Sensorized tissue analogues enabled by a 3D-printed conductive organogel

The field of healthcare simulation has grown rapidly to reduce medical errors, with high-fidelity simulations shown to improve performance-based outcomes. While current artificial tissues provide realistic appearance and feel, they largely lack integrated sensing for objective, real-time assessment. Quantifying tissue deformation is especially relevant for reconstructive procedures like skin flaps, where strain impacts healing and scarring. Piezoresistive strain sensors are preferred in medical simulation for their small footprint and insensitivity to stray capacitance; however, traditional composite-based sensors with solid conductors in elastomers suffer from drift and nonmonotonic responses. Alternative conductive media include liquid metals (e.g., EGaIn), hydrogels, and ionogels, each with drawbacks such as cost and uncertain biocompatibility (liquid metals), environmental instability due to water evaporation (hydrogels), and toxicity concerns (some ionogels). Deep eutectic solvents (DESs) offer a promising, low-cost, and potentially biocompatible ionic medium with low vapor pressure and wide liquid windows. This work introduces inexpensive, biofriendly, 3D-printable organogels that use a DES of choline chloride with PEG200 or propylene glycol (1:5 molar ratio) and fumed silica as a gelating agent, enabling monolithic integration of piezoresistive strain sensors into lifelike Y/V plasty training pads to objectively quantify soft tissue deformation during surgical simulation.

The paper surveys prior strain sensing approaches for soft, stretchable systems: composite piezoresistive sensors with dispersed solid conductors show high sensitivity but suffer from drift and nonmonotonic responses due to irreversible conductor rearrangement; liquid metal-based sensors (e.g., EGaIn) achieve high strains (≥100%) but are costly, involve critical materials (indium), and have uncertain biocompatibility; conductive hydrogels can reach very high strains (up to 1000%) but are prone to instability from water evaporation in air; ionogel-based sensors offer stable, low-hysteresis performance but raise toxicity concerns for many ionic liquids. Deep eutectic solvents (DESs) share favorable properties with ILs (low vapor pressure, wide liquid window, limited flammability) while being inexpensive and more biocompatible, motivating DES-based organogels. Prior rheological work shows that silica surface functionalization strongly influences glycol-silica mixture behavior, and cellulose- or starch-thickened DES gels have been reported, but shear-thinning DES organogels using fumed silica as the thickener had not been demonstrated.

