Opportunities at the Intersection of 3D Printed Polymers and Pyrolysis for the Microfabrication of Carbon-Based Energy Materials

Abstract

In an era marked by a growing demand for sustainable and high-performance materials, the convergence of additive manufacturing (AM), also known as 3D printing, and the thermal treatment, or pyrolysis, of polymers to form high surface area hierarchically structured carbon materials stands poised to catalyze transformative advancements across a spectrum of electrification and energy storage applications. Designing 3D printed polymers using low-cost resins specifically for conversion to high performance carbon structures via post-printing thermal treatments overcomes the challenges of 3D printing pure carbon directly due to the inability of pure carbon to be polymerized, melted, or sintered under ambient conditions. In this perspective, we outline the current state of AM methods that have been used in combination with pyrolysis to generate 3D carbon structures and highlight promising systems to explore further. As part of this endeavor, we discuss the effects of 3D printed polymer chemistry composition, additives, and pyrolysis conditions on resulting 3D pyrolytic carbon properties. Furthermore, we demonstrate the viability of combining continuous liquid interface production (CLIP) vat photopolymerization with pyrolysis as a promising avenue for producing 3D pyrolytic carbon lattice structures with 15 μm feature resolution, paving way for 3D carbon-based sustainable energy applications.

1. Introduction

In recent years, additive manufacturing (AM), or 3D printing, has enabled the development of highly replicable polymer structures with features on the scale of single microns.¹⁻³ AM is a software-controlled process for producing objects layer-by-layer that enables advanced fabrication of highly tunable and complex geometries, for example beam-based lattices, circumventing the constraints of traditional subtractive or molding methods.^4,5 Carbon is a low-cost element with high abundance in nature. While 3D printing enables a high degree of flexibility in the geometry of printed parts, 3D printing cannot produce pure carbon structures under ambient conditions directly without further processing steps.⁶ 3D carbon structures offering high electrical conductivity and chemical stability have demonstrated potential to be used as 3D electrodes.^1,7

Conducting a high temperature thermal treatment primarily under an inert atmosphere (argon or nitrogen gas) is called pyrolysis, and this process is often conducted on polymeric materials to initiate carbonization of polymer chains into fully carbonaceous structures.⁸ Carbon-based materials can be classified into three primary groups: amorphous carbon; semicrystalline carbon, also known as glassy carbon; and highly crystalline carbon, also known as graphitic carbon. Typically, the goal of pyrolysis is to eliminate noncarbon elements in a polymer (e.g., oxygen and hydrogen) and induce graphitization, the formation of highly stable sheets of trigonal planar sp² hybridized carbon.^6,9 The high electrical conductivity of graphitic carbon produced in this manner arises due to the presence of delocalized pi (π) electrons, which are free to move throughout the material.

The formation of structurally stable graphitic pyrolytic carbon requires a high surface area-to-volume ratio (S_V) to enable volatile byproducts created during carbonization to escape from within the depths of thicker features. Mechanical deformations and defects can result from built-in stress caused by trapped volatiles during the pyrolysis process. Heat transfer concerns during pyrolysis can also arise with thicker solid objects leading to heterogeneities.^8,10 Polymer structures with thinner features and high void fractions better facilitate volatiles escaping during heating without deformation due to shorter diffusion pathlengths.¹⁰ AM can generate high surface area structures such as beam-based lattices that achieve the desired high S_V characteristic needed for chemical conversion via pyrolysis.¹¹ Transforming 3D printed polymer lattices into carbon via subsequent pyrolysis therefore is an attractive approach for converting 3D printed structures into pure carbon. This technique has generated structures with high S_V suitable for applications in structural materials,^12,13 heterogeneous electrocatalysis,^14,15 biosensors^16,17 and renewable energy storage.¹⁸⁻²⁰ Thus, the convergence of AM with pyrolysis is poised to contribute to achieving a sustainable future.¹

Herein, we will review the current state of research into pyrolysis of 3D printed polymers to pyrolytic carbon structures. While there exists substantial interest in the generation of polymer-derived structural and functional ceramics (e.g., semiconductors, ceramic batteries) via 3D printing techniques such as precursor infiltration and pyrolysis (PIP), this review will focus on 3D carbon-based materials produced via carbonization of polymers, so the reader is directed to other reviews that focus on ceramics.²¹⁻²³ Previous reviews have also discussed the production of carbonaceous materials with 3D printing via carbon-filled composites, inclusion of graphene oxide in resins, and pyrolysis of 3D printed carbon precursors^6,24−26 and 3D printed aerogels.¹ Here, we add to this body of literature by focusing on the polymer resin compositions used for generating 3D carbon material, the geometry and feature size limits of 3D printed structures during pyrolysis, and the pyrolysis heat treatment conditions used to convert the polymeric structures to conductive carbon. To accomplish this, we begin with an overview of the 3D printing methods used with pyrolysis to date to generate carbon-based materials (Section 2). We then survey the current literature to elucidate the effects of polymer molecular composition (Section 3) and pyrolysis conditions (Section 4) on resultant carbon materials for use in several energy, catalysis, and biomaterial applications (Section 5). Finally, in Section 6, we introduce the use of continuous liquid interface production vat photopolymerization, which has already been scaled to an industry level, in conjunction with pyrolysis to generate high resolution, conductive carbon lattices, paving way for next generation carbon lattice structures for advancement of global sustainable energy and electrification.

2. 3D Printing Methods Used with Conventional Pyrolysis Thermal Treatments

Pyrolysis of polymeric materials has origins going back to the charcoal used in prehistoric times.²⁷ In the modern materials industry, pyrolysis has been used for over a century to produce high strength, conductive carbon structures, most prominently in the carbon fiber industry.^28,29 In the late 1800s through mid 1900s, the polymers used for carbon fiber production were initially sourced from natural feedstocks, such as bamboo shafts, cotton threads, and cellulose. In the 1960s, the industry transitioned from natural materials to synthetic fiber feedstocks, such as rayon and polyacrylonitrile, which are still used today.³⁰ Carbon fibers, with diameters of 5–10 μm, have been the state-of-the-art material in the pyrolysis field and are woven together for use in applications ranging from the electrical components in light bulbs to the heat-resistant fan blades in jet engines.³¹ However, pyrolysis of nonfibrous thicker 3D structures has been a relatively unexplored area due to the diffusion issues faced by volatile components escaping polymeric structures during pyrolysis. As illustrated in Figure 1, scalable AM in combination with pyrolysis has recently been utilized to overcome this challenge and generate thick (≥cm scale) pyrolytic carbon objects with micro- or nanoscale features that enable high S_V desirable in sustainable applications in the ongoing energy industry shift toward electrification and reliance on batteries for renewable energy storage.³²

Figure 1.

Historical context for pyrolytic carbon feedstock (top row) and production (middle) leading up to several modern applications (bottom).

The process of AM generates 3D objects using digital graphic data and is principally practiced using either extrusion techniques, powder sintering techniques, or light-based photopolymerization techniques.³³ Most approaches for pyrolysis of 3D printed structures to date are plagued by low carbon yields, high degrees of volume shrinkage, and structural warping during the polymer-to-carbon conversion.^{13,19,34−36} While some researchers have accommodated these challenges by designing the as-printed parts to account for large shrinkages,^32,37 others have sought to utilize this shrinkage to their advantage.^12,38 Strategies to mitigate shrinkage include incorporating carbon-based resin additives to increase char yields,^39,40 conducting preoxidation of the printed parts to reduce internal stresses developed during printing,^19,34 and designing custom resin formulations.^19,34,35 Interest in the combination of 3D printing and pyrolysis has increased rapidly over the past several years, evidenced by the increase in number of Google Scholar citations mentioning 3D printing, pyrolysis, and pyrolytic carbon (Figure 2).

Figure 2.

Histogram displaying the recent increase in prevalence of 3D printing and pyrolysis literature in Google Scholar from 2013 to 2023.

There are many methodologies that can accomplish 3D printing. The family of extrusion-based 3D printing techniques, such as direct ink writing (DIW)^1,18,41,42 and fused filament fabrication (FFF),^43,44 is widely employed for layer-by-layer printing to create 3D periodic lattices. For example, DIW using inks containing a high percentage of carbon nanostructures (e.g., graphene) has been used with pyrolysis to produce thick conductive carbon aerogels for energy storage.^1,18,42,45 Ink formulations for DIW are generally constrained to thixotropic and shear-thinning materials to prevent deformation of printed structures,⁴⁶ whereas ideal DIW inks have high viscosity and demonstrate solid-like behavior at zero shear. Further, such direct-write fabrication processes are challenging to scale due to their limited throughput at the necessary high resolution needed for achieving the S_V electrode manufacturing goals later described in Section 5, outside of certain material systems such as polyelectrolytes that are not applicable in pyrolytic carbon production.⁴⁷ Another AM method of note for carbon production is aerosol-jet microscale printing, which uses an ultrasonic actuator to atomize carbon precursors in an aerosol stream and deposit graphene ink via a nozzle with resolution down to 10 μm, without the need for a pyrolysis step.⁴⁸

