Chima
Anyaegbu
,
Gianluca
Vidali
,
Darsan
Haridas
and
Joel F.
Hooper
*
Department of Chemistry, Monash University, Clayton 3800, Victoria, Australia. E-mail: joel.hooper@monash.edu
First published on 13th May 2024
The synthesis of complex and highly functionalised polymer materials requires polymer chemists to search for versatile and efficient new methods for chemical synthesis. Recent years have seen the rise of radical decarboxylation as a novel approach to the initiation, functionalisation, crosslinking and degradation of polymers, allowing access to exciting new materials. This review describes the fundamentals of radical decarboxylation and its use in organic synthesis, along with the increasing application of this approach in polymer synthesis.
More recently, polymer chemists have exploited the versatility of radical decarboxylation for the initiation, functionalisation and degradation, cross-linking and depolymerisation of polymers. This review discusses the history and recent developments in each of these topics.
The choice of carboxylic acids as reactive functional groups in polymer chemistry is attractive due to both their ubiquity and their unique reactivity. Carboxylic acids can be found in many synthetic and natural polymers, especially polyacrylates and carbohydrate-based polymers. In addition, a variety of synthetic strategies exist to generate carbon centred radicals via radical decarboxylation. The versatility and functional-group compatibility of this chemistry allows for decarboxylation to be applied in a broad variety of contexts.
Radical decarboxylation reactions can be broadly categorised into three groups: oxidative reactions from free carboxylic acids, reductive decarboxylations from activated esters and redox neutral decarboxylation via homolysis (Fig. 1). Each of these approaches have a long history in small molecule synthesis, and have been applied more recently in polymer and materials science.
Fig. 2 Selected examples of radical decarboxylative couplings via oxidative single-electron transfer (SET). |
Oxidative decarboxylation using silver salts also originates in the 19th century, with Borodin reporting the decarboxylation of silver carboxylate salts in 1861.4 This idea was further developed by Minisci in 1971, who demonstrated the silver-catalysed decarboxylative coupling of carboxylic acids with nitrogen-containing heterocycles.5 This reaction has been well developed, and Minisci-type coupling reactions of carboxylates and a variety of coupling partners have been reported.6
A major resurgence in decarboxylative cross coupling began with the advent of photoredox catalysis. In 2014 Doyle and MacMillan reported the decarboxylative coupling of amino acids with aryl halides, using an iridium based photocatalyst in conjunction with nickel catalysis.7 In the subsequent decade, a huge number of decarboxylative transformations have been reported, including arylations, alkylations and heteroatom additions, using both metal-based and organic photocatalysts.8,9
In 1983, Barton reported the conversion of carboxylic acids to thiohydroxamate esters (known as Barton esters), which can undergo reductive decarboxylation to the alkane with a radical initiator and tin hydride.10
Okada and Oda demonstrated in 1988 that N-acyloxyphthalimides could undergo reductive decarboxylation under photochemical conditions in the presence of a thiol H-atom donor.11
The reactions of Barton esters and a variety of other activated esters was reinvigorated in 2016 by Baran and co-workers, who showed that activated esters could undergo single-electron transfer from a nickel catalyst. This results in decarboxylation of the activated ester and formation of an alkyl–nickel species, which can undergo cross-coupling with an aryl halide or organozinc reagent.12,13
Subsequent work has demonstrated the decarboxylation of redox-active esters catalysed by Fe, Co, Ru and organic catalysts, and has incorporated a wide variety of coupling partners, including sp2 and sp hydridised carbon-based nucleophiles, anilines and phenols (Fig. 3).13,14
Fig. 3 Selected examples of radical decarboxylative couplings via reductive single-electron transfer (SET) of activates esters. |
Other activated carboxylates such as hypervalent iodine species and N-hydroxyphthalimide esters have been utilised in decarboxylations, such as in the borylation reaction reported by Aggarwal.16
In 2001, Kamenska and coworkers showed that thermal and sonochemical activation of (diacetoxyiodo)benzene could also be used to initiate the polymerisation of methyl acrylate (MA) (Fig. 5A).18
In 2012, Han and Tsarevsky further exploited the decarboxylation of acyl iodoniums to synthesise branched and crosslinked polymers.19 This team used (diacetoxyiodo)benzene to initiate the polymerisation of methyl methacrylate (MMA) and methacrylic acid (MAA). In the presence of carboxylate groups from the methacrylic acid monomers, the hypervalent iodine reagent could undergo exchange reactions to produce activated methacrylate monomers. When incorporated into the backbone of the polymer, these activated acids can undergo thermal or photochemical decarboxylation to generate an sp3 hybridised radical, which serves as a branching or crosslinking site on the polymer (Fig. 5B).
