The impact of dynamic bonding associations on the rheology of amine containing polyolefins

Abstract

Polyolefins are utilized in a wide array of applications due to their desirable and predictable properties. The incorporation of polar functional groups into the nonpolar backbone imparts an opportunity to further tune the properties and broaden the application space. Among the library of H–bonding chemistries that have been implemented via post-polymerization modifications, amines are of particular interest due to their challenging syntheses. The discovery of a two-step catalytic combination of hydroaminoalkylation and ring-opening metathesis polymerization (ROMP) has opened access to a new class of amine containing polyolefins, transforming simple starting materials into polar-functionalized polymers with complete atom economy. New rheological properties emerge due to the dynamic associative interactions, including H–bonding, stacking, and aggregate clustering, which are highly tunable based on the molecular design. The amine containing polyolefins also demonstrate autonomous self-healing and adhesion to low surface energy substrates, with particular interest in advanced adhesive systems. This review article details the recent progress in the molecular design of amine containing polyolefins, and the diversity of linear, non-linear, and extensional rheological behaviors. A discussion into remaining challenges in our understanding of the complex microstructure of the amine containing polyolefins and future pathways to designer materials from simple starting feedstocks is also presented.

Introduction

Amine-containing polymers are unique due to their nucleophilicity and basicity, which enable their broad applications as antimicrobial agents [1, 2], compatibilizers in polymer blends [3, 4], materials for CO₂ capture [5, 6], and in water purification [7, 8]. Commercially available polymers with high content of amine functional groups include polyvinylamine (PVAm), polyallylamine (PAAm) and polyethylenimine (PEI) with structures shown in Fig. 1a. Owing to their high N/C ratio, these polymers are highly hydrophilic and remain water-soluble across a broad pH range of 2 to 11 [9]. These polymers are typically applied in the form of aqueous solutions. As a result, the study of their melt-state properties is limited. While some work has been done to address the melt-state rheological properties of PEI dendrimers, the focus has primarily been on the effects of molecular architecture rather than the role of amine functionality in the relaxation mechanisms and viscoelastic behavior [10].

Fig. 1.

a Representative structures of polyvinylamine (PVAm), polyallylamine (PAAm) and polyethylenimine (PEI). b Representative structures of amine-containing groups frequently involved in H–bonding interactions in associating polymers

Amine functional groups have also been incorporated into traditional polyolefin systems either through copolymerization or post-polymerization modification with H–bonding moieties, which are typically introduced as pendant groups along the polymer backbone or at the chain ends [11–19]. Common moieties, with structures shown in Fig. 1b, usually involve more than one type of hydrogen bonding element. Embedding these dynamic functional groups into the non-polar polyolefin chains transforms them into associating polymers and opens up exciting opportunities for stimuli-responsive behavior, enhanced adhesion, self-healing abilities, and tunable viscoelastic properties [20]. For example, through post-polymerization functionalization, Zhou et al. synthesized polyethylene-based materials bearing versatile functionalities including amine, carbonyl and halide groups [15]. These polymers exhibit improved ductility and toughness, along with markedly enhanced adhesion to metal surfaces. To better understand the impact of these functional groups to the polymer dynamics and their correlation with the newly introduced unique properties, rheology serves a useful tool.

As discussed, the functional groups usually involve multiple hydrogen bonds between two moieties and involve both oxygen and nitrogen atoms. The bonding strength for these moieties is much higher than simple amine-amine association [21, 22]. In addition, secondary amines may act as both donors and acceptors to form multiple hydrogen bonds within one functional group, in contrast to the directionality of carboxylic acid or hydroxyl groups. Despite extensive studies on oxygen and nitrogen containing hydrogen bonding, pure amine-amine associations have received relatively little attention. Fetters and coworkers functionalized linear and star-shaped polyisoprene (PI) chains with amines at the chain ends and studied their effect on the linear viscoelastic properties by comparing the melt viscosity to that of the unfunctionalized reference sample [23]. The zero-shear viscosity ratio varied from 0.8 to 3.2 depending on the molecular weight and structure of the chains, suggesting minimal or weak association between amine groups in the melt state [23]. Similar studies were conducted on mono-and bifunctional telechelic polyisobutenes (PIBs) with 2,6-diaminotriazine groups [13]. In this system, the viscosity increased by a factor of 40 for an unentangled mono-functionalized polymer with a molecular weight of 3.5 kg/mol. Given that a doubling of molecular weight typically results in only a twofold increase in viscosity, this nonlinear enhancement suggests the formation of aggregates in the melt state. In the bi-functionalized system, complex aggregations led to the formation of a plateau in the viscoelastic spectrum. Similar effects were observed when 6-diaminotriazine groups were grafted along the PIB backbone [24].

