Introduction

Material selection for specific applications necessitates careful consideration of both the required properties and the feasibility of manufacturing for the intended use. Materials with desirable properties (e.g., mechanical strength, chemical resistance, and environmental stability) often come at the expense of manufacturing aims (e.g., energy efficiency, cost effectiveness, environmental friendliness, rapidity, and simplicity). Polymer manufacturing exemplifies this challenge, specifically when considering a choice between thermoplastics and thermosets1,2. Thermoplastics are thermally processable materials due to their melting behavior and are generally straightforward to manufacture3. A broad range of thermoplastic materials are available, but most cannot maintain their mechanical properties in extreme environments, such as high temperatures. Conversely, thermosets provide desirable properties such as strength, thermal and chemical stability, and creep resistance due to the inclusion of permanent covalent crosslinks. As a result, they are more challenging to process and cannot be thermally reprocessed (i.e., melted) like thermoplastics. Thermosets also often require expensive equipment, such as ovens and autoclaves, and energy-intensive manufacturing processes with a significant environmental impact. Balancing manufacturing processes with achievable material properties is crucial for optimizing performance, driving innovation, and ensuring applicability across various industrial sectors.

Here, we focus on the industrially relevant emerging field of in-mold electronics (IME). IME leverages the melting behavior of thermoplastics to create structural electronics, integrating electrical components and connections into 3D structures through vacuum forming and injection molding4,5. Vacuum forming involves heating a thermoplastic sheet until pliable and using a vacuum to conform it to a mold, followed by cooling to retain the shape6. Vacuum forming has many advantages, including low-cost tooling, ease of updates and modifications, rapid turnaround for parts and tooling production, and the ability to produce durable and lightweight parts at affordable prices. IME applies electronic circuits and components to thermoplastic substrates, which are subsequently vacuum formed to create a functional 3D shape. Afterwards, more electronic components may be added, and injection molding is used to encapsulate the system7,8,9. IME is relevant in industrial applications, especially in the automotive sector, due to its ability to generate lightweight, space-saving, structural electronics10. Currently, IME relies on moldable thermoplastic substrates, such as polycarbonate and polyethylene terephthalate, for vacuum forming. However, thermosets would be advantageous for enhancing the robustness of these components against environmental factors, potentially expanding their use to other industries. Therefore, we aim to develop methods for implementing moldable thermoset substrates in the vacuum forming of structural electronics for IME applications.

With this goal in mind, we also want to be mindful of manufacturing considerations. Thermoset manufacturing often requires energy-intensive, expensive, environmentally harmful processes. Comparatively, frontal polymerization offers a unique pathway to rapid, energy-efficient curing for high-performance materials. In frontal polymerization, a reaction front is initiated by localized heating, generating a self-sustaining polymerization due to the heat released by the exothermic reaction11,12,13,14,15,16,17. The front propagates through the monomer or oligomer, converting the resin mixture into a solid polymer as it advances. This method can significantly reduce curing times, enabling rapid production. In addition, because the reaction is self-propagating, it requires minimal external energy input, leading to higher energy efficiency and lower energy costs.

Recent work on frontal ring-opening metathesis polymerization (FROMP) shows promise in the manufacture of poly(dicyclopentadiene) (pDCPD), a thermoset derived from an inexpensive byproduct of the oil and gas industry18. pDCPD possesses desirable properties such as high impact resistance, chemical corrosion resistance, high heat deflection temperature, and low shrinkage, making it attractive for industrial applications18,19. pDCPD is traditionally produced via reaction injection molding, wherein a mold is filled with a reactive mixture of monomer and catalyst, and the reaction is completed after injection19. In contrast, FROMP enables much more energy- and time-efficient manufacture of the same thermoset polymer. Recent developments in resin stability have enabled the processing of DCPD using FROMP via direct ink write (DIW) and the manufacture of carbon-fiber reinforced composites using vacuum-assisted resin transfer molding (VARTM), showing promise for industrial capabilities of FROMP20,21.

FROMP of free-standing, elastomeric gels has also been demonstrated20. Inspired by this work, we aim to vacuum form pDCPD in the gel state, followed by FROMP curing to achieve manufacturing of bespoke thermosets via vacuum forming for IME applications. In contrast to VARTM, which is a technique for molding thermoset composites, in which liquid resin is drawn into a dry fiber preform in a sealed mold using vacuum, vacuum forming, in which plastic sheets are softened and formed against a mold using vacuum, is conventionally restricted to thermoplastics due to the requirement of material softening from melting behavior and thus currently incompatible with typical thermosets. Notably, the need for solid substrates for application of electronic components prior to vacuum forming precludes the use of VARTM for IME applications and has also, prior to this work, precluded the use of thermoset resins for IME.

Herein, we develop a system capable of rapidly producing free-standing pDCPD gels while maintaining reactivity for practical processing times. We demonstrate that vacuum forming and subsequent FROMP of pDCPD generates thermosets with excellent material properties and benefits over traditionally vacuum-formed thermoplastics. We further employ pDCPD as a substrate for thermoformed, functional structural electronics. Overall, IMSE processes allow us to rethink the design and manufacturing of electronic products to reduce materials and costs while improving size, weight, and power (SWaP). The expansion of IME manufacturing to thermoset materials enables opportunities to incorporate devices with electrical functionality into demanding terrestrial and extraterrestrial environments and architectures.