In 2003, Yagci and co-workers20 built on their previous demonstration of thioxanthones as photoinitiators for (non-decarboxylative) free radical polymerisation, using thioacetic acid thioxanthone (Fig. 6) to initiate polymerisation. Upon excitation to the triplet state, this compound underwent decarboxylation to give the initiating radical, leading to the polymerisation of MMA.
Later studies by Arsu, Jockusch and coworkers21 on related photoinitiator systems provided a mechanistic framework for this reaction, where excitation of the thioxanthone to the π–π* triplet state is followed by intramolecular electron transfer from the carboxylate and subsequent decarboxylation.
In 2008, Ni and co-workers used TiO2 nanoparticles as photocatalysts in the polymerisation of vinyl acetate via decarboxylation of acetic acid under UV irradiation (Fig. 7).22 Monomer conversion in these reactions were shown to be proportional to acid concentration, and incorporation of the methyl radical to the polymer was confirmed by 13C NMR. The use of aqueous conditions was essential to this process, which the authors attribute to the requirement for ionisation of the acid to promote interaction with the TiO2 nanoparticles.
Building upon this research, Ni and co-workers extended the scope of this reaction to use longer chain carboxylic acids.23 In this work they showed that n-butyric acid's decarboxylation is more efficient than that of acetic acid, consistent with improved stability of the resulting radical. In addition, using FTIR they were able to correlate the decarboxylation efficiency with the carboxylate adsorption on the TiO2 surfaces.
In a subsequent study (Fig. 6),24 the group demonstrated the successful use of dicarboxylic acids in this polymerisation in both aqueous and bulk conditions. They suggest that the diacid aids in coordination to the nanoparticle photocatalysts, making the decarboxylation more favourable.
In 2017, Sugihara, Yoshimi and co-workers reported the first decarboxylative polymerisation that was applicable to a diverse range of carboxylic acids (Fig. 8).25 Their method used a dual organic photosensitiser approach, with 1,10-phenanthroline (Phen) and 1,4-dicyanobenzene (DCB). The authors have shown in their previous small-molecule studies that photoexcitation of the Phen catalyst to the excited state (Phen*) is followed by single electron transfer to DCB, giving the Phen radical cation.26 This species can accept an electron from the carboxylic acid, resulting in radical decarboxylation and subsequent polymerisation.
When 1 equivalent of the photosensitisers were used (with respect to carboxylic acid) the observed molecular weight was approximately 3 times higher than the theoretical value, suggesting a low level of initiator efficiency. Increasing this to a ten-fold excess of the photosensitisers gave improved initiator efficiency, albeit at the expense of reduced monomer conversion.
Analysis of the resulting polymers by 1H NMR and matrix-assisted laser desorption/ionization time of flight (MALDI-TOF) mass spectrometry showed clear incorporation of the carboxylate-derived end group, indicating that no competing initiation methods are occurring under the reaction conditions. The method was applied to primary, secondary and tertiary carboxylic acids, along with several amino acid and peptide substrates. The most efficient polymerisations were seen with substrates that generated more stable secondary, tertiary or α-aminyl radicals, while the primary acid gave the highest molecular weight polymer, indicating reduced initiator efficiency.
In 2022, Simon, Hooper, Garnier and co-workers reported a silver-mediated radical decarboxylation for the grafting of poly(N-isopropylacrylamide) (PNIPAM) to TEMPO oxidised cellulose nanofibers (Fig. 9).27 Utilising conditions developed by Minisci in the 1970s,5 a persulfate oxidant was used to oxidise Ag(I)NO3 to Ag(II), which can then promote oxidation of the carboxylic acid and allow for decarboxylation. The resulting radical on the surface of the cellulose fibre could then initiate polymerisation, leading to formation of a thermoresponsive cellulose/polymer conjugate in a water/acetonitrile mixture.
Fig. 9 Silver-mediated radical decarboxylation for polymer grafting from TEMPO oxidised cellulose nanofibers. |
While the persulfate oxidant was also able to initiate polymerisation in solution, a grafting efficiency of 55% was observed, which is a significant improvement over other direct grafting methods to cellulose nanofibers.
The same group also demonstrated in 2023 the application of this silver-catalysed method to the synthesis of hyperbranched polyacrylic acid conjugates of cellulose nanofibers.28,29 When acrylic acid was used as the monomer, the growing polyacrylic acid chains provided additional carboxylic acids for decarboxylation, leading to branching and cross-linking between the polymer chains. The monomer itself did not undergo decarboxylation under these conditions, due to the instability of the sp2 radical that would be formed.
