Harnessing Imine Diversity To Tune Hyperbranched Polymer

Dec 8, 2017 - ABSTRACT: Dynamic-covalent chemistry has enabled the facile synthesis of a new generation of degradable materials, but controlling the r...
6 downloads 10 Views 1MB Size
Article Cite This: Macromolecules XXXX, XXX, XXX−XXX

Harnessing Imine Diversity To Tune Hyperbranched Polymer Degradation Michael B. Sims, Kush Y. Patel, Mallika Bhatta, Soma Mukherjee, and Brent S. Sumerlin* George & Josephine Butler Polymer Research Laboratory, Center for Macromolecular Science & Engineering, Department of Chemistry, University of Florida, PO Box 117200, Gainesville, Florida 32611-7200, United States S Supporting Information *

ABSTRACT: Dynamic-covalent chemistry has enabled the facile synthesis of a new generation of degradable materials, but controlling the rate at which these materials degrade remains elusive. Using segmented hyperbranched polymers (SHPs) as model branched architectures, we demonstrate that SHPs containing imine crosslinks degrade under acidic conditions into well-defined linear chains at rates controllable via modification of the imine N-substituent. Imine-cross-linked SHPs were synthesized in a one-pot protocol by reversible addition−fragmentation chain transfer (RAFT) copolymerization of novel divinyl compounds containing dynamiccovalent oxime, semicarbazone, and acyl hydrazone moieties. The extent of SHP branching could be controlled through the relative stoichiometric ratios of crosslinker and chain transfer agent (CTA), and studies of the polymerization kinetics confirmed the growth of polydisperse branched species at high monomer conversions. When subjected to aqueous acidic conditions, the polydisperse branched architecture degraded into well-defined polymers, a process that was accelerated under more strongly acidic conditions and by incorporating less hydrolytically stable imine cross-links. Finally, we found that the rate of SHP degradation could be tuned with an unprecedented level of control by cross-linking the polymers with different proportions of multiple imines.



INTRODUCTION Recent advances in topological polymer chemistry have enabled the synthesis of complex architectures with unprecedented precision and functionality.1−4 Highly branched, or dendritic, polymers are particularly attractive targets due to their enhanced solubility, decreased solution viscosity, and extensive end-group decoration relative to linear analogues.5 Dendritic polymers can be generally subdivided into four major classes dendrimers, dendrons, dendrigrafts, and hyperbranched polymerswith each class varying in branching regularity and synthetic accessibility.6,7 Dendrimers, featuring perfect fractallike branching, are of extensive theoretical interest but are often highly tedious to synthesize. In contrast, hyperbranched and segmented hyperbranched polymers (SHPs) are the most irregularly branched of the dendritic architectures, but they can be synthesized very efficiently, often in one-pot protocols.8−10 SHPs, in which branch points are separated by linear segments, can be rapidly synthesized by the chain-growth copolymerization of vinyl monomers with a branching agent. This agent can be either an initiator-monomer compound (“inimer”) for self-condensing vinyl polymerization (SCVP)9,11−17 or simply a divinyl cross-linker. Compared to SCVP, the use of cross-linkers (also termed the “divinyl route”) requires more careful stoichiometric control to prevent gelation, but it advantageously enables the synthesis of branched architectures from virtually any combination of vinyl monomer and divinyl compound, many of which are © XXXX American Chemical Society

commercially available. In contrast, almost all SCVP inimers require some degree of synthetic effort.10 The synthesis of SHPs via conventional radical copolymerization of divinyl cross-linkers is often challenging due to the predisposition of these systems to rapid macroscopic gelation. The copolymerization initially produces linear copolymers (“primary chains”) that contain pendent double bonds derived from cross-linkers that have polymerized through a single vinyl group. At higher conversions, crosslinking of primary chains by propagation through these pendent double bonds produces progressively more branched species and, if the quantity of pendent double bonds is poorly controlled, eventually an infinite three-dimensional network.2 Gelation was originally a prohibitive drawback to this approach, but the discovery by Sherrington and co-workers10,18−20 that mercaptan chain transfer agents stalled the formation of a gel by limiting the molecular weight of primary chains enabled the application of radical polymerization techniques to the synthesis of discrete SHPs for the first time. Sherrington and co-workers later demonstrated that SHPs could be synthesized by atom-transfer radical polymerization (ATRP),21 and Perrier et al. found that reversible addition−fragmentation chain transfer (RAFT) polymerization was equally effective.22 Received: October 31, 2017 Revised: December 8, 2017

A

DOI: 10.1021/acs.macromol.7b02323 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Herein, we report the facile synthesis of SHPs cross-linked with a variety of imine derivatives and show that the degradation rate of these polymers can be tailored with a high degree of fidelity. To this end, novel cross-linkers containing dynamic-covalent oxime, semicarbazone, and hydrazone functional groups were synthesized in a mild and modular protocol. Reversible addition−fragmentation chain transfer (RAFT) copolymerization of imine-containing crosslinkers with N,N-dimethylacrylamide yielded high molecular weight SHPs that degraded under acidic conditions at rates dependent on the pH and imine functionality. We further utilized this approach to tailor the degradation rate of these materials through the strategic incorporation of different imine derivatives as cross-links.

