*#align(center)[#smallcaps[@highly-crosslinked-ultra-hard-networks-from-polymer-combs[Chapter]]]* = #smallcaps[Highly crosslinked, ultra-hard networks from polymer combs] #align(center)[Publication in preparation] == Abstract Comb polymers are branched macromolecules used as compatibilizers, surface modifiers, and, when polymerized, as strong networks. Comb polymers are typically synthesized to contain different backbone and sidechain chemistries, which can result in multifunctional reactivity and strong phase separation from different polymer domains. Here, we explored designing comb polymers that could result in high modulus, tough networks by synthesizing polymers with glycidyl methacrylate sidechains capable of participating in dense crosslinking reactions. We synthesized several comb polymers containing homopolymer, block, or random copolymer sidechains, and compared their mechanical properties and network structures to their linear counterparts. Comb polymers with random copolymer side chains had superior toughness, even compared to their linear counterparts. I hypothesize that statistical distribution of monomers along the grafts break up crosslinking sites, disfavoring intramolecular crosslinks and favor intermolecular crosslinks. I expect these polymers to be useful for multi-functional coatings and as adhesives with high modulus and toughness. == Introduction Comb polymers are macromolecules with polymers grafted onto a main polymer backbone. They are increasingly popular for their low viscosity,@manson1958 ease of side-chain functionalization, quick gelation times,@reynolds2020 complex phase behavior,@jiang2009 and large parameter design space.@zhang2016 Comb polymers have applications ranging from next-generation compatibilizers between immiscible polymer phases,@dong2015 tough networks and adhesives,@zhang2016a@ohnsorg2024 anionic exchange resins,@zhu2016 viscosity modifiers,@2009 tough resins with crystalline domains.@claesson2004@lorenzana2024 The low viscosity of these polymers and polymer solutions offers the additional benefit of requiring less energy to process.@kong2021 Previously, our group has created mechanosensitive networks from comb copolymers with reactive monomers along the main backbone.@lorenzana2024 Using poly(ethylene glycol) (PEG) grafts as molecular shields blocking the reactive sites on the main backbone, I produced crosslinked materials upon application of ultrasound. In that system, the material mass was dominated by inert PEG side chains. This resulted in low overall density of crosslinks, limiting the maximum modulus of the polymers. To address this limitation, here I have developed a suite of reactive comb polymers, where each sidechain is itself reactive and crosslinkable, significantly increasing the potential for new crosslinks and high modulus networks. I hypothesized that GMA containing grafts would allow for an increase in crosslinking density while still providing ultrasound activated crosslinking. This work was focused on the mechanical properties of these graft polymers. Most comb polymers used to create networks have crosslinking sites along the main polymer backbone, while the grafts do not contribute to the modulus. In contrast, I synthesized a well-defined methacrylic polyinitiator (pBIEM), where each monomer on the chain contains a pendent alkyl bromine capable of initiating atom transfer radical polymerization (ATRP). Homo- and co-polymers of GMA and BMA were then grown off the backbone using a \"grafting from\" approach to prepare densely grafted comb polymers. I then crosslinked the polymer combs to form networks with elastic moduli of 1 GPa and with toughnesses of 500kPa. I anticipate that this work will provide strategies for preparing high modulus, tough comb polymer coatings and adhesives. == Materials and methods === Chemical sourcing Materials were purchased from Sigma-Aldrich unless otherwise mentioned. 