Introduction

Additive manufacturing (AM) is a method for creating three-dimensional (3D) objects that facilitates rapid prototyping, reduced material waste, and access to parts with high geometric complexity1,2,3. Many AM methods are designed to create objects from a single material; however, significant work has been done to design and develop AM methods that incorporate different materials into a single part during the same production step4,5,6,7. Parallel deposition-based AM techniques, such as material jetting, direct write material extrusions, and melt material extrusions, were among the first techniques developed for multimaterial AM8,9,10,11,12. Vat photopolymerization (VP) AM methods offer benefits such as decreased print time, high feature resolution, and smooth surface finishes; however, VP poses a challenge for multimaterial AM, as the feedstock material for the final 3D part is a single homogenous liquid resin. Clever engineering solutions have been developed to achieve multimaterial parts through VP methods, such as fitting resin vats with fluidic systems to allow for resin exchange or building rotating tracks to switch between resins mid-print13,14,15,16,17,18,19. Although these methods are effective in creating multimaterial parts, the time required to exchange resins or clean previous resin from the part can significantly increase print times. Additionally, material heterogeneity in a single layer can be challenging to achieve, as each resin needs to be applied and cured sequentially. Controlled gradation in the transition from one material to another can also be difficult with such designs.

Chemistry-based approaches to multimaterial VP have been proposed as alternatives or complements that address these engineering challenges. Such approaches generally leverage the combination of two chemically orthogonal polymerization mechanisms, high selectivity between reaction mechanisms, or differing extents of reaction. Chemically distinct regions of the printed part are created through the application of light sources of differing intensities or wavelengths20,21,22. An early example of this is the work by Schwartz et al. on multimaterial actinic spatial control AM (MASC-AM), which combines wavelength-selective epoxide and acrylate polymerizations to yield multimaterial parts with domains of significantly different mechanical profiles23. Similar techniques have demonstrated a variety of multicomponent resin formulations for producing multimaterial parts with disparate mechanical properties as well as solubility characteristics24,25,26,27,28,29,30,31,32.

Despite the benefits of VP methods and the ability to use multiwavelength VP to create multimaterial parts from a single resin, there are still many material classes that are challenging to integrate with VP, such as resins with high viscosity or opacity, or resins that require thermal curing. Additionally, metals are not easily incorporated into VP. Most metal AM methods require intense stimuli such as directed energy deposition or laser powder bed fusion33,34,35,36,37. The intensity of stimuli supplied by accessible VP printers is simply insufficient for processing purely metallic materials. Previous solutions to this challenge have used organic polymeric binders within a metal-containing resin mixture that can be subsequently pyrolyzed to leave a solid metal 3D object. In this way, geometrically complex objects were produced with the engineering benefits of VP, and final parts were comprised of metallic species generally inaccessible with VP. Notably, in its current iteration, the overall process involves destruction to the polymer components and thus precludes creation of parts that combine organic and metallic domains.38,39,40,41.

Inspired by the pluripotency of hydrogel infusion additive manufacturing demonstrated by Saccone and colleagues, we were interested in leveraging MASC-AM technology to access a variety of multimaterial parts otherwise inaccessible through VP methods42,43. Herein, we present a VP-based AM technique, coined infusion multimaterial actinic spatial control additive manufacturing (iMASC-AM), which involves the application of wavelength-specificity to dictate material composition with spatial resolution that is determined by which wavelength(s) of light are applied to which volumes of the printed part (Fig. 1a). Since material compositions can be differentiated, physicochemical properties (e.g., swelling ratios) can therefore also be controlled along all three axes. We envisioned an integrated workflow that could enable selective infusion to specific volumes of a 3D part that is produced via MASC-AM. The iMASC-AM technique combines the straightforward ability to infuse solutes from a broad range of material classes with the multimaterial spatial resolution of MASC (Fig. 1b).

Fig. 1: iMASC-AM process.
Fig. 1: iMASC-AM process.The alternative text for this image may have been generated using AI.
Full size image

A From a single resin vat, radical polymerization can be initiated through visible light (405 nm) irradiation to yield elastomeric, acrylate-dominant regions, while cationic and radical polymerizations can be initiated through UV light (365 nm) irradiation to yield stiff, epoxide-dominant domains. B Idealized schematic showing selective infusion of infusion material into the acrylate-dominant region of a printed multimaterial part.