Results and discussion

Resin formulation and characterization

On addition of a ruthenium catalyst, a monomer mixture of primarily dicyclopentadiene (DCPD) monomer gradually reacts at ambient conditions to form a viscoelastic pDCPD gel, which can then rapidly undergo FROMP on thermal initiation to form fully cured pDCPD (Fig. 1a). Notably, DCPD is a bicyclic monomer with two double bonds (norbornene and cyclopentene) with differing reactivities via ROMP22,23. The ring-opening of the highly strained norbornene is exothermic, driving the propagation of frontal polymerization once initiated. In this way, the material transitions from a flexible, elastomeric gel with a glass transition temperature below room temperature to a fully cured, rigid thermoset with a glass transition temperature above room temperature, as indicated by the tan delta curve from dynamic mechanical analysis (DMA) experiments (Fig. 1b–f). In this work, we use a free-standing pDCPD gel for vacuum forming to a mold, followed by FROMP to lock in the shape and complete curing (Fig. 1g).

Fig. 1: Overview of vacuum forming and frontal polymerization of pDCPD.
Fig. 1: Overview of vacuum forming and frontal polymerization of pDCPD.
Full size image

a Ambient transition at room temperature (RT) of monomeric dicyclopentadiene (DCPD) via ring-opening metathesis polymerization (ROMP) to polydicyclopentadiene (pDCPD) gel in the presence of a ruthenium catalyst ([Ru]) over time (∆t), followed by frontal ring-opening metathesis polymerization (FROMP) to form fully cured pDCPD. Fully cured pDCPD has a greater chain length (N) and degree of crosslinking (Mm) than pDCPD gel (n and m, respectively). b Tan delta peaks of elastomeric pDCPD gel (solid red line) and fully cured pDCPD (solid yellow line), showing the difference in glass transition temperature. c pDCPD gel is flexible and conforms to surfaces. d Illustration of FROMP across a pDCPD gel, increasing crosslinking in the material as the front traverses it. e Photo of FROMP of a pDCPD gel, resulting in a noticeable color change. f After FROMP, fully cured pDCPD is rigid and maintains its shape. g A step-by-step vacuum forming concept for pDCPD gels. The gel is placed over a mold on a vacuum forming table and clamped into place. Vacuum is pulled to force the gel against the mold. FROMP is initiated locally on the gel, as indicated by the waved lines, and the front traverses across the gel to produce fully cured pDCPD.

Pot life, the length of time from formulation until frontal polymerization can no longer occur due to reagent decomposition or reaction, is one of the primary considerations for manufacturing with DCPD FROMP systems. To meet the needs of multi-step manufacturing processes such as IME, which demand long pot life (≥ 1 h), extending the pot life of FROMP systems is a necessary and an ongoing challenge24,25,26. Many proposed solutions involve using Grubbs 2nd generation catalyst (GC2), well known for its high reactivity, and incorporating inhibitors to reversibly or irreversibly curb catalyst activity27,28,29. For example, Robertson et al used alkyl phosphite inhibitors in DCPD resins with GC2 to extend the liquid state processing window30. Follow-up work focused on using the stabilized viscoelastic gel state for DIW capabilities and FROMP of free-standing, elastomeric gels20. However, reaching the stiffer gel state required long incubation times of 18 h. Using the same formulation would inevitably limit manufacturing throughput. One solution to increasing throughput, i.e., forming reactive gels faster, is to adjust the type and amount of inhibitor (e.g., alkyl phosphite)31. However, these inhibitors are generally toxic and introduce an added formulation space to explore. To reduce formulation complexity, we envisioned that thermally latent metathesis catalysts could also enable longer processing windows without the addition of inhibitors. Therefore, we selected a thermally latent ruthenium catalyst, UltraCat, which enables quick formation of elastomeric gels and practical working windows.

DCPD resins used in this work consist of the ruthenium catalyst UltraCat, 95 wt% DCPD, and 5 wt% 5-ethylidene-2-norbornene (ENB). ENB depresses the freezing point of DCPD, which is a solid at room temperature, and enables liquid phase processing at ambient conditions. While the final material is a copolymer of DCPD and ENB, there are minimal effects on the properties as compared with pure pDCPD20,30. For simplicity, we refer to gels and fully cured materials as pDCPD. Initial formulations vary the catalyst loading from 0.005 to 0.04 mol%. Rheological curves characterize the ambient-condition gelation behavior of resins over time (Fig. 2a and Supplementary Fig. 1). Modulus crossover approximately represents the time required to reach a viscoelastic gel state (Table 1). The modulus crossover for 0.02 mol% catalyst occurs approximately 10 min after resin preparation. Increasing the catalyst to 0.04 mol% results in a modulus crossover at 5 min after resin preparation, limiting the time available for liquid state processing. In comparison, decreasing the catalyst to 0.01 and 0.005 mol% results in a modulus crossover at 20 and 70 min after resin preparation, respectively. A catalyst loading of 0.04 mol% does not provide sufficient time for further characterization; therefore, a loading of 0.02 mol% is the maximum explored in subsequent investigations.