2-Hydroxyethyl methacrylate (HEMA, 98%), $alpha$-bromoisobutyryl bromide (99%), pyridine (anhydrous, 99%), dichloromethane (DCM, anhydrous, 99%), dioxane (99%), CuBr#sub[2] (99%), 1,5,7-triazabicyclo\[4.4.0\]dec-5-ene (TBD, 98%), ethyl $alpha$-bromophenylacetate (EBPA, 97%), N,N,N',N”,N”-pentamethyldiethylenetriamine (PMDETA, 99%), 1,1,3,3-tetramethylguanidine (TMG, 99%), and azobisisobutyronitrile (AIBN, 99%), were used as received. Glycidyl methacrylate (GMA, 99%), butyl methacrylate (BMA, 99%) were passed through a column of basic alumina to remove inhibitors. CuBr (99.9%) was purified by stirring in glacial acetic acid. Tris(2-pyridylmethyl)amine (TPMA, 98%), tris 2-(dimethylamino)ethyl amine (Me6TREN, 98%), cyclopentylmethyl ether (CPME, 99%), and 2-cyano-2-propyl dodecyl trithiocarbonate (CPDT, 97%) was purchased from TCI (Tokyo, Japan) and used as received. Methanol (99%), toluene (99%), dimethylformamide (DMF, 99%), isopropanol (99%), calcium chloride (99%) diethyl ether (anhydrous, 99%), tetrahydrofuran (THF, 99%), HCl (concentrated), reduced iron powder, copper turnings, basic alumina, and neutral alumina were purchased from Thermo Fisher (Waltham, MA) and used as received. === 2-(2-Bromoisobutyryloxy)ethyl methacrylate (BIEM) synthesis BIEM was synthesized according to a previously published protocol.@graff2015 Briefly, to a 250 mL three-necked round bottom flask (RBF), 50 mL of anhydrous DCM, 10 mL of HEMA (0.1 mol), and 10 mL of anhydrous pyridine (0.12 mol), and a magnetic stir bar are added. The RBF is then lowered into an ice bath and stirred for ten minutes. Meanwhile, a drying tube filled with calcium chloride, an addition funnel attached to a bubbler filled with silicone oil, and a glass adaptor attached to a nitrogen line are attached to each neck of the flask. 7 mL of $alpha$-bromoisobutyryl bromide (0.1 mol) is dissolved into 20 mL of anhydrous DCM and added to the addition funnel. The $alpha$-bromoisobutyryl bromide solution is added dropwise to the RBF over 15 minutes, after which the reaction is allowed to stir in the ice bath for a further 45 minutes, followed by removal from the ice bath. After stirring at room temperature for two hours, the crude product is filtered through cotton to remove pyridine salts and then concentrated under vacuum. The product is purified by flash chromatography over silica gel with chloroform, yielding a clear colorless oil after evaporation of the chloroform. === Copper activation and stirbar cleaning Copper turnings were activated by tying around a magnetic stir bar followed by immersion in conc. HCl for 15 minutes to remove copper oxides followed by rinsing with methanol. Stir bars were cleaned of residual metals by submerging in concentrated HCl overnight, followed by washing with DI water and ethanol. === CuBr purification To a 500 mL beaker was added 250 mL of glacial acetic acid followed by 25 g of CuBr and a magnetic stirbar. The beaker was covered with parafilm and allowed to stir vigorously for two days. After two days, the glacial acetic acid had turned blue as CuBr#sub[2] is dissolved by the acetic acid. The insoluble CuBr was collected by vacuum filtration, and the cake washed three times with -20 °C methanol, then three times with -20 °C diethyl ether. The now damp white powder was collected into a 50 mL RBF and dried under vacuum overnight. === PolyGMA synthesis We synthesized polyGMA according to a previously published protocol.