The limited structural diversity of amine-functionalized polyolefins stems from the inherent incompatibility of amine groups with standard polymerization techniques. To overcome these limitations, Schafer’s group used an approach for generating amine-containing polymers using a two-step synthetic protocol [25–27]. In this method, alkene substrates were first functionalized with unprotected secondary amines via a hydroaminoalkylation (HAA) reaction, catalyzed by N,O-chelated tantalum precatalysts. [28, 29] A wide range of amine and alkene substrates were explored, resulting in a diverse library of amine-containing small molecules. Among them, the amine-functionalized monomer M1 and M2 (Fig. 2) synthesized with cyclooctadiene and norbornadiene alkene substrates and various arylamines were identified as promising starting materials for the ROMP [25–27, 30]. Catalyzed by the second-generation Grubbs catalyst, the reaction gives poly(aryl amine-containing cyclooctene), abbreviated as P(ACC) and poly(aryl amine-containing norbornene), abbreviated as P(ACN) [25–27]. As evident from their chemical structures, these polymers exhibit a high density of amine groups—one functional group per monomer unit. Such a high level of functionalization is difficult to achieve in associating polymers synthesized via copolymerization or post-polymerization functionalization. For instance, in copolymerization approaches, the level of functionalization is typically limited to below 15 mol%, whereas in polymers prepared via post functionalization C–H amination [11, 16, 31], the achievable amine content is usually below 3 mol% [15, 17–19].

Fig. 2.

Catalytic process to produce P(ACN) and P(ACC), and structures of P(ACC)-H and P(ACC)-R. Figures reproduced from References [26, 27]

This review aims to summarize the recent progress in the rheological and mechanical characterisation of these novel amine-containing polymers synthesized via HAA and ROMP. We focus on changes in the chain relaxation process and flow behavior induced by the high density of amine associations, which underlie the materials’ self-healing and adhesive properties. The review begins with a summary of the rheological behavior of P(ACN) and P(ACC), with particular emphasis on their unexpected long-time equilibrium characteristics. The impact of this equilibrium behavior on mechanical strength, adhesion, and self-healing performance is then discussed. Following this, amine-containing polymers synthesized via post-polymerization HAA, using commercially available polymers as starting materials, are introduced, with a brief overview of their rheological behavior. The review concludes by outlining current research challenges and identifying promising directions for future investigation.

Linear and extensional rheology of P(ACN)

The successful synthesis of P(ACN)s with different N-methylaniline (NMA) substituents was first reported in 2016 along with the initial investigation of the thermal and rheological properties [25]. Future advancements to the HAA precatalyst enabled a faster reaction time (reduced from 20 h to less than 3 h), higher monomer yields (increased from 60 to 90%), larger scale synthesis and a simpler experimental setup to improve the accessibility of amine monomers for the ROMP polymerization [26]. The synthesized P(ACN)s, containing secondary amine groups, were compared to reference samples without hydrogen bonding, either an unfunctionalized polynorbornene (PNB) [25], or a tertiary amine-containing polynorbornene (structure P10 illustrated in Fig. 3) [26]. With the incorporation of NMA derivatives, P(ACN)s and the tertiary amine-containing polynorbornene exhibited lower thermal stabilities, evidenced by a more than 50 °C decrease of 5% weight loss temperature () compared to PNB [25, 26]. The glass transition temperature (), measured by DSC, increased from 48  °C for both PNB and tertiary amine-containing polynorbornene to 62  °C for P(ACN) with unsubstituted aniline and up to 82 °C for P(ACN) with bromo-substituents [25, 26]. These results suggest that the addition of intermolecular H–bonding restricts segmental motions and is a clear path to tuning the of these polymers.

Fig. 3.

a Master curves of storage and loss moduli for P1 with varying molecular weights. b Dependent of zero-shear viscosity on molecular weight for P1 and PNB. c First column: structure of amine-functionalized monomer before ROMP; second column: master curves of storage and loss moduli, along with the TMA model fit (solid lines), of the respective polymerized amine-containing monomer in column 1; third column: Uniaxial stress growth coefficient obtained using a Sentmanat extensional rheometer at different Hencky strain rates. Figures reproduced from References [25, 26]

The linear rheological properties, represented by the construction of master curves of storage () and loss () moduli, illustrated a significant difference between the reference samples and P(ACN)s. The entangled reference samples exhibited a classical frequency dependence expected for a monodisperse polymer melt, with an entanglement plateau regime at intermediate oscillatory frequency (ω) followed by a cross over and a terminal relaxation regime [25, 26]. The master curves of P(ACN) with methoxy substituent (P1) with varying molecular weight (Fig. 3a) revealed that with increasing molecular weight, the samples exhibited a transition from a typical Rouse relaxation ( 5.1 kg/mol) to a gel behavior ( 27.6 kg/mol), characterized by an overlap of storage () modulus and loss () modulus and eventually to a solid-like elastic behavior ( 94.2 kg/mol) with dominant over across the entire frequency range [25]. The solid-like behavior suggested the formation of microstructures of long lifetimes because of the H–bonding network. The difference was also revealed in the zero-shear viscosity () and molecular weight relation as shown in Fig. 3b [25]. The of PNB showed a common power-law relation with molecular weight, with the exponent equal to 3.77. However, of P(ACN) with methoxy substituent exhibited an exponential growth, and the authors attributed this to the presence of different relaxation mechanism from H–bonding networks.

