Introduction

A central goal in polymer synthesis is to directly convert simple chemical building blocks into useful materials. While a wide variety of interesting and potentially important structurally complex polymers have been discovered through recent research efforts (for example, biopolymers, advanced polymer networks, responsive materials and so on), their synthesis via traditional methods can be sufficiently involved to limit their accessibility, especially with the efficiency often demanded in polymer synthesis. One area where structure complexity has proven particularly powerful is in the field of π-conjugated polymers. The development of poly(heterocycles) (polypyrroles1, polythiophenes2,3 and others4,5,6) and their copolymers has sparked a renaissance in how scientists consider constructing a host of organic electronics, such as semiconductors, photovoltaic devices, or sensors7,8,9,10. A useful feature of conjugated polymers is their tunability. The modulation of substituents, conjugated heteroatoms or alternating backbone units can allow the construction of conjugated polymers with tailored electronic and other physical features. A number of powerful approaches have been developed to access conjugated polymers, including the now commonplace use of cross-coupling methodologies11,12. While very effective, these often achieve complexity from the monomers themselves, which can in some instances require a multistep synthesis, followed by halogenation and metallation, and can make accessing varied polymer structures an iterative process. The latter have made the development of alternative methods to construct conjugated polymers an area of growing relevance13,14,15.

In principle, an attractive synthesis of complex conjugated polymers would be to consider their structure as arising directly from available monomers. A challenge is in how to accomplish this in an efficient fashion. One possibility is offered by multicomponent coupling reactions. Multicomponent reactions have been heavily exploited in organic synthesis to increase molecular complexity without the need for multistep synthetic sequences, with high efficiency and minimal waste16,17. When coupled with transition metal catalysis, these can provide methods to both activate and selectively couple several simple substrates directly into complex products18,19,20. Although metal catalysis is an established tool for activating typically unreactive components towards efficient polymerization (for example, polyolefin synthesis, ring-opening polymerization and so on), the use of metal catalysis to control the coupling of multiple different monomers into new, well-defined and more complex polymer structures is much less explored. Block terpolymer synthesis is well established (I; Fig. 1a), and a number of intriguing examples have recently emerged in the metal-catalysed assembly of alternating multicomponent polymers II (refs 21, 22, 23, 24, 25, 26, 27), including the synthesis of conjugated22,23 and high-molecular-weight materials27. However, a method to directly convert multiple simple monomers into an entirely new polymer structure such as III has to our knowledge not been reported. Considering the variety of monomers relevant for polymerization (for example, diamines, diacids, alkenes, alkynes and so on), this method could provide an efficient route to access a diverse variety of structurally complex polymers, yet do so through combinations of available substrates, with high efficiency, and with facile access to structural diversity.

Figure 1: Multicomponent approaches to complex polymer synthesis.
figure 1

(a) Metal-catalysed multicomponent coupling approaches to polymers. (b) Palladium-catalysed Münchnone formation. (c) Previous work involving di-Münchnones in oligomer synthesis.

Full size image

A polymerization as in Fig. 1a requires a multicomponent reaction that is selective, high yielding, yet relies upon available monomers. A potential transformation that fulfils these features is the palladium-catalysed synthesis of heterocycles shown in Fig. 1b. We have recently reported a palladium-catalysed route to generate 1,3-dipoles 4 (Münchnones) from imines, acid chlorides and carbon monoxide (CO)28,29. This reaction proceeds in high efficiency and is equally important with substrates that are all easily accessible for a polymerization: imines, acid chlorides and CO. 1,3-Dipoles, including di-Münchnones, have been used in condensation chemistry to synthesize cross-conjugated materials in low molecular weights (Fig. 1c; refs 30, 31). A more attractive approach suggested by this palladium-catalysed synthesis would be to couple simple diimines, diacid chlorides and CO into a novel type of mesoionic, 1,3-dipole-containing conjugated polymer 4. Münchnones are known to undergo a variety of 1,3-dipolar cycloaddition reactions with unsaturated substrates (A≡B) to generate nitrogen-containing heterocycles32. As such, this platform would allow the assembly of various structurally distinct polymers from multiple combinations of available substrates.