@parkatzidis2023 Briefly, to a 20 mL scintillation vial with a magnetic stirbar, 2.84 g GMA (0.02 mol), 10 μL PMDETA (4.8e#super[-5] mol), 0.001 g CuBr (8e#super[-6] mol), 2.84 mL of DMSO and 70 μL of EBPA (4e#super[-4] mol) is added. The reaction was sealed with a rubber septum stirred at 60 °C overnight, after which it was diluted with THF, passed through a short plug of alumina, and precipitated into cold methanol. === pGMA_-b-_BMA and pGMA_-co-_BMA synthesis We synthesized linear copolymers by adapting a previously published protocol.@santos2019 Briefly, to a 20 mL scintillation vial 29 µL TMG (2.3e#super[-4] mol), 0.005 g CuBr#sub[2] (2.3e#super[-5] mol), and 1.65 g of BMA (0.011 mol) is added. For the block polymer, 1.65 g of pGMA (0.011 mol) and 0.625 mL of DMSO is added. For the random copolymer, 1.65 g of GMA (0.011 mol), and 1.25 mL of DMSO is added. The vial is capped with a rubber septum and nitrogen is bubbled through the solution for 15 minutes, after which a magnetic stirbar with an activated copper turning tied around it is added. The reaction is stirred overnight at 30 °C, after which it was diluted in THF, passed through a short plug of alumina, and precipitated into cold methanol. Spectra for all linear polymers are shown in @fig:linearcopolymer and @fig:linearblock. #figure(image("Images/C3S1.png", width: 100.0%), placement: auto, caption: [ #super[1]H NMR and GPC of pGMA_-co-_BMA. Spectra recorded at 500 MHz in CDCl#sub[3]. THF used as eluent. ] ) #figure(image("Images/C3S2.png", width: 100.0%), placement: auto, caption: [ #super[1]H NMR and GPC of pGMA and pGMA_-b-_BMA. Spectra recorded at 500 MHz in CDCl#sub[3]. THF used as eluent. ] ) === PolyBIEM (pBIEM) synthesis We synthesized polyBIEM by RAFT polymerization. To a 20 mL scintillation vial 5.56 g of BIEM (0.02 mol), 0.014 g CPDT (0.04 mmol), 0.0013 g AIBN (0.008 mmol), 5.56 mL dioxane, and a new magnetic stir bar is added. The vial is capped with a new rubber septum and degassed for 15 minutes by sparging with nitrogen using new needles. After degassing, the vial is added to an aluminum reaction block thermostatted at 60 °C to react overnight. The next day, the reaction was diluted with dioxane, precipitated into cold diethyl ether, and dried under vacuum at 0.001 mbar overnight. === Comb polymer synthesis The synthesis procedure for comb polymers follow similar reaction setup. Reagents and detailed conditions are described after the general procedure is outlined and in @combconditions. Synthesis was undertaken according to an adapted protocol for synthesizing polyGMA@catalao2015 _via_ SARA ATRP. For each reaction, reagents were prepared in a 20 mL scintillation vial. The vial was sealed with a rubber septum and after reagent dissolution, the vial was degassed by sparging with nitrogen for 15 min, after which either a stir bar with activated copper turnings wrapped around it or 0.011 g iron powder was added to the vial (@combconditions). The vial was added to an aluminum reaction block at varying temperatures for one hour, until the reaction had increased in viscosity. The crude polymer was diluted in THF and passed through a short plug of neutral alumina to remove dissolved copper salts, precipitated in -20 °C methanol, and dried under vacuum at 0.001 mbar for one hour. The polymer was used immediately after drying. PolyBIEM_-g-_(GMA_-co-_BMA): 2.8 g GMA (0.02 mol), 0.064 g pBIEM (4e#super[-4] mol), 0.061 g TPMA (2e#super[-4] mol), 0.0045 g CuBr#sub[2] (2e#super[-5] mol), and 2.8 mL of 70/30 v/v toluene/DMF. polyBIEM_-g-_(GMA_-co-_BMA): 2.8 g BMA (0.02 mol), 2.8 g GMA (0.02 mol), 0.064 g pBIEM (4e#super[-4] mol), 0.061 g TPMA (2e#super[-4] mol), 0.0045 g CuBr#sub[2] (2e#super[-5] mol), and 5.6 mL of 70/30 v/v toluene/DMF. PolyBIEM_-g-_(GMA_-b-_BMA): 2 mL of BMA (0.014 mol), 1 g of polyBIEM_-g-_GMA (0.007 mol), 0.0045 g CuBr2 (2e#super[-5] mol), 64 µL Me6TREN (2e#super[-4] mol), and 4 mL of 70/30 v/v toluene/DMF. PolyBIEM_-g-_BMA: 