Additionally, the work by Perrier importantly showed that reactive end-groups are retained at the SHP periphery, enabling facile chain extension and postpolymerization modification. Since these seminal reports, SHPs synthesized by RAFT polymerization have found numerous applications, such as therapeutic delivery23 and molecular imaging24,25 as shown by Thurecht and co-workers. Facilitating the degradation of branched architectures into lower-molecular-weight species is an important consideration, especially for potential biological applications. SHP degradation into linear polymers was investigated in the first report of the divinyl route, in which ozonolysis of SHPs cross-linked with but-2-ene-1,4-diacrylate resulted in cleavage of the branched architecture into linear polymers.10 However, this approach required harsh conditions that are unsuitable for more elaborately functional polymers. Therefore, the degradation of SHPs under milder conditions has been pursued using dynamic-covalent bonds26 that are reversible under a variety of stimuli. Reports to date include the reductive cleavage of disulfide-containing polymers,27−29 thermal cleavage of polymers containing furan−maleimide adducts,30 acidic degradation of acetals,31,32 and displacement of imine cross-links with small molecule aldehydes and amines.33,34 Additionally, although not strictly considered dynamic-covalent chemistry, our group has shown that hyperbranched polymers cross-linked with sugarderived glucarodilactones can be degraded under mildly basic conditions.35 These stimuli-responsive moieties facilitate efficient degradation of branched architectures, but the rate of degradation is generally constant under a given set of conditions. Imines offer an alternative avenue to most dynamic-covalent functional groups because the rate of hydrolysis at a given pH is strongly affected by the identities of their substituents, in particular that of the N-substituent.36 Interestingly, tailoring the degradation of SHPs and other branched architectures has received little attention, despite the recent utilization of imines as dynamic linkages for postpolymerization modification and materials synthesis.34,37−45 Thus, imine chemistry presents an interesting study into the methods by which dynamic-covalent chemistry can be exploited to rationally control the degradation behavior of a material. Schiff bases (where the N-substituent is an alkyl or aryl group) are generally the least robust of the isolable imine derivatives, whereas hydrazones and oximes are significantly more resistant to hydrolysis.46 Kalia and Raines determined rate constants for the acidic hydrolysis of small molecule Schiff bases, acyl hydrazones, semicarbazones, and oximes (among other derivatives) and observed increasing hydrolytic stabilities in this order.36 In a seminal report applying the disparate stabilities of imine derivatives to a macromolecular system, Maynard and co-workers47 prepared a series of PEG-based hydrogels containing acyl hydrazone, semicarbazone, and oxime cross-links. They observed that the degradation resistance of the gels generally increased as more stable imines were incorporated. Additionally, Connal and coworkers very recently showed that PEG-derived linear polymers containing a mix of oxime and Schiff base imines within the backbone degraded over the course of several days at rates that were tunable by increasing the oxime content.48 In view of these results, we were motivated to (1) more precisely explore the effect of cross-link stability on the degradation of macromolecular architectures and (2) to develop a route to imine-containing branched architectures that could be generalized to different vinyl monomers.



RESULTS AND DISCUSSION Cross-Linker Synthesis. Cross-linkers containing oximes, semicarbazones, and acyl hydrazones were prepared through direct condensation of the appropriate diamine with diacetone acrylamide (DAA) under acidic aqueous conditions (Scheme 1). Interestingly, water was found to be the ideal solvent for the

Scheme 1. Synthesis of Divinyl Cross-Linkers Containing Oxime (Top), Semicarbazone (Middle), and Hydrazone (Bottom) Functional Groups

synthesis of these imine-containing compounds despite their inherent hydrolytic lability. All three cross-linkers are hydrophobic and precipitate from the reaction solution, driving the imine equilibrium toward the products even in the presence of large amounts of water. This approach enabled very simple purification since the products could be obtained in high yields by filtration and washing with excess neutral water. 1H and 13C nuclear magnetic resonance (NMR) spectroscopy, highresolution mass spectrometry, and elemental analysis were used to confirm a high degree of product purity and the absence of side reactions such as conjugate addition to DAA or monocondensation of the diamine with DAA. The 1H and 13C NMR spectra of all cross-linkers were in good agreement with their assigned structures (see Figure S1 for DAA2o, Figure S2 for DAA2s, and Figure S3 for DAA2h), although some peak splitting was observed for the methyl and methylene groups near the imine moieties in DAA2s and DAA2h. We presumed this was due to the presence of both E and Z isomers about the CN bond, so variable-temperature 1H NMR (Figure S4) and 2D 1H−13C NMR spectroscopy studies (Figure S5) were B

DOI: 10.1021/acs.macromol.7b02323 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

in comparison to results that may be expected from an analogous linear polymerization. We also observed that decreasing the content of DMA relative to the cross-linker yielded polymers of progressively broader molecular weight distributions. We believe this is due to the greater favorability of branching reactions at shorter primary chain lengths (as dictated by the comonomer-to-CTA ratio). Shorter chain lengths are less sterically bulky and therefore better able to participate in polymer−polymer cross-linking reactions. Additionally, the concentration of pendent double bonds is greater in shorter chains, making the reaction of these moieties more statistically likely. To confirm our hypothesis that SHP polydispersity corresponds to more highly branched polymers, we calculated degrees of branching (DB) for SHPs 1−5 from the relationship defined by Fréchet and co-workers8 in eq 1:

conducted to confirm this hypothesis. A discussion of these data is included in the Supporting Information. Preparation of Degradable Segmented Hyperbranched Polymers by RAFT Polymerization. RAFT polymerization was chosen for the synthesis of segmented hyperbranched polymers (Figure 1) due to its robust functional

