Flash healing of laser-induced graphene

Graphene is a renowned two-dimensional material with exceptional mechanical, electrical, and thermal properties. Traditional synthesis routes (CVD, liquid-phase exfoliation, reduction of graphene oxide) face scale-up and process limitations. Laser-induced graphene (LIG) introduced in 2014 enables direct, mask-free conversion of carbon-rich precursors into 3D porous graphene using commercial lasers under ambient conditions, offering porosity, flexibility, and cost-effectiveness. However, the ultrafast laser pulse durations (microseconds to femtoseconds) promote amorphous, polycrystalline structures that degrade electrical conductivity, limiting LIG performance in electronics. Controlling defect density and crystal order in LIG patterns is therefore critical. Conventional defect-healing methods (high-temperature furnace annealing; chemical reduction) are unsuitable for patterned LIG on polymers due to substrate damage or inability to rearrange the carbon lattice. This work hypothesizes that flash Joule heating (FJH), with longer pulse durations than laser scribing yet still ultrafast, can locally and rapidly heal LIG’s topological defects while preserving pattern geometry and porosity, thereby improving crystallinity and conductivity for device applications including strain sensing and low-voltage sterilization.

Prior graphene production methods (CVD, exfoliation, reduction) demonstrated feasibility but suffer from specialized equipment needs, harsh chemicals, and energy-intensive steps. LIG expanded accessible fabrication using IR, visible, and UV lasers across diverse substrates and precursors, but short pulse durations induce amorphous structures favorable for catalysis yet suboptimal for conductivity. Conventional post-treatments such as furnace annealing (hundreds to thousands of °C) can carbonize or ablate polymer substrates, and chemical reductants (e.g., hydrazine, NaBH4, hydrohalic acids) remove oxygen functionalities without inducing atomic rearrangements to restore crystallinity. Flash Joule heating has been applied to carbon materials: ultrafast welding of carbon nanofibers to enhance graphitization (Hu’s group), gram-scale flash graphene from diverse feedstocks (Tour’s group), and tuning CNT/graphene hybrids via FJH; however, prior efforts emphasized bulk production rather than defect healing within pre-patterned LIG. Multiple laser passes can modestly improve LIG conductivity but risk structural collapse with repeated exposure. These gaps motivate a fast, localized post-treatment capable of healing defects in patterned LIG without damaging substrates.

F-LIG fabrication: LIG patterns were first written on 250 µm-thick PI films using a CO2 laser (10.6 µm; vector mode; power 4.8 W; 10 kHz; duty cycle 8%; scan speed 1000 mm/s; 5 pulses/dot; 30 µm line spacing). For FJH, pre-formed LIG patterns were shaped as dumbbells (effective region 1 mm × 10 mm for structure study; 1 mm × 20 mm for sensors) to localize heating in the narrower section. Silver paste was applied to the wider ends for electrical contact. FJH was conducted in a vacuum chamber using high DC voltage pulses (typically 20 ms), while recording instantaneous voltage and current to compute areal power density (PA = UI/A) and areal energy density (EA = ∫UI/A dt). Temperatures during FJH were determined by fitting emitted black-body radiation with an IR thermometer. Pulse duration effects were studied at constant 130 V with durations 20–60 ms. Characterizations: Resistance (two-point probe) was measured before/after FJH. Raman spectroscopy (532 nm) assessed defect density (ID/IG), crystalline domain size La via La = (2.4×10−10)×λ×(IG/ID), 2D band FWHM, I2D/IG, and peak shifts. XPS (survey and C 1s) quantified elemental composition and bonding changes. SEM observed morphology and porosity preservation. HRTEM (Cs-corrected JEOL ARM 300F2, 80 kV) imaged atomic-scale lattice order. High-energy synchrotron total scattering (APS 11-ID-C; λ = 0.1173 Å; Q up to 24.7 Å−1) provided PDFs via GSAS-II and PDFgetX3; G(r) was obtained via Fourier transform of S(q). XRD (Cu Kα, λ = 1.54 Å) evaluated interlayer spacing and (002) peak evolution (powdered samples). Adaptability tests: FJH was applied to LIG on different substrates (PES, PEEK, PES/lignin) and to LIG powders in a quartz tube (DC 70–130 V, ~100 ms). Strain sensor fabrication and testing: For bending sensors, LIG/F-LIG on PI (1 mm × 20 mm) were contacted with copper wires/silver paste; bending angles (25°–90°) were applied on an Instron 3382 UTM while monitoring ΔR/R0 at 1 V. For stretchable sensors, LIG/F-LIG were transferred to PDMS (Sylgard 184, 10:1 w/w, degassed, cured at 80 °C for 2 h) by casting onto LIG/PI, curing, peeling PI, wiring, and encapsulating with PDMS. Tensile tests applied strains up to 10% at speeds 5–35 mm min−1; signals were acquired with a Keithley 2612B. Gauge factor (GF) was obtained from ΔR/R0–strain curves. Human-motion, phonation, and robotic control: F-LIG-H/PDMS sensors were attached to body sites (wrist/face) or a phone microphone to capture signals. For robotic control, five bending sensors were mounted on glove fingers in a voltage-divider circuit (Vout = Vsource Rc/(Rs+Rc)), feeding an Arduino to actuate a robotic hand. Low-voltage sterilization: Square LIG/F-LIG films (10 mm × 10 mm) were incubated with E. coli (~1 h at 37 °C), rinsed, then subjected to DC biases (1, 3, 5 V) for 2 min. Bacteria were detached by sonication and quantified by CFU plating. Current density and surface temperatures were monitored to evaluate the sterilization mechanism.

