Exploring Inverse Vulcanized Dicyclopentadiene As a Polymer Matrix for Carbon Fiber Composites

Inverse vulcanization of waste or renewable dienes has generated materials with phenomenal properties across a spectrum of applications. Nevertheless, the use of these materials for structural applications remains a challenge. Here, the use of an inverse vulcanized cyclopentadiene polymer as a resin for carbon ﬁber reinforced composites is explored. The dynamic S ─ S bonds in the polymer matrix are used to repair composite specimens over 5 generations by heating the material to 140 ° C. A range of composites are manufactured and evaluated for their ﬂexural properties, using a range of ﬁber orientations. Finally, this polymer is used to reinforce a carbon ﬁber fabric composed entirely of reclaimed materials, constituting a composite entirely composed of waste materials and second life carbon ﬁbers.


Introduction
Composite materials in all their forms (micro-, nano-, etc.) have become an integral part of everyday life, and none more so than carbon fiber composites.These materials have advanced lightweighting in large structures for mass transport, facilitating unprecedented levels of fuel economy and flight range.For all their benefits, composites do suffer from limitations in terms of damage tolerance as, unlike metals, they are unable to be reheated and forged back to their original form. [1]For carbon fiber composites, microcracking will typically lead to the onset of further DOI: 10.1002/mame.202300298degradation and catastrophic failure.The ability to mend composites has been examined using a range of approaches, such as self-healing via the insertion of microparticles and patching provide useful solutions for composite repair. [2]These approaches provide a useful stop-gap for composite repair but are localized and may have limitations, such as the inability to repair deep internal defects.
To this end, the use of vitrimers and other polymers that possess dynamic covalent bonds, have become important materials being explored to create composites able to be reshaped, repaired, and recycled. [3]n these systems, the introduction of a stimulus (heat, light, pressure, etc.) causes specific chemical bonds to continuously break and reform.This feature is then used to create new polymer networks from existing structures, facilitating polymer reprocessing, repair, and thermoforming.While vitrimers present these beneficial properties they possess a significant short coming and this is typically in their physical performance relative to existing resins that are used in composite applications.To address this issue a significant research effort has been put into the development of thermoset (epoxy) polymers analogous to currently used resins by incorporating functional groups able to undergo dynamic exchange reactions. [4]Typical examples include transesterification reactions, imine formations, and disulfide exchange processes, among others.Several commercial products exist which include vitrimeric epoxy polymers such Vitrimax produced by Mallinda.
The dynamic nature of the S─S bond provides a useful chemical handle to develop reprocessable composites, at low temperatures.Odriozola et al. used a disulfide-derived epoxy polymer in a composite that was able to be healed, reshaped, and resin removed while maintaining the integrity of the underlying reinforcing fiber. [5]Similarly, Chen et al. examined the use of disulfide linkages in both the epoxy and hardener components of a resin and showed outstanding physical properties and the ability to repair these composites with minimal reduction in performance. [6]hese studies have highlighted the power of incorporating dynamic S─S bonds into a resin scaffold to imbue truly novel functionality into the final material.Typically, the current systems that include such S─S bonds require extensive synthesis (e.g., installation of the epoxy groups on a pre-existing disulfide scaffold).Also, sulfur is susceptible to oxidation when heated in air, drastically reducing the ability of the disulfide exchange Figure 1.Fundamental reaction of inverse vulcanization a), range of physically diverse copolymers synthesized using inverse vulcanization b). [9]actions-resulting in reduced repairability lifetime.Therefore, the inclusion of more S─S bonds within the polymer backbone should extend the repairability lifetime of the composite.
The enrichment of sulfur within the polymer can be achieved by inverse vulcanization, originally reported by Pyun et al. [7] In those systems, polymers can be created with very high mass components of sulfur (>50% by weight), leading to portions of the backbone possessing short chains of sulfur atoms ranging from 2 to 7 atoms long.The fundamental reaction of inverse vulcanization is the reaction of elemental sulfur (in radical or ionic form) with an alkene (Figure 1A), or multiple alkenes for cross-linking to occur.The plethora of organic compounds possessing alkene groups provides a blank canvas for the potential generation of many unique polymeric materials.The physical properties of the resulting polymer are typically dictated by the carbon portion of the co-polymer mixture.For example, inverse vulcanization of canola oil provides a soft sponge-like polymer as the highly flexible lipid chains allow a great deal of molecular mobility within the polymer structure. [8]Conversely, the same process carried out with rigid cyclic structures, such as dicyclopentadiene (DCPD) used in this study, results in a dark glass-like material (Figure 1B).
Another significant benefit of inverse vulcanization used in this work is that both sulfur and dicyclopentadiene are byproducts of oil refining.While the latter is used on relatively small scales for fine chemical synthesis and metathesis polymerization, both starting materials for this process are available on large scale at low cost, and require no synthetic elaboration to prepare them for the polymerization process.
Therefore, the goal of this work is to evaluate and demonstrate the potential that this unique polymer system has, specifically as a useful tool in the drive toward circular and sustainable polymer composites.

