Stretchable Complementary OTFT Neuron Circuit

• The paper demonstrates a fully photolithographic stretchable OTFT neuron circuit achieving 55,000 devices/cm², low-voltage operation, and up to 50% strain tolerance.
• It utilizes complementary p-type and n-type OTFTs with mobilities of 0.235 and 0.184 cm²/V·s, integrated via a transfer‐free photolithography process for high-resolution patterning.
• The neuron circuit exhibits integrate-and-fire behavior with firing frequencies ranging from 3 to 26 Hz, effectively mimicking biological neural activity under mechanical stress.

A stretchable complementary OTFT neuron circuit is a monolithically fabricated, mechanically resilient neuromorphic device in which every functional layer—including electrodes, dielectrics, semiconductors, and encapsulation—is intrinsically stretchable and compatible with high-resolution photopatterning. The architecture leverages complementary organic thin-film transistors (OTFTs) as critical elements to implement integrate-and-fire neuron-like behavior at low operating voltages, with the circuit output frequency directly modulated by input current. This technology addresses persistent challenges in skin-like electronics, such as compatibility across diverse polymer semiconductors and reliable complementary logic under mechanical deformation (Yuan et al., 16 Jan 2026).

1. Materials and Monolithic Fabrication Process

Intrinsic stretchability and broad semiconductor compatibility are achieved by leveraging a universal monolithic photolithography flow, in which all device layers are directly patterned and crosslinked, obviating the need for transfer or manual alignment steps. The process comprises:

• p-Type Semiconductor: Blend of diketopyrrolopyrrole–thieno[3,2-b]thiophene copolymer (DPPTT) with a perfluorophenyl-azide–end-capped polybutadiene (BA). UV irradiation at 254 nm and thermal annealing embed DPPTT fibers within a crosslinked rubber matrix, yielding $$\mu_p \approx 0.235\ \text{cm}^2\,\text{V}^{-1}\,\text{s}^{-1}$$, >10$^5$ on/off ratio, and 100% uniaxial stretchability.
• n-Type Semiconductor: Tetrafluorinated benzodifurandione oligo(p-phenylene vinylene) with 2,2′-bithiophene (F4BDOPV-2T), similarly crosslinked for $$\mu_n \approx 0.184\ \text{cm}^2\,\text{V}^{-1}\,\text{s}^{-1}$$ and on/off $>10^3$ under 100% strain.
• Dielectric Stack: Bilayer comprising nitrile butadiene rubber (NBR, $k \approx 25$) and photo-crosslinked SBS (styrene-butadiene-styrene, $k \approx 4$, $\sim$50 nm), ensuring a solvent-resistant, low-voltage interface.
• Encapsulation/Etch Mask: SBS ($\sim$3 µm), simultaneously serving as encapsulation and mask, forming "elastiff" islands for mechanical channel integrity.
• Electrodes: Crosslinked PEDOT:PSS for gates (patterned by lithography and O$_2$ plasma); CNT/Pd/Au source-drain patterned by PMMA/Cu lift-off, enabling minimum channel lengths down to 2 µm.

This fully photolithographic, transfer-free process achieves an unprecedented integration density of 55,000 OTFTs cm$^{-2}$, >95% yield on 4-inch wafers at 5 V operation.

2. OTFT Device Characteristics and Electrical Behavior

Both p- and n-type OTFTs in neuron circuits share matched W/L geometries and dielectric configurations. Typical device parameters include:

• Channel Dimensions: For logic, $W/L \approx 200\ \mu\text{m}/10\ \mu\text{m}$; standalone $W\approx 380\ \mu\text{m}$, $L\approx 35\ \mu\text{m}$
• Gate Capacitance: $C_i \approx 15\ \text{nF}\,\text{cm}^{-2}$
• Mobilities: $$\mu_p = 0.235\ \text{cm}^2\,\text{V}^{-1}\,\text{s}^{-1}$$; $$\mu_n = 0.184\ \text{cm}^2\,\text{V}^{-1}\,\text{s}^{-1}$$
• Threshold Voltages: $V_\text{th,p} \approx -0.06$ V; $V_\text{th,n} \approx +0.2$ V
• On/Off Ratios: p-type $>10^5$, n-type $>10^3$

In the saturation regime, transfer characteristics are governed by standard square-law equations: $$I_{D,p}(V_G) = \frac{W}{2L} \mu_p C_i (V_G - V_{th,p})^2$$

$$I_{D,n}(V_G) = \frac{W}{2L} \mu_n C_i (V_G - V_{th,n})^2$$

Measured I$_D$–V$_G$ curves fit these expressions up to $V_D = 5$ V, with off-currents below 1 nA and negligible hysteresis.

