Julia
Pribyl
a,
Kenneth B.
Wagener
a and
Giovanni
Rojas
*ab
aThe George and Josephine Butler Polymer Research Laboratory, Department of Chemistry and Center for Macromolecular Science and Engineering, University of Florida, Gainesville, FL 32611-7200, USA
bUniversidad Icesi, Facultad de Ciencias Naturales, Departamento de Ciencias Químicas, Calle 18 No. 122-135, Cali, Colombia. E-mail: grojas@icesi.edu.co
First published on 7th September 2020
Acyclic Diene Metathesis polycondensation produces materials with well-defined primary structures, perfection that is the consequence of symmetry imparted by monomer design. Diverse functionalities can be incorporated either in the polymer backbone or as pendant groups; chemical compatibility of the functional group with the catalyst is a requirement. Structural perfection led to structure/property investigations that would be difficult in imprecise systems. This comprehensive review summarizes the synthetic strategies for production of ADMET polymers. Further presented are the effects on secondary structure of controlling branch identity, in-chain functional groups, and frequency along the polymer chain. Results of spectroscopic, thermal and imaging techniques have been key for understanding the relationships between structure and properties.
The first well-defined metathesis catalyst was created by Schrock in 1986 based on tungsten1 after which molybdenum-containing Schrock catalysts became prevalent.2 Both are most effective for producing high molecular weight polymers with degrees of polymerization higher than 150 for hydrocarbon polymers, while being less useful for polymers containing functional groups. Catalysts containing late transition metals, primarily ruthenium, developed by Grubbs and coworkers are best used when functional groups are present;3–6 their use has expanded metathesis chemistry considerably. The first-generation ruthenium catalyst, known as Grubbs’ first-generation, led to a variety of new ruthenium structures featuring diverse reactivity and functional group tolerance, while maintaining low rates of olefin isomerization and consequently, chain walking. Fig. 1 shows the structures of Schrock's molybdenum and Grubbs’ first-generation catalyst.
ADMET monomers normally are α,ω-dienes that produce unsaturated ADMET polymers. The major advantage of ADMET chemistry is the formation of repeat units with symmetrically disposed functionalities (both hydrocarbon and otherwise) along the polymer backbone, a key feature that is a direct consequence of the monomer symmetry. The unsaturated polymer can be further saturated by exhaustive hydrogenation of the polymer chain. Fig. 1 shows a symmetric α,ω-diene that yields an unsaturated ADMET polymer with a generic functionality. Functionality is precisely placed along the polymer backbone by controlling symmetrically the length between terminal olefins. Depending upon the nature of the introduced functionality and the design of the α,ω-diene monomer, a diverse library of saturated and unsaturated ADMET polymers can be obtained with the functionality either incorporated in the polymer chain or pendant along the chain.
The first significant ADMET polymerization work appeared in 19914 followed by the creation of diverse polymers containing precisely placed in-chain or pendant functional groups. For example, when the R group is an alkyl chain, the resultant polymer can be seen as a model polyolefin; many other polymers containing sulfur, boron, silicon, and amino acids have been synthesized. The following sections review the synthetic methods used for the preparation of those monomers and polymers, along with discussion of secondary structure elucidation and applications of ADMET polymers as functional materials.
ADMET avoids the random nature of branching in polyethylene by generating precision repeat units, again, a consequence of step rather than chain polymerization methodology. The most significant early ADMET reference was reported in 1991 creating perfectly linear, nonbranched ADMET polyethylene (PE) via the polymerization of 1,9-decadiene, followed by exhaustive saturation with hydrogen.4 Since then ADMET has been used to produce functionalized unsaturated polyolefins8 as well as saturated PE models with perfect primary structures.7,9–11 The first ADMET synthetic and thermal studies yielded precision PEs with methyl, ethyl and hexyl branches;10–12 these ADMET polymers modeled commercial versions made via Ziegler–Natta or metallocene chemistry. Specifically they were precision examples duly named ethylene/1-propene (EP), ethylene/1-butane (EB), and ethylene/1-octene (EO).
ADMET monomer synthesis initially comprised reaction schemes requiring several steps.13 Since then, monomer synthesis for such alkyl branched precision polymers has been reduced to just two steps. The following sections describe the history of monomer synthesis including the preferred two step approach. The result has been the preparation of a family of alkyl α,ω-diene monomers for short chain branching (methyl, ethyl, etc.) to long chain branching (up to 21 carbons).
The same alkylation/decyanation approach was employed for the synthesis of α,ω-dienes having a long chain branch (21 carbons, Fig. 6).17 which model the copolymerization of ethylene with long α-olefins. These compounds play a significant role in metallocene PE as they can help ease processability while maintaining the mechanical properties.11,18
Overcoming this RCM phenomenon was achieved as illustrated in Fig. 7.15,19,20 Monomers contained two precisely substituted carbon centers with different length methylene run lengths present. The strategy creates an ADMET polymer that models a noncommercial LLDPE containing a methyl group on every 7th carbon (EP7, Fig. 7a), while Fig. 7b yields a methyl group on every 5th carbon (EP5). Synthesis of the requisite monomers started via monoalkenylation of diethylmalonate in the presence of sodium hydride, followed bis nucleophilic attack on 1,6-dibromohexane (Fig. 7a) or 1,4-dibromobutane (Fig. 7b) respectively. Hydrolysis using NaOH in water/ethanol yielded tetracarboxylic acids. Decarboxylation was achieved by heating the product in decalin, and further reduction with lithium aluminum hydride formed the respective diol. Reduction to the respective dialkyl-α,ω-diene was carried out by mesylation of the hydroxyl groups, followed by hydrogen incorporation with LiAlH4 in tetrahydrofuran.