We describe here our efforts towards the development of such a metal-catalysed multicomponent polymerization reaction. This has allowed the assembly π-conjugated poly(heterocycles) from multiple combinations of simple monomers: diimines, diacid chlorides, CO, alkenes and/or alkynes. In addition to demonstrating the feasibility of this strategy, this reaction allows access to a new class of conjugated polymer in the form of a mesoionic poly(1,3-dipole). The polymers can be easily modified by cycloaddition, providing access to families of conjugated materials in one-pot, metal-catalysed reactions.

Results

Multicomponent synthesis of conjugated polypyrroles

As model monomers for the polymerization, we examined the palladium-catalysed coupling of terephthaloyl chloride 1a, diimine 2a based on dialkylfluorene and CO (Table 1). The efficiency of this polymerization was determined by reacting 4a with the commercial alkyne dimethylacetylene dicarboxylate to form the polypyrrole 5a. Using simple Pd(II) sources, with or without the addition of phosphines ligands, leads to no polymerization (entries 1–4), presumably owing to their slow reduction under the mild reaction conditions. Conversely, the common Pd(0) catalyst Pd2dba3·CHCl3 (dba=dibenzylidene acetone) results in the rapid conversion of starting materials to form a poorly soluble, amorphous product (entry 7). Gel permeation chromotography (GPC) analysis shows that the tetrahydrofuran (THF) soluble fraction contains a polymer (Mn=7.8 kDa) but with a broad polydispersity (PDI=3.4). The addition of PPh3 or PCy3 in an attempt to attenuate the reactivity of the palladium catalyst leads instead to complete inhibition of the reaction (entries 5, 8 and 9) while P(o-tol)3 restores activity but yields a similarly poorly soluble product with a broad PDI (entry 10).

Table 1 A multicomponent synthesis of conjugated polypyrroles*.
Full size table

The low solubility and broad PDI in this product may be the alkene-based dba on the palladium precatalyst, which could react via cycloaddition with 4a and lead to crosslinking. To circumvent this reaction, Pd[P(o-tol)3]2 was used as a commercially available source of Pd(0) with only weakly associated, and unreactive, ligands. This catalyst leads to the near-complete consumption of the reagents under mild conditions (50 °C, 4 atm CO), and the generation of a polymer (5a) that is soluble in common solvents (entry 11). 1H and 13C nuclear magnetic resonance (NMR) and infrared (IR) analysis show the conversion of the monomers into the polypyrrole 5a, which has identical spectral features to independently prepared model compounds (Supplementary Figs 5 and 16). GPC analysis shows 5a to have a well-defined monomodal molecular weight distribution (PDI=1.6). The catalytic activity of Pd[P(o-tol)3]2 can be enhanced by the addition of CuPF6, which presumably acts as a phosphine scavenger to generate a mono-ligated palladium catalyst (entry 12). Alternatively, simply using high CO pressure leads to a rapid and near-quantitative polymerization, forming polypyrrole 5a as the only observable product (entry 13). Molecular weights as high as 22.7 kDa can be obtained by increasing the concentration and reaction time (entry 14).

Synthesis of poly(1,3-dipoles)

The reaction in Table 1 provides a new approach to synthesize pyrrole-based conjugated polymers from combinations of substrates that are either available or monomers themselves in other polymerizations (terephthaloyl chloride, CO, dialdehydes and alkynes). This platform can also be used to access new classes of conjugated materials. For example, the first step in the transformation in Table 1 also generates the 1,3-dipole-containing polymer 4a. 1,3-Dipoles such as Münchnones are typically considered reactive intermediates and used in situ in synthesis. However, while performing the palladium-catalysed coupling in the absence of alkyne, we noted the precipitation of a dark solid 4a that can be easily isolated by washing with acetonitrile. This polymer is surprisingly stable (<5% mass loss at up to 180 °C under nitrogen), and can be stored at low temperature in the absence of air and moisture, although it does hydrolyse in the presence of water. In order to fully characterize the moderately soluble 4a by NMR analysis, it was prepared in moderate molecular weight with an imine end-capping agent (4a′, Mn=6.7 kDa; Supplementary Methods). Spectral analysis show all the signals for a Münchnone, including characteristic carbonyl resonances in the infrared (1,710 cm−1) spectra, in the 13C NMR (δ 160.6) and others, all of which correlate with the model di-Münchnone 4a′′ prepared from diimine 2a, toluoyl chloride and CO. Therefore, 4a represents an unusual new class of mesoionic polymer, poly(1,3-dipoles).