When the optimal silver catalyst loading of 0.25 equivalents (with respect to the cellulose carboxylate groups) was used, the material achieved a free swell capacity of 240 times.
The first example of radical decarboxylation to initiate a controlled polymerisation was reported by our group in 2022 (Fig. 10).30 In this work, the decarboxylation of redox active esters was used to initiate cobalt mediated radical polymerisation (CMRP). Based on a previous report by Wang on the cobalt-catalysed decarboxylative Negishi coupling,31 it was observed that the in situ reduction of Co(II) to a Co(I) species could result in single electron transfer to a redox-active ester and subsequent radical decarboxylation. This gives a carbon-centered radical and a Co(II) complex, which is in equilibrium with the alkyl–Co(III) species.
This method allowed for the initiation of CMRP at low temperatures (0 to −40 °C), leading to low dispersities and allowing the polymerisation of dimethylacrylamide (DMA) using Co(acac)2 as mediator for the first time.
It was also shown that the reactive polymer–Co(III) complex could undergo Negishi-like coupling with an arylzinc reagent, leading to polymer functionalisation at both the α- and ω-ends.
Following this work on decarboxylative CMRP, our group also demonstrated the first RAFT polymerisation via decarboxylation of an unactivated carboxylic acid (Fig. 11).32 In this method, an acridinium photocatalyst33 promoted decarboxylation under green light irradiation, producing the initiating radical. The RAFT disulfide reagent intercepted this radical to deactivate the polymerisation, while also producing a sulfur-centered radical which could undergo reduction to turn over the catalytic cycle. Control experiments showed that the photocatalyst had a minimal effect on the rate of propagation, and that once the RAFT macroinitiator is formed, the polymerisation proceeds by a uncatalyzed photoRAFT mechanism involving direct C–S bond photolysis, as previously described by Matyjaszewski.34
Along with a number of small-molecule carboxylic acid initiators, this method was also applied to the synthesis of brush polymers, using an amino-acid functionalised backbone polymer to generate branching sites via decarboxylation.
The first polymer crosslinking via radical decarboxylation was achieved in 2008 by Kratochvil and Koros, who demonstrated the thermal crosslinking of polyimides (Fig. 12).36 They utilised a polyimide material with aryl carboxylate functional groups, which can undergo radical decarboxylation at elevated temperatures. When heated to 220 °C, the loss of CO2 from the polymer was observed, and a highly insoluble material was formed which showed resistance to plasticization under high pressures of CO2. The authors suggest several possible types of C–C linkages that may be formed under these conditions, but demonstrate by IR and 13C NMR analysis that formation of ester or anhydride links are not responsible for the cross-linking.
The exact mechanism of decarboxylation under these conditions is unclear, but oxidative processes seem unlikely due to the inert atmosphere used. Buchanan and co-workers have shown that thermal decarboxylation of aryl carboxylic acids can occur via formation of an anhydride intermediate, followed by C–O bond homolysis to generate acyl and carboxyl radicals which can undergo decarbonylation and decarboxylation respectively to generate aryl radicals.37
Cao, Li and co-workers synthesized a polyimide by combining a carboxylic acid containing diamine and dianhydride.38 The incorporation of a specific diamine monomer noticeably increased the chain distance of the crosslinked polyimides. The polyimide crosslinked at 425 °C, exhibited comparable CO2/CH4 selectivities but demonstrated superior resistance to plasticization under CO2 pressures up to 30 atm compared to the non-crosslinked counterpart. In the context of CO2/CH4 gas separation, the resulting material exceeded the performance specified by the “2008 Robeson Upper Bound”.39
The decarboxylative cross-linking of related polyimide materials has been well studied, and numerous applications have been reported in gas separation.40
In 2012, Guiver and co-workers applied thermal decarboxylation for the crosslinking of polymers of intrinsic microporosity (PIMS) based on a polyindane backbone (Fig. 13).41 This material contained nitrile groups that could be hydrolysed to the carboxylic acid in a controlled manner, and crosslinked by heating to 375 °C. Like in the work by Kratochvil and Koros, several different modes of chemical cross-linking were proposed in this process. The authors were able to tune the gas permeability of the cross-linked material by controlling the carboxylic acid content during the hydrolysis step, giving rise to higher selectivities in the separation of O2/N2, CO2/N2, and CO2/CH4 gas mixtures.
