Bulky magnesium(ii) and sodium(i) bisphenoxide catalysts for chemoselective transesterification of methyl (meth)acrylates

Given the industrial importance of (meth)acrylate esters, various groups have devoted considerable effort to investigating their chemoselective transesterification. In 2021, we developed magnesium(ii) and sodium(i) complexes derived from 2,6-di-tert-butyl-p-cresol (BHT-H) as chemoselective catalysts for the transesterification of methyl acrylate (MA) and methyl methacrylate (MMA), respectively. Based on our results, we report the discovery of magnesium(ii) and sodium(i) salts derived from 6,6′-(propane-2,2′-diyl)bis(2,4-di-tert-butylphenol) (PBTP-H2), i.e. Mg(PBTP) and Na2(PBTP), which are 41 and 81 times more effective catalysts than Mg(BHT)2 and Na(BHT) for the transesterification of MA and MMA, respectively. These new catalysts are highly effective across an extensive range of alcohols, including primary and secondary alcohols, diols, and triols. Overall, this efficient transesterification technology can be expected to find practical applications in industrial process chemistry.


Introduction
With industrial applications encompassing heat-resistant adhesives, varnishes, UV coatings, textile nishing, and polymeric plastics, 1 (meth)acrylate esters are produced at the million-ton scale 2 and are considered to be among the most important manufactured chemicals. The desirable physical properties of poly(meth)acrylate esters, such as their exibility, transparency, and weatherability, can also be controlled and ne-tuned to the requirements of their intended applications, most commonly via the functionalization of the ester group, which adds to their versatility. 2 While usually produced via the reaction of (meth)acryloyl chlorides with alcohols, the production of functionalized (meth)acrylates through a trans-esterication process 3-7 instead would theoretically reduce the amount of stoichiometric halide wastes produced. The use of methyl (meth)acrylates (MA and MMA) instead of their corresponding carboxylic acids is also advantageous in terms of handling given their superior solubility proles in organic solvents.
(Meth)acrylate esters are a,b-unsaturated esters capable of undergoing both Michael addition 8 and transesterication. [3][4][5][6][7] The extent of each of the two competing pathways depends on the character of the nucleophile and the mode of activation of the carbonyl group: nucleophiles with higher charge density, also known as "hard" nucleophiles, and carbonyl groups activated by a densely charged Lewis acid, such as that of a metal cation, tend to undergo nucleophilic substitution (Scheme 1a), whereas Michael addition is favored by "so" nucleophiles that have better molecular orbital matching with the carbon-carbon double bond (Scheme 1b). 9 As a result, a competent catalytic system for the transesterication of (meth)acrylate esters must exhibit excellent chemoselectivity toward the nucleophilic substitution reaction to maximize the transesterication Cite this: Chem. Sci., 2023, 14, 566 All publication charges for this article have been paid for by the Royal Society of Chemistry product yield. This is especially difficult in the case of MA, which has the least sterically hindered a,b-unsaturated end in comparison to other acrylate analogues, and is therefore more prone to undergo the undesired Michael addition reaction. In fact, the harsh reaction conditions employed in many acid/baserelated transesterication strategies invariably cause not only undesirable Michael addition reactions, but also the premature Michael-addition-initiated polymerization of MA or MMA (Scheme 1c), which results in low monomeric product yield.
Given the industrial signicance of (meth)acrylates, it is hardly surprising that various groups have devoted considerable effort to the investigation of their chemoselective trans-esterication. In line with the aforementioned Lewis-acid activation strategy to favor transesterication, much attention has been paid to the application of heterogenous metal catalysts such as Ti(V), 10 Zr(IV), 11 and Ca(NO 3 ) 2 /g-Al 2 O 3 , 12 to chemoselective transesterications. However, the main concerns in all the afore-mentioned studies are the physical and catalytic properties of the heterogenous catalysts; the substrate scopes of the systems are underdeveloped and remains unexplored. It is unclear whether these catalysts can efficiently catalyze the transesterication of acrylates with other structurally distinct alcohols, such as complex primary alcohols, secondary alcohols, and diols.