DB =

B+T B+L+T

(1)

where B is the number of branching units, T is the number of terminal units, and L is the number of linear units in the branched polymer. The area under the following peaks from the 1H NMR spectrum (Figure 1) was used to calculate DB: the methyl groups of DMA (peak f) and the pendent vinyl groups of the cross-linker (peak b, c, or d) correspond to L, the O-methylene groups of the cross-linker (peak e) correspond to 2B, and the methyl group of the CTA Z-group (peak h) corresponds to T/2. The results of these calculations confirm that more highly branched polymers were formed when the comonomer content was reduced and when cross-linker content was increased (Table 1). Branching can also be promoted by increasing the crosslinker-to-CTA ratio. Although the formation of an infinite network theoretically occurs when the number of cross-links per primary chain is unity (as predicted by Flory−Stockmayer theory53), the highest cross-linker-to-CTA ratio (in RAFT polymerization) or cross-linker-to-initiator ratio (in ATRP or NMP) one can use in practice while avoiding gelation at high monomer conversion is typically greater than this due to the presence of competing intramolecular cyclization reactions.29,54 Accordingly, discrete SHPs could be obtained near full conversion at equimolar concentrations of cross-linker to CTA (SHPs 1−3) and above (SHPs 4, 5). However, we observed gelation after approximately 93% DMA conversion

Figure 1. Schematic representation of the preparation of oximecontaining segmented hyperbranched polymers (top) and 1H NMR spectrum of p(DMA-co-DAA2o) (bottom).

group tolerance and applicability to a diverse range of architectures.49−52 We initially investigated the effects of varying DMA content relative to cross-linker (SHPs 1−3), cross-linker content relative to CTA (SHPs 4, 5), and crosslinker identity (SHPs 6, 7) on the molecular weights and molecular weight distributions of SHP products (Table 1). Branched polymers were successfully prepared in all iterations, as suggested by the observation of significantly higher molecular weights and broader molecular weight distributions

Table 1. Segmented Hyperbranched Polymers Prepared via Copolymerization of DMA and Imine Cross-Linkers GPC-MALLS entry SHP SHP SHP SHP SHP SHP SHP SHP

1 2 3 4 5 6e 7 8

cross-linkera

feed ratiob

time (min)

conv (%)c

DBd

Mn (g/mol)

Mw (g/mol)

Mw/Mn

DAA2o DAA2o DAA2o DAA2o DAA2o DAA2o DAA2s DAA2h

100/1/1 50/1/1 20/1/1 100/1.25/1 50/1.25/1 150/1.50/1 100/1/1 100/1/1

180 150 120 150 150 135 180 180

97 97 98 97 97 93 98 98

0.015 0.031 0.078 0.014 0.032 − −f −

35 600 18 100 37 300 26 500 48 600 55 000 21 600 21 500

106 000 101 000 309 000 189 000 685 000 375 000 56 600 32 100

2.99 5.56 8.13 7.13 14.2 6.81 2.63 1.50

a

DAA2o = oxime cross-linker, DAA2s = semicarbazone cross-linker, and DAA2h = hydrazone cross-linker. bFeed ratio is expressed as the ratio of monomer to cross-linker to CTA. [CTA]/[AIBN] = 1.0/0.2 for all polymerizations. cConversion of DMA was determined by GC. All polymerizations were conducted at DMA concentrations of 3.0 M. dDegrees of branching (DB) were calculated according to eq 1. eSHP 6 corresponds to the final time point from the below kinetics experiment. DB was not calculated because the reaction solution gelled shortly after this aliquot was sampled. fDBs could not be calculated for SHPs 6 and 7 due to the absence of clearly resolved branching unit resonances. C

DOI: 10.1021/acs.macromol.7b02323 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 2. (A) Pseudo-first-order kinetic plot of DMA (purple markers), DAA2o (orange markers), DAA2s (green markers), and DAA2h (blue markers) conversion (cross-linker conversion data stops after 75 min due to the concentration becoming too low to detect by NMR). (B) Evolution of molecular weight (blue markers) and dispersity (orange markers) for the copolymerization of DMA with DAA2o with conversion depicting exponential growth at later stages of the polymerization that is characteristic of branching reactions (theoretical molecular weight for an analogous linear polymerization is represented by the blue dashed line). (C) Overlaid GPC traces depicting the evolution of branching with polymerization time. The monomer/cross-linker/CTA/initiator ratio of the polymerization was 150/1.5/1.0/0.2.