Fig. 2: Catalyst loading and reactivity.
Fig. 2: Catalyst loading and reactivity.
Full size image

a Rheology of DCPD/ENB mixtures at ambient conditions with varying ruthenium catalyst loadings of 0.005 mol% (solid green line), 0.01 mol% (solid blue line), and 0.02 mol% (solid purple line). The horizontal, dashed line labels a storage modulus of 2 kPa. b Initial viscoelastic gel formation does not form a stand-alone gel. Scale bar is 20 mm. c Waiting for a suitable increase in storage modulus gives a stand-alone gel. Scale bar is 20 mm. d FROMP is initiated on a strip of gel with a soldering iron. The front traverses the entire strip. Scale bar is 10 mm. e FROMP front velocity in the gel and (f) residual exotherms from DSC measurements over time at varying ruthenium catalyst loadings of 0.005 mol% (green squares), 0.01 mol% (blue squares), and 0.02 mol% (purple squares), with normalized time indicating the time after the gel has reached a storage modulus of 2 kPa. Data are presented as the mean +/− one standard deviation for n = 4 replicates.

Table 1 Dynamic gelation behavior of DCPD/ENB resins with 0.005, 0.01, and 0.02 mol% catalyst. Usable gel working time corresponds to the amount of time available between a gel reaching a 2 kPa storage modulus and the end of its working time (i.e., pot life)
Full size table

While modulus crossover approximately indicates the initial point of viscoelastic gel formation, it does not necessarily correlate to the formation of a sufficiently robust gel necessary for handling and vacuum forming. Gels are formed by casting resins between two parallel plastic plates spaced to a desired thickness and incubating for a specified duration. Removal of the setting resin from the plates too early – specifically, immediately upon modulus crossover – reveals a tacky, gelatinous solid incapable of maintaining form and causing difficulties in handling (Fig. 2b). Through investigation, the earliest point at which gels can be easily manipulated by hand corresponds to a storage modulus of 2 kPa from rheological curves (Fig. 2a and Table 1). Catalyst loadings of 0.02, 0.01, and 0.005 mol% result in 25, 45, and 150 minutes to reach 2 kPa, respectively. After this, separating the plates results in a usable gel (Fig. 2c).

We want to ensure elastomeric gels remained malleable during the entire manufacturing process without vitrifying due to background polymerization or consuming so much monomer that frontal polymerization cannot be sustained after initiation. To process materials via vacuum forming, the working time of gels must extend through removal from the plates, transportation to the vacuum former, formation over a mold, and completion of the curing reaction across the entire gel. The working time is characterized for each loading of catalyst by creating a larger gel, removing it from the plates at the time corresponding to reaching 2 kPa from the rheological curves, cutting a moderately-sized strip, and frontally polymerizing by initiating with a heat source (e.g., soldering iron) from one end of the gel strip (Fig. 2d). These strips are tested for their ability to undergo FROMP after removal from the plates and every 15 min after until the front no longer propagates the entire length due to lack of sufficient heat generation. Time points are normalized to the time the gel reaches a 2 kPa storage modulus for each catalyst loading as determined by rheology. For catalyst loadings of 0.02 and 0.01 mol%, the working time of the gels is 45 and 75 min, respectively. Interestingly, this is in good agreement with the time at which the gels vitrify, or become glassy networks, as observed by parallel plate rheology (Supplementary Fig. 1 and Table 1). Comparatively, gels containing 0.005 mol% catalyst lose the capacity for FROMP prior to vitrification. This is likely due in part to catalyst decomposition, preventing sufficient exotherm generation at this low catalyst loading. For 0.02 mol% catalyst loading, a window of less than 1 h (45 min) exists between the formation of a usable gel and the end of its working time. Comparatively, catalyst loadings of 0.01 and 0.005 mol% result in more than 1 hour between the usable gel formation and the end of the working time. The temperature and velocity of the front, parameters commonly used to indirectly assess the reactivity of frontal polymerizations, are also evaluated (Fig. 2e and Supplementary Fig. 2) along with the residual exotherm of the gel as measured by differential scanning calorimetry (DSC) (Fig. 2f and Supplementary Fig. 3). At all catalyst loadings, the frontal velocity and the residual exotherm decrease over time, corresponding to the background conversion of DCPD/ENB during spontaneous polymerization21,30,32,33. At all catalyst loadings, the maximum temperature reached during FROMP remains relatively consistent over time (Supplementary Fig. 2). After undergoing FROMP, all materials are fully cured as indicated by the lack of a residual exotherm (Supplementary Fig. 4).

A catalyst loading of 0.01 mol% provides many benefits, including a practical 20 min liquid processing window, rapid 45 min time to usable gel formation (2 kPa storage modulus), and a 75 min working time. Comparatively, 0.005 mol% catalyst loading requires 150 min to form the usable gel phase with little benefit to extended working time and a lower frontal velocity. A 0.02 mol% catalyst loading gives quicker usable gel formation (25 minutes) but decreases working time significantly. Therefore, we explore 0.01 mol% catalyst loading for vacuum forming, followed by FROMP to enable pDCPD thermoset manufacturing.

