Generalised optical printing of photocurable metal chalcogenides

The study addresses the challenge that optical 3D printing (e.g., DLP/SLA) is largely limited to photocurable polymer resins, restricting functionality for electronic, optoelectronic, and energy applications. Conventional lithography is costly, multi-step, and not well-suited for 3D structures, while direct ink writing enables inorganic architectures but lacks the resolution/throughput of optical printing. The authors propose a generalised, mask-less DLP-based printing method for functional inorganic metal chalcogenides using photocurable chalcogenidometallate (ChaM) inks. By exploiting photoacid-generator-induced precipitation of ChaM anions under patterned UV exposure, they aim to directly form high-fidelity 2D patterns and layer-by-layer 2.5D/3D architectures of crystalline metal chalcogenides, expanding optically printable materials beyond polymers. The work evaluates print fidelity, universality across materials, crystallinity, electrical/thermoelectric properties, and demonstrates device-level feasibility via a micro-scale thermoelectric generator.

• Optical printing methods (DLP/SLA) provide high-resolution, high-throughput, mask-less additive manufacturing but have been limited to photocurable polymers or polymer–inorganic composites, constraining functionality.
• DIW (extrusion-based) enables printing of functional inorganic materials via viscoelastic particle-loaded inks, but optical printing can offer superior resolution and scalability; prior DLP-printed thermoelectric BiSbTe showed low ZT (~0.12) due to organic residues; calcination can remove organics but risks oxidation in chalcogenides.
• Prior optical lithography of inorganic nanomaterials: photocurable nanoparticle films and azide-based crosslinkers for CdSe QDs enabled 2D patterns via photolithography and lift-off, but still required multi-step processes and were not inherently 3D-printable.
• Chalcogenidometallates (ChaMs) are soluble inorganic molecular anions (M_xCh_y) that decompose thermally to crystalline metal chalcogenides, making them promising precursors for printed semiconductors. The present work leverages ChaMs with photoacid generators to achieve direct optical printing.

• Inks: Diverse ChaMs (Pt, Sb, Sn, Cu, Mo) were synthesized by dissolving corresponding metal chalcogenide powders using ethylenediamine/ethanedithiol alkahest in N2 glovebox or by coordination chemistry to yield ChaM anions with ammonium/ethylenediammonium counter-cations. Purification to remove unreacted solids/by-products and solvent exchange to polar solvents (NMF or DMSO) ensured compatibility with photoacid generators (PAGs).
• Photoactivation: Two PAGs were used: MFVT (2-[2-(5-methyl furan-2-yl) vinyl]-4,6-bis(trichloromethyl)-1,3,5-triazine) and IM-NIT (N-(trifluoromethylsulfonyloxy)-1,8-naphthalimide), absorbing ~350–400 nm, suitable for 365/405 nm sources. Upon UV, PAGs generate protons that protonate ChaM anions (high proton affinity of S/Se moieties), inducing precipitation (loss of charge, particle growth confirmed by ζ-potential change from −28.7 to −13.6 mV and DLS size increase 106→295 nm in MoS2 inks). Inks remained stable (>14 days) in dark.
• DLP printing: Two systems used: • Commercial DLP printer (Nobel Superfine, 405 nm UV LED, area 64×120 mm, X–Y resolution 50 µm) installed in UV-protected N2 glovebox for air-sensitive inks. • Custom DLP (Luxbeam Rapid System, 365 nm UV, max area 5×3 mm, max power 4 W, minimum resolution ~10 µm) with g-code automation and 100 nm-scale Z-stage control for layer-by-layer curing. Digital masks generated target geometries; inks were exposed to patterned UV to locally precipitate and fix ChaMs, followed by rinsing.
• Layer-by-layer/3D: Sequential deposition and curing of ~50 nm-thick layers enabled 2.5D and micrometre-scale 3D architectures (e.g., pyramidal stacks).
• Post-processing to crystalline chalcogenides (inert conditions unless noted): • Sb2S3: 573 K, 1 min (ambient sulfur). • Sb2Se3: 573 K, 1 min. • SnS: 623 K, 1 min. • SnSe2: 573 K, 10 min. • SnSe: 723 K, 3 min. • Cu2S: 723 K, 10 min. • MoS2: two-step—573 K, 5 min in ambient sulfur to form MoS3, then 723 K, 15% H2, 10 min. • PtS2: pre-heat 873 K, 1 h in N2; then 773 K, 30 min in ambient sulfur (N2 glovebox).
• Characterisation: SEM (FE-SEM Nano230), EDS; XRD (Rigaku D/MAX2500V/PC, Cu rotating anode); UV–Vis (Shimadzu UV-2600); Raman (WITec Alpha300R, 532 nm) for MoS2 and PtS2; interferometric 3D scanning (NanoSystem NV-3000) for thickness/roughness; ζ-potential/DLS (Malvern Zetasizer); OM (Olympus BX53M).
• Electrical/thermoelectric: Hall effect (HMS-5000) and Van der Pauw conductivities; temperature-dependent σ and Seebeck (Netzsch SBA 458) from 300–600 K (Cu2S) and 300–550 K (SnSe2). Annealing times varied (5–15 min) to study carrier properties.
• Device fabrication: • In-plane micro thermoelectric generator: DLP-patterned ten pairs of 300 µm-wide Cu2S (p-type, anneal 723 K, 15 min) and SnSe2 (n-type, 573 K, 15 min) legs directly connected without metal interconnects. Performance measured with ceramic heater (hot side) and Peltier cooler (cold side), logging ΔT, voltage, current, and power density. • Cross-plane device: Bottom electrodes (Cr 5 nm/Au 80 nm) thermally deposited; DLP-printed Cu2S and SnSe2 (~500 nm thick) films formed between Au and Ag paste top electrode; resistance measured; transient thermoelectric response recorded for small ΔT (<1 K).