A more comprehensive study on the linear rheological properties of P(ACN) was performed with varying aryl substituents, namely P1, P2 and P3 with structure shown in Fig. 3c [26]. The entanglement molecular weight () of these polymers was estimated based on the tertiary amine-containing reference sample, which had an of 28 kg/mol. Contrary to what was found in Reference 25, both unentangled ( 25.8 kg/mol) and entangled ( 119 kg/mol) P(ACN)s samples showed a solid-like behavior with a plateau regime spanning over more than seven decades in angular frequency. However, the horizontal shift factors () for different aniline substituents were nearly identical. Despite the similar temperature dependences of the rheological properties, variation in the aniline substitutes influences the electron density close to the amine and the H–bonding strength, which in turn affect both the and the plateau modulus. When comparing the plateau modulus for P2 and P3 of similar molecular weights, P2 had a higher chain mobility, which increased the intermolecular associations possibility and thus a higher plateau modulus (despite multiple H-bond sites provided by both substituents).

Extensional rheological properties for the reference and three P(ACN)s samples were also reported, illustrating the strong impact of H–bonding on the nonlinear rheological behavior of the materials [26]. The tensile stress growth coefficient of the highly entangled reference sample closely followed the LVE envelope across all applied Hencky strain rates. In contrast, the three P(ACN) samples showed strong strain hardening behavior after different durations of extension. While the types of substituents influenced the degree of strain hardening, the trend and underlying mechanism remains unexplored.

Characterization of P(ACC)

In contrast to the rigid backbone of P(ACN), the flexible cyclooctene backbone of P(ACC) significantly alters the material properties. The successful synthesis of P(ACC)s was first reported in 2021 for several associating polymers such as P(ACC), P(ACC)-H and a reference sample (P(ACC)-R) (Fig. 2), all with around 25 kg/mol [27]. Moreover, the study focused on the unique self-healing and adhesion properties of P(ACC), which promoted further investigation on this polymer. Compared to the reference sample P(ACC)-R with around − 25  °C, P(ACC) exhibited a higher of approximately − 18  °C, due to the presence of H–bonding [32]. The linear viscoelastic properties of freshly synthesized P(ACC) with molecular weight ranging from 35 to 144 kg/mol revealed thermorheologically simple behavior as described by the master curves in Fig. 4a and b [33]. At high angular frequency, all curves converge as the dynamics of these polymers are controlled by the relaxation of segments between entanglements/associations. The plateau regime is developed with increasing , and extended to lower angular frequencies. Terminal relaxation is observed for all samples at low angular frequencies as indicated by the and scaling.

Fig. 4.

The master curves of a storage modulus (ω) and b loss modulus (ω) of freshly synthesized P(ACC) polymer with different molecular weights at . c Storage modulus from master curves of fresh and partially aged P(ACC) polymer samples at . d Validation of using for aging process prediction by comparing samples aged at 110 °C and room temperature for 3 h and 477 days. The master curves of e storage modulus (ω) and f loss modulus (ω) of aged P(ACC) with different molecular weights at . Figures reproduced from References [32–34]

An interesting observation is that P(ACC) exhibited significant time-dependent behavior [34]. When stored in the bulk state at room temperature, P(ACC) gradually transitioned from a soft viscoelastic liquid to a gel-like state, followed by an eventual transition to a viscoelastic solid over long periods of time (from months to years). This process is referred as “aging” in the following discussion. The aging process can be accelerated with higher temperatures, allowing the entire process to be observed within an experimentally accessible timeframe [34]. Master curves before and after different aging time and temperatures were compared in Fig. 4c. The most significant change occurred at low angular frequency region where the power-law scaling in and decreased as the overall magnitude of the moduli increased. A secondary plateau emerged as the system approached equilibrium, similar to what was observed in P(ACN) systems. Another notable change was observed in the high angular frequency region, where the master curves continually shifted toward lower frequencies. This shift was attributed to the elevated glass transition temperature during the transition process, as the high frequency domain aligned again when the master curves were reconstructed at an iso-frictional state. The transition kinetic at different temperatures was found to be closely related to the horizontal shift factors () obtained from the master curves. Specifically, the time required to reach the same transition state at two different temperatures equals the ratio of their respective values. Figure 4d presents an example for the validation of the rule, where the master curves for samples aged under two equivalent conditions (25 °C for 477 days and 110 °C for 3 h) show excellent overlap. The rule is useful for the prediction of material state during long time storage.

