Mechanically interlocked 3D multi-material micromachines

The study addresses how to integrate intrinsically dissimilar materials—metals and polymers—into mechanically interlocked three-dimensional microdevices that preserve each material’s advantages. Existing micromachines often rely on single materials or planar/2D assembly methods that constrain geometry, adhesion, and functionality. The authors highlight key challenges: (i) shaping non-modified, water-soluble polymers like pure gelatin into 3D microscale structures can only be done via mold casting; (ii) direct metallization of soft hydrogels (e.g., gelatin) compromises their properties and cannot match the magnetic volume of fully metallic structures; and (iii) drug loading on metallic surfaces is limited without extra coatings. The research goal is to develop a fabrication strategy that mechanically interlocks independent 3D metallic and polymeric components at the microscale, enabling simultaneous exploitation of metal (e.g., strong magnetic response, biocompatibility) and polymer (e.g., drug loading, elasticity, shape transformation, degradability) properties. This capability is positioned as critical for advanced biomedical microrobotics requiring multi-functional components such as drug reservoirs, steering/locomotion units, and programmable connections/disconnections.

Mechanically interlocked systems are established at the molecular scale (catenanes, rotaxanes) and enable nanoscale machines, but scaling these to larger devices is limited by complex synthesis and materials integration. Microfabrication approaches such as soft-lithography microtransfer molding can produce closed-loop single-material structures, while origami-based methods produce compliant 3D devices but impose constraints on materials adhesion and allowable geometries and rely on compliant joints. Weaving/knitting/braiding concepts allow multifunctional architectures from varied materials but have not been realized below the micrometer scale. Template-assisted electrodeposition has produced fully metallic and multifunctional polymeric microstructures; fully iron microstructures are promising for biocompatible, high magnetic responsiveness but lack advanced functions like cargo delivery and shape morphing more easily achieved with soft materials. Indirect 3D/4D printing methods have shaped water-soluble materials with ~5 μm features. Prior interlocked metallic microstructures were limited to metal–metal via microcontact printing and electroplating on curved surfaces. Collectively, these works show the need for a process that interlocks metal and polymer components in true 3D with minimal design/material constraints.

Overview: The method combines two-photon absorption (TPA) 3D lithography in a positive-tone photoresist to create multiple, independent 3D vias, followed by selective electrodeposition of metal into vias that reach a conductive substrate and mold casting of polymers into electrically isolated vias. After dissolving the resist mold, interlocked, fully metallic and polymeric 3D components remain mechanically bound. Design constraint: avoid overlap between independently written vias; maintain minimum clearances to preserve separate pathways.

Two-photon lithography and template creation:

• System: Nanoscribe Photonic Professional GT (TPA, 780 nm fs laser), 63x objective, oil immersion with Immersol 518f; allows accurate detection of photoresist/substrate interface and use of opaque substrates and thick resist.
• Substrate: 170 μm coverslip (3 cm diameter) coated by e-beam with Ti (20 nm)/Cu (80 nm) to serve as conductive base for electrodeposition.
• Photoresist: AZ IPS-6050 positive resist, spin-coated to thicknesses of 32 μm (2400 rpm for 18 s) or 82 μm (1050 rpm for 13.5 s) with 300 rpm/s ramp; soft-bake to 125 °C (heating 15 °C/min), hold 5 min (32 μm) or 15 min (82 μm); quench to room temperature.
• CAD and writing parameters: For 32 μm structures, designed filament thicknesses (nominal) metallic/interlocking bar/other polymer parts = 0.5/2/3 μm; for 82 μm structures = 2/3/5 μm. Minimum metal–polymer via spacing 10 μm; polymer features ≥4 μm above substrate. Hatching/slicing: 0.3/0.3 μm (32 μm) or 0.3/1.0 μm (82 μm). Scan speed 2500 μm/s. Laser power 32% (32 μm) or 50% (82 μm) of system max. Empirical tuning recommended per geometry.
• Post-exposure processing: Rinse with DI water, nitrogen dry to remove oil; post-bake 100 °C for 140 s. Develop by vertical immersion in AZ 726MIF developer for 30 min (32 μm) or 90 min (82 μm), then transfer to DI water and keep wet to ensure channel wetting for subsequent plating.