Results

Infusion of multimaterial parts

MASC-AM uses a single resin formulation containing, for example, epoxide monomers and a photoacid generator, as well as acrylate monomers and a photoradical initiator. The resin is 3D printed via dual-wavelength vat photopolymerization. In design, this printer is capable of patterning light at single-pixel resolution (approximately 45 μm2). In practice, patterned features are visually discernable down to 50 μm resolution (Fig. S5). In regions exposed to UV light, the photoacid generator is activated, which primarily drives the cationic polymerization of epoxide monomers. In contrast, in regions where visible light is applied, the photoradical initiator is activated, which results in selective curing of the acrylate monomers (Fig. S6). The dual photochemical mechanistic approach of MASC is complementary to multimaterial VP that leverages greyscale light projection to accomplish spatially resolved degrees of crosslinking of the same resin monomers throughout the part, or approaches that implement a two-stage curing procedure in which one resin is thermally cured after the AM process is completed44,45,46,47,48,49,50.

To explore the possibilities of iMASC, we leveraged the disparate swelling profiles of epoxide- and acrylate-dominant regions to selectively infuse material within a patterned domain of a multimaterial part. We evaluated different solvents and assessed the swelling ratios of epoxide- and acrylate-dominant regions and found that across all solvents that we screened, the acrylate-dominant regions gave greater swelling percentages (determined gravimetrically) than did the epoxide-dominant regions (Fig. 2a). For example, when soaked in isopropanol (IPA) as a standard post-processing protocol, the epoxide- and acrylate-dominant regions experienced 5% and 14% increases in mass, respectively. We found that when soaked in water, the mass changes were more dramatic, achieving 15% mass increase for epoxide-dominant regions versus 130% for acrylate-dominant regions. Toluene was the least effective solvent for swelling the parts, increasing the mass of the epoxide- and acrylate-dominant domains by just 0.8% and 1.5%, respectively. The most effective solvent for swelling, as judged by the ratio of swelling percentages, was a 1:1 v/v solution of water and IPA, which increased the mass of the epoxide- and acrylate-dominant domains by 40% and 532%, respectively.

Fig. 2: Infusion behavior.
Fig. 2: Infusion behavior.The alternative text for this image may have been generated using AI.
Full size image

A Gravimetric swelling data for acrylate- and epoxide-dominant parts in various solvents (isopropanol abbreviated as IPA, tetrahydrofuran abbreviated as THF). Error bars represent standard deviation. B CAD file (top) and rhodamine 6G-infused smiley face (bottom). C CAD file (top) and Cu(NO3)2 -infused honeycomb-patterned part (bottom). D CAD file (top), control (left) and silicone-infused multimaterial part (right) being stretched with 100 g and 200 g weights, respectively.

These initial results led us to investigate the applicability of the hydrogel infusion technique across different materials (summarized in Fig. 2b–d). Organic dyes, such as rhodamine 6G, were readily infused into the printed parts. By keeping the concentration of the dye in the aqueous medium very low (e.g., <1 μM), we were able to visually observe the effects of selective swelling in the patterned part (Fig. 2b). Similarly, selective infusion was also visually observable with high concentration (≥ 2 M) infusion solutions of metal salts, such as copper (II) nitrate (Fig. 2c). The acrylate-dominant regions experience an approximately 50% volume increase after infusion with high concentration (≥ 2 M) solutions due to the added volume of the infusion material, bringing the visually-discernable feature resolution to 100 μm (Table S1, Fig. S7). Additionally, chemically distinct resin precursors, such as silicones, were successfully infused and later thermally cured to give rise to three-component multimaterial parts (Fig. 2d).