![]() | ||
Fig. 7 Preparation of short methylene run length methyl-substituted α,ω-dienes for modeling EP7 (a) and EP5 (b). |
The second approach to synthesize long methylene run lengths was reported in 2014,21 presenting a methodology for extending the distance between the halogen and the terminal olefin alkenylating reagent. Extension of the chain was achieved for methylene units longer than 18 of them, in particular 27 and 36. Fig. 9 illustrates the synthetic approach for extending the length of the alkenyl moiety, beginning with an α-olefin-ω-ol containing p methylene units extending to p + q units using an ω-bromoalcohol containing q carbons. The appropriate α-olefin-ω-alcohol was oxidized with pyridinium chlorochromate in dichloromethane to yield the respective aldehyde. A previously prepared Grignard reagent (a protected alcohol containing q methylene units) was added to the aldehyde to produce the respective alcohol and consequently extend the main chain. Once the chain was extended the resulting product contained two alcohol functionalities, one internal and free and another terminal and protected. In general terms, the product required reduction of the free OH and final deprotection of the terminal alcohol.
Reduction of the internal secondary alcohol was done by forming the respective xanthate with carbon disulfide and methyl iodide, followed by a further radical termination with phosphonic acid initiated by AIBN. Finally, the resulting α-olefin-ω-protected alcohol was deprotected with copper(II) chloride in a mixture of acetone and water. The significant advantage of this methodology is that the procedure can be repeated as many times as necessary for extending the main chain as desired.
Commercial LLDPE possesses two major structural defects: incorporation of the comonomer in a random fashion and chain walking promoted by the catalyst yielding structures with branches of different lengths in the main PE chain. The ADMET approach for modeling LLDPE circumvents both problems by introducing only and exactly the branch identity selected, and by placing the desired alkyl branch exactly in a specific place along the PE chain. ADMET polyolefins have the most perfect primary structure possible.
That said, more realistic models can be prepared by systematically introducing randomness. ADMET copolymerization chemistry allows this to be so. For example, copolymerization of alkyl-α,ω-dienes with 1,9-decadiene yields a model of LLDPE that statistically distributes the alkyl branch along the PE chain. This method of randomizing ADMET model polyolefins was introduced in 2003.11Fig. 10 shows the ADMET copolymerization of 6-methylundeca-1,10-diene with 1,9-decadiene to produce unsaturated EP models, and further exhaustive hydrogenation yielded nine EP models. Different materials were obtained by simply controlling the comonomer content during polymerization: EP0 is the ADMET polymer with no alkyl branch, while homopolymerization of 1,9-decadiene, and EP-111 corresponds to the maximum possible of branches obtained by this polymerization, corresponding to homopolymerization of 6-methylundeca-1,10-diene. The suffix numbers correspond to the number of methyl branches randomly incorporated in the PE chain per thousand polymer carbons. A series of ADMET models presenting 1.5, 7.1, 13.6, 25, 43.3, 55.6, and 111 methyl branches per 1000 of carbons was reported.
Using the same ADMET copolymerization approach models of ethylene/1-hexene (EH) copolymers were synthesized, i.e., polyethylene bearing a butyl branch randomly placed along the PE backbone.13,15Fig. 11 shows the copolymerization of 9-butylheptadeca-1,16-diene with 1,9-decadiene for the production of eight ADMET random EH models. By controlling the comonomer content during polymerization, model polymers were produced containing 2.5, 6, 11.5, 21.3, 37, 43.5 and 66.7 butyl branches per thousand PE carbons.
While these previous systematic studies of random methyl and butyl branches expanded the scope of ADMET polyolefins, it was important to prove that ADMET polymers containing longer branches could be modeled. Fig. 12 shows the same synthetic approach, but now containing a randomly placed pendant alkyl branch of 21 carbons. Thus an ADMET model of ethylene/1-tricosene (ETC) copolymer was generated in this manner.17 Control over the comonomer content during polymerization yielded materials containing 2.4, 5.8, 13.4, 22.2, 34.5 and 48 long branches (21 carbon branches) per thousand PE carbons.
All monomer syntheses started from a common material, an α,ω-diene-alcohol: the chlorinated monomer by substitution of the respective tosylated monomer with lithium chloride, the brominated monomer by addition of carbon tetrabromide and triphenylphosphine in dichloromethane, and the fluorinated monomer by adding diethylaminosulfur trifluoride in pyridine and dichloromethane at −78 °C.
![]() | ||
Fig. 14 Preparation of EAA copolymers via ADMET polymerization and their respective sodium and zinc ionomers. |
As suggested above, deprotections typically are done homogeneously in organic solution (e.g., tetrahydrofuran, toluene, dichloromethane, etc.). However, sulfonic ester polymers deprotected in this way produced a partially deprotected mixture of sulfonic ester and sulfonate functionality along the polymer backbone, resulting in a loss of precision. Partial deprotection occurred because while the sulfonic ester exhibits good solubility in organic solvents, as hydrolysis to the sulfonate occurred solubility decreased and the polymer began to precipitate, rendering full deprotection infeasible.
Due to the difficulty of fully removing DMSO after deprotection, a more convenient procedure involving refluxing a heterogeneous polymer/ethanol solution in the presence of sodium methoxide was devised to achieve the same result (emphasized in Fig. 16). Infrared spectroscopy of the deprotected polymer displayed no residual sulfonic ester stretches, indicating quantitative deprotection.