Comprehensive rheological characterization was performed on equilibrated P(ACC) materials [32]. Notable differences were observed in this set of measurements (Fig. 4e and f) compared to those for the freshly synthesized P(ACC) samples (Fig. 4a and b), particularly for P(ACC) with higher molecular weights. Samples with molecular weight lower than 30 kg/mol displayed viscoelastic liquid behavior similar to that of freshly synthesized samples. As molecular weight increases, the samples exhibited gel-like and viscoelastic solid behavior, resembling the trend reported by for P(ACN) samples.

The “aging” process was hypothesized to be driven by the polarity difference between the polymer backbone and the amine pedant groups, leading to physical aggregation through H–bonding and stacking interactions among secondary amines and phenyl rings [34]. Under this hypothesis, “aging” is essentially an equilibration process in which polymer chains rearrange to minimize the system’s free energy. Rheological characterization suggests that the primary differences between the freshly synthesized state and the fully equilibrated state lies in the strength of the associations [34]. In the fresh state, weak amine-amine associations with short disassociation time delay chain relaxation process but preserve the viscoelastic liquid nature of the system. In contrast, aggregated associations exhibited much longer disassociation time, leading to soft solid-like behavior within experimentally accessible timescales. This hypothesis was partially supported by FTIR results, which showed reducing band wavenumber for the associated amine peak as aging progressed, and by preliminary SAXS data, which revealed the emergence of a small shoulder in the aged sample indicating some level of phase separated aggregation [34].

Molecular modeling of association dynamics in amine containing polymers

Time-marching algorithms (TMA), which are powerful rheological modeling tools for probing polymer chain dynamics, were developed for both P(ACC) and P(ACN) with NMA pedant groups to capture their linear viscoelastic properties [26, 32]. The original TMA is based on tube-based reptation theory and incorporates polymer-specific parameters such as entanglement molecular weight rubbery plateau modulus and equilibrium time to describe polymer dynamics. To account for dynamic associations, additional parameters ( and ) were added [32, 35, 36]. was introduced to describe the probability of a monomer to create an interchain association, while accounted for the delay in the reptation process by the associated stickers. Since directly determines the molecular weight between associations through , with being the monomeric molar mass, either or can be used to describe the density of effective associations in the system, assuming not all polar amine groups are actively participating in an association event [32]. TMA functioned as an “inverse” model, where the optimal values of the parameters were determined to give the best fit to the experimental results [26, 32]. For the fully equilibrated P(ACN) samples, where associations primarily occur in the form of aggregation according to the hypothesis, was taken to be infinite due to the lack of terminal relaxation in the experimentally accessible timeframe [26]. In this case, associations function similarly to permanent crosslink before their disassociations, contributing to the plateau modulus. Therefore, and could be determined from the level of plateau modulus, where a rise in plateau modulus indicated a higher [26].

For non-equilibrated P(ACC) systems, a finite value is given to to account for an overall delay from the weak amine associations [32]. In this TMA model, is not determined from the plateau modulus since the weak amine associations is expected to have little effect on the plateau modulus. Instead, it was fixed to be for P(ACC) with varying due to the high density of associations in the system (approximately 1 sticker per Kuhn segment) [32]. It is interesting to note that the TMA model developed for P(ACC) provides a better fitting to samples with lower molecular weights compared to those with higher molecular weights [32]. This phenomenon can be reconciled with the recognition of the effect of “aging” [33]. In fact, the two TMA models effectively describe the two extremes of the equilibration process: the freshly synthesized and fully equilibrated state [26, 34]. To capture the behavior of partially equilibrated samples, the models can be combined by introducing two association possibilities: for weak, reversible associations and for stronger aggregated associations. A finite accounts for the delayed relaxation caused by weak associations, while is set to be infinity, reflecting the restricted chain motion between aggregated associations [33].

Tunable rheological and mechanical properties through sample formulation

Differences in amine structure has been shown to significantly influence the linear rheological properties as demonstrated above. For example, structure variations as simple as molecular weight can be leveraged to tune the material’s rheological behavior [37]. A low molecular weight P(ACC) ( 38 kg/mol) with minimal signature of aging was blended with a fully aged high molecular weight P(ACC) ( 134 kg/mol) [37]. The blends at various ratios spanned the full range from solid-like to liquid-like behavior observed in the two P(ACC) samples as illustrated in Fig. 5a, and the gel transition occurred close to a low/high polymer ratio of 1 as indicated in Fig. 5b.

