Self-energized multi-mode sensing electronic skin and preparation method and application thereof

Abstract

The invention discloses a self-energy-supply multi-mode sensing electronic skin based on ionic gel heterojunction array integration and a preparation method and application of the self-energy-supply multi-mode sensing electronic skin. The electronic skin adopts a multi-layer composite structure design and comprises an electrode array layer, a silver paste connecting layer, a nano material doped ionic gel array layer and a pure ionic gel continuous layer, and passive self-energized work is realized through a space ion migration effect at a heterogeneous gel interface. The method is characterized in that the design of a multi-layer flexible composite structure is adopted, space-time coded light, temperature, pressure and humidity four-mode stimulation can be synchronously converted into electroneurographic signals, and a collaborative identification strategy of response intensity dynamic range and synaptic learning / delay time difference is utilized; and high-precision decoupling of the multi-source signals is realized in combination with a classification decision tree algorithm. The electronic skin has the characteristics of flexibility, deformation, environment self-repairing and bionic synapse, and geometrical characteristics of a reconfigurable target source are fused through time-space domain multi-modal data.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of bionic electronic skin, and particularly relates to a multi-layer vertical flexible composite structure based on ion gel heterojunction array. The structure has self-powered multi-modal sensing function, and relates to a preparation method and application thereof. BACKGROUND

[0002] As the core organ of life to perceive the external environment, biological skin can analyze and distinguish various physical modal signals such as light, heat, pressure and humidity in real time by virtue of its complex neural network embedded in soft tissue, which provides an ideal biological prototype and functional paradigm for the design of bionic electronic skin. Although the cross-fusion of flexible electronic technology and neuromorphic computing has significantly promoted the development of skin-like electronic devices, there are still some key defects in the existing technology: firstly, traditional devices generally rely on rigid circuit modules and external power supply systems, which not only leads to bulky overall structure and complex preparation process, but also brings the problem of high energy consumption that cannot be ignored; secondly, due to the coupling interference effect between signals, most existing sensing systems can only realize the sensing of a single physical quantity (such as pure pressure or temperature), and it is difficult to reproduce the excellent ability of biological skin to analyze and respond to multiple environmental stimuli in parallel.

[0003] Under this background, ion gel material is considered as a potential solution to solve the above challenges due to its excellent high flexibility, convenient solution processing characteristics and ion conductance response behavior under external stimulation. The sensing mechanism of the material is that the external stimulation can drive the dynamic migration of anions and cations inside the polymer chain, thereby spontaneously generating a measurable induced current, which lays a physical foundation for realizing a self-powered sensing system. However, the current ion gel-based sensing devices have fundamental limitations: their functions are mostly concentrated in single-mode or double-mode sensing of pressure or deformation, and have not yet achieved the coordinated sensing of multiple physical quantities such as light, heat and humidity like biological skin; more importantly, there is a lack of integrated design scheme that can fully utilize the ion migration mechanism to simultaneously analyze and distinguish (decouple) these multi-dimensional signals. A particularly prominent constraint factor is that most of the existing sensors with neuromorphic computing function use silicon-based rigid materials, which cannot meet the needs of flexible wearable devices; at the same time, when multiple stimuli such as light, heat and pressure act on the system simultaneously, the traditional sensing architecture cannot effectively separate the characteristic parameters of each physical quantity, resulting in serious cross-interference of the sensing data. This limitation greatly limits the application performance of such technology in key scenarios such as intelligent robot modules that require fine environmental interaction, and wearable medical devices that require high-precision multi-parameter monitoring.

[0004] Therefore, it is an urgent need to develop an integrated skin-like sensing device that integrates self-powered characteristics, efficient multi-modal signal decoupling capability and excellent flexibility, in order to break through the bottleneck of the existing technology. SUMMARY

[0005] 1. Device structure design

[0006] The present application discloses a self-powered, multi-stimulus-responsive neuromorphic sensing electronic skin, which adopts a vertically stacked multi-layer composite array structure. The structure from top to bottom is: a high-transparency pure ion gel common cathode layer (transmittance > 90%), which ensures the light stimulus response without interference and provides certain external protection ability; a nano-particle doped photo-thermal responsive ion gel functional layer, which constructs a columnar array heterojunction structure array on the surface, realizes four-mode stimulus response by using photo-thermal effect and ion migration difference; a silver paste conductive connection layer with ultra-low impedance interface, which guarantees efficient and reliable electrical coupling between the electrode and the functional device; and a flexible electrode packaged substrate layer, which provides flexible support and integrated packaging for the entire device.

