Direct printing of functional 3D objects using polymerization-induced phase separation

The study addresses the challenge of integrating multiple, dissimilar materials into single 3D printed objects to achieve functional properties. The authors focus on vat polymerization as a platform to spatially control materials from the surface to the interior via polymerization-induced phase separation (PIPS). Prior light-based strategies have patterned materials by modulating wavelength, intensity, oxygen inhibition, and employing photochromic molecules. The hypothesis is that by formulating photoresins to balance gelation kinetics, crosslink density, and component diffusivity, one can direct non-polymerizable functional components (e.g., a silver precursor) to targeted regions, particularly the surface, during printing. This would enable coatings, gradients, and composites in a single print, overcoming limitations of multi-material printing and post-process coating. The purpose is to understand and control 3D PIPS to realize functional devices such as sensors, antennas, and antimicrobial surfaces.

The authors survey approaches in vat polymerization and related additive manufacturing for multi-material control: dual-wavelength and orthogonal chemistries to spatially control distinct polymerizations; modulation of crosslink density via light intensity and oxygen inhibition; use of photochromic molecules to achieve soft-hard segment materials; and phase-separating resins to print multicomponent glasses with domain sizes controlled by light intensity. They note in situ photoreduction to generate silver nanoparticles during printing and prior uses of PIPS in 2D holographic polymerization to create patterned phases. Nanoscale printing alternatives (two-photon polymerization, electrodeposition, ion reduction) provide fine features but lack the ability to combine macro/micro geometries with nanoscale phase control during standard vat printing. The literature indicates potential but lacks a generalizable, chemistry-driven method to direct functional material placement throughout 3D objects during printing.

Overall approach: Employ polymerization-induced phase separation (PIPS) in stereolithography (SLA) vat polymerization with purposefully formulated photoresins to direct a non-polymerizable functional component (silver neodecanoate complex, AgND) toward the surface during curing. Study how crosslinker molecular weight, crosslinker fraction, gelation kinetics, and diffusivity govern spatial distribution, and validate via microscopy, spectroscopy, electrical measurements, simulations, and application demonstrations. Materials and resin formulations: - Monomer: 2-ethylhexyl acrylate (EHA). - Crosslinkers: Poly(ethylene glycol) diacrylate (PEGDA) with Mn ~170, 250, 575, 700 g/mol, designated DA-170, DA-250, DA-575, DA-700. - Photoinitiator: Ethyl (2,4,6-trimethylbenzoyl) phenylphosphinate (TPO-L), 1 wt% of resin. - Functional component: Silver neodecanoate (AgND) complexed with 2-ethyl-2-oxazoline (82 wt% Ag neodecanoate in oxazoline). Functional resins contained 25 wt% AgND (equivalent to ~9.5 wt% Ag metal post-sintering), unless otherwise stated. - Resin compositions varied from 15–99 wt% crosslinker (balance monomer, 1 wt% initiator). 3D printing and sintering: - Printer: Peopoly Moai Laser SLA (405 nm laser; spot size ~70 µm; build volume 130×130×180 mm; layer height 10–200 µm). Default settings; laser power rating set to 75; measured intensity ~760 W/cm² for cylinders. Some optimization at 350 and 560 W/cm². - Printed test objects: Cylinders 1.5 mm diameter, 2 cm length; also truss lattices and dipole antenna elements. - Post-processing: Thermal sintering on Kapton at ~210 °C for 1 h under N2 (500 ppm O2) to decompose AgND to metallic Ag. Thermogravimetric analysis (TGA) confirmed minimal polymer mass loss and silver content (~9.5 wt% Ag post-sintering from 25 wt% AgND resin). Characterization: - Morphology and composition: SEM (Hitachi SU3500, 15 kV) and EDS (30 kV; spot size 30). EDS line scans across cylinder cross-section edges (~15 µm depth), and surface EDS probing ~first ~2 µm of surface for Ag wt%. - Electrical: Two-probe resistance along cylinder walls with 1 cm probe spacing; average of 10 cylinders; sheet resistance computed from R×W/L. - Gelation kinetics: Phase-contrast optical microscopy with focused 405 nm laser (~10 µW, 1–2 µm spot; ~566 W/cm²). Resins without AgND were loaded into capillary slides and exposed to determine formation and growth of polymer “islands.” Image analysis (ImageJ) extracted island edge positions vs time. Delay time t_d obtained by fitting Profile = ±(D_t − D_∞) e^(−t/τ), where t_d is delay before observable network. Trend compared with AgND-containing resins. - Coarse-grained simulations: Langevin dynamics (HOOMD-blue) of simplified systems to estimate probe (3-bead) diffusivity in unpolymerized resin vs fully formed networks for crosslinker lengths L = 3, 6, 9 beads. Beads with WCA interactions, FENE bonds; volume fraction 0.491 in a 20σ box. Networks formed by allowing 10% reactive crosslinkers to polymerize to saturation; probe MSD used to compute diffusion coefficients. Triplicate runs with different seeds; report mean and standard error. Applications and specific protocols: - Strain sensors: Truss objects (11.24 × 11.24 × 13.40 mm) printed from mixed DA-575/DA-250 resins to yield 39–58 wt% total crosslinker with 7.9 wt% Ag metal equivalent. After sintering, trusses mounted on programmable linear stage; resistance measured during compression cycles up to 250 µm at 625 µm/s; gauge factor GF = (ΔR/R0) × (L0/ΔL), with L0 = 13.4 mm, ΔL = 250 µm. - Dipole antennas: Arrays of 3D printed dipoles (designed for 2.4 GHz; 6.25 cm length) printed with functional Ag-containing resin and sintered. Array fed by microstrip; mounted in anechoic chamber; transmission measured with VNA and gain standard horn; radiation pattern vs angle and half-power beam width (HPBW) determined. - Antimicrobial tests: Objects printed with 0.0–1.0 wt% Ag (in 35 wt% DA-575 resin) to favor nanoparticle formation on surfaces. Halo inhibition zone test against E. coli TG1 (1×10^9 cfu/mL on LB agar, 18 h at 37 °C). Growth kinetics: immersion in E. coli suspensions (10^5–10^6 cfu/mL), incubation at 37 °C, 220 rpm; OD600 vs time measured. Comparative coating method: - For reference, electroless plating performed on silver-free printed antennas seeded by dip-coating with a diluted commercial Ag nanoparticle ink, then plated by dropwise mixing of glucose/tartaric acid/ethanol solution with [Ag(NH3)2]+ solution over 60 min at RT, followed by 140 °C drying; adhesion compared to 3D PIPS-derived coatings.

