 1-propanol favors the reaction equilibrium. But the catalytic nature of TiO2-A is not affected by the feeding ratio. We have added the results as Fig. S1C in the revised SI and the marked sentence in the manuscript as Response: Thank you for your comments. As shown in the following Scheme R1, we examined several arenes to react with 1-propanol under standard reaction conditions.
No product was detected, and therefore they would not participate in the reaction when used as the solvent.

Comment (5):
To demonstrate that the crystal structure did not change appreciably during the reaction, high-resolution electron microscopic images, such as those in Fig.   2b, should be presented.
Response: Thank you for your suggestion. The HRTEM image of used TiO2-A was added in Fig. 2c in the manuscript. The crystal structure of TiO2-A did not change obviously after the reaction.

Comment (6):
The characterizations of oxygen vacancies should be added. For example, O 1s XPS, EPR. Besides, it will be more convincible if the authors can present the semi-quantitative relationship between catalytic performance and the content of oxygen vacancies.
Response: Thank you for your suggestion. The characterizations of oxygen vacancy of TiO2 have been widely reported using XPS and EPR. In these works, the oxygen vacancy-rich TiO2 is synthesized by specific methods, such as hydrogen treatment, 1,2 solvothermal reaction, 3,4 anion doping, 5 and so on. However, in our work, the pristine TiO2-A was used as the catalyst, and the computational studies showed that diffusion of oxygen vacancy from the subsurface to surface was induced by substrate adsorption in the reaction. Therefore, in situ characterization methods are essential for the detection of surface oxygen vacancy of TiO2-A. Unfortunately, it is still challenging to conduct in situ XPS or EPR tests with substrate feeding at 300 ℃.
We also prepared a propanol adsorbed TiO2-A sample in the glovebox under Ar atmosphere and performed the XPS and EPR tests at room temperature. As shown in   Response: Thank you for your comments. As discussed above, the surface oxygen vacancy is induced by substrate adsorption, and then produces four coordinated Ti atom. The unsaturated Ti atom is favorable to coordinate with substrate and then involved in the subsequent reaction process quickly. The operando XPS test is necessary to monitor the valance state of Ti of TiO2-A surface during the reaction, but this technology is still limited now. Fig. R3 shows the Ti 2p XPS spectra of bare TiO2-A and propanol adsorbed TiO2-A at room temperature. And the peak shows no obvious shift in this ex-situ test. And it is still challenging now to detect because of the technical limitation. The TiO2-A catalyst exhibited similar reactivity to the corresponding primary alcohols, and more importantly, maintained the selectivity of α-C alkylation."

Comment (2):
The scope of phenols is inadequate. More functional groups need to be tested.
Response: Thank you for your suggestion. As shown in Scheme S1 in the revised SI, besides the 3-chlorophenol and 3-methylphenol in the previous manuscript, more substituent phenols were tested, including amino group, nitro group, methoxy group and dimethyl group as substituted groups at meta position of phenol (Scheme S1, Entry 1-4). Generally, the conversions of phenols are higher than 84%, with more than 80% selectivity to ortho-substituted α-C alkylation products. The effect of substituent groups on catalytic performance is slight. In addition, 4-chlorophenol and 4-methylphenol (Scheme S1, Entry 5-6) were tested to study the effect of substituent position. The para substituted phenols also exhibited good reactivity. We have added the results in Scheme S1 and the marked sentence in the manuscript as "Phenols with additional substituents were further tested by reacting with 1-propanol (Scheme S1, entries 1-6), and high conversions of phenols (more than 84%) and selectivity to ortho-substituted n-propyl products (more than 80%) were obtained."

Comment (3): What about the results with naphthols?
Response: Thank you for your comments. As shown in Scheme S1 in the revised SI, we tested several naphthols to react with 1-propanol under standard reaction conditions (Scheme S1, Entry 7-10). The 1-naphthol conversion rate reached 93.6%, with 87.9% selectivity to 2-n-propyl-1-naphthol. Substituted 1-naphthols with methyl, chlorine and methoxy groups were also studied and more than 90% conversions were obtained. It is worth noting that the selective alkylation at ortho position of hydroxyl group was still maintained for naphthols. Overall, the substrate scope of our TiO2-A catalyzed selective alkylation process can be extended to naphthols. We have added the results in Scheme S1 and the marked sentences in the manuscript as "In addition, naphthols as alternative substrates also showed similar alkylation selectivity in high conversions (Scheme S1, entries 7-10). These results demonstrate that the TiO2-A catalyzed selective α-C alkylation reaction applies to a broad scope of substrates."

Comment (4):
What about the reusability of the TiO2-A catalyst?
Response: Thank you for your comments. As shown in Fig. S2A in the revised SI, we studied the reusability of TiO2-A in alkylation reaction of phenol with 1-propanol.
After three-times reuse of the TiO2-A, the phenol conversion and product yields remained no obvious changes. The other example of reuse test is the alkylation reaction of phenol with 1-dodecanol (Fig. S2B). The improved Alc-TiO2-A catalysts also showed good catalytic stability. We have marked the sentences in the manuscript as "Its structure stability was maintained with multiple reuse tests. After three-times reuse of the TiO2-A, the phenol conversion and product yields remained no obvious changes (Fig. S2A). It is interesting to note that the used TiO2-A had a better mass balance than the fresh TiO2-A. This may be due to the adsorption of phenol on the fresh TiO2-A surface at the termination of the first reaction test and the surface has already reached saturated phenol adsorption in subsequent reuses of the TiO2-A catalyst." "The performance maintained after three-times reuse without additional treatment (Fig. S2B), making it a superior catalyst for industrial process."