[0007] 2. Material design

[0008] An ion heterogel (composed of polypyrrole nanoparticle doped ion gel) is used as the core sensing material and is synthesized by photo-polymerization reaction. The ion gel is copolymerized from 4-hydroxybutyl acrylate, 2-(2-ethoxyethoxy) ethyl acrylate monomer, crosslinking agent and photo initiator, and lithium bis-trifluoromethanesulfonimide is doped as an ion conductive medium.

[0009] 3. Self-powered mechanism and multi-modal sensing

[0010] By the difference in migration rate of lithium ions and bis-trifluoromethanesulfonimide anions in the heterogel, an ion concentration gradient is generated under external stimuli (light, temperature, pressure, humidity), forming a self-driven potential difference, which can output nerve electrical signals without external power supply.

[0011] Light, temperature, pressure stimulation: external stimulation directly causes ion migration, and the difference in the rate of anions and cations realizes signal conversion, and the addition of polypyrrole nanoparticles effectively enhances light absorption and photo-thermal conversion efficiency.

[0012] Humidity stimulation: gel swelling effect and molecular chain spacing expansion promote ion movement to realize signal conversion.

[0013] 4. Signal decoupling

[0014] Because of the significant distinction between the amplitude of humidity and pressure and the significant overlap of the response amplitude of temperature and UV stimulation, it is impossible to distinguish the two by amplitude alone. Therefore, the differences between the two stimuli in dynamic time characteristics were explored: UV light exhibits a stable learning time and a shorter memory time due to instantaneous stimulation, while temperature exhibits a fluctuating learning time and a significantly extended memory time due to the gradual heating process. Based on this, two types of classification decision tree models were designed: one combining response amplitude and learning time, and the other combining amplitude and memory time. The model uses least squares to optimize the segmentation threshold, recursively minimizes the mean square error loss function by traversing the feature space (such as amplitude, learning time) and scanning the segmentation point, until the stopping condition is met. The experiment uses 138 groups of data (including amplitude, learning time, memory time), and divides the training set and test set in a 7:3 ratio. The results show that the decoupling accuracy of both models is over 97.5%. This method breaks through the single amplitude limitation by fusing multi-dimensional dynamic features, and verifies the feasibility of using the multi-modal specificity differences of excitatory postsynaptic current to achieve high-precision signal differentiation.

[0015] 5. Self-healing and flexible electrode packaging design

[0016] Ionogels achieve self-healing function through dynamic hydrogen bonding and molecular chain reconstruction. After local damage, they can be repaired by coating with precursor solution, restoring the original sensing performance.

[0017] The device uses polyimide to package flexible electrodes, supporting bending, stretching, and twisting deformation, conforming to the human body or curved surfaces, and having high conformality.

[0018] 6. System integration and application

[0019] The visualization system architecture consists of three core modules: 1) signal sensing and conversion module; 2) data acquisition and processing module; 3) visualization interaction module. The system uses a 5x5 pixel array as the sensing terminal, which is based on a pure ionogel substrate and mechanically and electrically integrated with a custom printed circuit board.

[0020] In the signal conversion link, the array converts external stimulation signals into current signals, and then completes current-voltage signal conversion and amplification processing through a transimpedance amplifier. The transimpedance amplifier is commonly used in optoelectronic transmission communication systems to convert weak optical signals into electrical signals and amplify the signals to a certain intensity with low noise. The calculation formula of the output voltage V out and the input current I is: V out = VCC / 2 ± I*R. Where VCC is the starting voltage of the transimpedance amplifier, the signal acquisition system uses a data acquisition card, and its sampling frequency and resolution can meet the needs of multi-channel synchronous acquisition. The system power supply uses a multi-path cooperative power supply scheme: the main power supply provides working voltage for the transimpedance amplifier and the data acquisition card.

[0021] In terms of data flow processing, the parallel signals of 25 sensing pixels are transmitted to the transimpedance amplifier through the printed circuit board, amplified, and then converted into digital signals by the data acquisition card and transmitted to the computer terminal through the USB interface. The LabVIEW software is used to build a data processing platform, and a digital 45-55 Hz infinite impulse response band-stop digital filter is integrated to effectively filter the power frequency interference. In order to more intuitively display the response mode of the device array, the software's built-in graphical function is used to visualize the filtered signals. By mapping the signal intensity to color or grayscale values, the device array's patterned display is successfully achieved. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 is the physical map of the self-powered multi-modal neuromorphic sensing electronic skin of the present application and its mechanical performance.