Experimental Section
Carbon fiber mats including 0°, 0°/90°, and 90°weave orientations were obtained from Ironbark Composites, Victoria, Australia.The nonwoven discontinuous fiber mats were 300 g m −2 recycled carbon fiber sourced from Gen2Carbon, UK.All chemicals, reagents and solvents were purchased from Sigma-Aldrich Chemical Company and used as received.

S-DCPD Polymer Synthesis
Equal masses of elemental sulfur and dicyclopentadiene were placed in a reaction vessel and stirred (300 rpm) at 140 °C.The reaction time was observed to vary within the range of 2.5-3.5 h.The reaction vessel was removed from the heat based on the viscosity of the pre-polymer ideal for laminate lay-up and manufacturing.Following this, the obtained pre-polymer was then cured at 140 °C for 24 h to obtain the fully crosslinked polymer.Caution!Careful control of temperature of this polymerization is required.At temperatures higher than 140 °C there is a risk of a highly exothermic, runaway reaction.

S-DCPD Carbon Fiber Composite Manufacture
The process of composite manufacture is outlined below (Figure 2).Briefly, carbon fiber mats of varying fiber orientations were trimmed to 4 × 10 cm and weighed.To obtain composites of equal thickness as required for flexural strength testing (ASTM D2344) four layers for the continuous fiber weaves and two layers for the discontinuous nonwoven recycled fiber mats were used.Each of these layers were saturated with the synthesized prepolymer, placed on top of each other, and wrapped in aluminum (Al) foil to avoid slipping of layers and major loss of polymer during the curing stage.The Al foil containing the pre-impregnated laminates was placed between two 3 mm metal plates that were then screwed in place with a torque of 4.5 Nm in on the four corners.This was then placed in the oven at 140 °C for 24 h to facilitate full crosslinking of the polymer.

Flexural Strength Analysis
Flexural testing was performed in accordance with ASTM D2344.A minimum of five specimens 1 mm in thickness, 12 mm in width, and 35 mm in length were cut from the manufactured composites (Section 2.3) using the secotom 60 and allowed to dry.The specimens were then placed onto an Instron universal mechanical testing machine, attached with a load cell of 10 kN and a 3-point bend fixture.The effective flexural span was 30 mm between the two support pins and a central roller of 10 mm diameter was subsequently used to apply a bending force across the flexural specimen at a displacement-controlled rate of 2 mm min −1 .

Repair of Fractured S-DCPD Composite
To induce healing of the failed composite specimens postflexural testing, a simple method was adapted from the curing procedure detailed in Section 2.3.The fractured specimens were placed back between the two metal plates and screwed in place with a torque of 4.5 Nm on the screws.These plates were then placed in the oven at 140 °C for 2 h to allow facilitate repair.These repaired composite specimens were subjected to flexural strength testing as detailed in Section 2.4 and up to five repair cycles were conducted on the range of S-DCPD/CF composites.

Scanning Electron Microscopy
Zeiss Supra 55VP was equipped to obtain scanning electron micrographs used to study the surface of the fractured composite.Samples were adhered to aluminum stubs using carbon tape and gold coated using a Leica ACE600 high vacuum sputter coater.Imaging of the samples was conducted using an accelerating voltage of 5 kV and at an ≈10 mm working distance.