2.8 mL of BMA (0.02 mol), 0.064 g of polyBIEM (0.007 mol), 0.0045 g CuBr#sub[2] (2e#super[-5] mol), 64 μL Me6TREN (2e#super[-4] mol), and 2.8 mL of 70/30 v/v toluene/DMF. PolyBIEM_-g-_(BMA_-b-_GMA): 2 g BMA (0.014 mol), 1 g pBIEM_-g-_BMA (0.007 mol), 0.061 g TPMA (2e#super[-4] mol), 0.0045 g CuBr (2e#super[-5] mol), and 2.8 mL of 70/30 v/v toluene/DMF. #figure( placement: auto, align(center)[#table( columns: 4, align: (center,right,right,right,), table.header([Comb Polymer], [Reducing agent], [Temperature (°C)], [Methanol acidified?],), table.hline(), [PolyBIEM_-g-_GMA], [Iron], [30], [No], [polyBIEM_-g-_(GMA_-b-_BMA)], [Copper], [30], [Yes], [PolyBIEM_-g-_(BMA_-b-_GMA)], [Copper], [50], [Yes], [PolyBIEM_-g-_BMA], [Copper], [50], [Yes], [PolyBIEM_-g-_(GMA_-co-_BMA)], [Iron], [30], [Yes], )] , caption: [Synthetic setup for comb polymerizations.] , kind: table ) #figure( placement: top, align(center)[#table( columns: 4, align: (center,right,right,right,), table.header([Name], [GMA (%)], [Mn (g/mol)], [Đ],), table.hline(), [pGMA], [100], [5,907], [1.12], [pGMA_-b-_BMA], [45], [11,436], [1.12], [pGMA_-co-_BMA], [50], [11,942], [1.09], [pBIEM], [---], [57,641], [1.3], [pBIEM_-g-_GMA], [100], [275,294], [1.17], [pBIEM_-g-_BMA], [---], [415,543], [1.36], [pBIEM_-g-_(GMA_-b-_BMA)], [55], [491,777], [1.24], [pBIEM_-g-_(BMA_-b-_GMA)], [45], [551,826], [1.31], [pBIEM_-g-_(GMA_-co-_BMA)], [51], [423,522], [1.3], )] , caption: [Tabulated properties of polymers used in this study.] , kind: table ) === Polymer characterization Polymer molecular weight was determined by GPC. GPC was conducted on an Agilent 1260 with a PL gel 5 μm guard column and three 5 μm analytical mixed C columns (Agilent). THF was used as the eluent at a flow rate of 1 mL/min. The column was standardized with pMMA calibration standards and toluene was used as a flow marker. #super[1]H NMR spectra were measured on a Bruker Avance 500 spectrometer with deuterated chloroform as the solvent. Properties are compiled in @combproperties. === Sample preparation In general, to prepare a crosslinked thermoset, polymer was dissolved in acetone, and succinic acid was added to create a solution with a 1:1 molar ratio of oxirane:COOH. 0.05 eq of TBD was added as a catalyst and the solution was stirred at 100 °C until most acetone had evaporated and a partially crosslinked solid remained. The polymer was then pressed in a mold at 150 °C for 10 minutes, and the flexible solid was cut to shape and left in an oven at 120 °C overnight to complete the curing process. === DMA To determine the modulus of the crosslinked samples, I performed DMA measurements with a DMA 850 (TA Instruments) equipped with a tension clamp. Frequency sweeps were performed at 0.01 % strain with a preload of 0.01 N at room temperature. === Swelling ratio To determine their equilibrium swelling ratio, fragments of each polymer were weighed and added into excess xylenes, chosen to be a good solvent for these polymers.@kaya2002 The samples were then placed in an aluminum reaction block thermostatted to 100 °C for 48 hours to reach their equilibrium swelling ratio. After the 48 hours, the samples were carefully removed or filtered from the xylene and weighed, with the difference being used to calculate the swelling ratio. === Toughness measurements To determine the toughness of the crosslinked samples, I conducted uniaxial extension measurements with a Texture Analyzer (Stable Microsystems) equipped with a 50 N load cell. The crosshead displacement rate was controlled at a speed of 0.1 mm/s. The samples were stretched until the samples broke at the center of the sample in between the grips. == Results and discussion === Controlled synthesis of reactive combs We synthesized a suite of different comb polymer