• Generalised DLP optical printing of photocurable ChaM inks produced high-fidelity inorganic patterns and 2.5D architectures across multiple metal chalcogenides: compound semiconductors (Sb2S3, Sb2Se3, Cu2S, SnS, SnSe) and 2D TMDs (SnSe2, MoS2, PtS2).
• Patterning performance: • Line widths from 100 to 25 µm achieved; minimum near equipment limit (~10 µm). Arrays of hundreds of 100 µm squares over millimetre-to-centimetre scales printed with excellent fidelity. • Large-area 1.2 cm × 1.2 cm square exhibited RMS roughness ~4.66 nm across the full area. • Complex multi-material patterns (e.g., MoS2 maze with PtS2 solution line) with linewidths 85 µm and 55 µm. • Layer-by-layer printing enabled 2.5D stacks: ~50 nm per layer; architectures with circular base and triangular top layers reached ~500 nm thickness with accurate profiles; 1.5 µm-thick pyramids (30 layers) comprising 500/400/300 µm squares demonstrated.
• Microstructure/crystallinity: • SEM shows smooth, dense films without substantial voiding; SnSe, SnSe2, SnS exhibited 2D plate-like grains due to layered structures. • XRD confirmed high crystallinity and texture: SnSe showed c-axis oriented (200)/(400)/(800) peaks; SnSe2 showed a-axis-oriented (001)/(003)/(004) peaks; other phases matched bulk references. • Raman verified MoS2 (E2g ~380 cm⁻1, A1g ~403 cm⁻1) and PtS2 (Eg1 ~300 cm⁻1, Ag1 ~339 cm⁻1).
• Electrical properties: • Cu2S (p-type): With anneal time 5→15 min, hole concentration decreased 1.58×10^20→6.29×10^19 cm⁻3; hole mobility increased 1.29→3.89 cm^2 V⁻1 s⁻1. RT conductivity up to ~1000 S m⁻1 (≈20% of bulk SPS reference). Seebeck positive; maximum ~230 µV K⁻1 at 600 K (10 min anneal). Power factor up to 0.187 µW cm⁻1 K⁻2 at 600 K. • SnSe2 (n-type): Electron concentration ~1.0×10^17 cm⁻3 (nearly constant with anneal time); electron mobility increased 1.06→2.25 cm^2 V⁻1 s⁻1; RT conductivity up to 279 S m⁻1 (≈30% of undoped high-quality single crystal). Seebeck −290 to −194 µV K⁻1 (300–550 K). Maximum power factor ~0.37 µW cm⁻2 K⁻2 at 550 K.
• Thermoelectric devices: • In-plane micro generator (10 p–n pairs): Output voltage and power scaled linearly/quadratically with ΔT. At ΔT = 65 K, reached 223.5 mV and 0.564 mW cm⁻2 power density. • Cross-plane device (~500 nm films): Device resistance ~150 Ω; with ΔT increase from 0.2→0.4 K, output voltage doubled and power quadrupled, evidencing reliable thermoelectric response despite small ΔT.

The work demonstrates that photocurable ChaM-based inks combined with DLP enable direct, mask-less optical printing of functional inorganic metal chalcogenides, overcoming the long-standing material limitation (polymer-only) in optical 3D printing. The proton-induced precipitation mechanism via PAGs yields spatially selective solidification, supporting high-fidelity 2D patterns and layer-by-layer 2.5D/3D architectures. Post-annealing converts patterns into dense, crystalline semiconductors with controllable texture and composition (e.g., SnSe2 vs SnSe via temperature), preserving key electronic properties. Electrical and thermoelectric measurements show mobilities >1 cm^2 V⁻1 s⁻1 and conductivities that are significant fractions (20–30%) of bulk or single-crystal references, validating material quality suitable for device integration. The successful fabrication of a micro-scale thermoelectric generator with predictable scaling of output metrics and substantial power density underlines device-level feasibility. Compared with traditional lithography, this approach simplifies processing (no masks/lift-off), reduces cost/time, and offers multi-material, scalable printing, suggesting a promising alternative platform for patterning inorganic semiconductors in electronics, optoelectronics, and energy devices.

The study establishes a generalised DLP-based optical printing platform for inorganic metal chalcogenides by formulating universally photocurable ChaM inks activated by photoacid generators. It enables high-throughput, mask-less fabrication of high-fidelity 2D patterns and 2.5D/3D architectures across diverse chalcogenides, yielding dense, crystalline films with competitive electrical and thermoelectric performance. A micro-scale thermoelectric generator printed entirely from inorganic semiconductors demonstrates practical applicability. Future directions include integrating higher-resolution optical systems (e.g., two-photon lithography) to push spatial resolution, extending the material library, improving densification/crystallinity to approach bulk properties, and expanding device types (heterostructures, integrated electronics/energy modules).

• Electrical properties, while competitive, remain below bulk counterparts due to smaller grain sizes, higher defect densities, potential impurity residues from precursor chemistry, and porosity inherent to solution processing.
• Current DLP process is not directly applicable to electrode materials; auxiliary deposition (e.g., Au, Ag paste) was required for cross-plane devices.
• Printing resolution and maximum pattern size are currently equipment-limited (minimum linewidth near ~10 µm; area dependent on projector field).
• Some inks and processing steps require inert atmospheres (N2 glovebox) and controlled sulfur ambience to prevent oxidation or composition loss during annealing.
• Extremely thin films in cross-plane configuration limited sustainable ΔT (>1 K) under steady-state conditions, constraining measured power output in that geometry.