Alternatively, vat photopolymerization (VP)³⁷ 3D printing is a category of techniques that employs the selective curing of photoresins with light. VP approaches are split into three subclassifications: (1) raster-based methods including scanning stereolithography (SLA) and direct laser writing via two-photon polymerization (DLW-2PP),⁴⁹ (2) volumetric additive manufacturing including computed axial lithography (CAL)⁵⁰ and xolography,⁵¹ and (3) pattern projection-based methods including digital light processing (DLP),³² projection micro stereolithography (PμSL),⁵² and continuous liquid interface production (CLIP).⁵³ VP currently is a scalable technology for producing complex lattice structures that cannot be produced via traditional molding methods due to the high speed and resolution capabilities afforded by using light-based radical polymerization.⁵⁴ The DLP and CLIP techniques in particular offer excellent combinations of higher throughput, with linear speeds up to 17 mm/min, precision, and scalability, as the entire liquid vat surface is exposed to UV light patterns controlled via a digital micromirror device.⁵⁴⁻⁵⁶ For more information on printing rates and scalability of the aforementioned AM methods, the reader is directed to recent studies that focus on 3D printing speed and comparisons of photopolymerization approaches.^54,55,57

The prevalence of which AM methods have been used in combination with pyrolysis in recent literature is overviewed in Table 1. This summary demonstrates that there is a significant representation from both the extrusion (DIW, FFF) and the photopolymerization (SLA, DLW-2PP, DLP, PμSL, CLIP) methods. The reported SLA category serves as a catch-all classification for laser-based photopolymerization methods that do not fit into DLW-2PP or other VP categories. The extrusion techniques commonly generate structures with features on the millimeter scale, with one instance of DIW achieving 100 μm features,⁴⁵ while the photopolymerization techniques demonstrate feature sizes ranging from as low as 100 nm for DLW-2PP⁵⁸ to 2 μm for SLA,¹⁷ 28 μm for DLP,³² and 10 μm for PμSL.⁵² The pyrolysis furnace upper-end temperature for 3D printed structures ranges from 430 to 1300 °C in these studies, with the greatest abundance of studies using between 900 and 1000 °C as their final temperature.

Table 1. Summary of Current Additive Manufacturing (AM) Methods Used in Combination with Pyrolysis for Generation of Pyrolytic Carbon.

AM Method	Pyrolyzed Material Feature Size	Application	Resin	Furnace Conditions	Reference
Direct ink writing (DIW)	mm scale	Gas separation, adsorption, and desorption of CO2	Pluronic F127, activated carbon	900 °C final, 10 °C/min ramp	(59)
	100 μm	Electrodes for Energy Storage	Hydroxypropyl methylcellulose and GO	1050 °C final, 2 °C/min ramp	(45)
	mm scale	Capacitors and Electrodes	Cellulose and ZIF-8 MOF	1000 °C, 5 °C/min	(60)
Fused filament fabrication (FFF)	500 μm	Electrodes for Energy Storage	Cellulose, lignosulfonate	1200 °C final, various holds, 0.2 °C/m ramp	(44)
	mm scale	High Capacitance Supercapacitors	Polyacrylonitrile, solketal acrylate	1000 °C final, 200 and 260 °C air hold, 0.5 °C/min ramp	(43)
Direct laser writing via two-photon polymerization (DLW-2PP)	170 nm	Fiber optics	Organic–inorganic SZ2080	600 °C final	(61)
	170 nm	Semiconductors	Custom acrylates: PEGDA, PETA, ZIF-8 MOF	550 °C final, 20 °C/s ramp	(62)
	1 μm	High Strength Structural Applications	IP-Dip (Nanoscribe GmbH)	1000 °C final, 450 °C, 500 °C, 550 °C holds, 20 °C/s ramp	(36)
	250 nm	Electroforming, Template-assisted electrodeposition	IP-Dip (Nanoscribe GmbH)	500 °C final, 5 °C/min ramp	(63)
	100 nm	High Strength Structural Applications	Custom acrylates: PRM30 and PETA	1100 °C final, various holds, 3 °C/min ramp	(64)
	2–10 μm	Mechanics, Optics, Medical Detection	IP-Dip (Nanoscribe GmbH)	500 °C final	(65)
	100 μm	Neurotransmitter Detection	IP-S (Nanoscribe GmbH)	900 °C final	(58)
	135 μm	Batteries, material engineering, and catalysts	IP-Q (Nanoscribe GmbH)	Various temperatures up to 600 °C, 10 °C/min ramp	(49)
	200 μm	High Strength Structural Applications	IP-Dip (Nanoscribe GmbH)	900 °C final, 250 and 350 °C hold, 3 °C/min ramp	(12)
Stereolithography apparatus (SLA)	316 μm	Electrodes for Energy Storage	SU-8, High Temperature (HT) V1	900 °C final	(66)
	10 μm	Batteries, fuel cells, sensors, filters	SU-8	1300 °C, 1150 °C, 900 °C, 600 °C finals	(67)
	2 μm	Microelectrode chips for biochemical sensing	SU-8	900 °C final, 200 °C hold, 2 °C/min ramp	(17)
	135 μm	Energy storage, catalysis, adsorption	Pentaerythritol tetraacrylate (PETA), divinylbenzene (DVB), PEGDA	900 °C final, 300 °C air hold, 880 °C CO2 hold	(34)
	20 μm	Scaffold for cell growth	Boston Micro-Fabrication HTL yellow resin	900 °C final, 350 °C isothermal hold, 3 °C/min ramp	(68)
	200 μm	Semiconductors	PIC100 (EnvisionTEC) with graphene oxide	500 °C final, 150 °C, 300 °C, and 400 °C holds, 2 °C/min ramp	(69)
	700 μm	Heterogeneous catalysis and ferromagnetism	Sartomer SR-339, ZIF-8 MOF	950 °C final, 3 °C/min ramp	(70)
	∼500 μm	Electrodes for Biomedical Engineering	VisiJet FTX Green	850 °C final, 300 °C hold	(71)
	mm scale	Electrochemical Applications	Arkema Sartomer SR238, SR295	900 °C final, 30 and 400 °C holds, 2 °C/min ramp	(72)
Digital light processing (DLP)	133 μm	High Capacitance MnO2 Supercapacitors	Elegoo	1000 °C final, 900 °C CO2 hold	(37)
	200 μm	Polymer electrolyte membrane fuel cells	Elegoo	900 °C final, 10 °C/min ramp	(73)
	200 μm	Structural Energy Storage Applications	Elegoo	1000 °C final, 400 °C hold, 10 °C/min ramp	(74)
	33 μm	Sodium Ion Batteries for Energy Storage	Elegoo	1000 °C final, 400 °C hold, 10 °C/min ramp	(75)
	117 μm	Microbial Fuel Cell	Miicraft 2005T	800 °C final, 350 and 450 °C hold, 1.5–3 °C/min ramp	(15)
	28 μm	Electrodes for Energy Storage	PR48	1000 °C final, 300 and 400 °C hold, 5 °C/min ramp	(32)
	106 μm	High Strength Structural Applications	PR48	1000 °C final, 300 and 400 °C hold, 10 °C/min ramp	(76)
	60 μm	High Strength Structural Applications	PR48	1000 °C final, 300 and 400 °C hold, 10 °C/min ramp	(13)
	300 μm	Microsupercapacitors for Redox Reactions	Allnex Ebecryl 265	430 °C final, 150 °C hold	(77)
Projection micro stereolithography (PμSL)	∼10 μm	Aerogel foams, energy storage, separations, catalysis	PEGDA, bisphenol A ethoxylate dimethacrylate (BisA-EDMA), graphene oxide	1050 °C final, 1 °C/min ramp	(52)
	25 μm	Lithium-Ion Batteries for Energy Storage	PR48	1000 °C final, 300 °C air hold, 400 and 600 °C N2 hold, 2 °C/min ramp	(19)
Continuous liquid interface production (CLIP)	15 μm	Energy materials	PR48	800 °C final in N2, 300 °C hold in air, 5 °C/min ramp	This work (see Section 6)

3D printing techniques such as DIW and FFF are typically used to generate samples with features on the millimeter scale. Several of these extrusion-based 3D printing studies utilize biomass, such as cellulose-based compounds, due to evidence of these polysaccharide biomass materials generating high yield pyrolytic carbon outputs^44,60,78 However, structures produced using biomass components alone appear relatively granulated and inhomogeneous, and it has been reported that 3D printed biomass carbon precursors do not demonstrate degrees of graphitization as high as their synthetic polymer counterparts.⁷⁹ To bypass this challenge, various additives have been mixed into biomass-containing melt extrusion 3D printing materials for enhanced electrical properties and stability, for example metal organic frameworks (MOFs)⁶⁰ and graphene oxide (GO) nanoparticles.⁴⁵

Alternatively, stereolithography 3D printing methods have been used more prominently in conjunction with pyrolysis, due to the ability of these light-based printing techniques to achieve micrometer and nanometer feature sizes conducive to isotropic shrinkage and maintaining structural integrity during pyrolysis. DLW-2PP has been used to achieve the highest resolution features, down to 100 nm,⁶⁴ but method is magnitudes slower (10^–9–10^–3 mm³/s) than projection stereolithography methods such as PμSL (1 mm³/s) and CLIP (10³ mm³/s).⁵⁷ Most DLW-2PP studies have primarily used acrylate-based Nanoscribe resins for high strength, biosensing, and optics applications. Several of these investigations only conducted low temperature heat treatments (<600 °C) which begins the pyrolytic restructuring mechanisms of constituent polymeric materials but does not facilitate significant levels of carbon graphitization and conductivity, instead relying on additives such as MOFs⁶² and zirconium⁶¹ or electroforming after pyrolysis.⁶³ Other DLW-2PP studies investigating higher temperature pyrolysis for structural applications achieved among the highest strength values observed for pyrolyzed 3D printed structures, for example Desponds et al.⁶⁴ reported up to 1 GPa for a pyrolyzed zirconia and acrylate-base material and Bauer et al.¹² reported reaching 3 GPa for an IP-Dip honeycomb structure. However, these pyrolyzed structures with extremely fine features and high strength produced via DLW-2PP were less than 100 μm wide in total length due to small build area capabilities, indicating that this method would be time-intensive for generating macroscale load-bearing structures.

