I. Abstract of the Paper’s Core Research

• 9:1:1 ternary solvent system for stabilizing graphene aerosol ink

A ternary blend of ethanol, butanol, and terpineol contains no additional binders, exhibits a viscosity of 6.3 cP and a surface tension of 28.7 mN/m, has a particle size distribution centered at 295 nm, and remains free of sedimentation or nozzle blockage after standing for 5 weeks, making it ideally suited for ultrasonic atomization aerosol systems.

• Sub-5 μm ultra-high-resolution direct-write fabrication without post-processing

Using a 100-μm nozzle and tuning the focus ratio to FR = 2, it is possible to directly print 5-μm‑wide, high‑resolution graphene lines without requiring post‑annealing for pattern refinement. Switching to a 200-μm nozzle enables scalable, stable device fabrication, yielding uniform lines with clean edges and no filamentary burrs or material overflow.

• Multilayer Regulation Enables Low-Resistance Graphene Conductive Films

With 10, 20, and 30 printing cycles, the film thickness increases progressively; after 30 cycles, the sheet resistance of the thin film is reduced to as low as 110 Ω/□. A subsequent 350°C stepwise annealing process removes residual organic residues, resulting in densely packed graphene layers and a significant improvement in electrical conductivity stability.

• Dual-functional flexible devices exhibit excellent performance.

① Interdigital Capacitive Electrodes (IDE): A design without a dielectric layer, delivering a stable output of approximately 2 pF for touch response; after one thousand bending cycles, the capacitance retention remains above 88%, and the touch signal is reversible with no drift.

② Serpentine Joule‑heating thin film: 150 mW, ultra-low power consumption, with a maximum temperature rise of 56°C; exhibits uniform thermal distribution and maintains performance after thousands of thermal cycles and tens of thousands of bending cycles.

• Fully digital, maskless flexible manufacturing solution

CAD enables one-click switching between touch‑screen and heated‑device layouts, is compatible with thermally sensitive flexible substrates such as PI and PET, and streamlines the workflow while reducing R&D iteration cycles by 90%.

II. Complete Preparation Process Flow and Core Process Advantages of Aerosol Spraying

(1) Complete Fabrication Process for Flexible Graphene Devices

• Few-layer, high-quality graphene flakes were prepared via liquid-phase exfoliation, with large agglomerates removed by centrifugation.
• A ternary ink was formulated with an ethanol/butanol/terpinolene ratio of 9:1:1 and thoroughly dispersed by sonication.
• Plasma cleaning of Kapton polyimide substrates, with the workbench maintained at a constant temperature of 80–90°C;
• AJ ultrasonic atomization system, with layered printing paths available in 10-, 20-, and 30‑step configurations; small nozzles for high‑precision samples, and large nozzles for mass‑production devices.
• Stepwise heating annealing: pre-bake at 170°C, then hold at 350°C to remove residual solvents.
• Electrical characterization, thermal imaging, and bending‑cycle testing were conducted to complete device calibration.

( II) Five Core Process Advantages (Comparison of Inkjet, Lithography, and Screen Printing)

• Self-developed ternary ink compatibility ensures long-term atomization without clogging.

Unlike binary solvent inks, which are prone to aggregation and coffee-ring defects, this formulation employs a synergistic blend of three solvents to precisely regulate the evaporation rate, ensuring stability after five weeks of storage and trouble-free continuous printing in mass production.

• Up to 5 μm ultra-high printing accuracy, far surpassing the 30 μm lower limit of inkjet technology.

Focuses on sheath‑gas‑assisted beam compression of aerosol streams, enabling single‑channel direct‑write fabrication of ultrafine traces, and is compatible with high‑density microsensor arrays and narrow‑pitch flexible electrodes.

• Dual-device compatibility on a single production line, with multifunctional integrated manufacturing.

Using the same equipment and the same ink formulation, switching only the CAD layout enables the production of both touch‑sensing and heating‑film devices, resulting in an exceptionally high line‑utilization rate.

• Low-temperature, full-process technology, compatible with a wide range of heat-sensitive flexible substrates.

The printing substrate is heated to only 80°C, with a maximum annealing temperature of 350°C; PET, PI, and PU films remain undeformed, and the mechanical properties of the flexible substrates are not compromised.

• Lithography-free, etching-free, green, low-cost digital manufacturing

Eliminates multiple process steps—including mask fabrication, coating, etching, and stripping—produces no chemical waste, and is suitable for both small‑batch prototyping and large‑scale mass production.

III. Five Core Technological Innovations of This Study

Innovation 1: A completely new ternary composite graphene aerosol ink system

Innovation 2: Focused-beam direct-write process with sub-5μm feature size and no post‑processing.