Infusion of silicones

In general, the incorporation of Pt-cured silicones into vat-based AM, let alone multimaterial AM processes, is challenging. Therefore, we viewed patterned silicone multimaterial parts as an important target. Toward this end, we infused SYLGARD™ 184 elastomer precursors dissolved in 1:1 v/v IPA and tetrahydrofuran (THF) into the multimaterial parts for 24 h. As a technical note, our assessments of the infusion kinetics over time suggested to us that increased infusion time may lead to a greater mass of infused silicone monomers (Fig. S8). After infusion, the parts were then treated in an additional thermal post-processing step at 120 °C for 2 h to thermally cure the silicone networks via Pt-catalyzed hydrosilylation. Subsequent uniaxial tensile testing was found to be consistent with the successful formation of the silicone secondary network (Figs. S9–12). The effects were observed to differing extents across both epoxide- and acrylate-dominant parts. Specifically, tensile test specimens were prepared using either UV or visible light. Each was then subjected to infusion and thermal post-curing before tensile testing (Fig. 3a–e). On average, the epoxide-dominant parts experienced an 82% increase in yield stress (from 11.4 to 20.8 MPa, Fig. 3a) and no significant change in strain at break or toughness (defined here as strain energy density, Fig. 3b,c, respectively). On average, the acrylate dominant parts experienced a 105% increase in yield stress (from 0.21 to 0.43 MPa, Fig. 3a), no significant change in strain at break (Fig. 3b), and a 279% increase in toughness (from 0.52 to 4.16 MPa, Fig. 3c). When subjected to cyclic tensile loading to 50% of maximum strain, the parent acrylate-dominant material failed after 7 cycles, whereas the silicone-infused acrylate-dominant sample underwent 110 cycles without failure, indicating increased reversible deformation behavior of the silicone-infused material consistent with silicone network formation (Figs. S13–15).

Fig. 3: Mechanical analysis of silicone-infused parts.
Fig. 3: Mechanical analysis of silicone-infused parts.The alternative text for this image may have been generated using AI.
Full size image

A–C Mechanical analysis of silicone-infused epoxide- and acrylate-dominant parts. Error bars represent standard deviation. D–E Tensile analysis curves representative of the average mechanical performance of control and silicone-infused epoxide- and acrylate-dominant parts.

The silicone infusion significantly increased the tensile strength of both domains (Fig. 3a, d, e), increased the toughness of the acrylate-dominant domain (Fig. 3c, e), and maintained notable disparity between the two material regions. Additionally, the percentage increase of each mechanical property of the acrylate-dominant parts was found to be greater than that of the epoxide-dominant parts. These results are consistent with our understanding of the infusion behavior of the two materials and reflect the fact that the epoxide-dominant parts begin as a stiffer material than the polyacrylate parts. Finally, to illustrate the multimaterial functionality of iMASC-AM, mechanical analysis was done on printed samples with alternating acrylate- and epoxide-dominant domains (Fig. S16). This demonstrated the ability to program one or more localized stress responses within a printed multimaterial part (Fig. S17). After infusion, the multimaterial parts showed increased strength and toughness, which is consistent with previous, single-material silicone-infusion results (Fig. S18).

Infusion of metal salts

The potential versatility of iMASC was encouraging, and we next targeted the incorporation of inorganic species with particular interest in those that could ultimately lead to spatially resolved metallic and electrically conductive patterns. Multimaterial AM, combining materials of different classes (e.g., thermoplastics, thermosets, and metals), has historically been challenging to realize due to drastically different processing requirements and chemical incompatibilities across the different components. Notably, techniques that are commonly used for AM of metal parts tend to be destructive to organic components34,35,36,37.

We found that both copper(II) nitrate and silver(I) trifluoroacetate (AgTFA) were efficiently incorporated into acrylate-dominant regions during the infusion process. Spatial control of the Cu(NO3)2 infusion was easily discernable due to its deep blue color, indicating greater extent of infusion into the acrylate-dominant regions than the epoxide-dominant regions (Fig. 2d). Acrylate-dominant domains displayed a 156% increase in mass when infused with 2 M Cu(NO3)2 solution, compared to a 13% mass increase in the epoxide-dominant domains (Fig. S19), after removal of volatiles. Although the Cu(NO3)2 infusion was successful, to reach a functional metal-organic multimaterial part with electrically conducting and insulating domains, the copper(II) salt would need to be reduced to copper(0). This reduction can be achieved through high-temperature treatment ( > 700 °C); however, the organic components of the multimaterial part would be destroyed in the process, ultimately preventing the patterning of insulating domains.