growth under RAFT conditions, similar to that observed with DAA2o. Pseudo-first-order consumption of DMA was observed in all three polymerizations, and the identity of the cross-linker had relatively little effect on the overall rate of polymerization (Figure S9A). The growth of number-average molecular weight also showed little variation with cross-linker identity (Figure S9B), but the growth of weight-average molecular weight for semicarbazone SHPs and, even more so, hydrazone SHPs (Figure S9C) was significantly slower than that for oxime SHPs. The mechanism of branching can be inferred by examining the evolution of molecular weight with polymerization progress. First, the plot of molecular weight versus conversion shows relatively constant initial growth that is consistent with the molecular weight of a theoretical linear polymer (Figure 2B), indicative of a polymerization process that consists predominantlybut not exclusivelyof linear chain growth. With increasing monomer conversion, branching reactions via macromolecular coupling become progressively more probable, and molecular weight growth consequently deviates more extensively from that of an analogous linear polymer. This behavior is typical of SHP synthesis by the divinyl route.28 This hypothesis is supported by the change in molecular weight distribution over the course of the polymerization (Figure 2C). Early stages of the polymerization are dominated by growth of linear chains as evidenced by the initial development of a relatively unimodal molecular weight distribution that shifts toward earlier retention times with increasing conversion. The onset of branching is visualized by the appearance of a high molecular weight shoulder after approximately 30 min that later develops into a complex and multimodal molecular weight distribution as more highly branched species are formed. Finally, the molecular weight distribution at 135 min, while becoming broader, shifts to a later retention time than that observed at 120 min. This phenomenon, previously observed by Matyjaszewski et al., has been attributed to preferential gelation of high-molecular-weight, highly branched polymers, with lower-molecular-weight chains remaining in the sol.55 While the reaction medium at this time appeared to still be a highly viscous fluid, it is nonetheless possible that gelation had commenced but was not yet macroscopically apparent. Interestingly, we observed the formation of a macroscopic gel within minutes of sampling this final time point, suggesting that the critical gelation conversion is very near 93% for this system. Degradation Kinetics of Imine-Containing SHPs. Copolymerization of cross-linkers that contain labile functional groups yields polymers with a similarly degradable branched

when a 1.5/1.0 molar feed ratio of cross-linker to CTA was used (SHP 6). Attempts to further increase this ratio resulted in even more rapid gelation. Finally, the polymerization of semicarbazone and hydrazone cross-linkers (SHPs 7, 8) yielded similar, albeit lower, molecular weight polymers to those containing oximes. The molecular weight differences may arise from at least two factors during polymerization. First, the different sizes of the crosslinkers may result in different propensities of each toward intramolecular cyclization during polymerization, consequently affording polymers of different branching densities (and therefore molecular weights) at similar conversions. Second, some imine bonds in the cross-linker may degrade at the polymerization temperature, but the rate of imine bond cleavage is not easily quantifiable, particularly at the very low cross-linker concentrations that were employed. We also examined the possibility that the cross-linkers undergo propagation at inherently different rates, but this is not conclusively supported by kinetics data (vide inf ra). RAFT Polymerization Kinetics of Imine-Containing Segmented Hyperbranched Polymers. Kinetic studies of the polymerization were undertaken to study the process of SHP formation. For the copolymerization of DMA with DAA2o, the linear pseudo-first-order kinetic plot (Figure 2A) depicts a relatively constant rate of monomer consumption throughout polymerization even up to high conversions, indicating a constant radical concentration and limited number of termination events associated with this novel cross-linker. Furthermore, a comparison of the relative rates of DMA and DAA2o conversion suggests that each species reacts similarly, as would be expected considering the electronic similarity of both species’ vinyl groups. We also studied the rate of DAA2s and DAA2h incorporation under identical conditions to determine if the aforementioned cross-linker-dependent difference in SHP molecular weights was due to variances in polymerization kinetics. As shown in Figure 2A, DAA2s and DAA2h appear to be incorporated marginally slower than DAA2o. However, the concentration of cross-linker in these polymerizations is quite low (approximately 0.03 M), and this disparity may be due to error from integrating the very weak cross-linker NMR signal. Furthermore, the apparent trend in reactivity (DAA2o > DAA2h > DAA2s) does not agree with the trend in observed molecular weights, so we are reticent to draw significant conclusions from this data. Nevertheless, the kinetics of DMA conversion and growth of molecular weight for copolymerizations with DAA2s (Figure S7) and DAA2h (Figure S8) indicate typical SHP D

DOI: 10.1021/acs.macromol.7b02323 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules architecture. SHPs are ideal topologies for quantifying the kinetics of degradation, as they are fully soluble and therefore amenable to GPC characterization. Imine hydrolysis is promoted in acidic conditions (Scheme 2),46 but it has been Scheme 2. Degradation of Imine-Containing SHPs into Linear Polymers under Acidic Conditions

Figure 3. Kinetic plots of the degradation of semicarbazone-containing segmented hyperbranched polymers at pH values between pH 1.0 and 10. The degradation rate remained relatively constant as the pH decreased from pH 10 to pH 4.0, but further acidification resulted in significant acceleration of cross-link hydrolysis.