Organogel synthesis and characterization: Choline chloride (≥98%) was dried (130 °C, 72 h, vacuum) and combined with either propylene glycol (PG) or polyethylene glycol (PEG200) at a 1:5 molar ratio (ChCl:HBD). Mixtures were stirred for 2 h at 90 °C to form DES. Fumed silica (Aerosil) particles with different surface capping—R711 (3-trimethoxysilylpropylmethacrylate), R974 (methyl), and 200 (hydroxyl)—were added at varying wt.% and mixed with a planetary mixer (Thinky ARV-310, 2000 RPM, 10 min). Rheology was measured at 20 °C using 25 mm parallel plates (gap 0.5 mm) on an Anton Paar MCR302 rheometer: viscosity flow curves (shear rate 0.1–500 s−1); oscillatory amplitude sweeps (strain 0.01–100%, frequency 1 rad s−1). FTIR (ATR, Thermo Nicolet iS10) characterized silica; TGA (Mettler-Toledo TGA/DSC 3+) performed under N2 from 27–720 °C at 5 °C/min on ~76.6 mg gel. Selection of PEG-based organogel was guided by lower vapor pressure (PEG200 1.69×10−2 Pa vs PG 18.9 Pa) and high onset decomposition temperature (~227.4 °C). Sensor fabrication (freestanding): A nylon fabric sheet was mounted in a 3D-printed PLA mold. A skin-mimicking PDMS layer (Polytek PlatSil Gel-25: 20 g Part A + 20 g Part B + 20 g Deadener LV) was mixed, poured, stippled into the nylon, and cured (~30 min). Pairs of silver/nylon composite conductive threads (Agsis-Lite, Syscom) were stitched into the substrate (~15 mm separation for freestanding sensors; note: 15 cm is listed in Methods for substrate stitching, while electromechanical gage length used electrode tip spacing). The PEG-based organogel was loaded into a syringe, degassed by centrifugation (9000 RPM, 10 min), and 3D-dispensed (N-Scrypt 3Dn-300 with SmartPump) through a 20G blunt needle at 35 psi, 3 mm/s, 0.6 mm standoff to lay 25 mm lines bridging the electrodes. A subsequent PDMS layer (equimass Gel-25 A:B:Deadener LV) encapsulated the channels and cured overnight. Rectangular samples (10 × 64 mm) were cut for testing. Y/V plasty suture training pad fabrication: A skin layer (60 g Gel-25 skin formulation pigmented with 0.5 g Silc Pig Flesh) was cast over taut nylon. On the cured skin, two sets of three 20 mm organogel lines were printed; conductive thread electrodes were inserted ~5 mm into each line. A second skin layer (60 g) embedded the sensors. After cure, a thin petrolatum layer was applied to the skin surface with a 1 cm petrolatum-free perimeter. A “fat” layer (80 g PDMS, 1:1:2 Gel-25 A:Gel-25 B:Deadener LV, pigmented with 1 g Silc Pig Yellow) was cast and cured (~30 min), then the pad was demolded. A Y-shaped incision was cut (each segment 30 mm from nexus) guided by a stencil. Sensor placement: two perpendicular to the stem (10 mm from cut, 15 mm from bottom of Y) and one collinear with the stem (27.7 mm above the junction). Separate skin and fat layers were prepared for mechanical testing (ASTM D638 Type V dogbones). Electromechanical characterization: Samples were clamped in custom 3D-printed grips lined with sandpaper on a dynamic mechanical tester (TA ElectroForce TestBench). Preload 0.1 N; cross-sectional area by laser micrometers (Keyence IG-028). Gage length defined as distance between electrode tips (measured by calipers). Electrical stimulus: ±2 V square waves at 50 Hz; current measured at 500 Hz with an impedance analyzer (Ametek VersaSTAT 3) to compute resistance. Dynamic cyclic protocol: ten cycles at 0.05 Hz to strain amplitudes of 50, 100, 150, 200, 150, 100, 50% (total 70 cycles), followed by a 1%/s ramp to failure. Data processing: mean resistance per ± pulse pair (0.04 s), manual synchronization with mechanical strain, 20-point adjacent-averaging smoothing. Additional cyclic testing to 100% strain at 0.3 Hz for 1000 cycles. Mechanical testing of skin and fat layers at strain rates 0.16 mm/s and 3.0 mm/s. Sensor stability study: fabricated 12 sensors as above; continuously monitored one sensor’s resistance at 100 Hz during encapsulation (LCR-600) and stored sensors in a desiccator to assess resistance stability up to 1 week.