Moving away from the copper(II) experiments, we observed similar success with the selective infusion of aqueous AgTFA. We took an interest in silver, considering the ease with which silver-infused parts can be rendered electrically conductive via comparatively milder thermal post-processing. To explore the AgTFA infusion in more detail, we assessed a concentration range of 0.1 to 6 M and found a monotonic increase in the percentage change in mass as the concentration of AgTFA in the infusion solution was increased (Fig. 4a). At AgTFA concentrations greater than 2 M, however, salt precipitation on the surface was apparent. Salt precipitation on the part surface was also observable after long infusion times ( > 48 h). The infusion of 1 M AgTFA resulted in an average mass increase in the acrylate- and epoxide-dominant areas of 59% and 11%, respectively (Fig. S19), corresponding to an AgTFA concentration in the network of approximately 2.20 and 0.59 M, respectively. We found 24 h to be the optimal infusion time for 2 M AgTFA, as infusion kinetics experiments showed a stabilization of mass change due to infusion between 12 and 24 h (Fig. S20). Thermogravimetric analysis of acrylate-dominant domains infused with 2 M AgTFA showed similar residual mass from specimens collected from the part surface as those collected at 4 mm below the surface, indicating the infusion solution effectively penetrates the part (Fig. 4b).

Fig. 4: Silver trifluoroacetate (AgTFA)-infused parts.
Fig. 4: Silver trifluoroacetate (AgTFA)-infused parts.The alternative text for this image may have been generated using AI.
Full size image

A Gravimetric analysis of epoxide- and acrylate-dominant parts over a range of AgTFA solution concentrations. Error bars represent standard deviation. B Analysis of residual mass at the surface and interior of acrylate-dominant parts as determined through TGA over a range of AgTFA solution concentrations. C AgTFA-infused multimaterial part after infusion (left) and after subsequent thermal reduction (right). D CAD file (left) and thermal image (right) of a thermally reduced, AgTFA-infused part placed on a cold pack after removal from a hot plate. E CAD file (left), AgTFA-infused, thermally reduced part (center), and AgTFA-infused, thermally reduced part completing a circuit through the acrylate-dominant domain only (right).

Importantly, post-processing of silver salt-infused materials to achieve conductivity is straightforward and can be achieved via thermal processing at temperatures low enough to avoid thermal decomposition of organic networks. Encouraged by the prospects of patterned conductive and insulating domains within a single 3D part, we proceeded with thermally induced reduction of the AgTFA in parts produced via iMASC (Fig. 4c). After mild thermal processing ( ≤ 200 ˚C, Fig. S21), we found that the acrylate-dominant domains (high silver content) showed fast heat dissipation indicative of high thermal conductivity (Fig. 4d), and electrical conductivity between 4235 and 42 S m-1. In contrast, the epoxide-dominant regions showed slower heat dissipation (Fig. 4d) and electrical conductivity less than 2 × 10−5 S m−1 after the same thermal processing. There is notable variability in electrical conductivity at different regions of a single material, which could be attributed to variation in crosslink density, conversion, or light intensity. Therefore, further process development would likely be needed to determine the true cause of variability and standardize these values for specific applications. Nevertheless, we found that simple circuit designs were easily achieved via iMASC. For example, the silver salt-infused, thermally reduced acrylate-dominant domains were capable of completing a circuit with a 9-V battery to power an LED (Fig. 4e, Movies S1, 2). The equivalent epoxide-dominant domain did not complete a circuit.