Fig. 5.

a The master curves of storage modulus (ω) (closed symbols) and loss modulus (ω) (open symbols) for each bimodal molecular weight blend of short (S) (Mw = 38 kg/mol) and long (L) (Mw = 134 kg/mol) P(ACC) chains at . b Loss factor for each bimodal blend of short (S) and long (L) P(ACC) chains at . c Stress strain curves for two freshly synthesized P(ACC) with Mw=177 kg/mol (solid) and Mw=110 kg/mol (dashed) at four different linear strain rates and room temperature. d Stress strain curves for bimodal blends of aged short (S) and long (L) molecular weight P(ACC) at room temperature and 10 mm/min. Figures reproduced from References [37, 38]

Structure variation also exerts significant impact on the mechanical properties according to the stress-strain curves of two freshly synthesized P(ACC)s with 110 kg/mol and 170 kg/mol, shown in Fig. 5c as dashed and solid lines respectively [38]. An increase in led to an increase in engineering stress. However, for both samples and across all measured rates, strain softening was observed after the yield point, reflecting their viscoelastic liquid behavior. Samples are super ductile with ultimate strain higher than 3000%. Freshly synthesized P(ACC) with molecular weight lower than 40 kg/mol showed similar trends but exhibited a much lower (Fig. 5d), as the poorly entangled system could not sustain large deformations [37]. The stress-strain curves for fully aged P(ACC) with  134 kg/mol exhibited a completely different trend, as illustrated in Fig. 5d by pronounced strain hardening after the yield point, continuing up to the failure point at an ultimate stress of 1.06 MPa [37]. The engineering stresses were significantly higher than that of the freshly synthesized sample, while the ultimate strain was reduced to approximately 700%. The shape of the stress-strain curve resembled that of a rubber, suggesting that the strong aggregated associations act similarly to chemical crosslinks within the system.

As with linear viscoelastic properties, mechanical properties such as ultimate stress and strain, as well as the shape of the stress-strain curve (e.g., strain softening or hardening) can be tuned by adjusting the ratio of viscoelastic liquid and solid components in the blend system (Fig. 5d). An interesting case is that blending equal amounts of freshly synthesized low P(ACC) with fully aged high P(ACC) produced a stress-strain response that exhibited strain hardening after the yield point, followed by strain softening.

Impact of amine structure on adhesive properties

As first reported, P(ACC) exhibited adhesion to a wide range of substrates, even to low surface energy substrates such as PTFE, a material known for its resistance to adhesion [27]. In this study, the adhesion performance of P(ACC) was evaluated using the T-peel test, in which samples were sandwiched between two target substrates and the tensile force required to break the adhesive joint was measured. P(ACC) with a molecular weight of 27 kg/mol showed a peel strength of 19.2 N/m at a peel rate of 10 mm/min on non-surface-modified PTFE films, with cohesive failure observed. A comprehensive study on the peel performance of freshly synthesized P(ACC) was later conducted to investigate the impact of peel rate for a wide range of P(ACC) molecular weight [33]. Peel strength showed a strong positive correlation with molecular weight, increasing from 38 N/m for sample with of 36 kg/mol to 620 N/m for one with of 177 kg/mol at a peel rate of 10 mm/min, as presented in Fig. 6a. Cohesive failure was observed for all cases. Peel curves for two freshly synthesized P(ACC) samples (blue at 177 kg/mol and orange at 110 kg/mol in Fig. 6b exhibited increasing peel strength with peel rate in cohesive region, followed by a transition to adhesive failure, which caused a sharp drop in peel strength [38]. In addition to a higher peel strength in the cohesive failure region, higher molecular weight samples also showed an earlier transition to adhesive failure.

Fig. 6.

a Room temperature steady-state peel strength (in units of N/m) of freshly synthesized P(ACC) polymer at various molecular weight on untreated PTFE, rate of 10 mm/min. b Room temperature steady-state peel strength (in units of N/m) of two freshly synthesized P(ACC) polymer (Blue at 177 kg/mol and orange at 110 kg/mol) on untreated UHMWPE at various peel rates. Trends for c maximum, steady-state peel strength (in units of N/mm) and d lap shear strength (in units of N/mm²) of bimodal blends of aged P(ACC) polymer at various short (S) to long (L) chain ratios (S: L) (T = 25 °C) on untreated PTFE and rate of 10 mm/min. Error bars represent the standard deviation from replicates of 3–5 independent samples. Figures reproduced from References [33, 37, 38]