A similar strategy of cross-linking via thermal decarboxylation of aryl carboxylic acids was later used by Li and co-workers, who applied this method to the cross-linking of triptycene based polymers (Fig. 14).42
The resulting cross-linked materials showed increased gas permeability, and polymer membranes showed high levels of selectivity in the separation of gas mixtures including H2/CH4 and CO2/N2.
In 2019, Wang and coworkers applied the decarboxylation approach to the crosslinking of phenolphthalein-based poly(arylene ether ketone) (Fig. 15).43 After crosslinking, the polymer showed a 110 fold increase in CO2 permeability and an increase in the CO2 pressure required for plasticization from 2 atm to >30.
In 2017, He and co-workers demonstrated a practical solution for the low temperature cross-linking of carboxylate containing polymers via a Ag(I) catalysed radical decarboxylation (Fig. 16).44 Using polyacrylic acid (PAA) and other hydrophilic polymers, they observed gelation of aqueous solutions of the polymer in 20–30 minutes when treated with ammonium persulfate and catalytic AgNO3.
The resulting cross-linked hydrogel materials were highly stretchable, showing a maximum elongation up to 25 times the original length. The mechanical properties could be tuned by varying the cross-linking density, which could be controlled by altering the concentration of the silver catalyst. Higher loadings of Ag led to materials with higher tensile strength and lower maximum elongation.
Residual silver ions present in the materials could be utilised in photopatterning to produce images on the films, where the Ag(I) can be photochemically reduced to form silver nanoparticles.
In addition, this work also included a dehydrodecarboxylation reaction, where the inclusion of a cobalt-based catalyst resulting in formation of a polymer containing pendent alkene functional groups. This reaction utilised acridine (PC2) as the catalyst instead of the bulkier PC1, following a report by Larionov that demonstrated this as the optimal catalyst for dehydrodecarboxylation.64
More recently, Seidel and Sumerlin used the reductive decarboxylation approach to synthesise polystyrene (PSt)/polyethylene and PSt/polypropylene copolymers.46 In this work they utilised saccharin (meth)acrylamide as a precursor to methacrylic acid, allowing for the synthesis of alternating copolymers. Once the saccharin (meth)acrylamide was hydrolysed to the carboxylic acid, photochemical decarboxylation with the acridine-based photocatalyst in the presence of thiophenol resulted in alternating PSt/polyethylene and PSt/polypropylene copolymers. The group examined the difference in glass transition temperatures (Tg) between these alternating copolymers and their statistical analogues, with the well-defined alternating polymers showing a small breadth in Tg values.
In 2023, Seidel and Sumerlin demonstrated the degradation of PMA in a one-pot process that utilised radical decarboxylation and ozonolysis (Fig. 18).47 The PMA first underwent partial acidic hydrolysis to the statistical PMA/PAA copolymer, followed by their previously reported dehydrodecarboxylation process. In this step, the radical generated by photochemical decarboxylation is intercepted by a Co(II) species, followed by beta-hydride elimination to form the backbone alkene. Finally, a solvent switch to methanol allows for ozonolysis, (due to the increased solubility of ozone in this solvent) giving rise to small oligomers.
This method could also be applied to AA-containing copolymers, including polyacrylamides and PSt.
In an alternative strategy to that used by Seidel and Sumerlin, Theato and coworkers have described the use of RAFT (acryloyloxy)phthalimide based polymers for post-polymerisation functionalisation by radical decarboxylation (Fig. 19).48 Using a ruthenium-based photocatalyst in combination with the Hantzsch ester as a reducing agent, this reaction generated radicals on the polymer backbone that could react with a methyl acrylate acceptor, or be intercepted by a stable nitroxyl radical to form the TEMPO adduct.
The resulting TEMPO-functionalised polymers could be further elaborated in a nitroxide-mediated radical polymerisation with styrene to give graft polymers.
The same group also applied the radical decarboxylation of (acryloyloxy)phthalimide containing polymers for the synthesis of polyethylene-containing block copolymers.49 This approach utilised a ruthenium photocatalyst in conjunction with a tin-hydride reagent as a hydrogen atom source. Attempts were made to substitute the tin-hydride reagent for a less toxic silane, however only partial decarboxylation (up to 59%) was observed under these conditions.