To provide additional insight into gel formation, we characterize the conversion of monomer using 0.01 mol% catalyst at ambient conditions over time. The conversion of norbornene olefin bonds is tracked over time using FTIR experiments, and the gel fractions of formed gels are obtained (Supplementary Fig. 5). The results are in good agreement, indicating approximately 10–20% conversion of norbornene in the usable gels. However, a slight (~ 15 min) delay in the timing to achieve a usable gel is observed in these experiments compared to previously gathered data (Fig. 2) using different lot numbers of materials purchased from the same manufacturers (i.e., commercially available monomers and catalysts). In comparison, the conversion as determined by residual heat of reaction in DSC (Fig. 2f), carried out with the same lot of materials as other experiments in Fig. 2, appears to increase to higher values (20–40%) faster (Supplementary Fig. 5). We believe the difference in these results are due to lot-to-lot variations in the commercial monomers and catalysts. For the practical implementation of this system, we recommend evaluating the timing of gel formation for each new lot of materials, even from the same supplier. Regardless, within the same lot, gelation behavior is predictable and easily timed. While the variation we observe here does not result in drastically different timing of gel formation, future work should evaluate lot-to-lot variation and resulting gelation.

Vacuum forming and frontal polymerization of pDCPD thermosets

To examine the malleability of the pDCPD gels and use in thermoforming, several geometrically complex molds are additively manufactured for use in vacuum forming. Polymer gels are placed over molds and clamped into place on a vacuum table. Enabled by their elasticity, the gels successfully vacuum form to the molds and stretch without rupturing. A heat gun is used to initiate FROMP (Fig. 3a and Supplementary Movie 1). The front propagation is conveniently visible to the unaided eye by the change in material color from orange to transparent. Thermal imaging further visualizes the high-temperature front traveling across the mold. The time to complete the FROMP of the entire 10 cm × 10 cm area of the vacuum-forming window is about 5 min. After the reaction is completed, the fully cured pDCPD is separated from the mold and cut to size using a CO2 laser. This process is amenable to production of parts with distinctive features, including positive and negative curvatures and sharp edges and corners (Fig. 3b). The material is fully cured after the reaction, as indicated by the lack of a residual exotherm from DSC (Supplementary Fig. 4). Therefore, manufactured parts do not require post-processing and are immediately ready for use. While we opt to use the latent UltraCat catalyst formulation developed here, we note that this manufacturing process should be amenable to previously explored systems with a stable free-standing gel state with reactivity towards FROMP. To demonstrate broad applicability, we repeat the vacuum forming process with the commonly used Grubbs catalyst 2nd generation (GC2) and tributyl phosphite as a stabilizing inhibitor (Supplementary Fig. 6).

Fig. 3: Vacuum forming and FROMP of pDCPD gels for vacuum-formed thermosets.
Fig. 3: Vacuum forming and FROMP of pDCPD gels for vacuum-formed thermosets.
Full size image

a Photos and thermal images of vacuum forming and FROMP of pDCPD gels, initiated with a heat gun off-camera as indicated by the waved lines. The thermal gradient indicates temperatures from room temperature (RT, purple) to front temperature (light yellow). Scale bar is 20 mm. b pDCPD formed to a variety of shapes, including hills and valleys, a star with sharp corners, and a split flat-flex cable analog. Scale bars are 20 mm. c TGA of fully cured pDCPD. Inset of vacuum formed stars of thermoplastic polyethylene terephthalate glycol (PETG) and thermoset polydicyclopentadiene (pDCPD) before and after 1 h at 300 °C in an argon atmosphere. d Representative tensile test curve of fully cured pDCPD with break indicated by the crossed lines. e UV-Vis of fully cured pDCPD and an inset demonstrating visual transparency of fully cured pDCPD to visible wavelengths. The scale bar is 20 mm.

Vacuum forming of thermoset pDCPD results in benefits over traditionally used thermoplastics. Fully cured pDCPD retains mass up to 450 °C before thermally degrading, as characterized by thermogravimetric analysis (TGA) (Fig. 3c). Compared to an equivalent thermoplastic part manufactured via traditional vacuum forming, e.g., polyethylene terephthalate glycol (PETG), pDCPD retains its shape after spending 1 h at 300 °C, while PETG melts and loses shape (Supplementary Fig. 7). Notably, the fully cured pDCPD part discolors to dark brown. While oxidation is common in pDCPD materials34,35,36,37, in this case it is unlikely to be the singular cause of discoloration as the thermal treatment was performed in an argon atmosphere. Instead, discoloration is likely due to degradation of the remaining catalytic species in the material, as indicated by TGA (Supplementary Fig. 8). Although there are some changes to the network structure and mechanical properties, as discussed in the following section, the part retains its shape, in stark contrast to the thermoplastic, which inevitably melts.