• FJH induces rapid defect healing in LIG within milliseconds while preserving 3D porous architecture. Visual flashes indicated successful FJH; temperatures reached (20 ms pulses) increased with voltage: ~1300, 1700, 2100, 2300, and 2500 °C at 150, 160, 170, 180, and 190 V, respectively. A threshold in Joule heat/temperature was identified; too high voltage (e.g., 200 V) caused breakdown. • Electrical improvements: For 1 mm × 10 mm patterns with initial resistance ~590 Ω, FJH reduced resistance to ~120 Ω at 190 V, yielding ~5× higher conductivity. At 190 V, PA ~2100 W cm−2 and EA ~27.55 J cm−2. Pulse-duration study at 130 V: minimal change at 20–30 ms; resistance decreased from 600 Ω to 530 Ω at 40 ms and to 150 Ω at 50 ms; 60 ms caused layer breakage. • Raman: ID/IG decreased from 0.84 (LIG) to 0.33 (F-LIG-190V); La increased from 22.9 nm to ~60 nm (~2.6×). 2D band FWHM narrowed from 109.4 to 63.8 cm−1; I2D/IG increased from 0.73 to 1.05; 2D peak exhibited up to ~27 cm−1 blue shift at 190 V, consistent with improved quality and reduced stacking. • XPS: Carbon content increased from 96.28% (LIG) to 98.53% (F-LIG-190V); oxygen and nitrogen decreased to 0.78% and 0.69%, respectively, mainly via reduction of C–O bonds, indicating higher purity/fewer defects. • Atomic-scale order: HRTEM showed transformation from disordered pentagon–heptagon-rich structures to extended hexagonal lattices after FJH. PDFs revealed nearest-neighbor C–C distance narrowing and shortening: from 1.458 Å (LIG) to 1.444 Å (F-LIG-170V) and 1.425 Å (F-LIG-190V), with enhanced long-range order. • XRD (powdered LIG under FJH): (002) peak sharpened and shifted from 2θ = 25.9° (interlayer spacing ~3.47 Å) to 26.4° (~3.41 Å), consistent with increased crystallinity and reduced interlayer spacing due to removal of interlayer species. • Adaptability: FJH improved resistance and Raman signatures across varied LIG shapes/dimensions and substrates (PES, PEEK, PES/lignin) and in powdered LIG (ID/IG from 0.84 to 0.27; La from 23.0 to 71.9 nm at higher voltages). • Strain sensing (bending on PI): Distinct, repeatable ΔR/R0 signals; at 90° bending, ΔR/R0 reached 21.37% (F-LIG-H) and 8.82% (F-LIG-M) vs 1.91% (LIG), i.e., ~11.2× and ~4.6× greater than LIG. • Stretchable sensors (PDMS): At 10% strain, ΔR/R0 exceeded ~1180% for F-LIG-H/PDMS vs 163% for LIG/PDMS. Gauge factors: 129.3 (F-LIG-H) vs 16.4 (LIG). Sensors showed negligible hysteresis, speed-independent response (5–35 mm min−1), and stable performance over 2000 cycles (~13,000 s). • Application demonstrations: Monitoring eye blinks, mouth motion, phonation (“good”, “hello”, “graphene”, “sensitivity”), and wrist pulses (resolving P- and D-waves). Smart glove with F-LIG-H enabled accurate, rapid robotic hand gesture control and Morse code transmission (“SOS”, “HELP”). • Low-voltage sterilization: After equal initial bacterial adhesion (~8.5 × 10^5 CFU mL−1), under 5 V DC for 2 min, LIG achieved 76.3% killing whereas F-LIG reached 99.94% (viable counts ~2.09 × 10^6 CFU mL−1 on LIG vs ~53 CFU mL−1 on F-LIG). F-LIG had higher current density and achieved ~57 °C surface temperature at 5 V; current-driven mechanisms (electroporation and direct electron transfer) dominate at moderate temperatures and short durations.