Theato and coworkers also reported the use of thermally induced decarboxylation using a nickel catalyst for the synthesis of polyethylene copolymers and block polymers, using zinc as the reducing agent.50
Organo-photocatalysts were also explored for the single electron transfer induced decarboxylative post polymerisation modification and degradation of polymethacrylates by Sumerlin and coworkers (Fig. 20).51,52 They used Eosin Y as a photocatalyst in presence of green light for the synthesis of polyethylene based copolymers from poly(acryloyloxy)phthalimide based copolymers. They also applied this strategy to generate radicals in the polymer backbone, resulting in polymer degradation.
Fig. 20 Organocatalytic post-polymerisation functionalisation of (acryloyloxy)phthalimide containing polymers. |
Recently, Sumerlin, Evans and coworkers reported the use of electrochemical decarboxylation for the synthesis of poly(ethylene-co-methyl acrylate) and poly(propylene-co-methyl acrylate) from N-(acryloxy)phthalimide based copolymers (Fig. 21).53 They also used this strategy in the absence of H-atom donors for the degradation of polymethacrylate based copolymers with more than 95% efficiency.
Radical decarboxylation has also been used by Henderson and Hooper in the grafting of PMA-co-PAA polymers to the surface of carbon fibre (Fig. 22).54 Using the silver-catalysed radical decarboxylation approach, addition of the polymer radicals to the graphitic domains of the carbon fibre was observed. A high degree of polymer functionalisation (10 wt%) on the fibre surface could be achieved, compared with existing methods.55 This led to improved mechanical properties of the fibre composite materials due to enhanced adhesion between the functionalised fibre surface and the matrix material.
When a radical is generated on the PMMA backbone, depolymerisation can be favoured under conditions of high dilution and/or high temperature. Several recent studies have used the reactive chain ends of RAFT or ATRP polymers to generate these radicals under thermal, transition metal- or photo-catalytic conditions.
In two recent publications, Sumerlin and coworkers have shown the thermal decarboxylation of N-hydroxyphthalimide esters as an alternative method to generate reactive radicals for depolymerisation under mild conditions.
In 2023, the group reported the synthesis of endgroup-defined PMMA, with an N-hydroxyphthalimide ester at the α-end and a RAFT group or halide at the ω-end (Fig. 23A).58 Using these difunctionalised polymers, they could achieve up to 92% depolymerisation at 210–220 °C under vacuum, compared with the 375 °C temperature required to depolymerise unfunctionalized PMMA.
Fig. 23 Depolymerisation of PMMA via thermal decarboxylation of N-hydroxyphthalimide functionalised polymers. |
In 2024, this same group demonstrated the efficient depolymerisation of PMMA without functionalised endgroups, using polymer containing 1–5% N-hydroxyphthalimide-functionalised monomer (Fig. 23B).59 These thermally labile materials showed 95% mass loss after heating to 185–234 °C (depending on the degree of functionalised monomer). The authors demonstrated the depolymerisation of high molecular weight polymers, and showed for the first time that PMMA synthesised by free-radical polymerisation can be depolymerised at modest temperatures.
Thermal decarboxylation has become a well-established approach for the production of cross-linked polyimides for gas separation. This method is appealing as it produces only CO2 as waste, and doesn't require any other catalysts or reagents which may contaminate the resulting material.
While there are a few examples of decarboxylation for the synthesis of brush-like polymers, this area is still underdeveloped and opportunities exist for the synthesis of more complex architectures via this approach.
The use of electrochemical decarboxylation is also an area that remains relatively unexplored in polymer science, particularly the oxidative decarboxylation of free carboxylic acids. Recent advances in electrochemical Kolbe-like decarboxylations offer powerful and efficient new methods for polymer synthesis and modification. Applications of radical decarboxylation in biopolymers also offers a number of opportunities, due to the prevalence of carboxylates in polypeptides and polysaccharides.
There also exists significant opportunity to further apply decarboxylation in the field of radical depolymerisation. While this has attracted much attention recently, the low-temperature depolymerisation of unfuctionalised polyacrylates still remains elusive. Backbone decarboxylation may offer a strategy to generate radicals and promote depolymerisation in the future.
As discussed in this review, radical decarboxylation has many benefits that may translate into efficient industrial applications in the future. One issue to be addressed is the generation of waste CO2 and its effects on climate change. Many of the reactions discussed here generate negligible amounts of CO2 compared with the overall mass of the material, while other processes are more CO2 intensive. There currently exist a number of technologies for CO2 capture in industrial processes, largely driven by applications in the steel and cement industries.60,61 Due to the localised nature of CO2 generation in polymer synthesis, it is likely that some of these methods will be applicable. There also exist opportunities to utilise captured CO2 in the production of new polymers such as polycarbonates and polyurethanes.62,63
This journal is © The Royal Society of Chemistry 2024 |