In 2016, Ohshima accomplished the chemoselective trans-esterication of MMA and MA with a diverse array of primary alcohols, diols, and triols with excellent yield and selectivity using a zinc(II) cluster catalyst, 6 albeit that examples of secondary alcohols were lacking. 6h It should also be noted here that 20 mol% of toxic 4-(dimethylamino)pyridine (DMAP) is required as a ligand for this catalytic system to function optimally, which poses severe environmental concerns for its widespread adoption as a viable industrial solution. Thus, in 2017, Hashimoto and Ootsuka at Toagosei Co., Ltd. reported a new catalytic method using triethylenediamine (DABCO) and zinc(acrylate). 6i According to the patent, 6i Michael addition of DABCO to MMA gives an enolate anion, which is further transformed to a dimeric vinyl ester intermediate through transesterication with MMA. Subsequently the trans-esterication of this intermediate with primary alcohol occurs to give the corresponding ester and MMA.
In 2018, we developed tetramethylammonium methyl carbonate as an efficient, general and metal-free catalyst for transesterication reactions. 5b This catalyst has been successfully applied across a wide range of ester and alcohol partners, including examples using MMA as a substrate in reactions with secondary alcohols, diols, and triols; however, reactions using MA give complex mixtures.
It was evident that the chemical research community had made great headway in the development of catalysts capable of the efficient and chemoselective transesterication of (meth) acrylate esters. Nonetheless, a truly robust, economical, and environmentally friendly catalytic system capable of this feat remained elusive and warranted further research.
In 2021, we developed magnesium(II) and sodium(I) complexes (Mg(BHT) 2 and Na(BHT)) derived from 2,6-di-tertbutyl-p-cresol (BHT-H) as efficient and chemoselective catalysts for the transesterication of MA and MMA, respectively (Schemes 2a and b). 7 These catalysts are effective for various primary and secondary alcohols, diols, triols, and even tetraols. Interestingly, dimeric Mg(II) complexes, [Mg(OAr)(OR)] 2 , have been proposed as active species in the ring-opening polymerization (ROP) reactions. 13 Therefore, we considered a mechanistic possibility including such dimeric Mg(II) complexes based on DFT calculations (Scheme 2c) in our previous paper. 7 The observed energy proles showed that [Mg(OAr)(OR)] 2 favors the transesterication pathway over the Michael-addition pathway. Additionally, according to the DFT results, the dimeric Mg(II) complex (R = Bn) is by only 1.0 kcal mol −1 more stable than the monomeric Mg(II) complex.
Based on our previous study, 7 we report herein the discovery of magnesium(II) and sodium(I) salts derived from 6,6 ′ -(propane-2,2 ′ -diyl)bis(2,4-di-tert-butylphenol) (PBTP-H 2 ), i.e., Mg(PBTP) and Na 2 (PBTP), which serve as new catalysts for the transesterication of MA and MMA, respectively (Scheme 3). These catalysts are easily prepared from inexpensive chemicals, are much more active than Mg(BHT) 2 and Na(BHT), 7 and are effective across an extensive range of alcohols including primary and secondary alcohols as well as diols and triols. In general, this efficient transesterication technology can be expected to be attractive for practical applications in the context of industrial process chemistry.

Results and discussion
Initially, the chemoselectivity of several magnesium(II) aryloxides for the transesterication of MA with benzyl alcohol (1a) was estimated under equilibrium conditions without MS 5A. Representative results are shown in Table 1. When the in situgenerated Mg(BHT) 2 was used as the catalyst, the desired product (2a) was obtained in 40% yield together with side product 6a (1% yield) from the competing Michael addition aer 30 minutes (entry 1). Much to our delight, Mg(PBTP) was able to furnish 2a in 55% yield aer 1 hour without any formation of undesirable side products (entry 2).
Next, we attempted the transesterication of MA with isoborneol (1b) as a model reaction using magnesium(II) aryloxides as catalysts, and monitored the reaction progress at various time intervals to compare their catalytic activity. The respective results are summarized in Table 2. We rst tried Mg(BHT) 2 , and aer 6, 8, and 12 hours, 2b was detected in 62%, 85%, and 91% yield, respectively; aer 19 hours, 2b was obtained in 99% yield (entry 1). The use of Mg(PBTP) provided 2b in 90% and 99% yield aer 6 and 8 hours, respectively (entry 2). A comparison of entry 1 with entry 2 suggests that the bidentate ligation of Mg(PBTP) plays an important role in increasing the catalytic activity. Based on this inspired discovery, we then examined Mg (TBP), which was generated in situ from Bu 2 Mg and 3,3 ′ ,5,5 ′ -tertbutyl-1,1 ′ -biiphenyl-2,2 ′ -diol (TBP-H 2 ). However, only 35% and 42% yields of 2b were detected aer 6 and 8 hours, respectively; aer 24 hours the yield of 2b reached just 60% (entry 3). Interestingly, the ring size of the metalacyclic structure, in addition to the bidentate ligation of the magnesium(II) aryloxides, is also important for increasing the catalytic activity.