Fully cured pDCPD manufactured by vacuum forming and FROMP has good mechanical properties, with a Young’s modulus of 1.7 GPa and a tensile strength of 48 MPa (Fig. 3d and Supplementary Table 1). Glass transition temperatures, determined by DMA, are also high, within the range of 168-181 °C (Supplementary Fig. 9 and Supplementary Table 1). This illustrates the transition of the material from a soft, elastic gel with a low glass transition temperature of – 18 °C and low Young’s Modulus of 39 kPa (Supplementary Fig. 10) to a fully crosslinked, tough, thermoset network. Interestingly, the tan delta curve exhibits two distinct peaks, located close together. While we are unsure what causes this behavior, recent work reports this phenomenon occurring as a function of catalyst selection26. The materials properties obtained here are comparable to reported properties for pDCPD (Supplementary Table 2). In addition to its high glass transition temperatures and maximum operating temperatures, pDCPD also exhibits modulus and tensile strength competitive with thermoplastics commonly used for vacuum forming (Supplementary Table 2).

Exposure of pDCPD to 300 °C for 1 h in an argon atmosphere results in embrittlement of the material, as indicated by tensile testing results (Supplementary Fig. 11). Furthermore, the tan delta curves show a single thermal transition peak centered at 188 °C (Supplementary Fig. 11 and Supplementary Table 1), which could be reflective of increased crosslinking. Calculating the molecular weight between crosslinks of materials before and after heat treatment gives 1290 and 454 g mol–1, respectively, corresponding to approximately 10 and 4 repeat units between crosslinking sites and confirming the increase in crosslink density after high temperature exposure. Despite this, the Young’s modulus and tensile strength are largely unaffected (Supplementary Table 1), although the strain at failure decreases significantly. While oxidative crosslinking is known to occur in pDCPD38, again, we doubt its source as the singular cause of this phenomenon, owing to the inert atmosphere during heat treatment. Instead, we reason these property changes are due to increased crosslinking as a result of olefin addition22,23. As a comparison, we repeat the high temperature exposure in air and characterize the resulting material (Supplementary Fig. 12). While the Young’s modulus is maintained, the effect of oxidation on embrittlement of materials at such extreme temperatures is clear, as the tensile strength decreases significantly to 13 MPa (Supplementary Table 1). This provides confirmation that extensive oxidation likely does not occur as the crosslinking mechanism in materials heat-treated in argon.

Finally, while FROMP-manufactured parts may discolor to light brown over time due to oxidation at the surface (Fig. 3c), as-made parts are mostly transparent as observed visually and as measured by UV-Vis spectroscopy (Fig. 3e). Incorporating antioxidants can help to prevent oxidation and discoloration if desired39. However, pDCPD oxidation layers are generally thin (10–100 μm) and chemically inert, providing additional benefits such as resistance to bulk oxidation40. While pDCPD is susceptible to aging and subsequent property changes (e.g., embrittlement, color) due to oxidation even at ambient conditions, it is currently used in industrial applications, for example, heavy vehicles, where its excellent impact strength, stiffness, and toughness lends itself well for use in fenders and bumpers41.

Structural electronics using thermoset substrates

The manufacturing method presented in this study uniquely takes advantage of free-standing, solid gels. As such, these gels have the potential to act as 2D substrates for applying components or decorations prior to manufacturing to shape. As a demonstration, we decorated a flat gel with a diamond pattern using a stencil and marker prior to vacuum forming and frontal polymerization, enabling deformation of the pattern to adopt the 3D shape and curvature of the manufactured part (Supplementary Fig. 13). This introduces a manufacturing paradigm for high-performance, inflexible thermoset materials, where existing 2D patterning approaches are used to apply components to 2D substrates followed by processing to generate 3D shapes.

We expand this method to accommodate the manufacturing of structural electronics. Manufacturing 3D electronic components requires patterning processes for applying conductive and dielectric materials to substrates. While niche processes are able to pattern, for example, electrical interconnects over non-planar surfaces42, 2D patterning processes for electronics are a mature technology and offer an overall advantage for the IME manufacturing approach10. To this end, IME uses vacuum-formed thermoplastic substrates with applied electronic components, followed by injection molding to seal electronics4. We envision vacuum forming followed by frontal polymerization, enabling structural electronics with thermoset substrates, offering benefits such as improved thermal stability and chemical resistance. Furthermore, pDCPD has excellent dielectric properties (Supplementary Table 2), offering a promising option for IME43,44,45. In current practice, IME uses metal colloid inks that are specifically designed to withstand vacuum forming and injection molding processes to apply conductive traces to substrates. However, these inks have increased electrical resistance versus bulk metals and are potentially susceptible to aging and lifetime issues (e.g., oxidation), precluding their use in extreme environments. In addition, conductive traces must withstand FROMP conditions across the pDCPD gel, here reaching temperatures between 150–200 °C.