architectures with the goal of determining how graft density and structure affected eventual network modulus and toughness, as well as a set of analogous linear polymers. I kept monomer ratio and degree of polymerization consistent between linear and comb grafts to best compare resultant material properties. I synthesized four comb polymers: pBIEM_-g-_GMA (homopolymer GMA side chains), pBIEM_-g-_(GMA_-co-_BMA) (random GMA copolymer sidechains, inset of @fig:combsynthesis b), pBIEM_-g-_(GMA_-b-_BMA) ("hidden" GMA block copolymer sidechains, inset of @fig:combsynthesis c), and pBIEM_-g-_(BMA_-b-_GMA) ("exposed" GMA block copolymer sidechains, inset of @fig:combsynthesis d). I then synthesized three linear polymers analogous to the sidechains of the comb polymers: pGMA (compared to comb polymers with homopolymer side chains), pGMA_-co-_BMA (compared to comb polymer with random side chains), and pGMA_-b-_BMA (compared to comb polymers with both "hidden" and "exposed" sidechains). #figure(image("Images/C3F1.png", width: 100.0%), placement: auto, caption: [ SARA-ATRP produces narrowly dispersed, well defined comb polymers. (a) Illustrations of the monomers used in this paper. (b) GPC chromatogram showing the shifting of precursor polymer to earlier elution times, demonstrating the successful chain extension with GMA_-co-_BMA. (c) GPC chromatogram showing successful chain extensions of pBIEM with first GMA and then BMA to create block copolymer side chains. (d) GPC chromatogram showing successful chain extensions of pBIEM with first BMA and then GMA to create block copolymer side chains. For (b-d), insets depict the idealized structure of the initial and final polymer. ] ) First, I synthesized linear polymers _via_ ATRP, and determined they had well-defined molecular weights and narrow dispersities _via_ GPC (pGMA: Mn \= 5,907 g/mol, Đ = 1.12; pGMA_-co-_BMA: M#sub[n] = 11,942 g/mol, Đ = 1.09; pGMA_-b-_BMA: M#sub[n] = 11,436 g/mol, Đ = 1.12). I then created copolymers of GMA and BMA with matched lengths, and measured that all polymers had comparable amounts of GMA. In order to produce graft polymers, I used a \"grafting from\" approach to grow GMA and BMA from a pBIEM backbone (@fig:combsynthesis a). To accomplish this, I polymerized an inimer (a dual functional small molecule acting as initiator and monomer) of BIEM _via_ RAFT to create a linear polyinitiator $tilde.op$500 monomers long. I chose CPDT as the chain transfer agent as trithiocarbonates have been shown to not participate in the necessary ATRP to create the ensuing linear-block-comb architectures.@kwak2008 Successful synthesis of pBIEM showed a modest final Đ = 1.3, higher than would be expected for a RAFT polymerization (@fig:combsynthesis b). This was likely due to small amounts of contaminating crosslinking, as indicated by a small shoulder in the GPC chromatograms at lower elution times. The Đ was similar to other reported polymerizations of BIEM.@stals2013 Using two distinct reversible deactivation radical polymerizations, well-defined and narrowly disperse comb polymers were synthesized. First, I created a comb architecture with random side chains (@fig:combsynthesis b). pBIEM_-g-_(GMA_-co-_BMA) was synthesized by initiating polymerization from pBIEM using GMA and BMA as comonomers in a 1:1 molar feeding ratio (pBIEM_-g-_(GMA_-co-_BMA, M#sub[n] = 423,522 g/mol, Đ = 1.3). The final monomer ratio was confirmed _via_ #super[1]H NMR at 51 mol% GMA (@fig:randombrushnmr). I then created a second comb architecture with a \"hidden\" block architecture. In this structure, the reactive GMA monomers are closest to the backbone (@fig:combsynthesis c). pBIEM_-g-_(GMA_-b-_BMA) was synthesized in a sequential polymerization process. First, polymerization of GMA from pBIEM