![]() | ||
Fig. 17 Preparation of symmetric sulfone-α,ω-dienes and bulk high temperature ADMET polymerization of high Tm precision polysulfones. |
The first ADMET polymerizations of sulfone-α,ω-dienes, done in solution with Grubbs’ first-generation catalyst, yielded low molecular weight polymers due to limited solubility of the precision polysulfones in an organic solvent.35 Even so, the saturated precision polysulfones exhibited high melting points (up to 175 °C), inspiring further investigation.
![]() | ||
Fig. 18 Preparation of boronic ester-containing polymer (a), and phenyl boronic acid polymer (b) via ADMET polymerization. |
![]() | ||
Fig. 19 Preparation of telechelic polyoctenamer oligomers using ADMET chemistry (top) and copolymerization of the telechelic polymer with polydimethylsiloxane (bottom). |
ADMET polymers containing carbosilanes were obtained by addition of magnesium in ether to 5-bromopent-1-ene to prepare the respective Grignard reagent, which was added to silicon tetrachloride to form the disubstituted product, as shown in Fig. 20. Dichlorosilane-α,ω-diene monomer was polymerized using Schrock's molybdenum catalyst. After polymerization the material was transformed into the respective dimethyl, diphenyl, or dibutyl polycarbosilane by simply adding the corresponding alkyl lithium.44,45 Both Schrock's and Grubbs’ catalysts have been used for production of ADMET polycarbosiloxanes, yielding diverse materials and applications.46Fig. 21 shows an example of six carbosiloxane monomers, all synthesized using Grignard techniques and the appropriated chlorosilane, that were easily polymerized via ADMET.
Fig. 22 shows the incorporation of a generic O-protected amino acid to a carboxylic acid α,ω-diene, Fig. 22a, and the incorporation of a generic N-protected amino acid to an amino-α,ω-diene, Fig. 22b. In both cases, coupling the protected amino acid was carried out by mixing the α,ω-diene in a THF solution containing 1-hydroxybenzotriazole (HOBt) and diisopropylcarbodiimide (DIC). Both protected monomers yielded a set of ADMET polymers containing a variety of amino acids precisely distributed along the polymer chain.48,49,51 While previous examples showed the possibility of coupling amino acids along the polymer backbone, incorporation of amino acids into the main chain requires a different synthetic approach that has proved to be difficult, due to reactivity of both termini groups. Instead of using the respective amino acid, a different approach to synthesize chiral amino acid-based monomers was used starting from α-amino alcohols.48 Monomers shown in Fig. 23 were synthesized according to the procedure by Koyama et al.52 The alkenyl acid was reacted with the coupling reagents l-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC-HC1) and 1,3-dicyclohexylcarbodiimide (DCC) to form the activated ester. Then the amino alcohol was added, followed by the addition of 4-(dimethylamino)pyridine (DMAP) to catalyze the reaction, yielding a set of five monomers, as shown in Fig. 23.
In polymers synthesized by ADMET polymerization, three distinct crystal motifs are possible (Fig. 25). The thickness of thin crystals (Fig. 25a) is defined by the length of a single alkyl run length with the functional groups excluded from the crystalline domain. The rest of the chain may enter the amorphous phase or bend and adjacently re-enter the crystalline domain. Extended chain (EC, Fig. 25b) crystals have longer-range ordering of polymer chains, and the functional group is incorporated into the crystalline domain. Because the crystals must accommodate the functionality, the functional groups are generally aligned at an angle to the vector normal to the polyethylene crystal direction. In adjacent re-entry (AR) crystals (Fig. 25c) the polymer chain preferentially undergoes a hairpin turn at the location of the functional group, so that the chains predominantly crystallize intramolecularly with the functionality on the edges of the crystalline polyethylene regions. Note that in the cases of EC and AR crystals (Fig. 25b and c) that amorphous regions also exist but are not shown.
![]() | ||
Fig. 25 Three types of crystals formed by precisely functionalized PE (Reprinted from ref. 63.) (a) Thin crystal (b) extended chain (EC), and (c) adjacent reentry (AR). Black lines represent polyethylene segments, and blue dots represent the incorporated functional groups. (Reprinted with permission from ref. 63. Copyright©2017 American Chemical Society.) |
ADMET polymers prepared with alkyl groups pendant to the polyethylene backbone generally adopt the thin crystal motif if the pendant groups become sufficiently large that they cannot be reasonably incorporated into the crystal structure. The effect of branch size on crystal morphology was shown elegantly by comparing polyethylene with an ethyl or n-hexyl branch on every 21st carbon (EB21 and EO21, respectively).64 In the wide-angle X-ray diffraction (WAXD) pattern of EB21, a prominent peak located at 20.1° accompanied by a minor reflection peak at 22.5° (Fig. 26a) suggests a dominant hexagonal crystal structure which includes the ethyl branches but some triclinic character of the structure is also noted. It was reasoned that if the ethyl branches were all aligned horizontally in the crystal lattice, there would be crowding of these branches in the crystal lattice, leading to instability. Instead, it is likely that ethyl groups are staggered within the crystal structure to avoid lattice distortion. The observed effect of ethyl group staggering is tilting of the a–b plane relative to the CH2 sequences (in the c plane), leading to the observed triclinic character. Alternatively, if the pendant aliphatic group becomes so large that it cannot be accommodated within the crystal structure, it is simply excluded as shown for EO21. Fig. 26b shows the WAXD pattern of EO21. Sharp reflections at 19.6° and 23.1° differ only nominally from those recorded for monoclinic polyethylene (19.45° and 23.17°). The lack of an observed a-axis lattice expansion relative to monoclinic polyethylene is evidence that the n-hexyl branches must be excluded from the crystalline domain, resulting in thin crystals.