• Organogel rheology: Adding 12 wt.% R711 (3-trimethoxysilylpropylmethacrylate-capped) fumed silica to a PEG200:ChCl DES (1:5 molar ratio) produced a shear-thinning gel suitable for 3D printing, confirmed by inversion test and oscillatory sweep showing a linear viscoelastic region up to ~0.1% strain and G′/G″ crossover at higher strains. In contrast, R974 (methyl-capped) and Aerosil 200 (hydroxyl-capped) mixtures predominantly exhibited shear-thickening behavior at higher loadings (8–12 wt.%).
• Material stability and selection: PEG-based organogel chosen due to much lower vapor pressure (PEG200 1.69×10−2 Pa vs PG 18.9 Pa) and high onset decomposition temperature (~227.4 °C).
• Freestanding sensor performance: Sensors showed consistent, drift-free, monotonic piezoresistive response up to 300% strain across dynamic ramps. Exposure to large strains did not distort subsequent smaller-strain amplitudes.
• Response model: Relative resistance change followed a parabolic relation with strain up to 300% with excellent fit, Ravg = 0.998 ± 0.001 (n = 20). The fitting parameter β averaged 0.621 ± 0.059 at 300% strain cycles, indicating acceptable contact resistance with conductive thread electrodes; comparable wire-electrode fluidic sensors report β ≈ 0.746–0.797.
• Hysteresis: Low hysteresis with average 1.34 ± 0.76% (n = 20), superior to several wire-electrode fluidic sensors over similar strain ranges.
• Cyclic durability: Stable, drift-free operation for 1000 cycles at 100% strain and 0.3 Hz; amplitude and frequency tracking were excellent.
• Stability: Sensors stored in a desiccator exhibited stable resistance for at least one week post-fabrication.
• Electrode impact: Increased contact resistance from conductive thread electrodes relative to wires did not significantly impact sensor performance.
• Application demo: In a Y/V plasty suturing simulation, the flap sensor’s resistance rose with initial elongation during flap mobilization and remained constant as sutures were placed successfully, demonstrating actionable, real-time feedback of tissue strain.

The study addresses the need for objective, real-time quantification of soft tissue deformation in high-fidelity medical simulation by introducing a monolithically integrated, minimally intrusive strain sensor based on a biocompatible, low-cost, and 3D-printable DES organogel. By tailoring fumed silica surface chemistry (methacrylate-capped R711) and loading to achieve shear-thinning behavior, the authors create printable inks that form stable, conductive channels in elastomeric tissue analogues. The resulting sensors overcome key limitations of composite-based piezoresistive sensors (drift, nonmonotonic responses) and of alternative fluidic media (cost/biocompatibility for liquid metals, environmental instability for hydrogels, toxicity for many ionogels). Electromechanical characterization demonstrates monotonic response up to 300% strain, low hysteresis, negligible drift, and long cyclic life, ensuring straightforward signal interpretation without complex processing. The successful Y/V plasty training pad demonstration shows that strain readouts can reflect procedural quality (e.g., consistent tension after suturing), enabling objective performance assessment while preserving visual and tactile realism. Conductive thread electrodes, though more resistive than wires, maintained robust performance, facilitating easy integration with textile elements of training pads. This work establishes DES-based organogels as a tunable platform for embedded soft sensors in medical simulation and suggests broad applicability for quantifying tissue mechanics during training and assessment.

This work introduces a new class of shear-thinning, 3D-printable DES organogels (ChCl with PEG200 or PG plus methacrylate-capped fumed silica) and demonstrates their use in stretchable piezoresistive strain sensors monolithically integrated into lifelike Y/V plasty training pads. The sensors achieve up to 300% strain with monotonic, drift-free responses, low hysteresis (~1.34%), and durability over 1000 cycles at 100% strain, while using inexpensive, nonhazardous components and textile-compatible electrodes. The approach provides a minimal-footprint means to quantify tissue deformation for objective, real-time feedback in medical education. Future work will focus on improving long-term stability of the organogels, developing multi-sensor modules for simultaneous strain mapping, conducting validation studies with surgical trainees using the Y/V model, extending applications to other simulation scenarios, and exploring UV-enabled structures leveraging the methacrylate-functionalized silica surface (e.g., auxetic frameworks, UV-curable devices).

• Long-term stability: The authors note the need to develop organogel compositions with improved stability; stability was demonstrated over one week in a desiccator, but longer-term and varied environmental stability were not fully evaluated.
• Validation in training: The Y/V plasty model has not yet undergone formal validation with surgical trainees; future validation studies are planned.
• Multi-sensor integration: Demonstrations focused on individual sensors; simultaneous multi-sensor strain detection within complex modules remains future work.
• Material dependency: Achieving printable shear-thinning behavior required specific methacrylate-capped fumed silica (R711); other surface chemistries led to shear-thickening and may limit formulation flexibility without further optimization.