yielded pBIEM_-g-_GMA (Mn \= 275,294 g/mol, Đ = 1.17). Next, BMA was polymerized from the ends of the grafted GMA of pBIEM_-g-_GMA to create pBIEM_-g-_(GMA_-b-_BMA) (Mn = 491,777 g/mol, Đ = 1.24). Final monomer ratio was confirmed _via_ #super[1]H NMR at 51 mol% GMA (@fig:hiddenblocknmr).Finally, I synthesized a comb with an \"exposed\" block architecture, where the reactive GMA monomers were furthest away from the comb backbone _via_ sequential polymerizations of BMA to create pBIEM_-g-_BMA (Mn 415,543 g/mol, Đ 1.36), and then GMA to create pBIEM_-g-_(BMA_-b-_GMA) (Mn 551,826 g/mol, Đ 1.31) (@fig:combsynthesis d). Feed ratio was confirmed _via_ #super[1]H NMR at 51 mol% GMA (@fig:exposedblocknmr). All these \"grafting from\" reactions showed no measurable increase in the dispersity of the polymers relative to the pBIEM backbone. #figure(image("Images/C3S4.png", width: 100.0%), placement: bottom, caption: [ #super[1]H NMR of pBIEM_-g-_(GMA_-co-_BMA) after successive polymerizations. Spectra recorded at 500 MHz in CDCl#sub[3]. ] ) #figure(image("Images/C3S5.png", width: 100.0%), placement: auto, caption: [ #super[1]H NMR of pBIEM_-g-_(GMA_-b-_BMA) after successive polymerizations. Spectra recorded at 500 MHz in CDCl#sub[3]. ] ) #figure(image("Images/C3S6.png", width: 100.0%), placement: auto, caption: [ #super[1]H NMR of pBIEM_-g-_(BMA_-b-_GMA) after successive polymerizations. Spectra recorded at 500 MHz in CDCl#sub[3]. ] ) === Crosslinking and storage modulus of reactive comb polymers After successfully preparing the reactive comb polymers, determined the effect of polymer architecture on the bulk properties of their crosslinked resins. I began by crosslinking the thermosets with succinic acid, chosen due to its small molecular weight and functionality (@fig:combmodulus a). Small molecular weight crosslinkers enable high crosslink density, somewhat offsetting the diluting effect of BMA comonomer, and the carboxylic acid functionality facilitates step-growth polymerization with the epoxy groups of GMA. A 1:1 ratio of acid:epoxy (succinic acid:GMA) was utilized, again to achieve a high crosslink density. Additionally, a catalyst was used increase the rate of crosslinking (TBD at 5 mol% relative to epoxy). All linear and comb polymers were crosslinked identically. #figure(image("Images/C3F2.png", width: 100.0%), placement: auto, caption: [ Comb polymers crosslink to form high modulus materials. (a) Schematic of the crosslinking reaction between glycidyl esters and succinic acid, catalyzed by TBD. (b) Storage modulus of block polymers crosslinked with succinic acid, as measured by DMA. (c) Storage modulus of homopolymers crosslinked with succinic acid, as measured by DMA. (d) Storage modulus of random copolymers crosslinked with succinic acid, as measured by DMA. ] ) We first crosslinked and characterized the modulus of the block polymers (@fig:combmodulus b). Successful crosslinking was identified by visual inspection. Polymers were measured under tensile conditions at 1 Hz (6.28 rad/s) frequency, resulting in storage moduli of pGMA_-b-_BMA = 584.0 MPa, pBIEM_-g-_(BMA_-b-_GMA) = 410.0 MPa, and = 410.0 MPa. The linear copolymer had the highest modulus of the three, and shows some frequency dependency, overlapping with the comb block polymers at lower frequencies. Strong phase separation of the BMA blocks resulting in a dynamic glass transition is likely the cause of the frequency dependent behavior of pGMA_-b-_BMA. Both comb block polymers had nearly identical moduli. We then prepared crosslinked homopolymer and homopolymer graft thermosets in an identical manner to the block polymers. At 1 Hz, the measured storage moduli of pGMA = 2,484 MPa and pBIEM_-g-_GMA = 577.1 MPa (@fig:combmodulus c). In this case, the linear