![]() | ||
Fig. 26 WAXD patterns of slowly cooled EB21 (a) and EO21 (b). (Adapted with permission from ref. 64. Copyright©2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.) |
Another important consideration is the effect of precision spacing of the ethyl and n-hexyl groups compared to randomly incorporated alkyl branches (e.g., those in commercial LLDPE). Pristine branch spacing afforded by ADMET polymerization resulted in narrow lamellar thickness distributions (LTD), which are positively correlated to mechanical properties such as tensile strength. A comprehensive study of the effect of alkyl branch size on the crystal morphology found that all branches larger than ethyl groups are excluded from the crystalline domains, resulting in a thin crystal type secondary structure.13 In addition to WAXD analysis which showed the inclusion of methyl and ethyl branches in the crystalline domain and the exclusion of all others, differential scanning calorimetry (DSC) showed that all of the polymers with propyl and larger alkyl pendant groups exhibited a melting endotherm at essentially the same temperature (Fig. 27), evidence in further support that bulky groups are excluded from the crystalline regions. Further, the heats of fusion are all essentially the same. The trend of bulky alkyl group exclusion also holds when the methylene spacer length between groups is increased.65,66
![]() | ||
Fig. 27 DSC curves of linear alkyl-substituted ADMET polymers (R group is specified in parentheses next to each curve). (Reprinted with permission from ref. 13. Copyright©2009 American Chemical Society.) |
In addition to methyl- and ethyl-substituted precision polyethylenes, halogen-containing precision polymers form EC crystals. The crystal structures and crystallization thermodynamics of halogen-containing precision polymers have been investigated in detail.22,67–71 Some trends can be distilled from this work, particularly based on the halogen which is incorporated. When a halogen (F, Cl, or Br) is substituted on every 19th carbon, each of the WAXD patterns displays sharp diffraction peaks often associated with homopolymer crystallization rather than copolymer-like behavior (Fig. 28).22 A purely orthorhombic unit cell (isomorphous with that of linear polyethylene) was observed for polyethylenes with either F or O substituted for H on every 19th carbon (Fig. 28). Substitution with the larger halogens Cl and Br led to slightly broadened WAXD patterns and evidence of lattice expansion resulting in a less-symmetric triclinic unit cell.
![]() | ||
Fig. 28 WAXD patterns of linear polyethylene (top) and precision substituted polyethylenes prepared via ADMET polymerization. Pendant groups (X = H, O, F, Cl, Br) are indicated next to each pattern. (Reprinted with permission from ref. 22. Copyright©2006 American Chemical Society.) |
Interestingly, pendant groups larger than Cl, Br or ethyl branches can seemingly be incorporated into the crystal structure of precise polyethylenes, forming an EC crystal structure while retaining an orthorhombic unit cell. Precision poly(ethylene-co-acrylic acid) polymers were prepared by ADMET polymerization with the carboxylic acid pendant group located on every 9th, 15th, and 21st carbon (p9AA, p15AA, and p21AA, respectively).25 Based on DSC experiments of these polymers, the methylene spacer length had a profound effect on the thermal behavior. Similar to other ADMET polymers with the functional group pendant to the polyethylene backbone, placement of functional groups closer together yields less crystalline material because a higher fraction of steric bulk prevents crystallization. In this case, both p9AA and p15AA showed completely amorphous behavior (no melting endotherm seen via DSC). However, an ordered secondary structure for p21AA was evident from the sharp melting endotherm in the DSC trace. These DSC results agreed with the X-ray scattering data with respect to semi-crystalline versus amorphous behavior.
To induce further anisotropy in the p21AA sample, it was stretched and then analyzed by X-ray scattering. The scattering pattern of the stretched sample is shown in Fig. 29a. Drawing the sample in the direction of the arrow in Fig. 29a promoted alignment of the polyethylene segments (seen as multiple reflections of the (110) plane) which were oriented at a 90° angle to the low-angle acid reflections. A proposed schematic indicating the arrangement of the carboxyl groups in the crystal structure is presented in Fig. 29b. Based mainly on the X-ray scattering pattern in Fig. 29a, it was reasoned that the combination of low acid group frequency and crystalline stabilization led to hydrogen bonding between in-plane acid neighbors. A secondary structure consisting of EC crystals which do not undergo a crystal lattice change to accommodate the carboxylic acid groups was assigned in the absence of other evidence to suggest otherwise. This result is contrary to what one might predict based on previous studies with large aliphatic and large halogen pendant groups.13,22
![]() | ||
Fig. 29 (a) X-ray scattering pattern of p21AA stretched along the arrow direction. Multiple reflections of 2.48 nm−1 correspond to a layered structure with 2.53 nm spacing oriented perpendicular to the polyethylene backbone (b) Proposed schematic of the arrangement of p21AA chains within a crystalline domain showing hydrogen bonds between acid groups in the same plane. Hydrogen bonds may form with chains above and below the plane shown. (Reprinted with permission from ref. 25. Copyright©2007 American Chemical Society). |
Until recently, polymers with AR-type secondary structures (Fig. 25c) were thought to be rare in the large library of materials prepared by ADMET polymerization. In 2008, careful DSC studies of ADMET polymers with precision-spaced short pendant chains of poly(ethylene glycol) (PEG) were published. Modulated DSC and selective annealing experiments on an ADMET polymer with PEG pendant groups on every 21st carbon confirmed two distinct crystalline populations.53 It was reasoned that the bimodal melting behavior corresponded to nanoscale segregation between the polyethylene backbone and PEG side-chains. For this to occur, the pendant PEG groups (which are chemically immiscible with the PE backbone) must have induced a fold in the polyethylene backbone, and the pendant groups were crystallizing amongst themselves separately. In 2017, the Winey group performed a closer examination of the p21AA nanostructure (originally reported in 2007)25 using molecular dynamics simulations, wide angle X-ray scattering (WAXS), and Raman spectroscopy. Both EC and AR structures of p21AA were simulated, and the experimental data, particularly the structure factor peak deconvoluted from the experimental WAXS, matched closely with the calculated AR results (Fig. 30).