Next, the substrate scope of Mg(PBTP) as a catalyst for the transesterication of MA was investigated as shown in Table 3. Mg(PBTP) was more active than Mg(BHT) 2 , and the trans-esterication was completed much faster. The trans-esterication proceeded with primary alcohols such as a saturated fatty alcohol (1c), allylic alcohols (1e, 1f), and (hetero)arylmethyl alcohols (1a, 1d) to give the desired esters in quantitative yield. We were able to scale up the test reaction by a factor of ve to obtain 1.62 g of 2a (>99% yield, 4 h) using a lowered catalytic load of 1 mol% Mg(PBTP), demonstrating the viability of the catalytic system for the gram-scale synthesis.
Additionally, Mg(PBTP) was applicable to secondary alcohols, although longer reaction times were required. For example, (−)-menthyl acrylate (2g), which is potentially useful for the synthesis of pressure-sensitive adhesives (PSA), 14 and isoborneol-derived acrylate 2b were obtained in >99% yield. 5-Nonanyl acrylate (2h), a monomer for synthesis that is a common additive for UV-adhesives, was also synthesized successfully. 15 Trimethylolpropane acrylate 2l, which is commonly used in the production of UV coatings and has also recently found applications in the development of superswelling hydrogels 16 and as an alumina pigment modier, 5c was successfully obtained in 90% yield aer 24 hours. This result represents a successful chemoselective transesterication of an industrially useful triol. We increased the amount of 5A molecular sieves to 800 mg to account for the three-fold increase in equivalents of the side product methanol produced per molecule of target product. The transesterication of various  industrially useful diols (1i-1k) to produce 2i-2k was also successful with extraordinary yields (>99%) and reaction times of 6-24 hours. Once again, we used a higher amount of 5Å molecular sieves (600 mg) for these diols to facilitate the removal of methanol. Next, we focused on the chemoselective transesterication of MMA. As expected, Mg(PBTP), as well as Mg(BHT) 2 , were less effective for MMA. Fortunately, Na 2 (PBTP) (1.25 mol%) was highly effective and superior to Na(BHT) (2.5 mol%), as shown in Table 4. Under the standard reaction conditions, all the primary alcohols (1c, 1e, 1a) gave the corresponding products in quantitative yield within 10 minutes. Glycidyl methacrylate (7m), which is chelating and particularly susceptible to nucleophilic attack, which sometimes results in its decomposition and polymerization, was obtained in 87% yield. 5b-d Tri-uoroethyl methacrylate (7n), which is also unstable due to its electron-withdrawing nature was also obtained in 93% yield. 17 Quinine methacrylate (7p), a potentially useful monomer for the synthesis of molecularly imprinted polymers for the screening of halides and amino acids, 18 was successfully obtained in quantitative yield despite its notable steric demand. Similarly, other sec-alkyl methacrylates such as 7b, 7g, 7h, and 7o were successfully synthesized. The efficacy of Na 2 (PBTP) was Table 3 Substrate scope of Mg(PBTP) as a catalyst for the transesterification of MA a a Unless otherwise noted, the reaction was carried out using MA (14 mmol), 1 (2 mmol), Mg(PBTP) (5 mol%), Cu(CS 2 NMe 2 ) 2 (polymerization inhibitor, 0.2 mol%), and MS 5A (0.4 g) at 25°C. Isolated yields aer ash column chromatography on silica gel are shown. b Mg(BHT) 2 was used instead of Mg(PBTP) under otherwise identical conditions. See ref. 7. c The reaction was carried out using MA (70 mmol), 1a (10 mmol), Mg(PBTP) (1 mol%), Cu(CS 2 NMe 2 ) 2 (polymerization inhibitor, 0.2 mol%), and MS 5A (2 g) at 25°C. d Mg(PBTP) (10 mol%) was used. e MS 5A (0.6 g) was used. f MS 5A (0.8 g) was used. Table 4 Substrate scope of Na 2 (PBTP) as a catalyst for the transesterification of MMA a a Unless otherwise noted, the reaction was carried out with MMA (14 mmol), 1 (2 mmol), Na 2 (PBTP) (1.25 mol%), 4-acetamido-TEMPO (polymerization inhibitor, 0.1 mol%), and MS 5A (0.4 g) at 25°C. Isolated yields aer ash column chromatography on silica gel are shown. b Na(BHT) (2.5 mol%) was used instead of Na 2 (PBTP) (1.25 mol%) under otherwise identical conditions. See ref. 7. c MS 5A (0.8 g) was used. d MMA (28 mmol) and MS 5A (0.8 g) were used. e MS 5A (0.6 g) was used. f Na 2 (PBTP) (2.5 mol%) was used. g Na(BHT) (5 mol%) was used. highlighted by the facile transesterication of diols to furnish diesters 7r, 7i, and 7k quantitatively in a shorter time. In addition to these successful results, transesterication of trimethylolpropane (1l) quantitatively furnished triester 7l when an increased catalyst loading of Na 2 (PBTP) (2.5 mol%) was used.