The study demonstrates that flash Joule heating effectively heals topological defects in laser-induced graphene while preserving its 3D porous architecture and pattern fidelity. The longer yet still ultrafast heating durations of FJH (tens of milliseconds) enable atomic rearrangements that laser scribing alone cannot provide, reducing defect density, enlarging crystalline domains, narrowing and blue-shifting the 2D Raman band, shortening nearest-neighbor C–C distances, and enhancing long-range order. These structural improvements translate directly into markedly reduced resistance and increased conductivity, addressing a core limitation of LIG for electronic applications. A clear energy/temperature threshold for the flash-healing effect is identified, enabling process tuning; overly high energy leads to damage, highlighting the importance of precise control. The method is adaptable to diverse LIG geometries, substrates, and even powders, indicating broad applicability. In devices, the improved electrical properties and preserved porous scaffolds yield highly sensitive, stable strain sensors with large gauge factors and robust cycling, enabling precise human-motion monitoring, phonation recognition, and responsive human–machine interfaces. Enhanced low-voltage antibacterial performance further showcases the utility of F-LIG, where increased conductivity raises current density and contributes to superior bacterial killing at modest temperatures, likely via electroporation and electron transfer. Overall, FJH provides a fast, localized, and scalable route to tailor the crystallinity and performance of LIG for advanced soft electronics and antimicrobial interfaces.

Coupling laser-induced graphene with flash Joule heating yields flexible, patterned graphene with substantially reduced defect density and enhanced crystallinity and conductivity, achieved in milliseconds without compromising the 3D porous architecture. Comprehensive characterization (Raman, XPS, HRTEM, PDFs, XRD) confirms amorphous-to-crystalline transformation, shorter C–C bonds, reduced interlayer spacing, and higher purity. The improved material enables high-performance piezoresistive strain sensors with an ~8-fold sensitivity enhancement (GF ~129 vs ~16 for LIG), stable operation over thousands of cycles, and practical demonstrations in human-motion tracking, phonation recognition, robotic hand control, and Morse code communication. F-LIG also exhibits superior low-voltage antibacterial activity. The FJH approach offers a straightforward, rapid, and versatile solution to LIG’s intrinsic structural defects, with potential to advance high-performance graphene-based electronics and inspire further applications, including supercapacitors and chemical sensors.

• Process window constraints: A threshold in energy/temperature is required to trigger defect healing; insufficient energy (e.g., lower voltages or shorter pulses) yields minimal improvement, whereas excessive energy (e.g., ~200 V or overlong pulses such as 60 ms at 130 V) can cause LIG breakdown. • Equipment limitations: The power supply’s rising speed limited achieving set voltages above ~200 V (actual maxima ~370–380 V when set to 410–430 V), constraining precise comparison across settings. • Multiple lasing cycles provide limited improvement and can collapse LIG, underscoring the need for careful post-treatment selection. • XRD data were collected on powdered LIG; while trends likely reflect patterned LIG, direct diffraction on patterned samples was not reported. • Temperature estimates relied on black-body radiation fitting; absolute temperatures may be affected by emissivity assumptions.