Synthesis and Characterization of the neat S-DCPD Polymer
The preparation of the S-DCPD polymers is operationally very simple, as elemental sulfur is dispersed in an equal weight of DCPD and heated to 140 °C.At this temperature the elemental S 8 reacts with the strained alkene (C═C) of the DCPD, initiating the propagation of the inverse-vulcanized polymer. [10]The initially free-flowing mixture becomes more viscous over time (Figure 3a), correlating to the initial formation of the linear polymer, then progressively increased amounts of cross-linking.This is supported by 1 H NMR spectroscopy through the appearance of new peaks corresponding to the formation of sulfur bonds around 3.75-4.0ppm and the disappearance of the DCPD alkene peaks 5.5-6.0 ppm (Figure 3b).
To encourage good "wet-out" of the carbon fibers it is important to use the polymers, while they still possess low enough viscosity to penetrate the fibrous network.Herein lies an important balance, because insufficient reaction of the sulfur and DCPD materials will result in the reformation of S 8 due to significant amounts of polymer "backbiting." [11]On the other hand, significant cross-linking of the polymer will result in a thick gel that makes it a challenge to infiltrate the carbon fiber architecture.It was determined that heating the mixture at 140 °C for 2.5-3.5 h gave a thick mixture able to be processed similar to most liquid epoxy polymers.Considering the intended use of this material in structural applications, it is important to physically characterize the neat polymer (Figure 3).The degradation of S-DCPD polymer was determined to begin at 190 °C followed by a second mass loss around 375 °C (Figure 3f, see supporting information for larger images).The two-step pattern is also reported in other inverse vulcanized copolymers.The first step is related to the degradation of sulfur, and the second, associated with the loss of organic content or crosslinker component of the copolymer. [12]his may also indicate the presence of both long and short sulfur chains within the cured polymer.The glass transition temperature was estimated to be around 120 °C which was determined through differential scanning calorimetry, and although it was very broad and poorly defined, it was comparable to previously reported literature. [10]As reported for various inverse vulcanized materials, no melting peak was observed for the S-DCPD copolymer, confirming its amorphous structure [10] (Table 1).

