Selective Polymerization Catalysis from Monomer Mixtures: Using a Commercial Cr‐Salen Catalyst To Access ABA Block Polyesters

Abstract ABA triblock polyesters are synthesized using a commercially available chromium salen catalyst, in one pot, from monomer mixtures comprising epoxide, anhydride and lactone. The catalysis is highly selective and applies a single catalyst in two distinct pathways. It occurs first by epoxide/anhydride ring‐opening copolymerization and subsequently by lactone ring‐opening polymerization. It is used to produce various new ABA polyester polyols; these polyols can undergo post‐functionalization and chain‐extension reactions. The ability to use a commercial catalyst and switchable catalysis with monomer mixtures is expected to facilitate future explorations of new classes of block polymers.

Inpolymerization catalysis,t he ability to selectively control (block) sequence using monomer mixtures remains as ignificant challenge. [1] Nature overcomes this problem with exquisite selectivity and catalyzes thousands of reactions to form different biopolymers,i ncluding sugars,p eptides,a nd DNA, where precisely defined sequence determines function. Synthetic mimics of biosynthesis are most successful when sequential monomer coupling reactions are applied, analogous to artificial peptide synthesis. [2] Nonetheless,s uch processes are not especially suitable for synthetic polymer production-they are generally too labor intensive and timeconsuming to evaluate new block copolymers.
This work describes amethod to prepare block polyesters, which are relevant as degradable and, in some cases,b iorenewable materials useful as elastomers,f ibers,h ealthcare materials,a nd drug-delivery vectors and in electronics. [3] Generally,b lock polyesters are synthesized by sequential polymerization reactions and/or with macro-initiators-such methods can be limited by the conversion efficiency,i ntermediary purification steps,a nd by the nature of the repeat unit chemistry.Anattractive alternative would be to develop catalytic processes that selectively enchain particular blocks, but such switchable catalysis remains under-developed. In 1985, Inoue and co-workers reported an Al-porphyrin catalyst able to prepare block polyesters from mixtures of epoxide, anhydride,a nd lactone. [4] Nonetheless,u nderstanding the reaction was hindered as the sole lactone investigated was bbutyrolactone,which can ring-open at two sites complicating enchainment mechanisms.I n2 014, we reported switch catalysis using as ingle homogeneous dizinc catalyst which selectively polymerized mixtures of lactone (CL), epoxide (CHO), and carbon dioxide to form single-block polymer structures. [5] Experimental and theoretical studies suggested that the selectivity resulted from kinetic and thermodynamic control by the metal-chain end group. [6] This process is different from terpolymerization because two different polymerization cycles are accessed and the dominant catalytic cycle is switched by the chemistry of the catalyst-polymer chain end group.I ti sa lso preferable to tandem or multifunctional catalysis because asingle catalyst is active in both catalytic cycles. [7] Very recently,R ieger and co-workers demonstrated switch catalysis,u sing ad izinc b-diiminate catalyst, to prepare block/random copolymers from b-butyrolactone,c yclohexene,o xide and carbon dioxide. [8] It is important to understand the generality of this new switch catalysis,particularly its applicability beyond zinc complexes. Its uptake should be facilitated and accelerated by determining whether it also applies to commercially available polymerization catalysts.