To this end, we implement adhesive-backed copper foils to pattern conductive traces. Combining this with thermoset materials significantly widens the operational environment for IME beyond what is currently achievable using oxidation-prone nanoparticle inks and low-melt temperature thermoplastic substrates and encapsulants. Copper foil traces are demonstrated in previous work, in which serpentine patterns are used to accommodate stresses arising from the vacuum forming process46,47,48,49. These interconnects can flex in 3D by designing 2D traces with meanders, enabling proper trace placement upon vacuum forming without delamination from the substrate. Figure 4a summarizes the steps employed here. Copper tape is applied to the pDCPD gel. A UV laser marker cuts the desired trace design from the copper tape. Notably, the UV laser marker enables cold marking, relying on a photolytic degradation mechanism as opposed to pure thermal processing and avoiding initiation of FROMP on the gel. After cutting the traces, the excess copper tape is removed from the gel. The gel with traces is placed over the mold, aligning the traces with their desired placement. FROMP is then initiated across the gel to form a fully cured pDCPD part. Once complete, a CO2 laser cuts the part to shape. Using this method, we produce a structural electronic component from a fully cured pDCPD thermoset with conductive copper traces.

Fig. 4: Vacuum-formed structural electronics from frontally polymerized thermosets.
Fig. 4: Vacuum-formed structural electronics from frontally polymerized thermosets.
Full size image

a Step-by-step process for creating structural electronics from thermosets. Copper foil tape is placed on a polydicyclopentadiene (pDCPD) gel. A UV laser marker cuts the trace shape into the copper foil tape, and the residual is removed, leaving the conductive traces behind. The gel is vacuum formed to a mold, and FROMP is initiated as indicated by the waved lines. The fully cured part is cut from its surroundings with a CO2 laser, leaving a structural electronic with correctly placed conductive copper foil traces. b Trace, mold, and trace placement design for a simple power module. c Photo of the produced power module with copper traces. Scale bar is 20 mm. d Addition of spring contacts and surface mount electronic components to create a functional LED circuit operated by a switch.

Here, we demonstrate a battery power module with careful design of traces and mold (Fig. 4b). This part requires trace placement on the inside of the manufactured part to contact the batteries. Therefore, the gel is formed with traces facing towards the mold. On completion and removal from the mold, the part is successfully generated from the intended design with accurate trace placement (Fig. 4c). Holes are drilled for the placement of screws and battery spring contacts. Furthermore, electronic components, in this case surface mount switches, resistors, and LEDs, can be incorporated onto the trace path afterwards using traditional soldering methods. We envision this as an additional benefit of the thermoset material employed here, which can tolerate high-temperature soldering conditions without risk of deformation. Here, we demonstrate a functional device (LED circuit powered by batteries and operated by a switch; Fig. 4d), illustrating a practical use case for IME focusing on process compatibility with sharp curvature (~90°) and component integration. In addition, this process strategy has been validated via design and iteration of a human-machine-interface (HMI) exemplar (Supplementary Fig. 14), which is a common use case for IME in automotive and appliance applications10. Overall, we successfully demonstrate a system to create structural electronics from vacuum-formed thermoset materials with property benefits over traditionally employed thermoplastics (Supplementary Table 2).

In this work, we have successfully developed thermoset materials for compatibility with accessible vacuum forming processes using frontal polymerization, in this case, FROMP of DCPD. The system gives high throughput generation of structural electronics, provides advantageous material properties over traditionally used thermoplastics, and thus can address requirements for operation under extreme (mechanical, temperature) environments. The impact for IME—which seeks to incorporate electronic functions into 3D structural/mechanical objects—is significant, as IME provides greater design freedom while reducing connections between components and overall size weight and power (SWaP). The development of degradable and recyclable systems and foams using FROMP also offers opportunities for sustainability, part recovery, and mechanical and density tuning within the IME community50,51,52. Furthermore, opportunities exist to expand to other thermoset materials systems, including other frontal polymerization reactions16,17,53,54,55,56,57,58,59,60,61, B-staged epoxy resins62,63,64, dual-cure systems65,66,67,68, and covalent adaptable networks69,70,71,72. This expansion of materials that are compatible with vacuum forming capabilities represents a breakthrough for IME technologies, enabling 3D device integration into a wider range of durable products, environments, and architectures.

Methods

Materials

Dicyclopentadiene (DCPD, 94%) was purchased from Oakwood Chemicals. 5-Ethylidene-2-norbornene (ENB, 99%), Grubbs Catalyst 2nd Generation (GC2), tributyl phosphite (TBP, 90%), ethyl vinyl ether (EVE, 99%), and dichloromethane (DCM, ≥ 99.5%) were purchased from Sigma-Aldrich. UltraCat was purchased from Strem Chemicals Inc. All chemicals were used as received.

pDCPD Resin formulation and gel formation

A monomer solution of 95:5 DCPD:ENB (wt:wt) was used to create gels. DCPD was melted in an oven at 40 °C prior to adding ENB at the designated ratio. Mixtures remained in the liquid state after the addition of ENB. UltraCat was weighed into a cup at an amount corresponding to either 0.005, 0.01, or 0.02 mol% with respect to the monomers (0.4, 0.8, and 1.6 mg UltraCat per 1 g 95:5 DCPD:ENB, respectively). Minimal DCM was added (~ 1 wt% of the total mixture) to dissolve UltraCat before adding the monomer solution and mixing to achieve a homogeneous solution. The homogeneous precursor solution was filled between PETG plates separated by 2 mm thick polylactic acid spacer frames printed on an Ultimaker FDM printer and allowed to reach gelation over time at room temperature. After reaching a solid gel state, the plates were separated, and the gel was removed. Examples using 0.01 mol% GC2 (0.65 mg per 1 g 95:5 DCPD:ENB) were made similarly, except in this case, 0.004 mol% TBP (2:5 mol:mol with GC2, 0.08 µL per 1 g 95:5 DCPD:ENB) was added to the resin.