this interesting, as one would expect the transition from mildly acidic to basic pH to be accompanied by a sharp reduction in the rate of SHP hydrolysis. Yet, this again points to the criticality of the imine protonation step in the overall hydrolysis reaction. While we have not found a definitive literature value for the pKa of protontated semicarbazone, it can be reasonably inferred by considering that semicarbazones are intermediate in stability to oximes (pKa = −0.55−0.80)56 and alkylhydrazones (pKa = 5.8);36 therefore, one would also expect an intermediate pKa. Furthermore, the observation of a substantial change in degradation behavior between pH 3.0 and pH 4.0 also suggests that the semicarbazone pKa lies in this region. Therefore, any further basification beyond pH 4.0 or pH 5.0 would not greatly change the concentration of protonated iminium species and therefore would not significantly decelerate SHP degradation. Our final objective was to investigate the tunability of the SHP degradation rate by cross-linking the polymers with imines of different hydrolytic stabilities. Oximes are the most hydrolytically stable of the studied imine derivatives and should therefore yield the most robust materials, whereas semicarbazones and acyl hydrazones are less stable and should afford materials with enhanced degradability. Furthermore, we hypothesized that the SHP degradation rate could be tuned with even greater precision by incorporating various ratios of different imines (e.g., an SHP composed of one-half hydrazone cross-links and one-half semicarbazone cross-links). To this end, we prepared a series of six SHPs containing oxime, semicarbazone, and hydrazone cross-links in various proportions (Table 2). We next subjected each of these polymers to pH 4.0 acetate buffer and monitored the reduction in molecular weight by GPC-MALLS. SHPs containing oxime (SHP 8), semicarbazone (SHP 11), and hydrazone (SHP 13) linkages exhibited significantly disparate degradation rates (Figure 4). When SHPs 9, 10, and 12 (i.e., SHPs containing mixed imine compositions) were degraded under similar conditions, we found that the degradation rate of these mixed SHPs was proportional to the quantity of the incorporated imines. For example, the curves corresponding to the degradation of SHP 9 and SHP 10 (containing mixed ratios of oximes and semicarbazones) are between those of SHP 8 and SHP 11 approximately to the extent that oxime cross-links were replaced with semicarbazone cross-links. This demonstrates that the degradation rate of imine-containing materials can be precisely tuned with a remarkable degree of precision by only changing the stability of the cross-links. We emphasize that

shown from hydrolysis studies on small molecule imines that the addition of a scavenging agent to consume the liberated amine is necessary to promote complete degradation.36 Furthermore, the identity of the scavenging agent was found to have an effect on the apparent rate of hydrolysis in the referenced work. In view of this observation, we subjected semicarbazone-cross-linked SHPs to pH 4.0 acetate buffer in the presence of formaldehyde, acetone, and furfural to determine if a similar effect was observable for our system (Figure S10). Percent degradation was calculated according to eq 2, defined to account for the fact that there is residual molecular weight from linear chains even at 100% degradation: ⎛ M n − M n,PC ⎞ ⎟⎟ × 100 % degradation = ⎜⎜1 − M n,0 − M n,PC ⎠ ⎝

(2)

where Mn is the number-average molecular weight at a given time, Mn,0 is the initial SHP number-average molecular weight, and Mn,PC is the theoretical number-average molecular weight of the primary chains composing the SHP. Interestingly, the identity of the scavenger was found to have little, if any, effect on the apparent rate of SHP degradation. One could imagine several plausible explanations for this observation, one of which is that the imine equilibrium strongly favors dissociation in dilute solution due to the unfavorability of two polymeric species reacting with each other to re-form a cross-link. Having established that the identity of the scavenging agent does not affect the apparent degradation rate, we next sought to characterize the pH-dependent degradation kinetics of semicarbazone-containing SHPs. Formation of a protonated iminium species is a critical step in the hydrolysis of imines,36 so we reasoned that increasingly acidic conditions would result in faster degradation rates. A range of pH values between 1.0 and 10.0 were studied, and the rate of SHP degradation was again determined by monitoring the reduction in SHP molecular weight (Figure 3). Predictably, SHPs were observed to degrade most rapidly under the most acidic conditions studied (i.e., pH 1.0 hydrochloric acid/sodium chloride solution). Increasing the pH from 1.0 to 3.0 resulted in a marginal reduction in the degradation rate, but a substantially slower degradation was observed as the acidity was reduced from pH 3.0 to pH 4.0. Finally, the observed degradation rate changed little as the solution pH was changed from weakly acidic (pH 4.0) to neutral (pH 7.0) to basic (pH 10). We found E

DOI: 10.1021/acs.macromol.7b02323 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Table 2. Segmented Hyperbranched Polymers Prepared for Use in Degradation Studies GPC-MALLS entry SHP SHP SHP SHP SHP SHP

8 9 10 11 12 13

feed ratioa

cross-linker(s)b

time (h)

conv (%)c

Mn (g/mol)

Mw (g/mol)

Mw/Mn

100/1.25/1 100/1.25/1 100/1.10/1 100/1.25/1 100/1.25/1 100/1.10/1

O 2 /3 O/1/3 S 1 /3 O/2/3 S S 1 /2 S/1/2 H H

2.5 4 4 4 4 4

97 99 98 98 97 98

59 000 60 800 59 200 66 700 46 200 35 300

241 000 438 000 421 000 442 000 225 000 460 000

4.09 7.21 7.12 6.41 4.87 13.0

a Feed ratio expressed as [M]/[XL]/[CTA]. bO = oxime cross-linker, S = semicarbazone cross-linker, H = hydrazone cross-linker. cConversion of DMA as determined by GC. All polymerizations were conducted at DMA concentrations of 3.0 M.