The peel strength of a fully equilibrated P(ACC) was reported in 2023 [37]. With a molecular weight of 134 kg/mol, P(ACC) exhibited significantly lower peel strength—around 25 N/m at a peel rate of 10 mm/min—and failed adhesively. The discrepancy between these two sets of measurement originates from the differences in the stress-strain behavior. As Gent and Petrich [39] pointed out, peel strength in the cohesive failure region could be approximated by summing the tensile stresses from the stress-strain curve of the adhesive. For equilibrated P(ACC), adhesive failure occurred due to the significantly increased engineering stress during the deformation (Fig. 5d), which exceeded the strength of the bonding between P(ACC) and the substrate. This resulted in limited viscoelastic energy dissipation before the adhesive detached from the substrate, and therefore a significantly reduced peel strength in adhesive failure region. In contrast, the continued increase in peel strength for freshly synthesized P(ACC) with increasing molecular weight suggests that their engineering stress had not yet surpassed adhesive-substrate bonding strength. The high adhesion performance of P(ACC) on low surface energy substrates was attributed primarily to the soft and ductile nature of the materials, allowing sufficient energy dissipation from adhesive deformation before failure [38].

From the discussion, strain hardening is not a universally favorable characteristic for achieving high peel strength on low surface energy substrates. It must be balanced within the cohesive and adhesive failure regions. This conclusion is supported by the peel strength results for blends of fresh low and equilibrated high molecular weight P(ACC) [37]. In addition to the equilibrated high molecular weight P(ACC), blends containing 50 wt% and 75 wt% of the high molecular weight component, which exhibited strain hardening behavior (Fig. 5d), showed adhesive failure and reduced peel strength. High peel strength within the cohesive failure regime can be achieved with a high molecular weight component lower than 25 wt%, as suggested by the transition to mixed mode failure when 25 wt% of aged high molecular weight P(ACC) were blended to its low molecular weight counterpart (Fig. 6c). The lap shear strength exhibited a consistent increase with the addition of aged high molecular weight content (Fig. 6d) [37]. However, a transition from cohesive to adhesive failure could still be observed when the aged high molecular weight fraction reached 50 wt%.

Impact of amine structure on self-healing behavior

The self-healing ability of P(ACC) was reported early on for a P(ACC) with a of 27 kg/mol which fully recover mechanical properties in a stress-strain curve at room temperature after bulk samples were cut, rejoined and left for 20 min [27]. A systematic evaluation of self-healing ability was subsequently conducted by comparing their stress–strain curves before cutting and after sample healing (Fig. 7a and b) [40]. Freshly synthesized P(ACC) with higher molecular weight ( 120 kg/mol) achieved full recovery at room temperature, though it required a longer healing time of 40 min.

Fig. 7.

Self-healing performance of P(ACC) characterized by mechanical properties: stress strain curves of a freshly synthesized and b thermally aged P(ACC) at 90 °C after recovery for 5, 10, 20, and 40 min at room temperature. Self-healing performance of freshly synthesized low molecular weight P(ACC) characterized by peel strength: Comparison of peel displacement curves before and after healing times of c 0 s and d 1800 s for P(ACC) of 38 kg/mol. Figures reproduced from References [37, 40]

The self-healing ability of P(ACC) is hypothesized to result from a combination of H–bonding and high chain mobility (due to low ), which together promote interfacial chain reassociations and interactions between separated surfaces [32]. The effect of H–bonding was demonstrated by a comparison where full recovery could not be achieved by P(ACC)-R samples using the same healing time for P(ACC) [32].

Attempts were made to investigate the self-healing dynamics of freshly synthesized P(ACC) [32] using the theoretical framework of Stukalin et al. [41] and Hinton et al. [42], focusing on the proposed scaling laws that describe how the density of recovered polymer chain bridges across two interfaces evolves over time during the recovery process. Note that the theory was specifically developed for unentangled telechelic polymers, contrasting the entangled P(ACC) with pendent associations. After the creation of new surfaces and prior to their rejoining, the concentration of surface free-associating groups naturally decreases over time as polymer chains diffuse and form new partners, eventually reaching a new equilibrium state [41]. Based on this waiting time , (i.e., the time between surface creation and rejoining), the recovery process is classified into two regimes: self-healing, which occurs when the waiting time is shorter than the equilibrium time , and self-adhesion which takes place when the waiting time exceeds . Depending on the local density of available associating groups (“open stickers”), two distinct mechanisms of bridge formation were identified: anomalous diffusion, which occurs when two open stickers are within their diffusion range and can directly form a bridge; and the hopping process, which dominates when open sticker concentration is low, and chain segments must undergo partner exchange to eventually bridge across the interface.

Based on the presented theory, the equilibrium time for freshly synthesized P(ACC) was estimated to be approximately 0.64 s, which is faster than any experimentally accessible timeframe [32]. This suggests that all reported self-healing experiments on P(ACC) were conducted in the “self-adhesion regime”. When the equilibrium concentration of open stickers at the interface is lower than one per pervaded volume, bridge formation proceeds primarily through a hopping mechanism, with the surface density of bridges scaling linearly with healing time at a slope of 1. In contrast, when the open sticker concentration is higher than one per pervaded volume, recovery follows a more complex process: initially, bridge formation is governed by anomalous diffusion, with the bridge density again scaling linearly with healing time (slope = 1) before reaching the Rouse time of the sticker-bearing segment. After this point, the bridge density plateaus temporarily before resuming growth via the hopping mechanism to restore additional chain bridges across the interface.