Synthesis and Characterization of Unidirectional S-DCPD-CF Composites
In this section, our attention turned to the use of this polymer as a resin for carbon fiber composites.Three composite samples were manufactured with different continuous fiber orientations to establish, as per traditional composites materials, the reinforcing effect of the fibers depending on their orientation relative to the applied stress.Therefore, composites with 0°, 0°/90°, and 90°d irectional reinforcement were prepared, and their flexural properties were examined (Figure 4).
As expected, for the 90°samples, the fibers provided little-tono reinforcement as the stress was applied along the axis of the fiber and thus the physical properties were largely dictated by the polymer phase.Indeed, the flexural strength and modulus measured for the 90°sample was effectively the same as those found for the neat polymer (Figure 4).
Again, as would be expected in a traditional polymer system, the inclusion of fibers oriented in the 0°place in the 0°/90°sample showed significant improvement in both flexural strength and modulus (118.39 ± 7.06 MPa and 34.37±1.54GPa, respectively) compared to the neat resin and 90°samples as expected due to the active load absorbed by the reinforcing fiber.Similarly, further improvements were observed when only the unidi- rectional fibers in the 0°orientation were used, as all the included reinforcing fibers were being utilized for their load bearing capabilities allowing the composite to withstand an impressive 529.0 ± 37.61 MPa.After successfully manufacturing a reinforced composite using the polymer, the ability and extent of the composite for fracture repair through thermal healing was explored.
A major benefit of dynamic systems, such as inverse vulcanized polymers, is the ability to repair composite samples.This can be effected either using a chemical, for instance a disulfide  or nucleophilic solvents like triethylamine or pyridine to mend the fracture interface, [13] or in this instance simple heating.The broken specimens of all 3 geometries, once evaluated for their flexural properties, were repaired using the procedure detailed in Section 2.5.At this temperature (140 °C) the S─S bonds are known to undergo rapid metathesis, thus opening the possibility to repair the fracture surface by "knitting" closed. [14]It is important to note that in this instance, no additional chemical reagents are required for the repair and the high sulfur content in the polymer backbone enables repeated repair and evaluation cycles.To ensure that there was no degradation of the polymer occurring at the curing temperature (140 °C) thermogravimetric analysis was carried out on the neat polymer.This showed a degradation point at ≈190 °C, with a slight plateau at ≈450 °C.This suggested that the curing temperature was well within the stable thermal range of the polymer, and this was further confirmed by isothermal TGA at 140 °C showing no mass loss when subjected to this temperature for over 2 h (Figure 3g).This further affirmed the rehealing procedure used did not cause loss of mass or degradation of the polymer The samples were repaired and evaluated for their flexural properties for 5 generations each (Figure 5).Note that "generation 1" refers to the initial sample that has been fabricated and "generation 2" is the specimen having undergone 1 repair cycle.Despite performing poorly, the sample with 90°fiber orientation was evaluated over 5 generations, in the interest of thoroughness.No loss in flexural strength was observed and, although low, this eluded to the capability of the polymer to be repaired via heating without the loss of mechanical properties.Also, given the weakness of these samples, the data showed a significant amount of scatter and high variability.A much more obvious trend was observed when the 0°/90°samples were evaluated over the same repair cycle.As was expected a rapid decline in flexural strength is immediately observed after the first regeneration process, dropping from 118.39 ± 7.06 to 71.91 ± 5.38 MPa.This decrease is due to the breakage of reinforcing fibers that occurs during the first evaluation of flexural properties (Figure 5).
Presumably in this fiber layup orientation, only the 0°fibers are breaking, as the 90°oriented fibers are not undergoing any stress.Nevertheless, the repair of the samples was successfully performed, with all samples able to undergo at least 5 repair cycles with most of the flexural strength preserved from generation 2-5.Indeed, the difference of flexural strength from generation 2 (71.91 ± 5.38 MPa) is only ≈17.7 MPa higher than that of generation 5 (54.19 ± 8.75 MPa).The flexural modulus of these samples shows a consistent value (26.68 ± 2.68-28.86± 6.82 GPa), except for generation 4. Presumably, this lower result was due to poor consolidation of the composite after testing to failure in generation 3 as delamination of laminates were clearly visible for these 0°/90°orientation composites following flexural testing.This aberration from the other generations highlights the need for intimate physical contact of the fracture sites when undergoing repair.This may also be due to effects such as small movements of the sample pieces when being placed under pressure or when being moved.Impressively, in generation 5 the flexural modulus had returned to the original value for the previous generations, eluding improved pressure exertion during the rehealing process.
The trends observed for the 0°/90°samples were magnified in the 0°specimens (Figure 5).In this instance, a clear continual decrease in flexural strength was observed over each of the specimen generations.This effect is more noticeable in the 0°samples as all the fibers in this orientation are undergoing stress, and thus are undergoing breakage upon each successive repair iteration, as observed in other unidirectional composites undergoing flexural stress.Whereas in the 0°/90°samples, by the 3rd generation, it is likely that the 50% of fibers within the composite in the 0°orientation and are those withstanding the flexural load have broken, and thus a plateau is seen.The flexural modulus was initially very high for generations 1-3, at 103.18, 112.77, and 99.76 GPa, respectively.A decrease in flexural modulus was observed for generations 4 and 5 (86.45 and 61.81 GPa, respectively), even though an increase in modulus was expected, these changes are possibly due to the repair cycles inducing increased levels of sulfur-alkene cross-linking, causing the polymer to become progressively more brittle and thus failing at lower stress.In addition, fiber breakage is also likely to have had an effect, and a combination of these two scenarios is likely.
As noted previously, the polymeric system used in this study is derived from waste or by-products of petrochemical refining.The inclusion of virgin carbon fibers has shown significant physical benefits that can be installed into these systems, though the significant carbon footprint of the fibers themselves can undermine the gains achieved by using this polymeric system.