- Spatial control via 3D PIPS: By tuning crosslinker molecular weight and wt% (thus gelation rate and network density), the distribution of Ag can be programmed from surface coatings to gradients to bulk composites. - Surface Ag vs crosslinker: EDS shows surface Ag wt% decreases with increasing crosslinker fraction across all DA systems. Shorter PEGDA (e.g., DA-170) yields a broader tunable range: DA-170 surface Ag varies ~88% to ~18% when crosslinker increases from 25 to 99 wt%; DA-700 varies ~86% to ~40% over the same range. - Conductivity: Post-sintering, many samples exhibit conductive surface films. Resistance increases with crosslinker content, consistent with reduced surface Ag and percolation limits; above certain crosslinker fractions, conductivity is lost. Calculated sheet resistance of surface films is ~340 mΩ/sq, comparable to screen-printed traces using the same precursor (~200 mΩ/sq) and far lower than recent 3D printable conductive polymers (~6.6×10^5 mΩ/sq). - Minimum AgND for conduction: For a 35 wt% DA-250 resin, ~19 wt% AgND is the threshold for electrical conduction; 25 wt% AgND chosen as optimal (yielding ~9.5 wt% Ag post-sintering). - Gelation kinetics: Delay time t_d decreases with increasing crosslinker fraction and with increasing crosslinker molecular weight. Example (DA-170): t_d ~5.8 s at 15 wt% vs ~1.8 s at 99 wt%. Higher t_d correlates with higher surface Ag, indicating longer free-diffusion windows favor surface accumulation. - Diffusivity mechanism (simulations): In unpolymerized resins, shorter crosslinkers give higher probe diffusivity (lower viscosity). In formed networks, longer crosslinkers provide higher probe diffusivity due to lower crosslink density and larger mesh size. Diffusion reduction from resin to network is ~24× for L=3 vs ~2× for L=9, explaining why tight networks (short crosslinkers) impede AgND migration at short t_d and yield lower surface Ag under fast gelation. - Strain sensors: Trusses with graded Ag coatings show piezoresistive behavior. Greater crosslinker content (39→58 wt%) increases obstruction by polymer, generating larger ΔR/R under compression. Maximum ΔR/R at 250 µm: ~2.5% (39 wt%), 6.0% (46 wt%), 12.5% (52 wt%). Gauge factors: 2.3 (39 wt%), 3.2 (46 wt%), 5.1 (52 wt%), 15.7 (58 wt%). - Antennas: A 3D printed dipole antenna array (2.4 GHz) exhibits a measured radiation pattern closely matching theory: HPBW measured 45° vs theoretical 48° (difference 3°). Gain measurements comparable to literature. - Antimicrobial function: Objects with 0.5–1.0 wt% Ag display inhibition halos against E. coli on agar and suppress growth in liquid culture, whereas Ag-free controls allow growth, confirming antibacterial activity with low Ag loadings. - Integration advantage: 3D PIPS yields adherent, uniform metallic coatings directly during printing, avoiding adhesion issues seen with some post-print coatings.