In P(ACC) samples, the recovery of ultimate strain was found to follow a power-law scaling with healing time [32], but the failure of the theory to accurately describe the self-healing behavior of P(ACC) is not surprising, as (1) the model does not account for the effects of chain entanglement on polymer dynamics and (2) the strain recovery after stress-strain experiments may not have a simple linear correlation with the density of bridge formation. Hinton et al. further examined the self-healing dynamics of an entangled associating polymers using both extensional and creep protocols [42]. They noted that the stress or strain recovery after extensional flow primarily reflects entanglement dynamics, whereas creep tests probe the combined effects of both entanglements and associations. To our knowledge, scaling laws for self-healing kinetics of entangled associating polymers have not yet been rigorously investigated or theoretically established. The experimental self-healing results obtained represent a valuable dataset for validating future theoretical models. Addressing this knowledge gap presents a major opportunity for advancing the fundamental understanding of associating polymer networks.

In addition to extensional and creep protocols, self-healing ability could also be evaluated through peel experiments [37]. At the end of each peel test with cohesive failure, the substrates with residual adhesives were rejoined and allowed to heal for a designated period. The peel curve before and after healing were compared to assess the self-healing performance, as illustrated in Fig. 7c and d. The peel displacement curve in Fig. 7c indicated a small but an immediate self-healing of a freshly synthesized low molecular weight sample. With prolonged healing time, full recovery of the maximum peel strength was observed in Fig. 7d.

Post-polymerization functionalization and rheology of unentangled associating polymer

In addition to the direct polymerization of functionalized monomers, post-polymerization HAA has emerged as an alternate route for synthesizing amine-containing associating polymers. This single, atom-economic synthesis method was demonstrated for amine-terminated atactic polypropylene (ATPP) (Fig. 8a) with vinyl terminated polypropylene (VTPP) and various amine substrates, including NMA, N-methylbutylamine (NMBu) and N-methylcyclohexylamine (NMCy) [43]. This method was also extended to polyolefins such as vinyl-terminated polyethylene (VTPE) [44] and to polybutadiene as shown in Fig. 8a [4].

Fig. 8.

a Schematic for general hydroaminoalkylation reaction on vinyl-terminated polypropylene and polybutadiene. Naming convention for polybutadiene follows as PB50-n, where n is the number of functional groups per chain on the Ricon 150 starting polymer (originally containing 50 1,2 polybutadiene units). b Master curves of pure and functionalized high vinyl content polybutadiene at the reference temperature of . c Van Gurp-Palmen plot for Fig. 8b. Figures reproduced from References [43, 45]

The thermal properties of amine-terminated PP [43] and PE [44] were comprehensively studied. The introduction of NMA to VTPP with an of 3000 g/mol increased the from − 15.4 to − 14 °C by reducing chain mobility through H–bonding [43]. It also raised the by 1.2 °C and lowered the by 3.7 °C in VTPP with an of 2200 g/mol by disrupting chain ordering and crystalline formation [44]. However, no significant change was observed in the rheological properties. The low molecular weight telechelic PE remained a Newtonian liquid at room temperature, showing only a slight increase in viscosity [44]. An estimate of the fraction of dimer structure, formed by H–bonding between two amine functionalized PE chains, was made by analyzing the viscosity change relative to its PE precursor. The dimer fraction decreased with the molecular weight of PE precursor, from 16% for NMA-terminated PE with an of 2200 g/mol to 3% for an of 14,000 g/mol, due to the reduced chain mobility that limits optimal chain association.

In contrast to the telechelic structures of VTPP and VTPE, where functional groups are located at the chain ends, the amines in functionalized polybutadiene are distributed along the polymer backbone, and their concentration could be systematically controlled. We investigated the rheological properties of NMBu-functionalized polybutadiene containing up to 46 amine functional groups per chain [45]. Similar to P(ACC), samples with a high density of amine functional groups exhibited long-time equilibrium behavior. At equilibrium, pseudo master curves were constructed for samples with varying numbers of stickers as illustrated in Fig. 8b [45]. Notably, failure of time-temperature superposition was observed for samples with an intermediate number of stickers (15 and 20 per chain), as evidenced by the Van Gurp-Palmen plot in Fig. 8c, suggesting a highly temperature-sensitive network structure governed by H–bonding interactions. As the number of amine groups increased, the samples underwent a typical sol-gel transition, with the gel point occurring between 5 and 15 number of stickers per chain. At 46 stickers per chain, the unentangled polymer formed a full gel network as evidenced by the emergence of a plateau modulus at Pa. This value is comparable to the entanglement plateau modulus of high molecular weight polybutadiene, demonstrating that the potential of amine associations to significantly enhance the mechanical strength of low molecular weight polymers ( = 3900 g/mol), bringing them to performance levels similar to those of well-entangled systems.