Thep rocess requires as ingle catalyst active for both lactone ROPa nd epoxide/anhydride ROCOP (Scheme 1). We targeted ac ommercial Cr-salen catalyst [SalcyCrCl], which is applied with equimolar addition of co-catalyst (PPNCl). [7b,d,f,9] Thec atalyst system has precedent for other Scheme 1. The switch catalysis pathways proposed using mixtures of DL, NBA, and CHO. alternating copolymerizations (ROCOP), [7b,d,f,h, 9] and related systems are active in cyclic carbonate ROP, b-butyrolactone ROP, and for the copolymerization of dihydrocoumarin and propylene oxide. [9d, 10] Firstly,the Cr catalyst system was tested separately for ROCOP and ROP, under conditions relevant to subsequent switch catalysis (Table S1). Each polymerization occurred with high conversion and selectivity,b ut the resulting polyesters showed bimodal molar mass distributions; indeed similar bimodality was observed previously for CO 2 / CHO ROCOP using Cr-salen catalysts. [7h] To control the molar mass distribution, polymerizations were conducted under immortal conditions,t hat is,w ith the addition of various amounts of 1,2-cyclohexanediol (CHD,5 -20 equiv) ( Figure S1). Theexcess alcohol controls molar mass via rapid and reversible exchange equilibria;t he use of ad iol ensures formation of ABAt riblock copolymers.U nder optimized conditions (> 10 equiv CHD), both DL ROPand NBA/CHO ROCOP formed polyesters showing monomodal molar mass distributions with predominantly dihydroxyl chain end groups ( Figures S2 and S3). It should be noted that as mall fraction (< 10 mol %) of chloro-initiated chains should be present; these chains do not affect the GPC traces but were detected by MALDI-ToF ( Figure S3).
Having demonstrated that Cr-salen catalysts were active for the separate polymerizations,c atalysis using mixtures of DL/CHO/NBA, again with 10 equiv CHD,w as investigated. Ther eaction was successful and ABA-type polyesters were formed with narrow,m onomodal molar mass distributions ( Figure S1;A = PDL, B = PCHNBE). Ther eaction was monitored with regular removal of aliquots,w hich were analyzed using 1 HNMR spectroscopy to determine conversion;this was achieved by integration of monomer signals vs. an internal standard ( Figure 1, Figures S6 and S7). Over the first 1.5 h, only CHO/NBAR OCOP occurred producing the alternating polyester (PCHNBE). Theh igh selectivity was evidenced by ar apid reduction in anhydride concentration and the concomitant growth of signals assigned to PCHNBE. TheR OCOP catalysis was quite efficient, showing aT OF of ca. 67 h À1 .Importantly,over this time period there was almost no change in the signals assigned to the lactone (< 3% by NMR, Table S2) and no evidence for any epoxide homopolymerization. After 1.5 h, the anhydride was fully consumed and subsequently DL ROPo ccurred slowly,r eaching completion after 48 h( Figure 1, TOF % 3h À1 ). Ther esulting polymer contains 66 % cis ester functionalities,a nd 34 % isomerized trans ester linkages,a sd etermined by analysis of the polyester degradation products ( Figure S9).
Thes pectroscopic data and TOFv alues clearly indicate the high catalytic selectivity for CHO/NBA/DL, with ROCOP occurring before ROP, but such selectivity could  Table S2).
form either triblock polyester (PDL-b-PCHNBE-b-PDL) or am ixture of both polymers (PCHNBE + PDL). In order to characterize the polymer composition, aliquots were analyzed by GPC (Figure 1). Them olar masses increase with conversion and in all cases samples show narrow,m onomodal distributions (1.11 < < 1.32);such data are strongly indicative of block polyester formation. Signals corresponding to both blocks were observed by 1 HNMR spectroscopy and 13 C{ 1 H} NMR spectroscopy (Figures S10-S14). The 13 C{ 1 H} NMR spectrum also showed low-intensity signals at intermediate chemical shifts which are assigned to the minor epimerized isomer ( Figure S13). Such signals are not attributed to transesterification since reaction of the polymer with an efficient transesterification catalyst (DBU) significantly increased both the number and intensity of intermediary chemical shift signals ( Figure S15). Furthermore,t he switch catalysis was also generalized to produce ab lock polyester from PA /CHO/DL;i ts 13 C{ 1 H} NMR spectrum showed only signals for the two blocks with no intermediate signals, consistent with alack of transesterification reactions ( Figures  S26-S29). [6a] TheD OSY NMR spectrum of the polymer formed from DL/CHO/NBAs hows as ingle diffusion coefficient for all signals,c onsistent with block polyester formation. Theanalogous blend of constituent polymers shows two diffusion coefficients ( Figure S16). Ther eaction of its hydroxyl end groups with 2-chloro-4,4,5,5-tetramethyldioxaphospholane enabled analysis by 31 P{ 1 H} NMR spectroscopy to differentiate the end groups. [9b] Theh omopolymers show different signals (146.5 = PCHPE and 147.1 ppm = PDL) and the block polymer shows only one signal, at 147.1 ppm, consistent with its ABAs tructure (Figures S17 and S18 ,  Table S3). Finally,the crude block polymer was purified using hexane,which is known to dissolve the homopolymer (PDL). There was no detectable homopolymer in the hexane and the composition of the block polymer remained the same before and after purifications ( Figure S10). Thus,a ll the analytical tests confirmed the formation of single-block polymer structure,that is,PDL-b-PCHNBE-b-PDL.