For laser-based SLA, there is less consensus on commonly used resin precursors. Pan et al. demonstrated one of several instances of pyrolyzing SU-8 microstructures on silicon substrates, revealing a significant increase in anodic peak current and electrochemically active surface area (ECSA) for 3D printed electrodes versus 2D counterparts.⁶⁶ Toward electrode production, Steldinger et al. utilized a custom acrylate and aromatic vinyl-based carbon resin that undergoes CO₂ activation to achieve porous carbon structures with surface areas of 2200 m²/g.³⁴ Blyweert et al. combined acrylate components with mimosa tannin extract to produce pyrolytic carbon with thicker features (>4 mm) and mechanical strength of 215 MPa, but this system has not yet been shown to produce microscale features.⁷² Tian et al.⁷¹ used commercial VisiJet FTX Green resin to produce flexible carbon electrodes for skin biosensing applications for in vivo electromyography with areal capacitance of 8.9 mF/cm² and impedances at 1 kHz of around 2.5 kΩ, but they only produced flat structures that undergo significant warping and distortion.⁷¹

In the pattern projection light-based 3D printing space, DLP has been commonly used with commercial resin PR48, Elegoo, and other acrylate-based commercial resins to demonstrate feature sizes on a scale of 10–100 μm. Using PR48 resin, Narita et al. demonstrated the first instance of using DLP and pyrolysis to produce battery electrodes, achieving 28 μm carbon feature sizes, compressive strength of 27 MPa and specific strengths of 101 kN m/kg.³² Kudo et al. produced high strength pyrolytic carbon out of PR48 lattice structures that achieve up to 389.57 MPa in compressive strength and 468.62 MPa cm³/g in specific strength.^13,76 Using a liquid-crystal version of DLP and Elegoo resin, Katsuyama et al. showed that 3D printed pyrolytic carbon electrodes can facilitate the swift diffusion of carrier ions as anodes for sodium ion batteries (SIBs) and that reducing the diameter of the microscale lattice beams significantly improves rate performance and specific capacity.⁷⁵ They achieved SIB active material mass loading of 98 mg/cm² and areal capacity of 21.3 mAh/cm².⁷⁵ Kudo et al. included magnesium-oxide nanoparticles as porogens in Elegoo resin to generate carbon with hierarchical porosity including macropores (∼5 μm), mesopores (∼50 nm), and micropores (∼1 nm) with areal capacitance of 10 F/cm² and gravimetric capacitance of 105 F/g but lower mechanical strength compared to samples without MgO.⁷⁴ Large area PμSL, another recent VP method, has been used by Hensleigh et al. to produce microarchitected graphene aerogels from a custom acrylate resin with features sizes on the order of 10 μm with 60 nm pores that contributed to a surface area of 130 m²/g.⁵² Reale Batista et al. used large area PμSL with PR48 to generate electrodes for lithium-ion batteries.¹⁹ The large area and high resolution capabilities offered by PμSL particularly make it a compelling 3D printing method for carbon microstructure generation. Additionally, Section 6 of this perspective demonstrates the use of highly scalable CLIP VP alongside pyrolysis of PR48 resin to produce pyrolytic carbon lattices with 15 μm beam diameters.

These investigations advance the understanding of structural and material considerations for generation of pyrolyzed 3D printed structures. Of the methods overviewed, VP processes benefit most from inherent scalability enabling mass production of 3D pyrolytic carbon designs at length scales from cm to μm. Ultimately, for all 3D printing methods described herein, more promise arises with custom designing a resin tailored for optimal pyrolysis outputs rather than using existing commercially available resins not initially designed for pyrolysis.

3. Molecular Design Strategies for Polymeric 3D Printing Resins

A variety of commercial resins have been used in 3D printing and pyrolysis processes to produce 3D carbon structures; most are either acrylate or epoxy-based (Table 2). Acrylate monomers are commonly used due to their affinity for undergoing radical polymerization during AM.⁸⁰

Table 2. Composition and Characteristics of the Commercial 3D Printing Resins That Have Been Used in Tandem with Pyrolysis to Generate Conductive Pyrolytic Carbon Structures.

Commercial Resin Name	Commercial Resin Composition (Including Percentages When Provided)	AM Methods	Highest Resolution Pyrolyzed Feature	References
PR48	Acrylate-based: Allnex Ebecryl 8210 (39.8 wt %), Genomer 1122 (19.9 wt %), Sartomer SR 494 (39.8 wt %), 2,4,6-Trimethylbenzoyl-diphenylphosphineoxide (TPO, 0.4 wt %), Mayzo OB+ UV blocker (0.16 wt %), and undisclosed aliphatic urethane acrylate	DLP, PμSL, CLIP	15 μm	(32, 13, 76, 19), this work
Elegoo	Acrylate-based: trimethylcyclohexyl acrylate (53%), Isopropylidenediphenol (26%), (5-ethyl-1,3-dioxan-5-yl)methyl acrylate (14%), Ethyl phenyl (2,4,6-trimethylbenzoyl)phosphinate (5%), titanium dioxide (1.5%), carbon black (0.5%)	DLP	33 μm	(37, 74, 75)
Kayaku Advanced Materials SU-8	Epoxy-based: epoxy novolac resin with average epoxide group functionality of 8 (no wt % provided)	SLA	2 μm	(17, 66, 67)
IP-Dip (Nanoscribe GmbH)	Acrylate and polyether-based: aliphatic alcohol acrylate (60–80%), alicyclic hydrocarbon with acrylate and polyether components (<24%), additional hydrocarbon acrylate (<24%)	DLW-2PP	250 nm	(36, 63, 65, 12)
Boston Micro-Fabrication HTL yellow resin	Acrylate-based: Tris[2-(acryloyloxy)ethyl] isocyanurate (5–40%), Acylic Ester (20–80%), 4-(1-oxo-2-propenyl)-Morpholine (10–30%)	SLA	15 μm	(68)
IP-S (Nanoscribe GmbH)	Acrylate-based: methacrylate carbamate mixture (>95%), butyrolactone (<5%), aromatic ketone photoinitiator (<1%)	DLW-2PP	100 μm	(58)
VisiJet FTX Green	Acrylate-based: Triethylene glycol diacrylate (40–55%), Tricyclodecane dimethanol diacrylate (15–25%), TPO (1.5–2.5%)	SLA	500 μm	(71)
PIC100 (EnvisionTEC)	Acrylate-based: Methacrylated oligomer (50–90%), methacrylated monomer (5–40%), phosphine oxide photoinitiator (1–2%)	SLA	200 μm	(69)
Miicraft 2005T	Acrylate-based: acrylate compounds (90–98 wt %), photoinitiator (2–10 wt %), additives (0.2–5 wt %); no further details provided	DLP VP	117 μm	(15)

Several commercially available resins, such as PR48,^19,32,36 EnvisionTEC PIC100,⁶⁹ and Elegoo resins,³⁷ yield conductive carbon after pyrolysis, but they typically undergo substantial linear shrinkage of up to 70–90%. Achieving higher carbon yields leads to less shrinkage, which is desirable for material retention. Besides the challenges associated with geometry retention and char yields, pyrolytic carbon derived from commercially available resins often exhibits only modest electrical conductivity.^1,15,19,81 Modifications such as incorporating graphene oxide¹⁹ and tethering aromatic polymer precursors⁸² have been made to commercial resins to increase carbon content after pyrolysis, but these systems still have extraneous components in their composition that do not contribute to the carbon output.

Insights gained from the compositions of these commercial resins have been used to inform the development of custom resins designed specifically for high char yield pyrolysis output. (Table 3). The pyrolytic carbon structures resulting from these custom resin systems vary in shape and pore morphologies (Figure 3). We have grouped the structures into those that demonstrate high char yields ≥50 wt % (Figure 3A–D), those that report mechanical strength >100 kPa (Figure 3D,E), and those with high resolution features <100 μm (Figure 3F,G). Additionally, we highlight systems that incorporate important resin components including polyacrylonitrile (Figure 3H,I), biomass (Figure 3J–L), and MOFs (Figure 3M,N).

Table 3. Custom 3D Printing Resins Developed for Pyrolysis to Generate Conductive Carbon Structures^a.