Donor/acceptor-conjugated polymers have become an important thrust in the recent design of polymer-based photovoltaic materials7,8,33, although these do not typically incorporate formal charges into the backbone. Likely as a result of this charge separation, poly-Münchnone 4a is a dark purple solid, characteristic of low-bandgap materials. Ultraviolet/visible analysis shows intense (molar absorptivity up to 5.0 × 104 l mol−1 cm−1) and broad absorptions in most of the visible region (Fig. 2a). These are significantly red-shifted relative to the model bis-Münchnone, and indicate that 4a is highly conjugated. While the precise nature of this extended conjugation is still under investigation, the persistent charge-separated character of the mesoionic moieties creates a donor/acceptor motif on a single heterocycle, and may also provide a novel route to planarization as a mechanism to partially eliminate charge (for example, Fig. 2b). Estimation of the optical bandgap by absorbance onset gives values of 1.74 eV for 4a′ (1.59 eV for the high-molecular-weight 4a), and cyclic voltammetry shows a reversible reduction and an electrochemical bandgap of 1.84 eV. As such, these are a new class of low-bandgap-conjugated materials.

Figure 2: Ultraviolet/visible spectra and conjugation in poly-Münchnones.
figure 2

(a) Ultraviolet/visible spectra of 4a (22.7 kDa), 4a′ (6.7 kDa) and model dimer 4a′′ (n=1). (b) Potential resonance structures of 4a leading to planarization.

Full size image

In addition to their unusual electronic properties, the synthesis of 4 from diimines and diacid chlorides makes it straightforward to attenuate their structure and form a range of mesoionic polymers. Examples of the structural diversity available are shown in Fig. 3. Notably, each of the bis(acid chloride) monomers are either commercially available or easily prepared from the diacid compounds. For example, 2,5-thiophene dicarbonyl dichloride (precursor to 4c) can be synthesized in one step from adipic acid, a commodity chemical used in the production of nylon and other polyamides, whereas 2,5-furan dicarboxylic acid is derived from carbohydrates and identified by the US Department of Energy as one of the top 10 bio-based renewable chemicals34. The diimines used can be similarly altered to incorporate carbazole (4e), a common unit in conjugated polymer production8. In all cases, the catalytic coupling is clean and molecular weights are limited only by the solubility of the poly(1,3-dipole).

Figure 3: Diversity of Münchnone-containing polymers.
figure 3

Owing to the moisture sensitivity of 4, molecular weight determined by conversion to polypyrroles 5, as in Table 1.

Full size image

The structural manipulations translate into the properties of these polymers (Fig. 4a,b). For example, the disubstituted carbazole-containing polymer 4e displays visible absorbances that are blue-shifted (λmax of 539 nm) relative to 4a (λmax of 570 nm; Fig. 4a). Alternatively, the thiophene-containing material 4b is significantly red-shifted relative to these other polymers (absorbance onset ∼768 nm), corresponding to an optical bandgap of 1.58 eV. The latter is comparable to materials currently of interest as light harvesting materials in bulk heterojunction solar cells7,8. As such, this provides a method to both construct low-bandgap-conjugated polymers and manipulate or tune their electronic properties by choice of the constituent components.

Figure 4: Properties of polymers.
figure 4

Polymers were prepared with imine end-capping at 6–8 kDa to ensure full solubility for analyses (Supplementary Methods). (a) Ultraviolet/visible absorbance spectra of poly-Münchnones 4a–f in THF. (b) Polymers 5a, 4f′, 4e′, 4a′ and 4c′ in THF (from left to right). (c) Fluorescence spectra of select polymers 5–6 in THF.