Figure 4. Plots depicting the degradation kinetics of segmented hyperbranched polymers cross-linked with different imines in pH 4.0 acetate buffer. The degradation of polymers cross-linked entirely with oxime linkages (red markers/trace) was found to be the slowest of the studied polymers, whereas polymers cross-linked entirely with hydrazones degraded at the fastest rate (purple markers/trace). Intermediate degradation rates were observed by cross-linking the polymers with imines of different hydrolytic stabilities at defined ratios.

changing the identity and quantity of cross-links in an SHP using this chemistry is very facile, as it is accomplished by changing the reaction stoichiometry during SHP synthesis.





CONCLUSIONS Divinyl cross-linkers containing oxime, semicarbazone, and hydrazone linkages were synthesized and used to prepare segmented hyperbranched polymers with tailorable hydrolysis rates. Importantly, this chemistry can easily be applied to the synthesis of other polymeric architectures. The polymerization kinetics of each cross-linker was characterized, and it was shown that higher molecular weight SHPs with greater degrees of branching could be synthesized by increasing the cross-linker content of primary chains. Finally, the degradation rate of imine-containing SHPs was found to increase both with solution acidity and as more labile imines were incorporated into the polymer. While not studied herein, one could envision that a judicious combination of pH gradients and imine composition could even enable control over the exact profile of the degradation curvea highly useful ability if outcomes that require precise degradation behaviors are desired.



Materials, experimental procedures, spectroscopic characterization data, and GPC characterization data (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected]fl.edu; Fax +1 352 392 9741 (B.S.S.). ORCID

Brent S. Sumerlin: 0000-0001-5749-5444 Funding

This material is based upon work supported by the National Science Foundation (DMR-1606410) (B.S.S.) and the National Science Foundation Graduate Research Fellowship (DGE1315138) (M.B.S.). Notes

The authors declare no competing financial interest.



REFERENCES

(1) Fréchet, J. M. J. Dendrimers and other dendritic macromolecules: From building blocks to functional assemblies in nanoscience and nanotechnology. J. Polym. Sci., Part A: Polym. Chem. 2003, 41, 3713− 3725. (2) Gao, H.; Matyjaszewski, K. Synthesis of functional polymers with controlled architecture by CRP of monomers in the presence of crosslinkers: From stars to gels. Prog. Polym. Sci. 2009, 34, 317−350. (3) Konkolewicz, D.; Monteiro, M. J.; Perrier, S. Dendritic and Hyperbranched Polymers from Macromolecular Units: Elegant Approaches to the Synthesis of Functional Polymers. Macromolecules 2011, 44, 7067−7087.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.7b02323. F