The results confirm that polymerization-induced phase separation during vat polymerization can be harnessed to deterministically position functional materials within printed objects by balancing gelation kinetics, crosslink density, and diffusivity. Longer gelation delay times allow AgND to migrate toward the advancing polymer–resin interface, increasing surface accumulation. Concurrently, the evolving polymer network modulates diffusivity: tight, short-spacer networks rapidly constrain AgND mobility, limiting surface migration under fast gelation; longer-spacer networks remain more permeable, enabling higher surface Ag even at shorter t_d. This interplay explains the observed trends across PEGDA lengths and loadings and provides a predictive framework to design resins for desired morphologies (coatings, gradients, composites). Functionally, this control enables single-step fabrication of conductive, piezoresistive surfaces; RF conductors forming efficient antenna elements; and antimicrobial surfaces with minimal additive loading. The approach overcomes limitations of multi-material hardware and post-process metallization by embedding the functional phase during printing with nanoscale domains and strong adhesion. More broadly, these findings illustrate a route toward smart, structurally integrated electronics and surfaces via chemistry-driven spatial control during standard SLA printing.

The study introduces a general, single-step strategy to directly print functional 3D objects by exploiting polymerization-induced phase separation in vat polymerization. By tuning resin formulation—crosslinker molecular weight and fraction—together with understanding gelation delay and diffusivity, the spatial distribution of a non-polymerizable functional component (AgND) can be programmed to yield coatings, gradients, or composites. Demonstrations include: conductive silver-coated lattices functioning as piezoresistive strain sensors with tunable gauge factors; a 2.4 GHz dipole antenna array with radiation characteristics matching theory; and antimicrobial objects exhibiting E. coli inhibition with low Ag content. The universality of 3D PIPS suggests applicability to a wide range of functional fillers (e.g., catalysts, bioceramics, antiviral agents), enabling structural electronics, soft robotics, and embedded sensing. Future work could extend to other chemistries and functional phases, optimize light exposure profiles to further refine phase separation, integrate multi-component systems for hierarchical gradients, and develop quantitative predictive models coupling reaction kinetics, phase behavior, and transport for broader material sets.

- Gelation time measurements were performed without AgND to improve interface visibility; while trends matched AgND-containing resins for DA-575, quantitative differences may exist when AgND is present. - The EDS surface analysis probes approximately the first ~2 µm, limiting depth resolution of surface composition gradients. - Coarse-grained simulations modeled simplified systems with crosslinker lengths up to L=9 beads and did not replicate experimental crosslinker ratios; networks were idealized (fully formed) without ongoing reaction kinetics, providing qualitative rather than quantitative predictions. - Functional demonstrations primarily used silver precursors; generalization to other functional materials, sizes (e.g., nanoparticles), and chemistries, while promising, was only briefly illustrated. - Sintering at ~210 °C is required to convert AgND to metallic Ag, which may limit compatibility with some substrates or embedded components. - Printing and characterization were conducted on a specific SLA platform and parameter set; scalability and robustness across different printers and exposure conditions may require further validation.