![]() | ||
Fig. 30 Structure factor comparison between simulations and experimentally obtained WAXS profile of p21AA. (top) Experimental WAXS profile deconvoluted into amorphous and crystalline peaks (black = experimental, green = total fit, dashed gray = amorphous halo, solid gray = crystalline peak). (middle) Calculated AR structure factor peak (blue). (bottom) Calculated EC structure factor peak (red). Both calculated structure factors were overlaid onto the experimental crystalline peak (solid gray) for comparison. (Reprinted with permission from ref. 63. Copyright©2017 American Chemical Society.) |
The agreement between the modeled AR data and experimental data rectified the puzzling result that the bulk-crystallized p21AA seemed to perfectly integrate the carboxylic acid groups into an orthorhombic crystal structure despite the large size of the functional group. Winey and co-workers noted that further comparison between p21AA and similar ADMET polymers support the AR structural hypothesis. For example, the layer spacings of p21AA were compared with those of the ADMET polymer with two geminal acid groups on every 21st carbon (p21gAA) as well as the polymer with a phosphonic acid on every 21st carbon (p21PA). Despite the very large van der Waals volume differences between the functional groups on these three polymers, their layer spacings are within 7% of one another. If these polymers were each arranged in an EC structure, it would stand to reason that the more space a functional group occupies, the larger the required tilt angle between the normal to the functional group layer and the chain axes for efficient packing (decreasing the layer spacing). This is not the observed trend. Also, the melting temperatures of p21AA, P21gAA, and P21PA are similar (47.7, 46.1, and 47.4 °C, respectively),60 and very similar to the melting point of PEG-grafted polyesters with 20-carbon-long alkyl segments (between 44 and 46 °C).72 In contrast, in purely aliphatic systems, increasing the functional group size in EC drastically lowers the melting point.
The exciting implications of the AR-type secondary structure lead to consideration of how this nanoscale arrangement might result in channels of ion-conducting functionality for efficient shuttling of protons or other ions in electrochemical devices. Use of these types of materials as ion-transport membranes is only possible if long range ordering of the polymer chains can be achieved. Yan and co-workers demonstrated long-range ordering through the carefully controlled solution growth of single crystals made from precise acid and ion-containing polymers.73 Specifically, p21AA, p21gAA and a precise polymer with imidazolium bromide on every 21st carbon (p21ImBr) were chosen to demonstrate the growth of large single crystals from dilute solutions at appropriate crystallization temperatures. The p21AA single crystals were quite uniform in size and generally square-shaped (Fig. 31a and b), while the p21ImBr polymer formed remarkably large crystals with a few different shapes (one is shown in Fig. 31c and d). In addition to ample Raman spectroscopy and scattering data which suggested an AR-type secondary structure, the P21ImBr crystal thicknesses exceeded the fully extended contour length of the polymer (based on average molecular weight) by as much as a factor of three. Since it is incredibly unlikely that the polymer chains would be stretched out end-to-end in an EC-conformation to form these large crystals, this serves as another piece of evidence in support of the AR-type secondary structure in these polymer crystals.
![]() | ||
Fig. 31 (a) AFM image of a p21AA single crystal (polymer structure and schematic of p21AA crystal showing 80° and 100° growth sectors inset). (b) Height profile corresponding to the path drawn in white in image (a). (c) AFM image of p21ImBr (polymer structure inset). (d) Height profile corresponding to the path drawn in white in image (c). (Adapted with permission from ref. 73 Copyright©2017 Elsevier Ltd.) |
Recently, an exciting advance showing the advantages of an AR-type secondary structure was observed for precision polyethylenes with a sulfonic acid on every 21st carbon (p21SA).34 The folded chain structure formed channels of sulfonic acid groups which are highly hygroscopic and lead to swelling of the acid-enriched regions in the presence of humidity (Fig. 32). Sulfonic acids are favored for many ion-conducting membranes mainly due to their hygroscopicity. Perhaps the most prevalent commercial sulfonic acid-containing membranes are made of Nafion®, which has been used as the benchmark material in fuel cells and other electrochemical devices for decades. However, the microstructure and the path that protons take through the material is not without controversy.74 To demonstrate whether the AR-type ordered morphology of p21SA membranes would lead to desirable proton conductivity, electrochemical impedance spectroscopy (EIS) was performed at varying degrees of humidity on both Nafion® 117 and a membrane made of p21SA. The conductivities of the two materials were very similar above 60% relative humidity, and this was the first reported instance of such proton conductivity in a polyethylene-based crystalline structure.