DOI: 10.1021/acs.macromol.7b02323 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (4) Zheng, Y.; Li, S.; Weng, Z.; Gao, C. Hyperbranched polymers: advances from synthesis to applications. Chem. Soc. Rev. 2015, 44, 4091−4130. (5) Inoue, K. Functional dendrimers, hyperbranched and star polymers. Prog. Polym. Sci. 2000, 25, 453−571. (6) Fréchet, J. M. J.; Tomalia, D. A. Dendrimers and Other Dendritic Polymers; Wiley: West Sussex, 2001. (7) Tomalia, D. A.; Fréchet, J. M. J. Discovery of dendrimers and dendritic polymers: A brief historical perspective. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, 2719−2728. (8) Hawker, C. J.; Lee, R.; Fréchet, J. M. J. One-Step Synthesis of Hyperbranched Dendritic Polyesters. J. Am. Chem. Soc. 1991, 113, 4583−4588. (9) Fréchet, J. M. J.; Henmi, M.; Gitsov, I.; Aoshima, S.; Leduc, M. R.; Grubbs, R. B. Self-Condensing Vinyl Polymerization: An Approach to Dendritic Materials. Science 1995, 269, 1080−1083. (10) O’Brien, N.; McKee, A.; Sherrington, D. C.; Slark, A. T.; Titterton, A. Facile, versatile and cost effective route to branched vinyl polymers. Polymer 2000, 41, 6027−6031. (11) Hawker, C. J.; Frechet, J. M. J.; Grubbs, R. B.; Dao, J. Preparation of Hyperbranched and Star Polymers by a “Living”, SelfCondensing Free Radical Polymerization. J. Am. Chem. Soc. 1995, 117, 10763−10764. (12) Gaynor, S. G.; Edelman, S.; Matyjaszewski, K. Synthesis of Branched and Hyperbranched Polystyrenes. Macromolecules 1996, 29, 1079−1081. (13) Vogt, A. P.; Gondi, S. R.; Sumerlin, B. S. Hyperbranched Polymers via RAFT Copolymerization of an Acryloyl Trithiocarbonate. Aust. J. Chem. 2007, 60, 396−399. (14) Vogt, A. P.; Sumerlin, B. S. Tuning the Temperature Response of Branched Poly(N-isopropylacrylamide) Prepared by RAFT Polymerization. Macromolecules 2008, 41, 7368−7373. (15) Alfurhood, J. A.; Bachler, P. R.; Sumerlin, B. S. Hyperbranched polymers via RAFT self-condensing vinyl polymerization. Polym. Chem. 2016, 7, 3361−3369. (16) Bachler, P. R.; Forry, K. E.; Sparks, C. A.; Schulz, M. D.; Wagener, K. B.; Sumerlin, B. S. Modular segmented hyperbranched copolymers. Polym. Chem. 2016, 7, 4155−4159. (17) Alfurhood, J. A.; Sun, H.; Bachler, P. R.; Sumerlin, B. S. Hyperbranched poly(N-(2-hydroxypropyl) methacrylamide) via RAFT self-condensing vinyl polymerization. Polym. Chem. 2016, 7, 2099−2104. (18) Costello, P. A.; Martin, I. K.; Slark, A. T.; Sherrington, D. C.; Titterton, A. Branched methacrylate copolymers from multifunctional monomers: chemical composition and physical architecture distributions. Polymer 2002, 43, 245−254. (19) Isaure, F.; Cormack, P. A. G.; Sherrington, D. C. Facile synthesis of branched poly(methyl methacrylate)s. J. Mater. Chem. 2003, 13, 2701−2710. (20) Isaure, F.; Cormack, P. A. G.; Sherrington, D. C. Synthesis of Branched Poly(methyl methacrylate)s: Effect of the Branching Comonomer Structure. Macromolecules 2004, 37, 2096−2105. (21) Isaure, F.; Cormack, P. A. G.; Graham, S.; Sherrington, D. C.; Armes, S. P.; Butun, V. Synthesis of branched poly(methyl methacrylate)s via controlled/living polymerisations exploiting ethylene glycol dimethacrylate as branching agent. Chem. Commun. 2004, 1138−1139. (22) Liu, B.; Kazlauciunas, A.; Guthrie, J. T.; Perrier, S. One-Pot Hyperbranched Polymer Synthesis Mediated by Reversible Addition Fragmentation Chain Transfer (RAFT) Polymerization. Macromolecules 2005, 38, 2131−2136. (23) Coles, D. J.; Rolfe, B. E.; Boase, N. R. B.; Veedu, R. N.; Thurecht, K. J. Aptamer-targeted hyperbranched polymers: towards greater specificity for tumours in vivo. Chem. Commun. 2013, 49, 3836−3838. (24) Munnemann, K.; Kolzer, M.; Blakey, I.; Whittaker, A. K.; Thurecht, K. J. Hyperbranched polymers for molecular imaging: designing polymers for parahydrogen induced polarisation (PHIP). Chem. Commun. 2012, 48, 1583−1585.

(25) Rolfe, B. E.; Blakey, I.; Squires, O.; Peng, H.; Boase, N. R. B.; Alexander, C.; Parsons, P. G.; Boyle, G. M.; Whittaker, A. K.; Thurecht, K. J. Multimodal Polymer Nanoparticles with Combined 19F Magnetic Resonance and Optical Detection for Tunable, Targeted, Multimodal Imaging in Vivo. J. Am. Chem. Soc. 2014, 136, 2413−2419. (26) Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K. M.; Stoddart, J. F. Dynamic Covalent Chemistry. Angew. Chem., Int. Ed. 2002, 41, 898−952. (27) Li, Y.; Armes, S. P. Synthesis and Chemical Degradation of Branched Vinyl Polymers Prepared via ATRP: Use of a Cleavable Disulfide-Based Branching Agent. Macromolecules 2005, 38, 8155− 8162. (28) Pal, S.; Hill, M. R.; Sumerlin, B. S. Doubly-responsive hyperbranched polymers and core-crosslinked star polymers with tunable reversibility. Polym. Chem. 2015, 6, 7871−7880. (29) Liang, S.; Li, X.; Wang, W.-J.; Li, B.-G.; Zhu, S. Toward Understanding of Branching in RAFT Copolymerization of Methyl Methacrylate through a Cleavable Dimethacrylate. Macromolecules 2016, 49, 752−759. (30) Sun, H.; Kabb, C. P.; Sumerlin, B. S. Thermally-labile segmented hyperbranched copolymers: using reversible-covalent chemistry to investigate the mechanism of self-condensing vinyl copolymerization. Chem. Sci. 2014, 5, 4646−4655. (31) Zou, L.; Shi, Y.; Cao, X.; Gan, W.; Wang, X.; Graff, R. W.; Hu, D.; Gao, H. Synthesis of acid-degradable hyperbranched polymers by chain-growth CuAAC polymerization of an AB3 monomer. Polym. Chem. 2016, 7, 5512−5517. (32) Amato, D. N.; Amato, D. V.; Mavrodi, O. V.; Martin, W. B.; Swilley, S. N.; Parsons, K. H.; Mavrodi, D. V.; Patton, D. L. ProAntimicrobial Networks via Degradable Acetals (PANDAs) Using Thiol−Ene Photopolymerization. ACS Macro Lett. 2017, 6, 171−175. (33) Jackson, A. W.; Stakes, C.; Fulton, D. A. The formation of core cross-linked star polymer and nanogel assemblies facilitated by the formation of dynamic covalent imine bonds. Polym. Chem. 2011, 2, 2500−2511. (34) Mukherjee, S.; Bapat, A. P.; Hill, M. R.; Sumerlin, B. S. Oximes as reversible links in polymer chemistry: dynamic macromolecular stars. Polym. Chem. 2014, 5, 6923−6931. (35) Pal, S.; Brooks, W. L. A.; Dobbins, D. J.; Sumerlin, B. S. Employing a Sugar-Derived Dimethacrylate to Evaluate Controlled Branch Growth during Polymerization with Multiolefinic Compounds. Macromolecules 2016, 49, 9396−9405. (36) Kalia, J.; Raines, R. T. Hydrolytic Stability of Hydrazones and Oximes. Angew. Chem., Int. Ed. 2008, 47, 7523−7526. (37) Fukuda, K.; Shimoda, M.; Sukegawa, M.; Nobori, T.; Lehn, J.-M. Doubly degradable dynamers: dynamic covalent polymers based on reversible imine connections and biodegradable polyester units. Green Chem. 2012, 14, 2907−2911. (38) Du, Z. X.; Xiang, S. N.; Zang, Y.; Zhou, Y.; Wang, C. D.; Tang, H. L.; Jin, T.; Zhang, X. L. Polyspermine imine, a pH Responsive Polycationic siRNA Carrier Degradable to Endogenous Metabolites. Mol. Pharmaceutics 2014, 11, 3300−3306. (39) Wang, J. J.; Zhao, D. P.; Wang, Y. N.; Wu, G. L. Imine bond cross-linked poly(ethylene glycol)-block-poly(aspartamide) complex micelle as a carrier to deliver anticancer drugs. RSC Adv. 2014, 4, 11244−11250. (40) Binauld, S.; Stenzel, M. H. Acid-degradable polymers for drug delivery: a decade of innovation. Chem. Commun. 2013, 49, 2082− 2102. (41) Grover, G. N.; Lam, J.; Nguyen, T. H.; Segura, T.; Maynard, H. D. Biocompatible Hydrogels by Oxime Click Chemistry. Biomacromolecules 2012, 13, 3013−3017. (42) Lee, K. Y.; Bouhadir, K. H.; Mooney, D. J. Degradation behavior of covalently cross-linked poly(aldehyde guluronate) hydrogels. Macromolecules 2000, 33, 97−101. (43) Lehn, J.-M. Dynamic Combinatorial Chemistry and Virtual Combinatorial Libraries. Chem. - Eur. J. 1999, 5, 2455−2463. G