3. Neuron Circuit Topology and Operational Principles

At its core, the stretchable neuron circuit is based on a complementary inverter in a relaxation-oscillator configuration. Key features include:

• Complementary Inverter: Both p- and n-channel OTFTs share gate input $V_i$, supply $V_{DD}$, and matched geometries. The static voltage transfer curve (VTC) is determined by the equality of saturation currents: $$\frac{1}{2}\mu_p\frac{W_p}{L_p}C_i(V_{DD}-V_{in}-|V_{th,p}|)^2 = \frac{1}{2}\mu_n\frac{W_n}{L_n}C_i(V_{in}-V_{th,n})^2$$
• Neuron Relaxation Oscillator: The inverter output is connected to a node $V_m$ with membrane capacitor $C_m$ and high-resistance leak $R_\text{leak}$. The input current $I_\text{in}$ charges $C_m$ until $V_m$ reaches the inverter's switching threshold ($V_M$), at which point feedback causes rapid discharge (reset). The process then repeats, producing periodic spike outputs.

4. Neuron Dynamics and Integrate-and-Fire Behavior

During each cycle, the circuit performs temporal integration and spike generation analogous to biological neurons. Neglecting leak during the integration phase, the node equation is: $I_{in} = C_m\,\frac{dV_m}{dt}$ With $V_m$ integrating from $V_{reset}$ to $V_{th,on}$, the period for charging is

$$T_{charge} = \frac{C_m (V_{th,on} - V_{reset})}{I_{in}}$$

and thus the output frequency is

$$f_{out} = \frac{1}{T_{charge} + T_{reset}} \approx \frac{I_{in}}{C_m\,\Delta V} \quad \text{where} \quad \Delta V = V_{th,on} - V_{reset}$$

Empirically, with $C_m \approx 1$ nF and $\Delta V \approx 3$ V, the neuron fires at $f_{out}$ spanning 3–26 Hz as $I_{in}$ varies between 9 nA and 500 nA, consistent with linear theoretical prediction. The mechanism implements three principal neuronal behaviors: input integration, sharp thresholding (via inverter VTC), and regenerative reset through inverter feedback.

5. Performance Metrics and Mechanical Robustness

Key performance and mechanical attributes are summarized as follows:

Metric	Value/Range	Conditions
Supply Voltage ($V_{DD}$)	3–7 V (logic), 7 V (neuron operation)	3–7 V static; 7 V for neuron spiking
Output Frequency ($f_{out}$)	3–26 Hz (neuron), >3.3 kHz (ring osc.)	Neuron: $I_{in}=9$–500 nA; Ring osc.: $V_{DD}=60$ V
Power per Event	<0.1 μJ/spike	Event-driven regime
OTFT Density	55,000 cm$^{-2}$	2 μm min. channel length
Mechanical Stretchability	Up to 50% strain with <35% freq. loss	Neuron circuit

Circuits tolerate up to 50% tensile strain, with the neuron’s firing frequency degrading by <35%, primarily attributed to increased interconnect resistance. At the device level, individual OTFTs retain >60% of their mobility under 100% strain. The principal trade-off involves channel length: reducing $L$ to 2 μm raises integration density, but elevated contact resistance ($R_c$) limits kHz operation to $L \geq 10\ \mu$m. Thicker SBS encapsulation enhances robustness but marginally increases parasitic capacitance.

6. Significance and Broader Context

The implementation of a monolithic, stretchable complementary OTFT neuron circuit constitutes the first demonstration of a skin-like neuromorphic device with biologically relevant integrate-and-fire properties resilient to mechanical deformation (Yuan et al., 16 Jan 2026). The platform’s universal, photolithographic process for crosslinked dielectric/PSC interfaces and self-aligned encapsulation enables adaptation to new polymer semiconductors without bespoke process development, addressing longstanding limitations in material specificity and compatibility.

By achieving high device densities, low-voltage operation, and robust mechanical performance, this technology accelerates the integration of complex neuromorphic functions into skin-conformal, event-driven electronics, directly supporting the future implementation of distributed on-skin and soft robotics systems requiring high mechanical compliance and functional mimicry of biological circuitry.