Full size image

Multicomponent synthesis of poly(heterocycles)

These poly-Münchnones 4 also offer access to another deeper level of molecular complexity via their backbone reactivity. This can provide, to our knowledge, a unique route to convert one conjugated organic polymer into other backbone-conjugated polymers35,36,37. For example, the addition of phenyl methylpropiolate to 4a results in the transformation of the purple, low-bandgap poly-Münchnone into a moderate bandgap, blue-emitting polypyrrole 5b (Fig. 5). 1H NMR and infrared analysis suggest the complete disappearance of 1,3-dipole unit in this reaction, with no observable byproducts. In order to quantify the efficiency of cycloaddition, 13C-labelled polymer 4a was generated from isotopically enriched terephthaloyl chloride 4-C6H4(13COCl)2. NMR (13C) analysis shows the quantitative reaction with alkyne to form polypyrrole (>95%; see Supplementary Methods for details), and is consistent with the high cycloaddition reactivity of the 1,3-dipoles. A range of pyrrole-based polymers can be formed by this reaction. For example, changing the alkyne used can provide access to a wide variety of pyrrole-based polymers. Representative examples of these include the diester-containing polymer 5a or the diketone-substituted 5c. In addition to alkynes, electron-deficient alkenes are suitable dipolarophiles to form pyrroles. This can allow the facile formation of fused ring pyrrole-based polymer 6a (from cyclic alkene cycloaddition), the mono-substituted 6b (from chlorocyanoethylene) or the unsubstituted polypyrrole 6c with high yield. By using different diimines and diacid chlorides, various other polypyrroles can be obtained (for example, 5d–5 h). As such, this provides a platform to readily incorporate desired functionalities onto conjugated polymers, all of which emanate from a single polymer 4. The reactivity of Münchnones is also not limited to pyrroles. The addition of N-tosyl imine forms imidazole-based-conjugated polymer 7, and backbone conjugation in these polymers can be easily quenched by the addition of alcohols, leading to the generation of polyamides 8. Poly-Münchnones 4 can therefore be considered highly reactive and versatile conjugated materials, where a single polymer 4 can be transformed into entire families of new conjugated materials. In the case of phenyl methylpropiolate, this polymerization can be performed in a single step (Fig. 6), thereby allowing the orthogonal, four-component synthesis of a conjugated polymer.

Figure 5: Transformation of poly-Münchnones into families of conjugated polymers.
figure 5

Derivatization performed on poly-Münchnone samples of 8–12 kDa to ensure full solubility of 4a–e, and allow the quantification of the cycloaddition.

Full size image
Figure 6: One-pot, four-component polymerization.
figure 6

See Supplementary Methods for experimental details.

Full size image

This structural diversity can provide a further platform to modulate properties. For example, between polymers 4 and 8, optical absorbance and polymer bandgap can be tuned across the visible spectrum, and materials can be formed that fluoresce anywhere from blue to green to not at all (Fig. 4c and Table 2; Supplementary Tables 1 and 2 for the full list of properties). Similarly, electrochemical studies show that this cycloaddition (or lack thereof) can be used to tune HOMO energies by over 1 eV. While the primary thrust of these studies was not in product design, several of the polymers reported here display notable properties. Polypyrroles and their copolymers have attracted significant interest as electronic materials1,38. Three of the pyrrole-based polymers (5a, 6a and 6b) have good photoluminescent quantum efficiency (>35%) with emission maxima in the range of blue light (413–459 nm). Alternatively, the pyrrole-based imide-functionalized unit has been identified as a promising electron acceptor unit in conjugated polymers, and can be readily generated by this approach (6a; ref. 39), while alternating pyrrole-thiophene materials such as 6d represent new variants of materials found to be of use in field-effect transistors40. The range of properties observed is a direct result of their structural diversity, which varies from the backbone-conjugated heterocycles, the spacer units, to the substituents. This level of structural attenuation would require an individual synthesis for each new monomer via typical methods. In this case, each polymer is generated in one pot, from a small pool of monomers, and with minimal waste (often only HCl and CO2).