![]() | ||
Fig. 32 (a) Intermediate- and wide-angle X-ray scattering profiles of p21SA at 20% relative humidity (RH) at room temperature (intermediate X-ray scattering of commercial Nafion® 117 is shown in light gray for comparison). (b) Primary layer peak of p21SA as a function of RH at 40 °C and immersed in water (far right vertical strip). Each vertical strip is an I versus q curve, with the far-right vertical strip displaying the primary layer peak of p21SA submerged in water. Images (c) and (d) are schematic representations of the layered structure with increasing humidity. (Reprinted with permission from ref. 34. Copyright©2018 Springer Nature Limited.) |
Based on the work discussed in this section, the implications of harnessing predictable secondary structures are becoming clearer. The ability to extract the effects of chemical structure on nanoscale morphology has enabled the construction of a set of design rules for precision materials with predictable structures. These rules will serve as a powerful toolbox for the design of precision polymers which can help solve contemporary problems as well as those yet undiscovered (Table 1).
Polymer structure | Secondary structurea | Functional group frequency (on and every # of carbons) | T m (°C) | Ref. |
---|---|---|---|---|
a TC = thin crystals, EC = extended chain crystals, AR = adjacent reentry. b Secondary structure originally assigned as EC, but updated to AR after additional molecular dynamics simulations, Raman spectroscopy, and WAXD experiments.63 c Based on selective annealing DSC experiments. d X-ray scattering performed at room temperature which was above the melting point of the polymer. e Determined based on DSC results. | ||||
![]() |
EC | 15 | 397 | 75 |
![]() |
EC | 21 | 63 | 13 |
![]() |
EC | 21 | 2413 22.264 | 13 and 64 |
![]() |
TC | 21 | 12 | 13 |
![]() |
TC | 21 | 12 | 13 |
![]() |
TC | 39 | 75.3 | 65 |
![]() |
TC | 75 | 104 | 66 |
![]() |
TC | 21 | 14 | 13 |
![]() |
TC | 21 | 1213 10.364 | 13 and 64 |
![]() |
TC | 21 | 11 | 13 |
![]() |
TC | 21 | 9 | 13 |
![]() |
TC | 21 | 13 | 13 |
![]() |
TC | 21 | 9 | 13 |
![]() |
TC | 21 | −8, 17 | 13 |
![]() |
EC | 9 | 124.3 | 67 |
![]() |
EC | 15 | 125.7 | 67 |
![]() |
EC | 19 | 127.5 | 22 |
![]() |
EC | 21 | 127.7 | 67 |
![]() |
EC | 19 | 134.7 | 22 |
![]() |
EC | 9 | 15 | 68 |
![]() |
EC | 15 | 54 | 68 |
![]() |
EC | 19 | 72.7,22 6368 | 22 and 68 |
![]() |
EC | 21 | 70 | 68 |
![]() |
EC | 9 | −13 | 71 |
![]() |
EC | 15 | 49 | 71 |
![]() |
EC | 19 | 61.5 | 22 |
![]() |
EC | 21 | 70 | 71 |
![]() |
Amorphous | 9 | — | 25 |
![]() |
Amorphous | 15 | — | 25 |
![]() |
ARb | 21 | 45 | 25 |
![]() |
Amorphous | 15 | — | 28 |
![]() |
AR73 | 21 | 48 | 28 |
![]() |
Amorphous | 15 | — | 28 |
![]() |
AR73 | 21 | 87 | 28 |
![]() |
AR | 21 | 90 | 73 |
![]() |
ARc | 21 | Approx. −47 and −25 | 53 |
![]() |
ARc | 21 | Approx. −47 and −25 | 53 |
![]() |
Amorphous | 8 | — | 76 |
![]() |
Amorphousd | 14 | 7 | 76 |
![]() |
EC | 20 | 37 | 76 |
![]() |
EC | 8 | 175 | 35 |
![]() |
EC | 14 | 162 | 35 |
![]() |
EC | 20 | 150 | 35 |
![]() |
Amorphouse | 9 | — | 33 |
![]() |
AR34 | 21 | 65 | 33 |
![]() | ||
Fig. 33 Chemical structures of partially fluorinated and/or sulfonated precision polymers as precise analogs of commercial proton-conducting membranes. (Reprinted with permission from ref. 77. Copyright©2017 Elsevier Ltd.) |
ADMET polymers have also found use as platforms for controlled molecule release and/or stimulus-triggered degradation. In one such study, ADMET polymerization was used to prepare polymeric prodrugs consisting of polyethylene backbones with pendant PEG or methylene spacer groups bearing a therapeutic agent (the non-steroidal anti-inflammatory drugs ibuprofen or naproxen in this case, see Fig. 34).78 Surprisingly, each of the tested variations of pendant type (hydrophilic or hydrophobic and length) seemed to generate similar drug release profiles characterized by an initial rapid release of the drug followed by a tailing-off of release. It was hypothesized that solubility of the polymer was the main determining factor in the release, where surface available drug groups were cleaved first. Perhaps the polymer chains collapse as a result of reduced solubility which contributed to the decrease in drug release after the initial burst.
![]() | ||
Fig. 34 Chemical structures of prodrug precision polymers. (Adapted with permission from ref. 78. Copyright©2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.) |
A similar area of interest consists of polymers which degrade in response to environmental stimuli into small molecules which are ideally benign and/or simple to re-use. This field is generally divided among two strategies: self-immolative or chain-shattering degradation. Self-immolative polymers are defined by an incorporated group at the chain end that, once activated, triggers a cascade reaction causing the polymer chain to “unzip” into small molecules that may or may not be the same as the monomer used to construct the chains. Chain-shattering approaches require triggering reactions to occur at each repeating unit, often requiring stoichiometric amounts of the stimulating molecule for full degradation. Miller and co-workers recently published an approach for a self-amplifying chain-shattering system which needs only small amounts of weak acid to begin degradation.79 Their synthetic approach is shown in Fig. 35a. Briefly, the synthesis of the iodo acetal monomer (1) started with the treatment of acrolein (2) with TMSBr. Then acetalization with alcohol (3) resulting in bromo acetal (4). The bromo acetal was simply converted to the iodo acetal (1) under standard Finkelstein conditions in good yield. ADMET polymerization afforded structure (5), and the internal olefins were hydroxylated via Upjohn-dihydroxylation to yield (6).