Highest Resolution Feature	Custom Resin Components with Percent Composition (wt %) when Provided	3D Printing Method	Char Yield	Application	Reference
250 μm (Figure 3A)	· Cellulose nanocrystals (162.5 g/mL suspension in water; 67.9%)	DIW	50%	Energy Storage	(60)
	· ZIF-8 metal–organic framework (MOF) (32.1%)
500 μm (Figure 3B)	· Pyromellitic dianhydride with 4,4′-oxidianiline (PMDA-ODA) (18–24%)	DIW and SLA	75%	Energy applications	(82)
	· TPO photoinitiator (2.5%)
	· N-methyl-2-pyrrolidone solvent (remainder of composition)
mm scale (Figure 3C)	· Poly(4-vinylphenol)-plutonic F127 in a 2:1 by mass ratio (36.3%)	DIW	58%	Gas Separation, adsorption of CO2	(59)
	· Activated carbon (48.5%)
	· Bentonite binder (15.1%)
>100 μm (Figure 3D)	· Phthalonitrile (55 wt %)	PμSL	64%	Energy, electrochemistry, and aerospace	(83)
	· Bisphenol A ethoxylate diacrylate (BisA-EDA) (small undisclosed amount)
	· DMF solvent (remainder)
135 μm (Figure 3E)	· Pentaerythritol tetraacrylate (PETA) (1:1 volume ratio with DVB, with the monomer mixture being 50 vol % total)	DLP	48%	Energy storage, catalysis, and adsorption	(34)
	· Divinylbenzene (DVB) (1:1 volume ratio with PETA)
	· DEP porogen (50 vol %)
10 μm beams, 60 nm pores (Figure 3F)	· Bisphenol A ethoxylate dimethacrylate (BisA-EDMA) (Cumulative 12 wt % with PEGDA)	PμSL	40–50%	Energy storage, separations, catalysis	(52)
	· Polyethylene glycol diacrylate (PEGDA) (Cumulative 12 wt % with BisA-EDMA)
	· Graphene oxide (GO) (1 wt %)
	· Photoinitiator (2–4 wt %)
	· N,N-Dimethylformamide (DMF) solvent (remainder of composition)
25 μm (Figure 3G)	· Trimethylolpropane triacrylate (TMPTA) (95 wt %)	PμSL	17%	Energy storage	(19)
	· GO (3 wt %)
	· Phenylbis (2,4,6-trimethylbenzoyl) phosphine oxide (2 wt %)
mm scale (Figure 3H)	· Polyacrylonitrile (PAN) (90 mol %)	FFF	77% by volume	Energy Storage	(43)
	· Solketal acrylate (SKA) (10 mol %)
	· DMSO and methanol solvents
500 μm (Figure 3I)	· Poly(n-butyl acrylate) with trithiocarbonate terminus (PBA-CTA) (28.8 wt %)	LCD VP	40%	Energy Storage	(84)
	· Acrylonitrile (54.9–66.4 wt %)
	· TMPTA (0.8–15.3 wt %)
	· TPO photoinitiator (1.0 wt %)
	· Sudan II UV blocker (0.03 wt %)
700 μm (Figure 3J)	· Microfibrillated cellulose hydrogel (1.6%)	FFF	Up to 50%	Energy Storage	(44)
	· Lignosulfonate (76.6%)
	· Cellulose powder (CP) from cotton (21.8%
400 μm (Figure 3K)	· Hydroxypropyl methylcellulose (5%)	DIW	No mention	Energy Storage	(45)
	· GO (40 mg/mL suspension in water, 95%)
500 μm (Figure 3L)	· Starch-based packaging waste (20%)	DIW	No mention	Energy Storage	(85)
	· Sodium hydroxide (80%)
700 μm (Figure 3M)	· 2-Phenoxyethyl acrylate (41.4 wt %)	SLA	11%	Catalysis	(70)
	· TMPTA (41.4 wt %)
	· ZIF-8 MOF (5 wt %)
	· Ni-BTC MOF (5 wt %)
	· Photoinitiator (7.2 wt %)
mm scale (Figure 3N)	· Mimosa tannin extract (25 wt %)	SLA	22%	Electrochemistry	(72)
	· Acrylate resins (HDDA, PETA) (74.7 wt %)
	· ZIF-8 MOF (0.3 wt %)

^aResin compositions are listed as reported or calculated from methods sections.

Figure 3.

Carbon structures produced via 3D printing of custom polymer systems followed by pyrolysis corresponding to the Table 3 studies. These structures are grouped into high char yield samples (green), samples reporting mechanical strength (orange), samples with high resolution features (blue), and samples produced from polyacrylonitrile-containing resins (gold), biomass-containing resins (red), and MOF-containing resins (purple). These optical and SEM images are reproduced (adapted) with permissions when necessary from (A) Catarineu et al.⁶⁰ Copyright 2023 Elsevier, (B) Arrington et al.⁸² Copyright 2021 American Chemical Society, (C) Comroe et al.⁵⁹ CC BY, (D) Lu et al.⁸³ Copyright 2024 Elsevier, (E) Steldinger et al.³⁴ CC BY, (F) Hensleigh et al.⁵² CC BY-NC, (G) Reale Batista et al.¹⁹ Copyright 2023 Elsevier, (H) Usselmann et al.⁴³ CC BY, (I) Bobrin et al.⁸⁴ CC BY, (J) Shao et al.⁴⁴ Copyright 2018 Elsevier, (K) Yao et al.⁴⁵ Copyright 2019 Elsevier, (L) Idrees et al.⁸⁵ Copyright 2020 Elsevier, (M) Cherevko et al.⁷⁰ CC BY, and (N) Blyweert et al.⁷² CC BY.

Developing an effective formulation for 3D printing polymeric structures that can be pyrolyzed to graphitic carbon with high-resolution features (<10–100 μm), high char yield (>50 wt %), good conductivity (>1 S/cm), and structural integrity (strength >100 kPa, no distortion) relies on intentionally tailoring the molecular structure of the constituent components. Polymer components used in literature to achieve high char yields have primarily consisted of functional groups that either contain 6-membered carbon rings, for example aromatics, or stabilize via cyclization to 6-membered carbon rings during the pyrolysis process, for example nitrile groups. Aromatics, such as phenyl or naphthyl groups, provide stability to the carbon structure and are less likely to decompose into volatile compounds, resulting in a higher char yield.⁸⁶ Nitrile groups can stabilize the polymer structure during pyrolysis by participating in the formation of stable nitrogen-containing cyclic compounds during the heating procedure.⁸⁷ This mitigates the decomposition of polymer into volatile gases during pyrolysis, increasing char content.

Several components that have been incorporated into 3D printed polymer formulations that have shown promise for pyrolysis to carbon are included in Figure 4. These molecular structures have been separated into acrylate-based cross-linker monomers typically used in VP 3D printing, monomers that increase pyrolysis char yield, comonomers that improve formulation printability, and nonpolymerizing additives that increase char yield.

Figure 4.

Molecular structures of several components used in 3D printing polymer resins designed for improved pyrolysis output.

3.1. High Char Yield

Focusing first on high char yield resins, often these materials are highly cyclic compounds such as biomass-containing materials or polyimides, often with additives included to bolster mass yield. Catarineu et al. developed a cellulose-based resin for DIW with ZIF-8 MOF additives to produce a carbon cathode with >50 wt % char yield for a zinc-ion hybrid supercapacitor (Figure 3A).⁶⁰ After pyrolysis at 1000 °C, this material underwent graphitization and achieved surface areas of 660 m²/g, but demonstrated low-resolution lattice features due to the DIW nozzle being 600 μm. Other biomass-containing studies that did not report char yield are discussed later. Arrington et al. achieved a char yield of 75 wt % by developing an SLA resin with polyimide pyromellitic dianhydride with 4,4′-oxidianiline (PMDA-ODA), poly(amic acid) salts, and solvent pyrolyzed at 1000 °C to produce carbon with high conductivity over 1 S/cm (Figure 3B).⁸² During Raman spectroscopy, their rigid 3D organogels displayed D and G bands characteristic of polycrystalline pyrolytic carbon.⁸² Comroe et al. achieved a high char yield of 58 wt % by incorporating activated carbon particles into poly(4-vinylphenol) in N,N-dimethylacetamide solvent.⁵⁹ After FFF and pyrolysis at 800 °C, this formulation generated structures that demonstrated viability for gas separation applications but were printed with mm-scale resolution and underwent significant deformation (Figure 3C). Lu et al. printed structures with features on the 100 μm scale that achieved 62% char yield using phthalonitrile, containing both aromatic and cyclization-prone nitrile groups, along with cross-linker BisA-EDA via PμSL (Figure 3D).⁸³

3.2. Mechanical Integrity

Looking to mechanical strength, the phthalonitrile structures printed by Lu et al. also demonstrated storage modulus of 3.7 GPa and flexural strength of 156 MPa.⁸³ Steldinger et al. produced hierarchically structured porous carbon (Figure 3E) strong enough to withstand crushing pressures up to 100 kPa via custom resin formulations with monomers that included PETA, DVB and PEGDA and dibutyl phthalate as a porogen. Due to its high degree of aromaticity, DVB enabled char yields up to 48% at 900 °C, but electrochemical characterization was not reported and the resin required long solidification times of 240–420 s per layer.³⁴ Limited studies report mechanical properties, indicating a need for the field to generate tougher structures and more widely report mechanical characterization.

3.3. High Resolution Features

PμSL VP printing was used in two custom resin studies to achieve high resolution features <100 μm. Hensleigh et al. developed a resin that combines BisA-EDMA, which alone pyrolyzes to brittle char, and PEGDA (700g/mol), which alone produces a mechanically weak gel, in a 1:1 weight ratio in DMF solvent to produce a robust carbon aerogel after pyrolysis (Figure 3F).⁵² Their microarchitected graphene aerogel lattices demonstrated beam diameters on the order of 10 μm and attained mechanical elastic modulus normalized by material density comparable to graphene.⁵² Reale Batista et al. incorporated GO sheets into TMPTA resin that they printed using PμSL and pyrolyzed in N₂ up to 1000 °C after several isothermal holds (Figure 3G).¹⁹ These GO/TMPTA lattices achieved significantly greater gravimetric capacitance than analogous commercial PR48 samples. PμSL and similar high resolution VP 3D printing techniques have demonstrated the most promise for achieving desired high-resolution features.