Characterization of S-DCPD-CF Composites Using Nonwoven Recycled Carbon Fiber
Therefore, as a concluding study in this work, a composite material was fabricated that uses nonwoven carbon fiber fabrics that are made from reclaimed carbon fibers.These nonwoven fabrics have undergone a carding (entanglement via needle punching) process and have a random distribution of orientations providing reinforcement in all directions.Therefore, repeating the same fabrication methodology as reported above on the nonwoven fabrics (300 g m −2 ) gave a composite material that was suitable for flexural characterization.This constitutes a carbon fiber composite that is entirely made from reclaimed or waste materials and provides context to the samples manufactured with virgin fiber presented above.
Similar to the work outlined above, these composites containing reclaimed carbon fibers were physically evaluated and repaired 5 times, again to provide a direct comparison to the virgin fiber composites presented earlier (Figure 6).The physical properties are most comparable to the 0°/90°reinforcement samples, which is logical as these are the most representative of isotropically reinforced materials that we examined.The initial flexural strength (172.6 ± 12 MPa) was significantly higher than the first generation of the 0°/90°samples (118.4 ± 7.1 MPa) though did have a significant reduction of 33% upon the first repair cycle (generation 2, 116.2 ± 17 MPa).After this the flexural strength of generations 3-5 plateaued and were statistically indistinguishable ranging from 95 to 82 MPa.
The flexural modulus, again similar to that of the 0°/90°specimens, showed a significant reduction from generations 1-2, 21.3 ± 1.2 and 15 ± 1.6 GPa, respectively.This was similarly attributed to fiber breakage in the initial test.Nevertheless, for generations 2-5 flexural strength remained very consistent through the repairing processes, being within 2 GPa (14-16 GPa) of each other showing a good retention of properties (Figure 6).Scanning electron microscopy was used to observe the fracture surface of the recycled carbon fiber composite (Figure 7).At the point of fracture, fiber breakage and crack propagation (Figure 7d) were both evident.While some wet out of fibers was obtained, fiber debonding and dry fibers were also observed at the fracture site (Figure 7c), a commonly attributed to the inert nature of carbon fiber surfaces.Modification of the fiber surface through electrochemically initiated polymerization has previously been employed to improve the surface chemistry and may be a useful avenue to explore in the future.

Conclusion
In conclusion, we have demonstrated the use of a repairable polymer derived from petrochemical waste and by products in carbon fiber composites.The polymer itself can be prepared and processed similar to a traditional thermoset epoxy resin, within a specific reaction time frame (2.5-3.5 h).The fabricated composites showed the expected effects with respect to fiber reinforcement geometry as is observed in traditional thermoset composites, albeit with reduced physical properties compared to commercial epoxy polymers.Nevertheless, these composites when tested under flexural load were able to be repaired by re-fusing the polymer back together using only thermal treatment (140 °C for 2 h) and retested over 5 generations.Varying degrees of property retention, dependent of carbon fiber layup orientation were observed over the generational evaluation of these composites.Finally, a composite was prepared from reclaimed carbon fiber nonwoven materials with this unique polymeric system.This constitutes a composite entirely derived from second life or waste materials.Again, the composite was evaluated for flexural properties, being similar to a traditional isotropic 0°/90°fiber weave layup.For this final composite system, a 30% reduction in flexural strength and modulus was observed from generation 1-2, though subsequent generations remained consistent.
Future studies for this area will include new composite fabrication techniques to better control fiber wet-out, the addition of co-monomers to the DCPD monomer portion to increase the ductility and challenges with processing, and finally surface modification of the carbon fibers to maximize fiber matrix adhesion.

Figure 2 .
Figure 2. Flow chart demonstrating the fabrication of S-DCPD-CF composites.a) S-DCPD oligomer poured and evenly spread on carbon fiber laminates.b) Saturated laminates were stacked and wrapped in aluminium foil to avoid loss of polymer during cure.c) The wrapped laminates were sandwiched between to steel plates and bolted using 4.5 Nm torque using a torque wrench.d) Steel plates were then placed in the oven to cure at 140 °C for 24 h.e) Final composite.

Figure 3 .
Figure 3. a) The progressive reaction of elemental sulfur and dicyclopentadiene to give S-DCPD prepolymer.b) 1 H NMR spectroscopy of the inverse vulcanization reaction showing consumption of the strained alkenes followed by reaction of the lower energy alkenes.c) Postcuring the neat polymer to cross link the polymer giving flexural specimens.d) XRD of the fully cured samples 1 day and 1 week after postcure, showing no reformation of crystalline sulfur.e) Removal of surface defects via simple exposure to heat via heat gun.f)Thermogravimetric analysis of cured polymer showing mass loss initiating at 190.1°C.g) Isothermal TGA of cured polymer at 140 °C showing no loss in mass over 2.5 h period.

Figure 4 .
Figure 4. Flexural strength and flexural modulus of the neat polymer and when reinforced with 3 orientations of carbon fiber, inset is the neat and 0°fiber reinforcement samples.

Figure 6 .
Figure 6.Flexural properties of composites made from S-DCPD and reclaimed carbon fibers over 5 generations of repair.

Table 1 .
Characterization of S-DCPD polymer properties