![]() | ||
Fig. 35 (a) Synthetic approach to form 3-iodopropyl acetal-containing polymer; (b) mechanism of 3-iodopropyl acetal hydrolysis with self-amplified cleavage after generation of HI in step 3 (this process happens within the polymer chain, depicted here as R1 and R2 for clarity). (TMSBr = bromotrimethylsilane, PhH = benzene, TsOH = p-toluene sulfonic acid, NMMO = N-methylmorpholine N-oxide). (Adapted with permission from ref. 79. Copyright©2019 American Chemical Society.) |
A proposed general mechanism of self-amplifying chain-shattering is shown in Fig. 35b. The reaction begins with activation of the 3-iodopropyl acetal with weak acid, eventually releasing HI which then serves as the self-amplifying degradation reagent. The chain-shattering of structure (6) was studied by monitoring the disappearance of the acetal proton via1H NMR spectroscopy in a 40% (v/v) CD3CN in D2O solution (pD0 = 5.5). A sigmoidal disappearance of the acetal proton along with a sigmoidal shift in the GPC trace to lower MW species suggested that the proposed self-amplifying mechanism was correct. A hydrogel made with a similar acetal motif also showed similar sigmoidal decomposition behavior. It is important to note that the two molecules released from the decomposition are hydroiodic acid, a very strong acid that accelerates further degradation, and (2) acrolein, a potent biocide. These aspects provide the potential for these polymers to serve as benign vehicles for controlled release of biocidal acrolein, for instance.
![]() | ||
Fig. 36 Synthesis of conjugated polymer poly(3-decylselenylene vinylene) (P3DSV) via ADMET polymerization. (Reprinted with permission from ref. 92. Copyright©2015 American Chemical Society.) |
Hillmyer and co-workers investigated how the optoelectronic characteristics (e.g., band gap, photoluminescence, absorption strength, charge transport) of poly(thienylene vinylenes) could be tuned through statistical ADMET copolymerization between structurally varied thienylene vinylene monomers.93 A study by Zhang and Qin demonstrated similar optoelectronic tunability by copolymerizing thienylene vinylenes with selenylene vinylenes.92 They proposed that by replacing S for Se in the CP structure, additional tunability of optoelectronic properties would be accessible due to the larger atom size, enhanced polarizability, and stronger spin−orbit coupling of Se. Further work from the same group presented a versatile synthetic route toward derivatized selenylene vinylene monomers resulting in some fundamental understanding of structure/property relationships of these unique polymers.88
A novel CP which functioned as a photoswitchable polyelectrolyte was recently reported by Joo and Bielawski.94 The α,ω-diene monomer was derived from a 4,5-dithienylimidazolium salt (see resulting ADMET polymer structure before and after UV-irradiation in Fig. 37a). The resulting polymer displayed impressive emission properties resulting from the conjugated structure. Fig. 37b and c show the absorption spectra of the monomer and polymer prior to irradiation and the emission performance of each after irradiation, respectively. These results are presented photographically in Fig. 37d. The authors noted that this photochromic polymer motif may serve as an interesting alternative to other CPs such as poly(thienylene vinylenes) which have already been explored in-depth. For example, adding donor/acceptor groups or changing the conjugation length of photochromic monomers may provide tunability in the resulting polymeric optoelectronic properties.
![]() | ||
Fig. 37 (a) Structure of photoswitchable CP upon irradiation with UV or visible light, (b) comparison of UV-Vis spectra for CP monomer and polymer of structure shown in (a), (c) emission spectra for monomer and polymer after excitation at 365 nm, and (d) photographs of vials containing solutions of monomer or polymer before and after excitation at 365 nm. (Adapted with permission from ref. 94. Copyright©2019 Royal Society of Chemistry.) |
ADMET polymers with varying amino acids incorporated at every 19th carbon in the main chain were investigated by Freudenberg et al.96 The effect of chiral, achiral, polar, and non-polar amino acids on the thermal properties was examined (see Fig. 38 for a schematic of materials tested). It was found that polymers bearing non-polar amino acid side chains formed only amorphous materials. Polar amino acid side chains in the glutamic acid and aspartic acid residues resulted in semi-crystalline polymers with sharp melting endotherms observed via DSC. WAXS analysis suggested a lamellar structure for both glutamic acid and aspartic acid systems, but the authors could not make a more specific assignment of secondary structure based on the data presented.