Metal electrodeposition (iron):

• Electrolyte: FeSO4·7H2O 250 g/L, FeCl2·4H2O 42 g/L, NH4Cl 20 g/L; pH 2; total volume 200 mL in 400 mL beaker; stir at 100 rpm with 2.5 cm magnetic stir bar.
• Electrodes: Counter = 5 cm × 5 cm Pt sheet; reference = double-junction Ag/AgCl. Potential applied with potentiostat (PGSTAT204): −0.95 V vs Ag/AgCl starting immediately upon immersion to avoid substrate etching. Stop at current onset, then rinse with DI water.

Polymer casting (examples):

• Gelatin: 10 wt% in DI water typical; 0–20 wt% tested. Plasticizer glycerol 0–10 wt% for increased elongation. For demonstration, gelatin included 0.1 wt% rhodamine B (drug model). Procedure: maintain above gelling temperature; degas ~5 min; immerse sample; store at −4 °C ≥2 h; remove excess by mechanical rubbing.
• PDMS (Sylgard 184): Base:curing agent 5:1 (82 μm structures) or 10:1 (32 μm) to tune stiffness. Mix thoroughly up to 5 min, degas. Ecoflex 00-30: mix equal volumes of parts A/B. Casting: place sample in shallow container, pour ~5 mL polymer; vacuum infiltration by repeated pump/vent cycles until no surface bubbles remain (indicating complete infiltration). Remove excess polymer, place on sulfonated paper. Cure PDMS at 65 °C for 4 h; ecoflex cure at room temperature overnight; mechanically remove thin surface film.
• UV-curable polymer (NOA63): Drop-cast, vacuum infiltrate for a few hours. Use a 1 mm PDMS slab to squeegee excess and ensure conformal contact. UV cure at 365 nm, 2.7 mW/cm^2 for 2 h. Remove PDMS slab; remove AZ mold by washing with acetone, isopropyl alcohol, and water.

Imaging and materials verification:

• SEM: Leo 1530 field-emission SEM; samples sputter-coated with thin Au to reduce charging. EDX used for Fe/polymer structures.

Magnetic actuation:

• System: 8-coil planar electromagnetic setup (MFG-100, Magnebotix) to generate rotating fields and gradients. Samples placed at workspace center. For characterization on silicon wafer: forward motion at 2 mT; rotational motion at 4 mT. Additional demonstrations: DI water rolling/tumbling and silicone oil tumbling at specified fields/frequencies (e.g., 20 mT at 50 Hz for floating forward motion with ecoflex frame; 20 mT at 20 Hz for rolling with PDMS frame; 20 mT at 5 Hz tumbling in silicone oil; 17 mT at 15 Hz rolling in DI water). Frequency increased until lift-off at water surface for floating cases.

Design motifs and interlocking strategy:

• Two via types: (i) vias contacting conductive substrate enable metal electrodeposition; (ii) electrically isolated cavities remain empty after plating and are later filled by polymer casting. Interlocking is achieved by designing the metal and polymer channels to pass through or around each other (e.g., bar-through-cage, ring-through-helix) forming closed loops that mechanically bind dissimilar materials without chemical bonding.