3.4. Polyacrylonitrile-Based Resins

Looking to the carbon fiber industry history (Figure 1), one of the most widely used materials in pyrolysis is PAN.⁸⁸ The pyrolysis of PAN to conductive carbon via cyclization of linear chains has been extensively studied and is promising for use in conductive electrode applications.^87,89 In particular, there has been a recent push in late 2023 and 2024 for inclusion of polyacrylonitrile into new 3D printing resin systems for pyrolysis. Usselmann et al. demonstrated a FFF method copolymerizing PAN with solketal followed by pyrolysis to generate open mesh carbon electrodes for supercapacitors (Figure 3H).⁴³ However, this melt extrusion method printed features on the mm scale rather than μm scale, which limits the achievable high end S_V. A study by Chazot et al. introduced a new strategy for 3D printing PAN using an interfacial vat photopolymerization method but has not yet been combined with pyrolysis.⁹⁰ In 2024, Bobrin et al. demonstrated generation of pyrolyzed PAN-based structures via VP 3D printing for the first time (Figure 3I), using a solvent-free microphase separated resin comprising of a sacrificial polymer poly(n-butyl acrylate) with a trithiocarbonate terminus, TMPTA, and carbon precursor acrylonitrile.⁸⁴ After pyrolysis at 1200 °C, they produced structures with mm-scale printed features and nanoscale porous texture that achieved conductivity of 0.025 S/cm.⁸⁴ Bobrin et al. have also used this pyrolyzed 3D printed PAN microphase chemistry in a recent carbon-ceramic system.⁹¹ However, no existing VP literature demonstrates generation of PAN structures with <100 μm printed features, which is a key target area for future work due to the affinity for PAN for pyrolysis of highly graphitic carbon.

3.5. Biomass-Based Resins

Several more studies include biobased components, for example, Shao et al. incorporated cellulose and lignosulfonate into FFF filament to generate electrode disks with 700 μm features (Figure 3J)⁴⁴ and Yao et al. incorporated hydroxypropyl methylcellulose into a graphene oxide DIW ink to produce cubic lattice electrodes with 400 μm beams (Figure 3K).⁴⁵ The Shao et al. carbon structures demonstrated electrical conductivity of 47.8 S/cm.⁴⁴ Alternatively, Idrees et al. produced FFF resin consisting of starch-based packaging waste with sodium hydroxide for solid-state supercapacitor electrodes (Figure 3L).⁸⁵

3.6. MOFs and Other Nonpolymerizing Additives

In addition to tailoring the polymer backbone, numerous studies have increased the char yield, porosity, surface area of their pyrolytic carbon structures by incorporating nonpolymerizing additives into the resins, for example MOFs. In 3D printing and pyrolysis literature, the MOF usually incorporated is zeolitic imidazolate framework ZIF-8 due to its low cost, water stability, and compatibility with zinc ion capacitor systems.⁶⁰ Blyweert et al. added ZIF-8 MOF into an acrylate-based resin comprised of HDDA and PETA mixed with mimosa tannin extract, yielding low density carbon structures after pyrolysis at 900 °C that exhibit electrical conductivity of 7 S/cm (Figure 3N).⁷² Blyweert et al. later improved this approach to achieve a 27% char yield by integrating vanillin with tannin into the acrylate polymer network, which improved thermal stabilization and decreased the depolymerization of free chains in the polymerized resin.⁴⁰ Cherevko et al. incorporated ZIF-8 and the nickel-based Ni-BTC MOFs into TMPTA and 2-phenoxyethyl acrylate resin for SLA printing of structures for nickel-based catalysis.⁷⁰ After pyrolysis at 950 °C, their structures (Figure 3M) demonstrated isotropic shrinkage, surface areas of 145 m²/g, and magnetic coercivity of 36 Oe, on the order of nickel metal. Catarineu et al. also utilized ZIF-8 as described above and showed via elemental analysis that the zinc introduced by ZIF-8 additives volatilizes at temperatures greater than 1000 °C.⁶⁰ Together, these studies have shown that MOFs can enhance the porosity, surface area, and electrochemical properties of 3D-printed carbon structures.

Outside of MOFs, other nonpolymerizing additives added to custom 3D printing resins have included porous activated carbon particles^34,59,85 and GO⁶⁹ to enhance conductivity. Hensleigh et al. and Reale Batista et al. both incorporated GO into their PμSL resins as mentioned above to enhance the porosity and gravimetric surface area of their structures (Figure 3E-F).⁵² Yao et al. incorporated GO into their hydroxypropyl methylcellulose resulting in graphene aerogel structures with surface area of 100 m²/g after pyrolysis at 1050 °C with high loading capacity for MnO₂ active material at 182.2 mg/cm² (Figure 3J).⁴⁵ The choice of additive to complement polymer resin composition plays a pivotal role in the outputted surface area and strength of a material. Commonly, these additives add to the porous nature of the outputted carbon, increasing material surface area and resulting conductivity, but decreasing the mechanical strength.

4. Pyrolysis Conditions and Geometry Parameters

During pyrolysis of polymeric structures, gaseous byproducts such as CO, CO₂, hydrocarbons, and H₂O^65,92−94 are formed and liberated at high temperatures, leaving behind a carbon skeleton comprised of fused ring structures of sp² hybridized carbon commonly accompanied by disordered defects and sp³ hybridization.^36,95,96 The molecular behavior during the pyrolysis process can be categorized into three stages: monomer evaporation and cyclization, chemical decomposition, and carbonization.⁴⁹ This process must be controlled to maximize the graphitic yield for high conductivity and robust mechanical integrity, and to minimize geometric distortions. Several polymer pyrolysis methods exist beyond traditional tube furnace pyrolysis, such as fast joule heating,^13,97 rapid thermal processing via IR lamps,⁸⁶ microwave-assisted pyrolysis,⁹⁸ and inductive heating.¹¹ These nontraditional methods involve rapid heating to a final temperature and are commonly used in the pyrolysis of biomass and waste plastic, where retention of a specific part geometry is not necessary. The focus of this section is therefore limited to the parameters of conventional tube furnace pyrolysis, namely temperature ramp rates, maximum pyrolysis temperature, chamber pressure, and atmosphere, and their effect on the properties of pyrolytic carbon generated from 3D printed polymer resins.

Polymer degradation to carbon usually occurs between 300 and 550 °C, as volatile byproducts form and diffuse out, leaving behind a disordered carbon-rich framework.^36,96,99,100 Rapid heating rates can cause the formation of byproducts to outpace their diffusion out of the polymer network, causing pockets of gas to become trapped in the bulk or on the surface of the carbon matrix (Figure 5A).^96,101 Fast heating rates also prevent adequate relaxation of internal stresses developed in the polymer during the printing process, contributing to structural distortions in the printed parts.^19,36,102

Figure 5.

Pyrolysis condition effects on structure including (A) pyrolyzed SU-8 material after undergoing heating rates of (i) 25 °C/min, (ii) 50 °C/min, and (iii) 75 °C/min demonstrating more gas pockets forming at higher ramp rates (Reproduced with permission from Sharma reproduced with permission from Sharma et al.¹⁰¹ Copyright 2014 Elsevier), (B) beams of higher thicknesses, lower S_v, and high vacuum (HV) experiencing less warping during pyrolysis than smaller thicknesses and low vacuum (LV)³⁶ (CC BY-NC), and (C) the temperature profile of a furnace run with isothermal holds⁸¹ (CC BY-NC-ND).

Key structural changes impacting electrical conductivity and mechanical strength of pyrolytic carbon are determined by the final pyrolysis temperature and dwell time. Cleavage of hydrogen from carbon, heteroatom liberation from the carbon matrix, the cyclization of nonaromatic components, and compactification and rearrangement of C–C bonds in sp²-hybridized domains during the pyrolysis are all impacted by the maximum temperature and dwell time.^16,95,96,102 Between 550–700 °C, after the primary thermochemical degradation of the polymer, the remaining disordered carbon material contains a high fraction of unsaturated dangling bonds and active radicals.¹⁰³ At temperatures above 700 °C, hydrogen, nitrogen, and any remaining oxygen are expelled and the aromatic network becomes interconnected, increasing the stacking thickness and in-plane crystallite size of the newly formed graphene-like structures.^99,100,104 Pyrolysis at temperatures above 800 °C has been observed to increase the modulus, hardness, and electrical conductivity compared to pyrolysis at lower temperatures due to enabling increased carbon content.^16,96 For example, after pyrolysis of SU-8, at 900 °C, the elemental composition of the material reports 90 wt % carbon content by mass whereas when the pyrolysis temperature is increased to 1300 °C, the carbon content reaches 99 wt %.¹⁰⁰ The formation of gaseous byproducts and the coalescence of the aromatic network are primarily responsible for the large volume changes observed in pyrolyzed polymers.