![]() | ||
Fig. 38 Schematic of ADMET polymers containing amino acid residues in the main chain. (Reprinted from ref. 96. Published by The Royal Society of Chemistry.) |
Glycolipids, generally found decorating the outside of cell membranes, are important for maintaining membrane stability and for cell recognition. A clever enzymatic approach to prepare α,ω-unsaturated trehalose diesters was described by Hibert et al.97 When crystallized from the bulk, the resulting ADMET polymer of the pure trehalose-derived glycolipid exhibited an EC ordered morphology driven by the association of carbohydrate and lipidic regions, respectively. Nanoscale self-assembly into micelles was observed in water due to the amphiphilic nature of the synthetic glycolipid (Fig. 39a and b). In another study, oligodeoxynucleosides linked with olefin bridges via head-to-tail ADMET polymerization were shown to form remarkably long fibrils via self-assembly similar to that which drives the formation of the double helix found in deoxyribonucleic acid (DNA, Fig. 39c and d).98
![]() | ||
Fig. 39 Schematic of synthesized oligodeoxynucleosides from regio-irregular ADMET polymerization (top) or head-to-tail ADMET polymerization (bottom), and TEM images of corresponding self-assembled nanostructures. (Reprinted with permission from ref. 98. Copyright©2019, American Chemical Society.) |
For aliphatic polyesters and polyamides, precision spacing of the carbonyl moieties is essential, because dipole–dipole or hydrogen bonding interactions are responsible for the famously high melting points and good mechanical properties of these materials. Sustainable polyesters and amides are not generally aliphatic, so structure/property relationships are the focus of the cutting-edge research in this area. Sustainably derived α,ω-diene ester monomers were synthesized from glucose derivatives isosorbide and glucarodilactone and 10-undecenoyl chloride (a castor oil derivative).102 The two monomers (glucarodilactone undecenoate (GDLU) and isosorbide undecenoate (IU)) were homopolymerized and copolymerized, and Mn values of 50–60 kDa were achieved (see monomer and polymer structures in Fig. 40). The two homopolymers and two copolymers with very similar GDLU/IU compositions were analyzed for their thermal and mechanical characteristics. Poly(IU) exhibited the highest thermal stability (Td = 369 °C) and all of the polymers exhibited melting points between 22 and 59 °C. Tensile testing of homopolymers poly(GDLU) and poly(IU) revealed brittle behavior, while both copolymers exhibited rubbery behavior and shape-memory properties. Further work by the same group found that the percentage of GDLU incorporated was positively correlated with desirable degradation properties, and rheological measurements on GDLU-containing polymers suggested that a transient network was responsible for the observed elastic and shape-memory properties.103
![]() | ||
Fig. 40 Synthesis of GDLU and IU and subsequent ADMET polymerization to form glucose and castor oil-derived polyesters. Conditions: (a) 4-(dimethylamino) pyridine, 10-undecenoyl chloride, tetrahydrofuran/chloroform, 0 °C to RT, 16 h; (b) 1.0 mol% Grubbs’ 2nd generation catalyst, 1.0 mol% methyl 10-undecenoate, toluene (0.33 M), 80 °C, 16 h, vacuum. (Reprinted with permission from ref. 102. Copyright©2015, American Chemical Society.) |
Barbara et al., reported the synthesis of poly(ester-alkenamers) prepared by ADMET polymerization of ferulic acid (from lignocellulosic biomass), biosourced diols (isosorbide and butanediol), and bromo-alkenes.104 Thermal stability up to ∼350 °C was reported and the Tgs were tunable by adjusting the nature of the internal diester and the alkene length of the α,ω-dienes. A structurally complex bisphenol polyester derived from syringaresinol which followed similar structure/property trends was prepared by Hollande and co-workers.105 Mutulu and Meier reported an ADMET polymerization approach to prepare unsaturated precision polyamide-X,20 (X corresponds to the number of carbons in the aliphatic diamine) from methyl-10-undecenoate (derived from castor oil).106 The symmetric amide-containing α,ω-diene was synthesized by treating 2.5 equivalents of methyl-10-undecenoate with the diamine of interest (1 equivalent) with 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) as a catalyst. ADMET polymerization of the prepared monomers yielded unsaturated polyamides with modest Mn values.
A large library of phenolic polyethers were prepared by Vlaminck and co-workers from lignin-derived aromatic diols.107 A summary of the α,ω-dienes prepared is presented in Fig. 41. Diols were created by either a Dakin reaction (n = 0) or reduction (n = 1) of the aldehyde, and the dienes were formed by a Williamson ether synthesis with allyl bromide (A, m = 1), butenyl bromide (B, m = 2), or pentenyl bromide (P, m = 3). Several structure/property relationships in the ADMET polymers were observed due to the large structural variety including phenyl substitution pattern, alkyl spacer length, and varying methoxy substituents (0, 1, or 2). As might be expected, ortho-substitution on the aromatic ring resulted in lower Tg values compared to para- and meta-substituted rings. Chain flexibility afforded by longer alkyl spacers between aromatic groups also resulted in lower Tgs. Added methoxy substituents on the aromatic ring generally increased the Tg values. More structurally complex polyethers and polyesters were constructed from vanillin and eugenol derivatives as well as polyethers from L-ascorbic acid.108,109
![]() | ||
Fig. 41 Overview of synthesized diene monomers starting from hydroxybenzaldehydes (H), vanillins (V) and syringaldehyde (S), with different substitution patterns (ortho (o), meta (m), para (p)). (Reprinted with permission from ref. 107. Copyright©2017 Elsevier Ltd.) |
It is clear from this body of work that the precision afforded by ADMET polymerization has led to structure/property understanding that would be impossible to study in imprecise systems. Effect of functional group frequency and size on precision polymer secondary structure is perhaps the most well-studied vignette in this area. Major advances in both identifying and leveraging highly ordered secondary structures in ADMET polymers has occurred in only the past few years. Many functional polymer systems that make use of the olefin bond left behind after ADMET polymerization have been explored and are discussed in this work, among other functional polymers. Still, much of the future research potential in ADMET polymers lies in taking advantage of well-controlled nanoscale secondary structures. The design rules to achieve each possible semi-crystalline structure are becoming clearer, and this understanding will lead to the design of precision polymers which may contribute greatly to the devices of the future.
This journal is © the Partner Organisations 2021 |