Table 2 Properties of polymers 5–8.
Full size table

Synthesis of polymers from vanillin

Finally, we have examined the potential of using other renewable materials as precursors to conjugated polymers. Lignin is a major component of lignocellulosic biomass and the world’s largest renewable source of aromatic compounds, thereby making it a potentially attractive, bio-based feedstock for π-conjugated polymers41. Lignin depolymerization yields a variety of aromatic building blocks, including the dialdehyde 9 (a dimer of vanillin). As this multicomponent polymerization uses aldehydes and carboxylic acids as monomer feedstocks, 9 can be incorporated in this palladium-catalysed polymerization to generate the conjugated poly-Münchnone 4f (Fig. 7). Ultraviolet/visible and electrochemical studies show that 4f is a moderately low-bandgap polymer (1.9 eV). As above, the dipole in 4f can undergo cycloaddition reactions to generate the polypyrroles 5i and 5j, each of which are blue-emitting materials. These polymers are hybrid materials derived from the following four simple substrates: vanillin, terephthaloyl chloride, a primary amine and CO, and represent, as far as we are aware, the first use of lignin in cross-conjugated polymer formation. Considering the versatility of this reaction, it should prove relevant for the controlled assembly of a range of renewable-based conjugated polymers.

Figure 7: Conjugated polymers from vanillin.
figure 7

Multicomponent synthesis of polymer 4f and its derivatization to 5i and 5j.

Full size image

Discussion

In summary, we have described a new type of metal-catalysed multicomponent polymerization, which provides a method to convert combinations of monomers, such as diimines, diacid chlorides, CO, alkynes, alkenes and alcohols, into structurally well-defined conjugated polymers. In addition, new backbone-conjugated mesoionic polymers (4) can be prepared via this coupling, which can undergo efficient post-polymerization cycloaddition to generate families of conjugated materials. In light of the variety of dipolarophiles possible for cycloaddition, as well as diimines and bis(acid chloride)s available, this can be used to construct arrays of conjugated materials, yet without the typical need to preassemble each new conjugated unit, and with the efficiency often desired in polymer synthesis. This multicomponent catalytic polymerization approach could prove equally applicable for the controlled assembly of a range of new materials from established monomers and/or other inexpensive building blocks. Experiments directed towards the latter are currently underway.

Methods

General procedure for the synthesis of poly-Münchnones and polyheterocycles

To diacid chloride 1 (0.1 mmol) and diimine 2 (0.100 mmol) in THF (0.6 ml) in a 5-ml vial under N2 was added N,N-diisopropylethylamine (51.7 mg, 70.0 μl, 0.400 mmol), and Pd[P(o-tol)3]2 (7.2 mg, 0.010 mmol) in 1.3 ml THF (1.3 ml)/MeCN (0.6 ml). The vial was placed in a 40-ml Parr steel autoclave, charged with CO (20 bar) and heated to 45 °C. (For the imine endcapped poly-Münchnone, p-tolyl(H)C=NC16H37 was added; see Supplementary Methods for details.) The CO was evacuated, the vessel was brought into a glovebox and the appropriate dipolarophile in 1.5 ml THF was added. The reaction was stirred at room temperature or 50 °C for 16 h, 0.2 ml water was added and the polymer product was extracted with o-dichlorobenzene using a Soxhlet extractor. The solvent was removed in vacuo, and the residue was dissolved with a minimum amount of hot chloroform (∼1 ml) and added dropwise into methanol (∼20 ml) to precipitate the polymer. The suspension was centrifuged and the methanol layer was decanted. The polymer was washed with methanol (3 × 2 ml) before drying in vacuo at 50 °C. See Supplementary Methods for further information. For NMR analysis, ultraviolet/visible, fluorescence and cyclic voltammetry of the molecules in this article, see Supplementary Figs 1–7.

Additional information

How to cite this article: Leitch, D. C. et al. A palladium-catalysed multicomponent coupling approach to conjugated poly(1,3-dipoles) and polyheterocycles. Nat. Commun. 6:7411 doi: 10.1038/ncomms8411 (2015).