Thec atalysis is proposed to occur via two different polymerization cycles that are linked by ac ommon Cralkoxide intermediate (Scheme 1). It is important to note that at hird cycle,i nvolving epoxide ring-opening polymerization (forming ether linkages), is not accessed even over prolonged reaction times,a sc onfirmed by 1 HNMR and MALDI analysis ( Figures S4 and S5). Our hypothesis is that large differences in the rate of anhydride vs.l actone insertion and ahigh barrier to Cr-carboxylate reaction with lactone (and Cralkoxide with epoxide) control the selectivity.More generally, the above rationale also allows for scenarios where both polymerization rates are similar and as tatistical copolyester might form (vide infra).
Next, these hypotheses were specifically tested using other monomers and conditions.F irstly,e xperiments were conducted to confirm the reactivity of the Cr-carboxylate intermediate.T he preliminary kinetic analyses indicate ROCOP rates that are zero order in anhydride concentration ( Figures S7 and S8)-that is,d uring ROCOP the catalyst resting state is the Cr-carboxylate species.T otest the stability of this intermediate towards reaction with lactone,apolymer-ization reaction was conducted using excess anhydride and lactone vs.e poxide (DL/NBA/CHO,2 00:200:125). ROCOP proceeded until the epoxide was consumed ( % 60 %N BA conversion) producing only alternating polyester,P CHNBE. At this point no further conversion of any monomer occurred even over prolonged reaction times (4 days) and despite the presence of excess DL (200 equiv) ( Figure S19). In order to confirm the catalyst had not decomposed, the polymerization was "switched on" by the addition of af urther 200 equiv of epoxide.A tt his point, ROCOP was resumed until complete anhydride consumption which was followed by DL ROP leading to triblock polyester formation (Figures S20 and S21). It is important to note that the proposed pathway does not examine the intimate mechanism or active catalyst structure-(s) which have been the subject of extensive investigations. [7b,d,9b,c, 11, 12] To test the catalytic scope,aseries of triblock polyesters of differing compositions were prepared by controlling the relative amounts of monomers in the mixture.I na ll cases, high monomer conversions were achieved and well-defined, low-molar-mass block polyesters formed (Figures S22 and S23). DSC analyses showed amorphous structures with as ingle glass-transition temperature,w hich is indicative of block miscibility.T he block polyester composition was directly controlled by the monomer composition in the mixture and glass-transition temperatures could be tuned from À30-111 8 8C (Table S4 and Figures S24 and S25). Several factors contribute to (micro)-phase separation including polymer architecture,b lock volume fractions,t he Flory-Huggins interaction parameter (c), and degree of polymerization. [13] Here,t he block miscibility is most likely ac onsequence of the low molar masses,w hich are expected to fall below the entanglement molar masses.S imilar effects were observed for ABApoly(styrene-b-isoprene-b-styrene): at low M n ,asingle glass-transition temperature was observed (M n = 11 000 gmol À1 ,T g = 38 8 8C), at intermediate M n values,t wo intermediate glass-transition temperatures occur (M n = 16 000 gmol À1 ,T g,1 = À48 8 8C, T g,2 = 43 8 8C) whilst at higher M n ,t here are two glass-transition temperatures,s imilar to the homopolymers (M n = 161 000 gmol À1 ,T g,1 = À65 8 8C, T g,2 = 105 8 8C). [14] TheC r-salen switch catalyst system was also applied to ar ange of different monomer mixtures;t he influence of the anhydride structure was investigated (Table 1). Each reaction and polymer product was fully characterized using arange of techniques,including NMR analysis ( 1 H, 13 C, COSY,HSQC), conversion vs.t ime plots (using both IR and NMR data), GPC,and MALDI, and by comparison of composition before and after