The chamber pressure and atmosphere during pyrolysis affects both the rate of outgassing of byproducts, which can cause second-order reactions with the carbon backbone, and the rate of heteroatom diffusion within the polymer during the polymer-carbon conversion.^36,99 Vacuum conditions have been shown to enable the formation of higher molecular weight byproducts compared to ambient pressure.¹⁰⁵ Low chamber pressure in vacuum pyrolysis enhances the transport of volatile byproducts within the polymer, allowing their evacuation to occur at lower temperatures resulting in larger linear shrinkage (Figure 5B).⁹⁹ Samples undergo slower carbonization in inert atmospheres, such as argon (Ar) or nitrogen (N₂), compared with vacuum pyrolysis, as the slower outgassing of byproducts hinders the conversion of polymer to carbon. The use of inert carrier gases such as argon or nitrogen are common to avoid oxidation of carbon. Slower carbonization also occurs in an air or oxygen-containing atmosphere, which can also generate an oxygen-rich skin on samples pyrolyzed at high pressures. Compared with those pyrolyzed under vacuum conditions, the residence time for adsorbed oxygen-containing byproducts in samples under pressure is higher, increasing the probability of second-order reactions.³⁶ In nonvacuum pyrolysis, using a gas with a high heat transfer coefficient, such as hydrogen (H₂), more readily transfers heat to the polymer, increasing the rate of byproduct formation which causes higher shrinkage and mass loss.⁹⁹

To overcome challenges associated with geometric distortion during the polymer carbonization, researchers often include one or more isothermal stages at temperatures corresponding to significant polymer degradation prior to the final pyrolysis temperature (Figure 5C).^{12,13,19,32,34} This allows for the controlled outgassing of byproducts while the material is still malleable prior to carbonization. Slow ramp rates, typically between or below 1–5 °C/min, are preferred to reduce geometric distortion and mass loss.^{12,19,32,35,81,106} Since electrical conductivity is enhanced by an interconnected network of sp²-hybridized carbon, final pyrolysis temperatures are typically above 900 °C with dwell times of 1 h or longer in inert or vacuum environments (Table 4). When reporting conductivity of carbon structures, most studies focus on device metrics and less on material properties of the carbon itself, as evidenced by less than half of the studies in Table 4 reporting material electrical conductivity. It is difficult to compare material properties when only device metrics are provided, so we recommend that the future papers in this field be more diligent about reporting carbon conductivity via electrochemical impedance spectroscopy.

Table 4. Overview of Pyrolysis Parameters and Furnace Conditions Used in Combination with Resins Used in Conductive Carbon Production.

Resin	Ramp Rate (°C min–1)	Pyrolysis Temperature (°C)	Pyrolysis Atmosphere	Pyrolysis Time	Char Yield (wt %)	Linear Shrinkage (%)	Electrical Conductivity	Reference
Commercial Resins
IP-DIP	3	900	Vacuum	1 h	20%	80%	Not Reported	(12)
	1200	450–550	Vacuum	Not Reported	Not Reported	70%	Not Reported	(36)
	1200	450–550	N2	Not Reported	Not Reported	55% - 68%	Not Reported	(36)
Elegoo	Not Reported	1000	Vacuum	2 h	Not Reported	66% - 67%	Not Reported	(37)
PR48	2	190 postcure; 300 in Air; 1000 in N2	N2	1 h; 3 h; 1 h	16%	30%	5 Ω (EIS)	(19)
	10	1000	Vacuum	4 h	∼4%	66%	Not Reported	(13)
	5	1000	Vacuum	4 h	∼5%	66%	Not Reported	(32)
	5	300 in Air; 800 in N2	N2	3 h; 2 h	16%	65%	Not Reported	This work; see Section 6
PR48 with titanium(IV) ethoxide	2	1000	Ar	1 h	25.80%	39%	Not Reported	(108)
FullCure 705	1	1000	N2	2 h	16.5% - 22.3%	N/A	Not Reported	(35)
	1	1000	N2	2 h	15.2% - 19.7%	N/A	Not Reported	(35)
Formlabs HTR	10	1000	N2	1 h	7%	60% - 62%	25 Ω (From EIS)	(81)
Clear 2005T	0.4–3	800	N2	2 h	7%	43%	23 Ω (From EIS)	(15)
SU-8	5	1200	Vacuum	N/A	Not Reported	N/A	Not Reported	(104)
Custom Resins
GO/TMPTA	2	1000	N2	1 h	17%	19%	6 Ω (From EIS)	(19)
CN154/PETA/HDDA/BAPO	1–2	900	N2	1 h	28%	25%	Not Reported	(40)
	1.5–2.5	900	N2	1 h	21%	25%	7 S cm–1	(39)
	1.5–2.5	900	N2	1 h	11.70%	40%	Not Reported	(39)
PEGDA/PETA/DVB	3.3	900	N2	15 min	48%	35%	Not Reported	(34)
GO/PEGDA/BisA-EDMA	1	1050	N2	3 h	40% – 50%	50% – 60%	64 S m–1	(52)
ZIF-8/2-Phenoxyethyl acrylate/TMPTA	3	950	H2, 7.03 vol % Ar	5 min	11%	50%	Not Reported	(70)
Poly(4-vinylphenol)/ Pluronic F127/activated carbon/bentonite	10	900	N2	N/A	58%	N/A	Not Reported	(59)
PAN/Solketal acrylate	0.5–5	1000	N2	30 min	Not Reported	23%	∼9 Ω (From EIS)	(43)
GO/hydroxypropyl methylcellulose	2	1050	N2	3 h	Not Reported	Not Reported	∼1.2 Ω (From EIS)	(45)
ZIF-8/cellulose nanocrystals	5	1000	N2	2 h	50%	<10%	∼5 Ω (From EIS)	(60)
Microfibrillated cellulose/lignosulfonate/cellulose powder	0.2–1	600–1000	N2	10 min	40% – 45%	41% – 48%	47.8 S cm–1	(44)

Due to the dependence of shrinkage on the volatilization of side products, the S_V of printed parts becomes important. Higher S_V enables facile mass transport of volatile gases out of the polymer network due to corresponding with thinner diffusion pathways, thus higher S_V typically corresponds to more rapid and higher degrees of shrinkage.^36,107 However, Sun et al. observed that while this trend is consistently observed in low and no vacuum systems, the dependence of part shrinkage on S_V was not observed under high vacuum systems on a single-digit micron scale when using DLW-2PP 3D printing of IP-Dip.³⁶ Under high vacuum, beams with features between 1 and 4 μm exhibited similar shrinkage rates regardless of S_V (for example linear shrinkage of 45% at 450 °C and 30% at 500 °C). Conversely, under low vacuum (nitrogen atmosphere), beams with higher S_V underwent greater shrinkage, for example at 500 °C, beams with diameters of 1 μm, 2 μm, 3 μm, and 4 μm (decreasing S_V) underwent 32%, 39%, 42%, and 45% linear shrinkage.³⁶ Further investigation must determine whether this trend continues for larger features sizes (≥100 μm scale), as more findings like these would enable nuanced control over outputted pyrolytic carbon feature sizes.

4.1. Future Outlook on Catalytic Graphitization of 3D Printed Structures

To utilize 3D printed carbon structures as conductive electrodes for energy storage and conversion, they must offer excellent conductivity and mechanical integrity. Pyrolysis to achieve graphitization is a critical step for achieving these properties. However, the conversion of carbonaceous precursors to graphitic carbon occurs at very high temperatures (>3000 °C).^109−111 One method for avoiding high temperatures that require significant energy to maintain is through catalytic graphitization, which is when polymers are converted with catalyst assistance into carbon nanomaterials, such as carbon nanotubes^112−115 and graphene,^116,117 through dissolution and precipitation mechanisms. Catalysts used in catalytic graphitization of polymers serve two functions: (1) facilitating effective degradation and volatilization of noncarbon elements and (2) enabling arrangement of graphitic carbon on the structure.^118−120 Similar to carbon nanotube synthesis by chemical vapor deposition method, the catalysts for carbonization are usually transition metals, such as nickel,¹²¹ iron¹²² or metal compounds like nickel oxides,¹¹³ ferrocene¹²³ or ferrous chlorides.¹²⁴ Furthermore, conducting carbonization under high pressure environment can generate carbon materials with textured micro or nanostructures and a relatively high carbon yield compared to atmosphere carbonization.¹²⁵ Besides the selection of catalysts, other critical factors that may significantly affect carbonization results include heating rate of pyrolysis,¹²⁶ pressure¹²⁵ and the S_v of the structure. Incorporating catalytic graphitization into the pyrolysis of 3D printed polymers is a critical future focus for the field that has been underutilized to date to achieve desired graphitic carbon structures at lower energy costs.

5. Applications for Pyrolysis of 3D Printed Structures

3D printed carbon structures have the potential to address several critical global climate change challenges including as electrodes for energy applications such as renewable energy storage devices (Figure 6A), as conductive components supporting electrification of petroleum-heavy thermochemical processes, and as substrates for electrocatalysts supporting recovery of resources from waste or CO₂ capture. They also show merit in lightweight mechanical applications (Figure 6B) and biomedical applications like biochemical sensing (Figure 6C). Their high surface area, electrical conductivity, and tunable surface chemistry make them ideal candidates for addressing pressing global climate change mitigation challenges.

Figure 6.

Diagrams showing example uses for pyrolyzed 3D printed structures in (A) an energy storage application with a carbon lattice being assembled into a Li-ion coin cell (from Narita et al.³² Copyright 2021 John Wiley & Sons), (B) a structural application showing a carbon lattice undergoing 21.9% strain without breaking (from Crook et al.¹⁰ CC BY), and (C) a biomedical application specifically targeting customizable neurotransmitter detection implants (from Yang et al.⁵⁸ Copyright 2018 John Wiley & Sons).