Rheology

Rheology was used to characterize the gel behavior of resins at room temperature. An ARES-G2 (TA Instruments) operating in small-amplitude oscillatory shear mode was used with a 20 mm diameter parallel plate assembly. A frequency of 2 Hz and an amplitude of 1% strain were used with a gap size of 1.5 mm between plates.

Rate of FROMP Propagation

To determine the FROMP rate of gels over time at different catalyst concentrations, a larger gel was made (7.8 cm × 6.6 cm), and approximately 10 mm wide x 40 mm long strips were cut from the gel for each desired time point. FROMP was initiated at one end of the strip, sitting on a glass slide, by contacting it with a soldering iron set to 450 °C. A video recording was taken, and Tracker software was used to determine the position of the front over time. The front position was plotted with respect to time, and the slope of the steady state region was used to determine the rate. This was done in triplicate for each catalyst concentration and time point, starting from the time the gel reached a storage modulus of 2 kPa as determined by rheology and every 15 min after until propagation of the front was no longer observed along the entire length of the gel strip.

Maximum temperature during FROMP

To determine the maximum temperature reached during FROMP of gels, the resin was mixed and divided into glass test tubes (approximately 4 g per test tube). Thermocouples interfacing with a temperature logger were placed in the resin such that the thermocouple did not touch the test tube walls and only contacted the resin. The resin was allowed to form a gel around the thermocouple over time at room temperature. FROMP was initiated from the bottom of the test tube using a soldering iron set to 450 °C, contacting the outside of the test tube. The front was observed, and the maximum temperature as it passed over the thermocouple was recorded. This was done in triplicate for each catalyst concentration and time point, starting from the time the gel reached a storage modulus of 2 kPa as determined by rheology and every 15 min after.

DSC

Differential scanning calorimetry (DSC) was used to determine the residual exotherm of gels over time. A Q200 DSC (TA Instruments) was used. A larger gel was made, and ~ 5 mg samples were cut from the gel for each desired time point and loaded into hermetic aluminum pans. DSC was run on samples at a ramp rate of 10 °C min−1. This generated an exothermic peak in the DSC curve corresponding to the ring-opening of unreacted groups during ROMP. The residual heat of reaction of the exotherm was determined as the area under the peak. This was done in triplicate for each catalyst concentration and time point, starting from the time the gel reached a storage modulus of 2 kPa as determined by rheology and every 15 min after. Exotherms from DSC experiments were used to calculate the conversion of gels with 0.01 mol% catalyst over time via Eq. (1).

$${{C}}_{{DSC}}=\frac{{{H}}_{{i}}-{{H}}_{{t}}}{{{H}}_{{i}}}$$
(1)

Where CDSC is the relative conversion of the gel at time t, Hi is the integration of the exotherm from the initially prepared liquid (time = 0), and Ht is the integration of the residual exotherm from the gel at a time t.

FTIR

Fourier-Transform Infrared Spectroscopy (FTIR) was used to determine the conversion of norbornene bonds in gels containing 0.01 mol% catalyst at ambient conditions over time. A Nicolet iS50 FTIR (Thermo Scientific) was used. Resin was casted between glass plates with 20 μm spacers and placed in the measurement fixture. Spectra were recorded for the material approximately every 3 s. Afterwards, Gaussian fitting was used on the peak occurring at 3140 cm−1, corresponding to the norbornene C=C-H vibration, to obtain the area under the peak over time. The conversion of norbornene was calculated via Eq. (2).

$${{C}}_{{NB}}=\frac{{{A}}_{{i}}-{{A}}_{{t}}}{{{A}}_{{i}}}$$
(2)

Where CNB is the conversion of norbornene at time t, Ai is the initial area under the peak (time = 0), and At is the area under the peak at time t.

Gel Fractions

Gel fractions were used to determine the gel content at ambient conditions over time of materials using 0.01 mol% catalyst. A larger gel was made, and ~ 500 mg samples were cut from the gel at each desired time point and placed into a vial. The vial was then charged with 20 mL of DCM and 100 µL of EVE to stop further metathesis polymerization. The solutions were stirred at room temperature for 7 days, after which the remaining solid gel samples were filtered and allowed to dry at room temperature for 3 more days. The solid samples were weighed before and after this process. Gel fractions were calculated via Eq. (3).

$${{f}}_{{g}}=\frac{{{m}}_{{f}}}{{{m}}_{{i}}}$$
(3)

Where fg is the gel fraction, mf is the final mass after treatment and drying, and mi is the initial mass.