DOI: 10.1021/acs.macromol.7b02323 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (44) Blasco, E.; Sims, M. B.; Goldmann, A. S.; Sumerlin, B. S.; Barner-Kowollik, C. 50th Anniversary Perspective: Polymer Functionalization. Macromolecules 2017, 50, 5215−5252. (45) Collins, J.; Xiao, Z.; Mullner, M.; Connal, L. A. The emergence of oxime click chemistry and its utility in polymer science. Polym. Chem. 2016, 7, 3812−3826. (46) Carey, F. A.; Sundberg, R. J. Advanced Organic Chemistry, Part A: Structure and Mechanisms, 5th ed.; Springer: New York, 2007; pp 645− 653. (47) Boehnke, N.; Cam, C.; Bat, E.; Segura, T.; Maynard, H. D. Imine Hydrogels with Tunable Degradability for Tissue Engineering. Biomacromolecules 2015, 16, 2101−2108. (48) Collins, J.; Xiao, Z.; Connal, L. A. Tunable degradation of polyethylene glycol-like polymers based on imine and oxime bonds. J. Polym. Sci., Part A: Polym. Chem. 2017, 55, 3826−3831. (49) Moad, G.; Rizzardo, E.; Thang, S. H. Living Radical Polymerization by the RAFT Process. Aust. J. Chem. 2005, 58, 379− 410. (50) Moad, G.; Rizzardo, E.; Thang, S. H. Living Radical Polymerization by the RAFT Process − A Second Update. Aust. J. Chem. 2009, 62, 1402−1472. (51) Moad, G.; Rizzardo, E.; Thang, S. H. Living Radical Polymerization by the RAFT Process − A Third Update. Aust. J. Chem. 2012, 65, 985−1076. (52) Hill, M. R.; Carmean, R. N.; Sumerlin, B. S. Expanding the Scope of RAFT Polymerization: Recent Advances and New Horizons. Macromolecules 2015, 48, 5459−5469. (53) Flory, P. Principles of Polymer Chemistry; Cornell University Press: Ithaca, NY, 1953. (54) Landin, D. T.; Macosko, C. W. Cyclization and reduced reactivity of pendant vinyls during the copolymerization of methyl methacrylate and ethylene glycol dimethacrylate. Macromolecules 1988, 21, 846−851. (55) Gao, H.; Min, K.; Matyjaszewski, K. Determination of Gel Point during Atom Transfer Radical Copolymerization with Cross-Linker. Macromolecules 2007, 40, 7763−7770. (56) O’Ferrall, R. A. M.; O’Brien, D. Rate and equilibrium constants for hydrolysis and isomerization of (E)- and (Z)-p-methoxybenzaldehyde oximes. J. Phys. Org. Chem. 2004, 17, 631−640.

H

DOI: 10.1021/acs.macromol.7b02323 Macromolecules XXXX, XXX, XXX−XXX