homopolymer had a far higher modulus than the comb homopolymer, showing a roughly five-fold increase. Finally, I prepared crosslinked random copolymer thermosets in the same manner. At 1 Hz, the measured storage moduli of pGMA_-co-_BMA = 881.0 MPa and = 515.0 MPa (@fig:combmodulus d). Here, the linear polymer has a modulus almost twice as high as the comb polymer. The highest modulus value belonged to pGMA and pBIEM_-g-_GMA for linear and comb polymers, respectively, is expected, as the overall crosslinking density of these samples should be the highest. As crosslinking was mediated _via_ epoxy groups, the GMA homopolymers have the greatest degree of epoxy crosslinks without other comonomers to dilute it. More surprising is that all comb polymers consistently show a lower modulus compared to their linear counterparts. I hypothesized that this was due to increased intramolecular crosslinking occurring between comb hairs of block copolymer and homopolymer samples. In densely grafted comb polymers, GMA monomers would be forced very near to GMA on combs within the same polymer, favoring intramolecular crosslinks. These intramolecular crosslinks would create loop defects that do not contribute to the modulus of the final thermoset. === Network architecture effects on material properties To interrogate the network structure of each crosslinked polymer, we conducted swelling studies. Fragments of all samples were submersed in mixed xylenes at 100 °C for 48 hours to reach equilibrium swelling, and swelling ratio was determined by the difference in mass before and after swelling. Resultant values for block copolymers were pGMA_-b-_BMA = 1.85, pBIEM_-g-_(BMA_-b-_GMA) = 1.91, and pBIEM_-g-_(GMA_-b-_BMA) 1.2 times. For homopolymers, swelling ratios for pGMA = 1.34 times and pBIEM_-g-_GMA = 0.85. Random copolymers pGMA_-co-_BMA = 1.95 and pBIEM_-g-_(GMA_-co-_BMA) = 1.68 (@fig:combswelling a). All comb polymers, with the exception of pBIEM_-g-_(GMA_-co-_BMA) crumbled into small insoluble particles, indicating either that the solid was held together primarily by intermolecular interactions, or that the network topology was uneven, resulting in rapid material failure.@xu2020 Linear polymers retained their original shape (@fig:combmicro). Additionally, all comb polymers, with the exception of pBIEM_-g-_(BMA_-b-_GMA), showed lower swelling ratios than their linear counterparts. #figure(image("Images/C3F3.png", width: 100.0%), placement: auto, caption: [ Swelling of thermosetting polymers. (a) Swelling ratios of crosslinked polymers in mixed xylenes. n = 3 (b) Cartoon depiction of comb copolymers with random (left) and block (right) sidechains. ] ) #figure(image("Images/C3S3.png", width: 100.0%), placement: auto, caption: [ pBIEM_-g-_(BMA_-b-_GMA), pBIEM_-g-_(GMA_-co-_BMA), and pGMA_-b-_BMA swelling in xylenes. ] ) According to Flory,@flory1943 swelling ratio is directly related to degree of polymerization between crosslinks, which in turn is directly related to modulus. According to the swelling data then, each comb polymer should have a higher modulus. To explain this, after comparing the storage modulus data to the swelling data, I hypothesized that comb polymers contain a high degree of intramolecular crosslinking compared to linear polymers. The intramolecular crosslinks would form microgels and small loop defects, which would not contribute to the formation of elastically active network chains that are able to dissipate stress and increase the storage modulus (@fig:combswelling b). Additionally, these intramolecular crosslinks contribute to a more inhomogeneous network structure, which reduces the swelling ratio of crosslinked networks, as the network is less easily able to expand to accommodate additional solvent.