Vacuum forming

For all vacuum forming experiments, a catalyst concentration of 0.01 mol% was used, and materials were gelled for 60 min before further usage. Desired mold shapes were designed in SolidWorks and 3D printed from ST45 resin on an Origin One SLA printer. Molds were sprayed with PTFE mold release coating before vacuum forming. A Formech 450DT was used to carry out vacuum forming of pDCPD gels (15.2 cm × 15.2 cm). Molds were placed over the vacuum channel on the stage, and pDCPD gels were draped over the mold and clamped into the frame of the machine. Vacuum was pulled to force the gel against the mold at room temperature. A temperature-controlled heat gun set to 450 °C was used to initiate FROMP from one side of the formed gel from a distance of approximately 12 inches. After initiation, the heat gun was removed, and the material was allowed to complete FROMP before turning the vacuum off. After allowing the part to cool down, it was removed from the machine, and the mold was removed. Video was taken of the process normally and with a FLIR C2 camera. Designed parts were cut from the surrounding material using an Epilog 50 W CO2 laser cutter.

DMA

Dynamic mechanical analysis (DMA) was performed on samples using a DMA 850 (TA Instruments). Testing specimens were laser cut from pDCPD using an Epilog 50 W CO2 laser cutter from samples that underwent FROMP during vacuum forming. Specimens were approximately 15 mm long, 3 mm wide, and 2 mm thick. Samples were loaded into tensile clamps with a gauge length of approximately 10 mm. The samples were oscillated at a frequency of 1 Hz and an amplitude of 0.1% strain at a temperature ramp rate of 5 °C min−1. Tests were done in triplicate to confirm the repeatability of obtained materials properties.

DMA data were used to calculate the molecular weight between crosslinks for fully cured pDCPD and heat-treated pDCPD (300 °C for 1 h in argon) via Eq. (4).

$${{M}}_{{c}}=\frac{3{pRT}}{{E}^{\prime} }$$
(4)

Where Mc is the molecular weight between crosslinks in g mol−1, p is the density of pDCPD (taken as 1 g cm−3), R is the universal gas constant (8.314 J mol−1 K−1), T is 50 K after the glass transition temperature, and E’ is the storage modulus in Pa at T, indicative of the rubbery plateau.

Tensile testing

Tensile testing was done using a Mark-10 tensile tester equipped with a 1500 N load cell. Type-V dogbone testing specimens were laser cut from pDCPD using an Epilog 50 W CO2 laser cutter from samples that underwent FROMP during vacuum forming. The gauge of dogbone samples was marked with a black marker, samples were deformed at a rate of 10% min−1 of the original gauge length, and the deformation of samples was tracked using a FLIR X-One video extensometer. The video extensometer data was used to obtain the deformation of samples in the form of strain, and the tensile tester data was used to calculate the stress on the sample during deformation. This was done in triplicate to confirm repeatability of materials properties. The Young’s Modulus was taken from the initial linear slope of the stress-strain curve. Ultimate tensile strength was determined as the maximum stress reached by the sample from the stress-strain curve.

TGA

Thermogravimetric analysis (TGA) was done to determine the thermal stability of samples using a Discovery SDT650 (TA Instruments). Samples were cut from pDCPD that underwent FROMP during vacuum forming and loaded into alumina pans. The temperature was ramped from room temperature to 800 °C at a ramp rate of 15 °C min−1. The sample mass was tracked during the ramp and plotted as the percentage of the original mass remaining versus temperature.

UV-Vis

UV-Vis spectra were obtained from samples that underwent FROMP during vacuum forming. A Cary UV-Vis Compact Peltier (Agilent) was used. Samples fitting the dimensions of the loading chamber were cut using an Epilog 50 W CO2 laser cutter. Samples were loaded into the chamber, and the spectrum of the sample was measured. Results are reported as transmittance across the measured wavelengths.

High temperature exposure of PETG and pDCPD

To demonstrate the utility of a thermally stable thermoset over a traditionally vacuum-formed thermoplastic, melting demonstrations were carried out using an environmental oven with an argon atmosphere. Vacuum-formed star shapes made from commonly-vacuum formed thermoplastic PETG and thermoset pDCPD were placed in the oven on a Pyrex dish. The oven was cycled between pulling a vacuum and filling the oven with argon gas for 3 cycles. The samples were then ramped from room temperature to 300 °C at a ramp rate of 5 °C min-1 and held for 1 h before cooling down and removing. For the sample treated in air, the process was repeated in an air atmosphere.

Battery power module

Gels with conductive copper foil traces were made by placing a square of copper conducting tape (3 M) onto the gel. The copper tape was laser cut with a UV laser marker (Keyence, MD-U1000), and excess tape was removed. The gel was vacuum formed with the traces facing inward towards the mold and lined up with the desired location using registration markings. FROMP was initiated from one end of the gel with a heat gun, and the final part was removed afterward. Springs and screws were set into the piece to hold batteries. A simple LED circuit was made by mounting a switch, resistor, and LED to the desired location along the trace path with SnPb solder.