5.1. Energy Storage

Intensive research has shown promise for using 3D architected electrodes in battery applications, as increasing electrode thickness enhances energy density.^1,127−130 Industry standard planar 2D electrodes are hindered by tortuous ion diffusion pathways, low electrical conductivities due to the inclusion of binders, and high mass fractions of inactive components like separators and current collectors.^37,131 3D architected carbon electrodes can improve upon 2D electrodes through high surface area (approaching graphene surface area of 2800 m²/g) and open structures to accommodate volume changes during cycling, enabling higher active material loadings (≥70–180 mg/cm²), high energy density (approaching 70 Wh/kg), and facilitating ion transport through electrolyte that mitigates the solid state diffusion limitations that often arise with high active material loadings.^{1,20,32,132,133}

VP 3D printing techniques, such as PμSL, DLP, and CLIP, expand the landscape for optimized geometries to overcome the increased ion transport lengths associated with thick, 3D electrodes. For example, Reale Batista et al. used simulation-driven geometry design to optimize topology of supercapacitor electrodes for energy density via PμSL and pyrolysis at 1000 °C in N₂ of PR48 and custom GO-containing TMPTA.¹⁹ Alternatively, Katsuyama et al. printed supercapacitor electrodes with hierarchical porosity from Elegoo using SLA followed by pyrolysis at 1000 °C under vacuum and CO₂ that they coated in MnO₂ and incorporated into an asymmetric cell demonstrating a wide potential window of 1.8 V in NaSO₄.³⁷ A similar approach was taken by Rezaei et al. to produce SLA printed 3D pyrolytic carbon/Mn₃O₄ hybrid supercapacitor electrodes that demonstrated a gravimetric capacitance of 186 F/g with 92% capacitance retention after 5000 cycles.¹⁰⁶

3D printed pyrolytic carbon can also serve as anodes in Li⁺- and Na⁺-batteries.^32,134 Commercial LIB anodes are typically fabricated via slurry casting of graphite with binders, which is limited by low electrical conductivity of the binder and tortuous ion diffusion pathways between active materials and binders.^1,127,135 In contrast, a monolithic 3D architected carbon electrode removes these factors, and 3D printing can enable uniform porosity and current distribution, which can aid in alleviating dendrite formation arising from locally high potentials and Li⁺ concentrations.³² For example, Narita et al. produced carbon-based 3D Li-ion battery anodes via pyrolysis of PR48 at 1000 °C, printed using a DLP technique (Figure 6A).³² Upon galvanostatic cycling, the initial charging process exhibited a Coulombic efficiency of 78%, which increased upon subsequent cycles to >99%. The electrode displayed a 25% capacity retention after 500 cycles at a high current density of 100 mA/g and their electrodes withstood a uniaxial compressive stress of 27 MPa, indicating their viability to withstand packaging pressures during commercial cell fabrication.

5.2. Electrocatalysis

3D printed carbon structures have been used as substrates for electrocatalysis due to their high specific surface areas enabling increased loading capacity of catalyst active material similar to energy storage application standards.^136,137 Peng et al.¹³⁸ utilized 3D printed graphene electrodes with hierarchical porous structures integrated with NiFeP nanosheets for water splitting, which exhibited high flexural strength and surface areas enabling a voltage range of 1.58 V at 30 mA/cm and improved ion transport rates for electrocatalysis compared to nonarchitected systems. Kou et al.¹³⁹ utilized 3D printed Ni electrodes with periodic structures to suppress gas bubble coalescence, jamming, and trapping. These electrodes demonstrated high areal current density of 1000 mA/cm² compared to 79.5 mA/cm² for commercial 3D Ni foam at an overpotential of 245 mV. 3D printed carbon materials can also be valuable for CO₂ conversion to valuable products due to their controlled structural and electronic properties and high CO₂ affinity.¹⁴⁰

5.3. Mechanical Applications

Pyrolytic carbon materials fabricated by 3D printing have also emerged as high-performance mechanical materials. A target goal is achieving compressive strength equal to or greater than the 4 GPa demonstrated by high strength carbon fibers.³¹ Bauer et al. fabricated honeycomb carbon nanolattices using DLW-2PP and pyrolysis that exhibit high mechanical strength, up to 3 GPa.¹² Unit cell design plays a critical role in high strength 3D carbon lattice production. An important consideration for mitigating distortion during pyrolysis is whether the lattice structure experiences stretching-dominated or bending-dominated deformation, as defined by Deshpande et al.¹⁴¹ Triangular cells are considered the most rigid and stretching-dominated unit cell type.¹⁴² Stretching-dominated lattices exhibit greater modulus and initial yield strength than bending-dominated lattices with the same relative density. This causes stretching-dominated structures to be more ideal for structural applications and isotropic shrinkage during pyrolysis conversion due to struts maintaining tension in line with the lattice nodes during load bearing, whereas bending-dominated structures are more ideal for energy absorption as their nodes bend away from notes toward the cell edges, causing softening.¹⁴¹ Additionally, closed cell configurations, such as shell or plate unit cells, can afford higher strength and stiffness compared to beam-based structures by enhancing the interconnectivity between unit cell, reducing structural entropy and increasing the storage of strain energy.¹⁴³ For example, Crook et al. 3D printed a closed cell cubic and octet nanolattice structure via DLW-2PP printing followed by carbonization at 900 °C that exhibited a compressive strength of 3 GPa at 0.792 g/cm³ and stiffness of 21.6 GPa, a > 600% improvement over analogous octet beam-based lattices (Figure 6B).¹⁰

5.4. Biomedical Applications

3D carbon structures have been incorporated into micro- and nano- electromechanical systems for biomedical in vitro applications, such as biosensing, dielectrophoretic¹⁴⁴ and cell-sorting or trapping^145−147 systems. 3D printed pyrolytic carbon structures are valuable for biomedical applications due to their biocompatibility in addition to their conductive nature, enabling incorporation into devices that are promising candidates for drug delivery¹⁴⁸ and biosensing for human-machine interfaces such as neural probes and health-monitoring implants.¹⁴⁹ For example, Mishra et al. 3D printed hollow carbon microneedles with SU-8 and pyrolyzed at 900 °C under inert conditions.¹⁴⁸ The carbon needles exhibited promising hardness, biocompatibility, and thermal and chemical stability, enabling penetration of mouse skin repeatedly without breakage. Moreover, 3D carbon structures have shown potential as scaffolds for tissue engineering for both neural¹⁵⁰ and musculoskeletal applications.¹⁵¹ For example, Islam et al. fabricated 3D carbon architectures by SLA 3D printing epoxy polymer followed by pyrolysis at 900 °C in nitrogen.¹⁵¹ The carbon microlattices exhibited an elastic modulus around 2.28 MPa, which is acceptable for use in human tissue repair applications. Furthermore, Fuhrer et al. utilized 3D pyrolytic carbon structures as platforms for culturing neural stem cells.¹⁵⁰ Their structures facilitated electrical stimulus to cells, allowing the observation of cell responses in a system that is compatible with magnetic resonance imaging.¹⁵⁰ Additionally, Hemanth et al. and Yang et al. (Figure 6C) both generated 3D carbon microelectrodes from SU-8 photoresist using two-photon lithography with electrochemical sensitivity that enabled neurochemical dopamine detection.^17,58

These advancements underscore the potential of these 3D microelectrodes in applications like sensitive electrochemical sensing and bioelectrochemistry, benefiting from high surface area and efficient electron transfer properties.

6. Combining CLIP and Pyrolysis

The DLP method of continuous liquid interface production (CLIP) has established itself as a 3D printing industry standard due to its exceptional combination of rapid printing (linear speed up to 17 mm/min) and smooth surface finish.^2,53,55 CLIP technology exploits a unique approach involving a continuous liquid interface, known as the dead zone, created by oxygen, which serves as a polymerization inhibitor that prevents adhesion to the window as a printed part is cured.⁵³ Recent advancements in high-resolution CLIP have additionally demonstrated the ability to mass produce high-resolution objects and resolve voxels as small as 2× 2 μm² in the print plane with 1 μm unsupported thicknesses.¹⁵² Integration with injection CLIP (iCLIP) has yielded microfluidics. iCLIP mitigates overcuring via injecting resin into void spaces during printing, enabling negative space microchannel diameters as small as 50 μm.^2,153,154 These advancements have positioned CLIP as an ideal candidate for manufacturing 3D polymer templates with complex geometries.

CLIP has not previously been used in combination with pyrolysis to generate conductive carbon lattice structures. Herein we demonstrate the viability of combining CLIP and pyrolysis by using a commercial resin found in prior pyrolysis literature, PR48.³² Using a CLIP printer with 5.4 μm pixel size, 8 mm diameter PR48 octet lattice disks with 50 μm diameter struts and void fractions of 88% (based on 500 μm³ unit cells) designed with nTopology software were generated (Figure 7A). These PR48 disks were pyrolyzed using a thermogravimetric analysis (TGA) furnace at 800 °C in N₂ with a 5 °C/min ramp to achieve pyrolytic carbon lattice disks with 15 μm strut diameters (Figure 7B–C, Figure S1). The lattice features printed via CLIP remained intact and underwent approximately 65% linear shrinkage after conversion to carbon. Raman spectroscopy was conducted on a representative PR48 structure printed via CLIP and pyrolyzed at 800 °C. The PR48-derived carbon lattice structures demonstrated characteristic D and G bands of graphitic carbon with a I_D1/I_G ratio of 0.94 based on peak intensity after baseline subtraction (Figure 7D). See Supporting Information S1 for full experimental section.