Innovative Method for Preparing Thin, Controllable-Thickness, Low-Resistance Graphene Films

Innovation 4: Integrated Flexible Capacitive Touch Electrode Without a Dielectric Layer

Innovative 5‑ink dual‑function device integrated manufacturing paradigm

IV. Existing Process Challenges and Corresponding Solutions for Overcoming Them

Challenge 1: Conventional graphene inks tend to settle, and prolonged aerosol printing can clog the nozzle.

Challenge 2: The minimum line width for inkjet printing is 30 μm, making it impossible to fabricate micro-sensor arrays.

Challenge 3: Volatilization imbalance in binary inks leads to coffee-ring effects and uneven line widths during printing.

Challenge 4: Carbon films have poor conductivity, requiring a high number of processing steps and elevated annealing temperatures, which can easily damage the substrate.

Challenge 5: Touch‑sensing devices require an additional dielectric layer, increasing process complexity and the risk of bending‑induced failure.

V. Future Diversified Industrial Application Scenarios

• Wearable health monitoring devices

• Curved Smart Appliance Touch Panel

• Soft robotic bionic electronic skin

• Flexible, Lightweight IoT Sensing Tag

• Flexible Electronics Laboratory Rapid Prototyping Platform for Universities

Figure 1. Schematic diagram of aerosol jet printing based on graphene patterns.

Figure 2. Characterization of exfoliated graphene. (a) AFM image showing the surface topography of the exfoliated flakes. (b) Histogram of flake thicknesses derived from AFM data, with most flakes thinner than 50 nm. (c) Distribution of lateral dimensions of the flakes as measured by AFM, with an average size of approximately 300–400 nm. (d) Raman analysis reveals the D, G, and 2D spectral features; the I₂D/IG intensity ratio indicates multilayer graphene. (e) Low-magnification TEM image showing stacked graphene layers. (f) High-resolution TEM image clearly displays lattice fringes with an interlayer spacing of 0.35 nm; inset: SAED pattern confirming the hexagonal structure of graphene.

Figure 3. Properties of graphene ink and the influence of printing parameters: (a) Dynamic light scattering (DLS) results for graphene ink used in particle-size analysis; (b) Time-dependent variation of surface tension; (c, d) Shear‑thinning behavior of viscosity as a function of shear rate, along with the corresponding relationship between shear stress and shear rate. Effects of nozzle size and focus ratio (FR): (e) Under fixed carrier‑gas flow of 25 sccm and a platform‑travel speed of 1 mm·s⁻¹, the line width and thickness vary with FR when using a 100 µm nozzle; (f–h) Optical micrographs of single‑pass graphene lines deposited with a 100 µm nozzle at a platform‑travel speed of 1 mm·s⁻¹, showing that line width decreases as the focus ratio increases (FR = 0.75, 1.0, and 2, respectively); (i) Deposition instability observed outside the stable focusing range (FR > 2): excessively high back pressure from the sheath‑flow disrupts aerosol‑jet stability, leading to diffuse and nonuniform splatter. All images include a 40 µm scale bar.

Figure 4. Effects of nozzle size, FR value, and printing speed on print line width and thickness: (a) Variation of line width with FR value for different nozzle sizes (150 and 200 µm); (b) Variation of thickness with FR value for different nozzle sizes; (c) Dependence of line width and thickness on printing speed when using a 200‑µm nozzle. Error bars represent the standard deviation obtained from multiple measurements taken at different locations along each printed line.

Figure 5. Electrical characteristics of the TLM structure: (a) Schematic diagram of the TLM structure; (b) Thermogravimetric analysis (TGA) showing the relationship between weight percentage and heat flow as a function of temperature; (c) Current–voltage (I–V) characteristics of the printed TLM structure; (d) Dependence of total resistance on channel length; error bars represent the standard deviation obtained from multiple measurements of printed TLM devices; (e) Extracted sheet resistance as a function of printed film thickness for graphene films deposited 10, 20, and 30 times.

Figure 6. Joule heating effect: (a) Experimental setup and optical image of the serpentine structure with an active area of 55.5 mm²; (b) Thermal imaging showing that as the voltage increases, the power gradually rises, enabling uniform heat dissipation and sustained heating within the serpentine structure; (c) Temperature–power relationship curve; (d) Temperature–time relationship curve; (e) Continuous heating at a power level of 84 mW.

Figure 7. Capacitance measurement results for the IDE structure: (a) Optical image of the printed IDE structure, with a line width of approximately 130 µm; (b) Time-dependent capacitance curve; (c) Temporal evolution of capacitance under applied stimulation.

Figure 8. Capacitive touch-sensing performance of AJP‑printed graphene IDEs. (a) Variation of the baseline capacitance with the number of bending cycles for devices printed 10, 20, and 30 times. (b) Trends in normalized capacitance change (ΔC/C0) and capacitance retention as a function of the number of bending cycles, demonstrating the sensor’s stable response under repeated mechanical deformation.