To determine the print parameter threshold for achieving conductive or insulating domains, we investigated the effects of UV exposure on conductivity outcomes. In the resin system used for these prints, UV light initiates the epoxide polymerization at high exposure and the acrylate polymerization at low exposure. It was found that parts irradiated with UV light for 15 s or less showed electrical conductivity measurements consistent with acrylate-dominant material, and parts irradiated with UV light for 25 s or more showed electrical conductivity measurements consistent with epoxide-dominant material (Fig. S22). This can likely be attributed to a stark difference in swelling behavior between samples that experience 15 and 25 s of UV irradiation (Fig. S23). The irradiation time required to create insulating domains (25 s) is significantly less than is required to achieve the mechanical properties typical of an epoxide-dominant material in our MASC setup (75 s). This finding indicates that even low conversion of epoxide is enough to create insulating domains in the final infused part, suggesting print times for these multimaterial parts could be significantly shorter than print times traditionally required for epoxide VP, while still achieving spatially resolved conductive and insulation regions.

Discussion

The development of iMASC-AM offers a pivotal advancement in achieving multimaterial 3D printing capabilities through multiwavelength VP by eliminating the need for complex reformulation of wavelength-specific resin chemistries. The MASC-AM product is a divergent intermediate that can be used to create a variety of material combinations via iMASC-AM. By leveraging wavelength-selective acrylate and epoxide polymerizations, we demonstrated that regions of a printed object can be spatially patterned to allow for disparate amounts of infusion of a secondary material. This selective infusion greatly expands the material scope for multiwavelength VP, allowing access to opaque, thermally-responsive, or viscous materials that are challenging to process through VP. The iMASC-AM approach enables fabrication of a wide variety of chemically diverse multimaterial architectures through a single streamlined printing process.

Although the technique is capable of producing compelling multimaterial architectures, it is important to note that this process is highly dependent on the permeability of the original printed part, as well as the size and solubility of the infusion material. Large or poorly soluble particles may reduce the efficiency of the infusion or result in aggregation on the surface of the part instead of infusion. Large ( > 10 mm) or geometrically complex parts may require different processing techniques or geometric modifications to ensure consistent levels of infusion throughout the full depth of the part, as well as geometric fidelity throughout the infusion process. Addressing these considerations, as well as exploring the possibility for automation, will help to realize the scalability and industrial potential of this platform.

Methods

Materials

2-Hydroxyethyl acrylate, 3,4-epoxycyclohexylmethyl-3,4-epoxycyclohexanecarboxylate, phenylbis(2,4,6-trimethylbenzoyl)phosphine oxide, silver trifluoroacetate, rhodamine 6G, tetrahydrofuran, triarylsulfonium hexafluorophosphate salts (50 weight % in propylene carbonate) and copper(II) nitrate hemi(pentahydrate) were purchased from Sigma-Aldrich. Epoxycyclohexyl ethyl functionalized polyhedral oligomeric silsesquioxane cage mixture (30% in 3,4-epoxycyclohexylmethyl-3,4-epoxycyclohexanecarboxylate) was purchased from Hybrid Plastics. Hydroquinone (99.5%) was purchased from Acros Organics. 2-Propanol was purchased from Fisher Scientific. SYLGARD® 184 silicone elastomer base and curing agent were purchased from Dow.

Resin formulation

25 mL 2-Hydroxyethyl acrylate, 20 mL 3,4-epoxycyclohexylmethyl-3,4-epoxycyclohexanecarboxylate, 0.2 g phenylbis(2,4,6-trimethylbenzoyl)phosphine oxide, 0.06 g hydroquinone, 1.75 mL epoxycyclohexyl ethyl functionalized polyhedral oligomeric silsesqioxane cage mixture (30% in 3,4-epoxycyclohexylmethyl-3,4-epoxycyclohexanecarboxylate), and 6 mL triarylsulfonium hexafluorophosphate salts (50 weight % in propylene carbonate) were added to an amber borosilicate vial in the specified amounts (order of addition is not important). The resin was stirred in the dark at room temperature for 24 h before manufacturing.

Additive manufacturing

Resin was manufactured using a custom-built dual-wavelength vat photopolymerization printer. Epoxide-dominant regions were irradiated with UV light for 75 s per 0.1-mm layer. Acrylate-dominant regions were irradiated with white light for 15 s per 0.1-mm layer. Emission spectra for the light sources used have been previously published49,51. After printing, multimaterial parts were soaked in 2-propanol for 12–24 h to remove unreacted monomer, then dried under vacuum at room temperature for 24 h.