Figure 7.

Optical images of a representative 8 mm diameter PR48 octet lattice disk with 50 μm diameter struts and 500 μm³ unit cell (A) (i) after CLIP printing with (ii) inset showing the lattice pattern, and (B) after pyrolysis via a ramp in N₂ to 800 °C at 5 °C/min ramp rate; (C) scanning electron microscope image of a representative pyrolytic carbon lattice with 15 μm beams; (D) Raman data of 800 °C pyrolyzed PR48 lattice.

To investigate the effect that heating treatment has on carbon structure output, lattice disk samples were subjected to a range of furnace conditions. Prior studies indicate that PR48 undergoes the bulk of its mass loss decomposition around 400 °C before arriving at complete carbonization at 650 °C.³² Additionally, the highest char yield previously achieved using pure PR48 by Reale Batista et al. involved an hold at 300 °C in an oxidative environment.¹⁹ Therefore, we investigated the impact of this condition and other conditions involving isothermal holds, including a hold in N₂ at 400 °C and holds in air at 300 and 260 °C all with 800 °C final pyrolysis temperature and 5 °C/min ramp rates. The 800 °C pyrolysis run in N₂ with no holds was used as a control. Modifying the furnace procedure affected the surface morphology of the lattice beams, with the addition of the oxidative holds (Figure 8A–C) leading to increased surface texture topology compared to smoother samples produced in an inert atmosphere (Figure 8D,E). Additionally, the runs with longer isothermal holds in similar conditions lead to less surface deformities, for example adding a 1 h 400 °C hold (Figure 8D) yields smoother surface than the baseline N₂ ramp (Figure 8E) and holding at 260 °C in air for 3 h (Figure 8B) versus 1 h (Figure 8C) causes fewer rough edges around the outside of the beams. Heat treatment also affected the char yield, with the control N₂ ramp run yielding 3.1 wt % char yield while the greatest char yield of 16.2 wt % was achieved via the procedure with a 3-h isothermal hold at 300 °C in air (Figure 8F, Tables 1 and 4). We hypothesize that different heat treatments result in different surface topographies and char yields by affecting the rate and pathway of off-gassing pyrolysis byproducts and the degree of carbon graphitization. Conducting isothermal holds around temperatures where rapid mass loss occurs may enable volatiles to escape more gradually through available pathways in the polymer template while the polymer structures rearrange. Additionally, it has been shown that an oxidative treatment before inert pyrolysis causes dehydrogenation, aromatization, and oxidation of the polymer functional groups that help stabilize the polymer structure before cyclization and graphitization in inert gas at higher temperatures.⁸⁷

Figure 8.

Scanning electron microscope images of PR48 octet lattices that were pyrolyzed at 800 °C in N₂ with different heat treatments, including (A) a 3 h air 300 °C hold with (i) image of several nodes and (ii) image of a single node, (B) 3 h air 260 °C hold, (C) 1 h air 260 °C hold, (D) 1 h N₂ 400 °C hold, (E) control with no isothermal holds; (F) TGA curves displaying mass loss for each of these heat treatments with char yield being representative of the final mass yield at 800 °C.

Further investigation into pyrolysis of 3D printed polymers could advance understanding of this process by using an evolved gas analysis technique, such as mass spectrometry, to observe of the type, rate, and residence time of volatile gases produced at different temperatures.¹⁵⁵ Future studies should also prioritize conducting (electro)chemical and mechanical characterization on samples pyrolyzed under different pyrolysis conditions to connect processing conditions to properties like electrochemically active surface area, electrical resistance, strength, degree of graphitization, and resulting surface topography. One important consideration for sample mechanical integrity is that samples can be handled and assembled into electrochemical cells systems without breaking or cracking, which requires compressive strength on the order of 100 kPa or higher.^34,83 Mitigating material loss by achieving higher char yields is also desirable to produce larger objects.

This work demonstrates the viability of combining CLIP and pyrolysis for production of well-defined 3D carbon structures with high surface area features for energy storage. The 15 μm lattice feature sizes are, to the best of our knowledge, the highest resolution pyrolytic carbon achieved via a DLP-based printing technology. Future work should aim toward development of tailored custom resins designed for printability via CLIP and refinement of the pyrolysis process to achieve graphitic carbon structures that experience high mechanical strength with feature dimensions of 5–7 μm, the width of carbon fibers.^31,156

7. Conclusion

Herein we have reviewed the existing literature on pyrolysis of 3D printed polymers to carbon structures. One significant shortcoming of existing work is that the commercial resins often used in these studies, such as acrylate-based PR48 and epoxy-based SU-8, were not originally designed for pyrolysis^17,32,74 and therefore inherently are not tailored for the most optimal pyrolysis char yield output. Several custom polymer resins have been developed to be more intentional about the molecular composition of the pyrolysis precursor structure, for example by incorporating carbon-containing additives into acrylate-based monomers and generating aromatic polyimide structures.^52,82 Among the custom resin compositions, the structures with aryl cyclic functional groups and nitrile-containing groups that cyclize easily demonstrate the most promise in terms of high char yield and structural integrity and should be investigated further. Most often, acrylate groups are incorporated into these resins to enable radical polymerization during 3D printing. Additives such as graphene oxide also show promise for enhancing porosity and conductivity. Regarding the furnace conditions used to pyrolyze these materials, reaching temperatures greater than 900–1000 °C while heating in inert environments after stabilization steps in oxygen most often facilitates production of carbon with useful structural and electrochemical properties.

While several tailored custom resin systems developed to date are promising, they still often fall short when compared to the electrical conductivity (>100 S/m) and structural integrity (mechanical strength ≥4 GPa) of industry standard carbon fiber precursors that are used outside of the 3D-printing field.^31,157 Therefore, future studies must improve upon existing 3D printing resin compositions to generate graphitic carbon structures that have electrochemical characteristics conducive for applications in energy materials. Researchers should take direct inspiration from the polymers typically used in carbon fiber production, for example polyacrylonitrile, and should aim to advance scalable 3D printing procedures to achieve features like lattice beam diameters on the scale of carbon fiber precursor diameters, which are approximately 5–7 μm.¹⁵⁶ Focusing on light-based 3D printing methods that demonstrate both the speed and resolution capabilities for microfabrication of energy materials at industry scale is recommended. For example, in this work we report the fabrication of pyrolytic carbon structures with 15 μm beams with CLIP; to our knowledge, these are the smallest lattice features produced via a digital light projection-based 3D printing method. It is also necessary to further understand the carbonization process of 3D printed polymer during pyrolysis, which can be accomplished via by establishing new in situ characterization capabilities inside furnace systems to monitor degradation processes real time. Further investigation into incorporating catalytic graphitization into the pyrolysis process is also recommended due to the opportunity to more effectively volatilize noncarbon elements in the 3D printed precursor structure and mitigate the high energy usage caused by heating at temperatures greater than 1000 °C.

Embracing the fusion of 3D printing and pyrolysis toward development of groundbreaking carbon structures not only opens new horizons for sustainable energy storage solutions but also marks our collective stride toward an innovative and environmentally conscious future. This endeavor supports the mitigation of fossil fuel usage and facilitates the incorporation of electrification and renewable technologies to redefine society’s current energy landscape.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.4c00555.

• Detailed experimental section for fabrication, pyrolysis, and characterization of the carbon lattice disks generated for Section 6; lower magnification scanning electron microscope image of the pyrolytic carbon lattice sample included in Figure 7C (PDF)

Author Contributions

^¶ P.R.O. and S.C. contributed equally to this work. CRediT: Philip R. Onffroy conceptualization, data curation, methodology, project administration, validation, visualization, writing - original draft, writing - review & editing; Samuel Chiovoloni conceptualization, data curation, methodology, visualization, writing - original draft, writing - review & editing; Han-Lin Kuo conceptualization, data curation, methodology, visualization, writing - original draft, writing - review & editing; Max A. Saccone conceptualization, data curation, investigation, methodology, supervision, validation, writing - review & editing; Jennifer Q. Lu conceptualization, funding acquisition, investigation, project administration, resources, supervision, writing - review & editing; Joseph M. DeSimone conceptualization, funding acquisition, investigation, project administration, resources, supervision, writing - review & editing.

The authors declare the following competing financial interest(s): Joseph M. DeSimone reports financial support to the DeSimone research group was provided by startup funds from Stanford University School of Medicine and the School of Engineering. Philip R. Onffroy reports financial support was provided by the Knight-Hennessy Scholarship and National Science Foundation (NSF) Graduate Research Fellowship Program Grant #DGE-2146755. Jennifer Q. Lu reports financial support to the Lu research group was provided by National Aeronautics and Space Administration (NASA) Grant #NNH18ZHA008CMIROG6R and NSF Grant #CHE-1900647. Joseph M DeSimone was a cofounder of, and has financial stake in, Carbon, Inc., a CLIP-based 3D printing company. Joseph M DeSimone declares that he is an inventor on related US Patent 9,040,090, U.S. Patent 9,216,546, U.S. Patent 9,360,757, and others. Joseph M DeSimone declares that as a co-founder, board member, and former CEO of Carbon, Inc., and as an active researcher in additive manufacturing, that he has had hundreds of social media and journalist interviews about Carbon and additive manufacturing. Max A. Saccone advises 3D Architech, a VP-based metal 3D printing company. The authors declare that they have no other competing interests.

Special Issue

Published as part of JACS Auspecial issue “Polymers for the Clean Energy Transition”.