• The process fabricates fully interlocked 3D metal–polymer micromachines by combining TPA lithography, electrodeposition (Fe), and polymer casting in a single template, yielding two independent geometries composed of different materials.
• Demonstrated independent design freedom: polymeric geometries can be created without imposing the path-length/continuity constraints required for electrodeposition, enabling complex soft structures interlocked with metallic cages/helices. SEM confirms high-resolution features and separable, interlocked components (scale bar 50 μm in examples).
• Materials versatility: Interlocked structures produced with shape-memory polymers (SMPs), PDMS (highly flexible silicone), and pure gelatin (no crosslinkers), including gelatin loaded with rhodamine B (0.1 wt%) to model drug cargo. Feature sizes targeted in the 5-μm range via prior casting method; hatching/slicing as small as 0.3 μm.
• Swarm aggregation control: Microparticles (metallic elements) mechanically stitched via compliant polymer links into meshes/filaments (scale bar 500 μm), preserving individual mobility while preventing loss/detachment and enabling tunable collective behavior via link geometry/density.
• Buoyancy control via hydrophobic/hydrophilic frames interlocked around helical swimmers: ecoflex (hydrophobic) frames cause migration to the air–water interface and floating forward motion under rotating fields; PDMS (hydrophilic) frames sink and enable rolling/tumbling locomotion. Demonstrations: ecoflex frame in DI water at 20 mT, 50 Hz shows linear floating motion; PDMS frame at 20 mT, 20 Hz shows rolling.
• Multiple locomotion modes from cage–bar and multi-cage designs: In-plane cage rotation around the bar yields no net translation; out-of-plane interception induces tumbling with net forward motion. Multi-element designs enable forward translation along long axis, rolling along short axis, and rapid reorientation/obstacle avoidance by switching magnetic field direction. Example conditions include silicone oil tumbling at 20 mT, 5 Hz and DI water rolling at 17 mT, 15 Hz.
• Process reproducibility: Frequency-dependent motion analyses (Supplementary) indicate consistent behavior across fabricated samples.
• Biocompatibility and magnetic responsiveness of fully iron elements are retained while adding polymer-enabled functions: enhanced drug loading (bulk hydrogel), elastic compliance, and on-demand shape transformation (SMPs).

The results validate that mechanical interlocking can resolve processing incompatibilities between metals and diverse polymers at the microscale, enabling simultaneous exploitation of each material’s strengths in a single machine. By decoupling metal and polymer via networks in a single positive-resist template, the authors achieve truly independent 3D geometries that remain mechanically bound after mold removal. This addresses the central challenge of integrating soft materials for drug loading, elasticity, and shape change with metallic components that provide high magnetic responsiveness and biocompatibility. The demonstrated devices exhibit controllable buoyancy, multiple magnetic locomotion modes (floating propulsion, rolling, out-of-plane tumbling), and strategies for controlled aggregation in swarms without sacrificing mobility. Compared to origami, 4D DLW, or microtransfer molding, the method imposes fewer geometric/material constraints and uniquely enables hard–soft interlocking at this scale. Although the TPA process is comparatively slow, emerging high-throughput femtosecond projection systems promise improved scalability. Overall, the approach expands the microrobotics design space by introducing topological interlocks as a robust, fabrication-compatible means to integrate multifunctional materials.

This work introduces a modular 3D fabrication strategy that mechanically interlocks fully metallic and polymeric microcomponents, producing multifunctional micromachines with unprecedented design freedom at the microscale. Using positive-resist TPA lithography to define independent via networks, electrodeposition for metal filling, and mold casting for polymers, the authors demonstrate devices combining iron’s strong magnetic response and biocompatibility with polymer-enabled functionalities (drug loading, elasticity, shape morphing). Applications shown include buoyancy-tunable swimmers, multi-mode locomotion (floating, rolling, tumbling), and swarm aggregation control using compliant links. Future research directions include: scaling throughput with projection-based TPA; integrating additional functional materials (e.g., degradable hydrogels, metal–organic frameworks, multiferroic composites); incorporating stimuli-responsive interlocks for on-demand disassembly; and advancing in vivo-relevant tasks such as targeted drug delivery, navigation through complex biological environments, and real-time imaging/contrast agent integration.

• Throughput: Two-photon lithography is relatively slow, limiting mass production; although faster femtosecond projection approaches may mitigate this.
• Geometric constraints from exposure voxel: The elliptical laser voxel necessitates minimum spacing between adjacent vias to avoid unintended fusion and complicates writing large continuous surfaces (e.g., spherical domes), reducing the available design workspace.
• Design clearances: Practical minimum distances (e.g., ≥10 μm between metal and polymer vias; polymer ≥4 μm above substrate) constrain the tightness of interlocks and feature packing.
• Process sensitivity: Parameters (laser power, slicing/hatching, baking and development times) require empirical tuning per geometry; variability may affect fidelity and yield.
• Material behavior: Soft-magnetic Fe components and polymer mechanics impose actuation and durability limits; interactions with substrates or interfaces (without buoyancy frames) can induce drift in fully metallic swimmers.
• Current demonstrations are benchtop; in vivo biostability, degradation, and long-term functionality of interlocked assemblies were not assessed.