Outlook

Throughout the discussion on rheological, mechanical, and adhesive performance, we have summarized the current state-of-the-art in the emerging field of amine-containing associating polymers, which incorporate simple secondary amine functional groups either along or at the termini of traditional polyolefin backbones. The introduction of a high density of weak amine groups into non-polar polyolefins induces notable changes in material properties, particularly the emergence of long-term equilibrium behavior. Although a hypothesis involving the evolution of H–bonding architecture was proposed, direct physical evidence of these structural changes remains limited, and the precise mechanisms of aggregation are still not fully understood. These knowledge gaps present major challenges for ongoing research. To address them, new characterization techniques and theoretical models should be developed. In particular, the TMA model incorporating both weak binary associations and strong aggregations is a promising method. By using the model to reverse fit the rheological data collected during the equilibrium process, we may gain valuable insights to the evolving structural changes of the amine functionalities.

In addition to the aminated polymers discussed above and their interesting properties, the existing systems and the functionality of amines could be modified and further explored to achieve better properties or even wider applications. We herein present an overview of the future investigation directions and application opportunities of these materials.

The presence of secondary amine in these polymers provides valuable opportunities for further functionalization of the materials. Their pH responsive nature allows easy protonation, resulting in cationic surfaces that are attractive for antimicrobial applications [46]. Preliminary antibacterial tests have been performed by our group [44], showing promising inactivation of bacteria strains for VTPE. Moreover, the reactive amine groups enable post-polymerization functionalization with small molecules or polymers bearing complementary groups, such as hydroxyls, facilitating the introduction of fluorescence or enhancing compatibility with a wide range of polymer matrices [43].

Another research direction lies in the copolymerization of aminated polymers with various backbone structures and pendant groups to synergistically combine desirable properties from different polymer systems. For instance, random copolymers of cyclooctene and norbornene polymers with ethylene glycol functionalized NMA derivatives have demonstrated good performance as a binder material for MnO₂/Zn batteries [47]. In this application, high hydrophilicity is essential to promote wetting of the material and ensure durable cycling stability, which was realized by tunning the length of the ethylene glycol chains in the pedant groups. Additionally, the balance between the mechanical strength and flexibility of the copolymer was adjusted by varying the ratio of the cyclooctene to norbornene in the backbone components.

The properties of these aminated polymers can also be tailored through physical blending with other compatible polymers [48]. For example, the weak mechanical strength of low molecular weight P(ACC) can be significantly enhanced by blending it with a semi-crystalline polymer such as polycaprolactone (PCL). In comparison to higher molecular weight P(ACC), the blend offers improved performance by diminishing the time-dependent relaxation behavior inherent to the presence of high density of amines and providing superior resistance to creep deformation.

Abbreviations

ATPP

Amine-terminated atactic polypropylene

HAA

Hydroaminoalkylation

LVE

Linear viscoelastic

NMA

N-methylaniline

NMBu

N-methylbutylamine

NMCy

N-methylcyclohexylamine

PAAm

Polyallylamine

P(ACC)

Poly(aryl amine-containing cyclooctene)

P(ACN)

Poly(aryl amine-containing norbornene)

PCL

Polycaprolactone

PE

Polyethylene

PEI

Polyethylenimine

PI

Polyisoprene

PIB

Polyisobutene

PNB

Polynorbornene

PP

Poypropylene

PTFE

Polytetrafluoroethylene

PVAm

Polyvinylamine

ROMP

Ring-opening metathesis polymerization

TMA

Time marching algorithm

UHMWPE

Ultra-high molecular weight polyethylene

UPy

Ureidopyrimidinone

VTPE

Vinyl-terminated polyethylene

VTPP

Vinyl-terminated polypropylene

Author contributions

Z.Z.—conceptualization, main manuscript text, reviewing and editing. N.M.—figure preparation, reviewing and editing. A.M.—figure preparation, reviewing and editing. S.A.—conceptualization, figure preparation, reviewing and editing. E.M.—reviewing and editing, L.L.S.—project management, funding acquisition, conceptualization, reviewing, editing. S.G.H.—project management, funding acquisition, conceptualization, reviewing, editing. B.M.Y.—project management, manuscript management, conceptualization, main manuscript text, reviewing, editing.

Funding

The authors thank NSERC and UBC for financial support of this work. Additional support from MITACS has facilitated collaborative research.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.