General multimaterial infusion

Silver(I) trifluoroacetate, copper(II) nitrate, and rhodamine 6 G infusions: 2, 2, and 10-5 M solutions of silver(I) trifluoroacetate, copper(II) nitrate, and rhodamine 6G, respectively, were prepared in deionized water. Printed parts were swelled in infusion solutions for 12-24 h, then dried under vacuum at room temperature for 24 h.

Thermal processing of silver trifluoroacetate-infused parts

Dried, silver trifluoroacetate-infused parts were heated at a rate of 1 °C/min to 100 °C, then held at 100 °C for 60 min. They were then heated at a rate of 10 °C/min to 200 °C, then held at 200 °C for 60 min.

Silicone Infusion

SYLGARD™ 184 elastomer base was mixed with SYLGARD™ 184 curing agent in a 10:1 ratio, respectively. This mixture was dissolved in 1:1 isopropanol:tetrahydrofuran (10% v/v). Printed parts were swelled in the SYLGARD™ 184 solution for 24 h and then dried under vacuum at room temperature for 24 h. The dry parts were thermally cured at 120 °C for 2 h.

Nuclear magnetic resonance

The uncured resin sample was packed in a Bruker 4-mm HRMas sample holder inside a ZrO2 rotor. The cured parts were frozen in liquid nitrogen and ground with magnesium oxide. These were weighed into 4-mm Doty pencil rotors with glass-filled Torlon caps. The 13 C cross-polarized (CP) MAS NMR spectra were obtained using an Avance III Bruker instrument using a 4-mm broadband MAS probe at a resonant frequence of 125.79 and 500.22 MHz for 13 C and 1H, respectively. Each sample used high power (100 kHz) 1H Spinal 64 decoupling, a 3 s recycle delay. A 1.2-ms contact time employing a 70–100% ramped shape on the 1H channel. For the resin sample, a 5 kHz spinning speed and 18k scan averaging were used. For the visible light-irradiated sample, an 11 kHz spinning speed and 16k scans were employed, whereas for the UV light-irradiated sample, an 8 kHz spinning speed and 20 k scan averaging was used. The number of scans varied for the length of time available for the experiments in a shared-use facility, and the spinning speed was the maximum attainable with the individual rotor.

Instruments

Dual wavelength vat photopolymerization

Dual-wavelength vat photopolymerization was performed on a custom 3D printer equipped with an Acer X152H visible light projector and a PDCO5-5 365-nm light projector purchased from Xiamen Zhisen Electro. Equip. Co., LTD. Print files were prepared in MATLAB. MATLAB was used to run the printer by controlling projectors via Psychtoolbox and sending G code commands via the serial port51.

Thermogravimetric analysis

Thermogravimetric analysis was conducted on a TA Instruments Q50 thermogravimetric analyzer under air.

Scanning electron microscopy/ energy dispersive X-ray spectroscopy

Scanning electron microscopy and energy dispersive X-ray spectroscopy were conducted on a Zeiss Gemini SEM 450 at an accelerating voltage of 5 kV.

Tensile analysis

Tensile analysis was conducted on an MTS Criterion Electromechanical Testing System using ASTM D638 type V specimens. The strain rate was set to 0.1 mm/s and 0.8 mm/s for epoxide and acrylate-dominant materials, respectively. Tensile strain of acrylate-dominant materials was measured through video analysis by calculating the change in the number of pixels in the gauge region.

Contact angle measurements

Contact angle measurements were performed on a Dataphysics OCA 15EC Optical contact angle goniometer. A dosing volume of 4 μL, dosing rate of 1 μL/s, DS 500/GT (500 μL) syringe, and SNS 050/026 syringe were used.

Viscosity measurements

Viscosity measurements were performed on a TA Instruments HR 20 rheometer. Experiments were performed at 25 °C with a 90-s soak time. Data was collected between shear rates of 10 and 103 s-1 with 5 points per decade, 5-s equilibration time, and 30-s averaging time.

Conductivity measurements

Bulk resistance was measured with a Hewlett-Packard 33401 A Multimeter. Conductivity was calculated using the equation σ = L/(SR), where L is the distance between electrodes, S is the cross-sectional area of the sample, and R is the bulk resistance.