To numerically verify the higher catalytic activity of Mg(PBTP) compared to that of Mg(BHT) 2 , density functional theory (DFT) calculations were performed for the trans-esterication of MA with 1a catalyzed by [Mg(PBTP)] 2 based on its crystal structure shown in Fig. 1. 19 The potential energy prole is explained in Fig. 2 and 3. The pathway shown is the most plausible in the present study. [Mg(PBTP)] 2 is stabilized by the coordination of 1a. 1a is then activated as int1 by intramolecular proton transfer in [Mg(PBTP)] 2 $1a via TS1. In addition, MA is also activated as Int2 by coordination of Int1. Nucleophilic addition of BnO − to MA in Int2 then occurs via TS2 to form Int3. This is the rate-determining step (RDS), and the calculated activation energy is 10.8 kcal mol −1 (E EDS a2 ). Thus, this elementary step includes both proton transfer from 1a to OAr and C-O formation in a stepwise manner. Subsequently, methanol is released from Int3 via TS3. As shown in Table 5 2 enhances the rate constant of RDS for the former, which is by a factor of 41 higher than that of the latter. 20 This estimation provides a reasonable interpretation for the experimental activity enhancement of [Mg(PBTP)] 2 as shown in Table 3.   Next, DFT calculations were performed for the trans-esterication of MMA with 1a catalyzed by Na 2 (PBTP) based on its crystal structure shown in Fig. 1. In this case, 1a and MMA are activated at the same time by coordination of Na 2 (PBTP). Subsequently both proton transfer from 1a to OAr and nucleophilic addition of BnO − occur via TS4 to give Int5 in a concerted manner. This is the RDS, and the calculated activation energy is 18.8 kcal mol −1 (E EDS a ). As shown in Table 5, the E EDS a value using Na 2 (PBTP) is 2.6 kcal mol −1 lower than that using Na(BHT). This difference in E EDS a enhances the rate constant of RDS for Na 2 (-PBTP), which is by a factor of 81 higher than that for Na(BHT). 20 This estimation provides a reasonable interpretation for the experimental activity enhancement of Na 2 (PBTP) shown in Table 4.
Since Na 2 (PBTP) has two Na sites, another possibility is sequential transesterication at the other Na site. This indicates one Na site has MMA and 1a and ready for the reaction. The other Na site is coordinated by the product of the reaction. In fact, this structure corresponds to Int5 in Fig. 4. The Int5 state is 15.4 kcal mol −1 higher than the reactant state, Na 2 (PBTP)$ 2(MMA)$2 (1a). This result indicates that 7a and MeOH are predominantly produced from Int 5, MMA, and 1a, and Na 2 (-PBTP)$2(MMA)$2(1a) is regenerated at the same time. Thus, the possibility of the sequential transesterication at the other Na site of Int 5 is ruled out.

Conclusions
In summary, we have developed the catalysts [Mg(PBTP)] 2 and Na 2 (PBTP), which are highly effective for the chemoselective transesterication of MA and MMA, respectively, under mild conditions at 25°C. These catalysts are superior to Mg(BHT) 2 and Na(BHT), respectively, which we had previously developed. 7 The results of DFT calculations strongly support our experimental results. Overall, based on the observed chemoselectivity, high yields, mild conditions, and lack of toxic metal species, the catalytic methods reported here represent new practical, green, and sustainable catalyst candidates for the industrial synthesis of acrylates.

Data availability
The data supporting this study is available within the main text and the associated ESI. †

Conflicts of interest
There are no conicts to declare.