@cuthbert2021 If these materials do contain a high degree of intramolecular crosslinks, they may contain fewer intermolecular covalent crosslinks or entanglements, resulting in reduced bulk mechanical properties. Intramolecular crosslinks would form highly crosslinked microgels, but adjacent polymers are loosely connected. Then, in addition to dissociating in solvent, I would expect these materials to have much lower toughness compared to samples that do not contain as many of these defects. To examine this, I conducted tensile tests to determine the toughness of the crosslinked thermosets. Each sample was cut into rectangles of approximately 6.5 mm in width and 1.5 mm in thickness. The samples were pulled until they broke in half and the toughness calculated from the measured force, elongation at break, and sample cross-sectional area. Block polymers, pGMA_-b-_BMA had a toughness of 180.0 kPa, pBIEM_-g-_(BMA_-b-_GMA) had a toughness of 95.60 kPa, and pBIEM_-g-_(GMA_-b-_BMA) had a toughness of 93.30 kPa. In contrast, random copolymers, pGMA_-co-_BMA had a toughness of 578.0 kPa, and pBIEM_-g-_(GMA_-co-_BMA) had a toughness of 616.0 kPa (@fig:combtoughness). Homopolymers, pGMA had a toughness of 248.0 kPa, and pBIEM_-g-_GMA had a toughness of 62.30 kPa. #figure(image("Images/C3F4.png", width: 100.0%), placement: auto, caption: [ Random comb sidechains suppress intermolecular crosslinking and create tougher materials. Tensile toughness of comb and linear polymers. n = 3 ] ) In the case of both block polymers and homopolymers, the comb polymers had lower toughness. These results are somewhat contrary to conventional block polymers, as the phase separation between segments of block polymers and resultant self-assembly is routinely exploited to create rubbery microdomains capable of dissipating stress and increasing the toughness of thermosetting resins. Linear block polymers@he2016 and comb block polymers@moon2021 have been demonstrated to improve the toughness of epoxy resins, and it was expected that comb polymers with a rubbery core would show superior toughness. I attribute this to intramolecular crosslinking forming loops and microgels that do not create elastically active network chains. Unsurprisingly, the comb homopolymer showed the worst toughness. BMA is known to increase toughness of crosslinked networks,@johnson1994@kim1977 and its absence in the homopolymers leads to increased brittleness. Most surprisingly, the random copolymers showed the highest toughness, with the comb polymer even achieving higher toughness compared to the linear polymers. It is worth noting that the highest measured toughness of pBIEM_-g-_(GMA_-co-_BMA) was 1,088 kPa, versus 883.0 kPa for pGMA_-co-_BMA. I hypothesize that the randomly distributed BMA monomers space apart the reactive GMA monomers, making it less probable that they will react to form loops with adjacent sidechains. This tensile data shows that I are ultimately able to create high modulus materials with toughness comparable to their linear analogues using highly crosslinked comb polymers. == Conclusion We synthesized novel thermosetting comb polymers containing both reactive glycidyl groups and rubbery butyl methacrylate. Block and homopolymer comb polymers showed low modulus and toughness due to intermolecular crosslinking creating loop defects. Using a random copolymerization of reactive and rubbery monomers, I was able to suppress intermolecular crosslinks and increase the toughness of highly crosslinked comb polymers. Through this, I demonstrated a method for creating very high molecular weight thermosetting polymers with high toughness and modulus. High molecular weight and a tunable density of epoxy functionality make this material and attractive option for fast setting coatings and adhesives.