isolation. Table 1p rovides an overview of composition and relative rate data;t he complete characterization data sets for each block polymer are provided in the Supporting Information (Figures S26-S75). Them ajority of monomer mixtures resulted in selective formation of only ABAt riblock polyesters (this applies to mixtures of PA , THPA, TCA1, or TCA2w ith CHO/DL). In all these cases, fast and selective ROCOP is followed by slow ROP (< 5% DL conversion at > 99 %conversion of anhydride). Analysis of the relative rates reveals that selective catalysis occurs when the rate of ROCOP is > 20 times that of ROP,i nl ine with the mechanistic hypothesis (Scheme 1). Thep olymerizations all proceed with high monomer conversions and form block polyesters with monomodal molar mass distributions. ABA-type block polyester structures were confirmed through multiple experiments,a nalogous to the range of characterizations of PDL-b-PCHNBE-b-PDL. Moreover,inevery case the molar mass increased continuously throughout the reaction. In particular,t he polymerization of DL/CHO/PA results in formation of ab lock polyester containing an aromatic backbone group.T hus,t he aliquots were analyzed by GPC with both RI and UV detectors:t here was ac lear evolution in molar mass using both detection methods ( Figure S33). This finding confirms covalent bonding between the alternating semi-aromatic polyester and the PDL blocks. In contrast, the monomer mixture comprising DL/CHO/ camphoric anhydride (CA) showed similar rates of ROCOP and ROP and, in line with the switch hypothesis,s tatistical copolyesters formed ( Figure S66-75).
To highlight the potential for the new materials,p ostfunctionalization reactions,using the thiol-ene reaction, were conducted upon PDL-b-PCHNBE-b-PDL. Thea lkene functional groups were successfully substituted with either hydrophilic or hydrophobic side chains,a so bserved by 1 HNMR spectroscopy (Figures S76,S78,and S79). In both cases,t he functionalization reactions occurred without disrupting the block polymer structure as indicated by similar molar mass distributions before and after the reaction ( Figure S77). Although thiol-ene reactions are well established in polymer post-functionalization, this proof-of-concept highlights the future potential for these materials in coatings applications, where low molar masses would benefit processing and where multifunctional thiols are popular cross-linking agents. [15,16] All new block polymers also have hydroxyl-telechelic structures,that is,they are new polyester polyols.Polyester polyols are used in polyurethane production and as proof of chain extension, PDL-b-PCHNBE-b-PDL was reacted with 4,4'methylene diphenyl diisocyanate (MDI). Thep recursor material showed one glass-transition temperature at 26 8 8C and contained 39 wt %PDL. After chain extension, the molar mass increased from % 4t o% 70 kg mol À1 and DSC analysis indicated phase separation, as two glass-transition temperatures at 46 and 88 8 8Cw ere observed ( Figures S80 and S81). Such multiblock polymers,w ith controllable compositions, may be interesting as rigid plastics or thermoplastic elastomers. [17] In conclusion, monomer mixtures can be selectively reacted with ac ommercially available chromium salen catalyst, in one pot, to form well-defined ABAt riblock polyesters.T he scope of the catalysis is demonstrated using ar ange of different mixture compositions and monomers to deliver new polyester polyols.T hese polyesters can undergo post-functionalization reactions to modify the side-chain substituents or chain-extension reactions to produce multiblock polyesters.S witch catalysis,u sing Cr-salen catalysts,i s expected to be applicable to the preparation of other block and multiblock polymers.T he method should be applied using mixtures of carbon dioxide,l actones,a nhydrides,a nd epoxides to produce new block polycarbonates,e sters,a nd ethers.  (Tables S5-S9).