[0023] Figure 2 is the graph of the relationship between the postsynaptic current and time under different stimuli (light, temperature, pressure, humidity) of the present application.

[0024] Figure 3 is the real-time recording of the change and excitatory postsynaptic current of each stimulus when humidity, temperature, pressure, and ultraviolet light stimuli are applied simultaneously.

[0025] Figure 4 is the decision criterion diagram for realizing four-modal signal decoupling of the present application.

[0026] Among them, a, the amplitude interval diagram under four modes. b, the average learning and forgetting time under light and temperature stimuli. c, the decision tree model constructed based on amplitude and forgetting time.

[0027] Figure 5 is the schematic diagram of the vertical multi-layer composite structure of the electronic skin of the present application.

[0028] Among them, 1, pure ionogel continuous layer, 2, nanomaterial-doped ionogel array layer, 3, silver paste connecting layer, 4, electrode array layer, 5, flexible substrate layer. DETAILED DESCRIPTION

[0029] As Figure 5As shown, the electronic skin adopts a vertical multi-layer composite structure, which includes, from top to bottom: a high-transparency pure ionic gel continuous layer 1 with a light transmittance > 90% as a common negative electrode; a nanomaterial-doped ionic gel array layer 2 as a sensing layer; a silver paste connecting layer 3 with an ultra-low impedance interface; a flexible substrate-encapsulated electrode array layer 4; and a flexible substrate layer 5. The common negative layer constitutes the basis of the common cathode to realize the transmission cycle of electrical signals, which provides the ability to protect against external damage; the sensing layer realizes multi-modal response through the ion migration mechanism; and the silver paste connecting layer ensures good electrical coupling between the electrode and the device, which enhances the light response sensitivity by inducing local surface plasmon resonance.

[0030] As shown in FIG. 1, the electronic skin is a self-powered multi-modal neuromorphic sensing electronic skin. Figure 1 As shown in FIG. 2, it is a physical diagram of the self-powered multi-modal neuromorphic sensing electronic skin and its mechanical performance. In the bending, twisting, folding, and stretching states, it shows very good performance.

[0031] The core components of each layer of the electronic skin and the preparation method of the electronic skin are as follows:

[0032] 1. Synthesis of polypyrrole nanoparticles: In a 40 mL mixed solvent composed of ethanol and water in a volume ratio of 1:4, 3 moles of polyvinylpyrrolidone and 14 moles of iron trichloride hexahydrate are added. The solution needs to be stirred at a speed of 1000 rpm for 2 hours at room temperature to ensure that the polymer surfactant and catalyst are fully dissolved. Next, 6 moles of pyrrole are dissolved in 10 mL of water and slowly added to the above mixed solvent, and the stirring is continued at a speed of 400 rpm for 4 hours at room temperature to promote the reaction. After the reaction is completed, the resulting solution is treated by washing with ethanol and centrifuging 5 times to remove the residual surfactant and catalyst, and finally a black product, polypyrrole nanoparticles, is obtained.

[0033] 2. Synthesis of ionic gel and doped polypyrrole nanoparticle ionic gel precursor ink: The ionic gel is synthesized by one-step co-photopolymerization process, using 4-hydroxybutyl acrylate and 2-(2-ethoxyethoxy)ethyl acrylate as monomers, and ethoxylated trimethylolpropane trimethacrylate as a crosslinking agent. The specific steps include: mixing the monomers 4-hydroxybutyl acrylate, 2-(2-ethoxyethoxy)ethyl acrylate and the crosslinking agent in a volume ratio of 1:4:0.05, then adding 1% of the mass of the mixed solution of the photoinitiator and 0.5 mol of lithium bis(trifluoromethanesulfonyl)imide. Stir these ingredients at a speed of 500 rpm for 30 minutes to form a uniform precursor ink.

[0034] 3. Electronic skin preparation method: First, embed the cylinder into a polytetrafluoroethylene mold filled with polydimethylsiloxane. Then heat-cure the assembly at 60 °C for 6 h to obtain a semi-cylindrical polydimethylsiloxane mold. Inject the ionic gel precursor ink into this customized semi-cylindrical polydimethylsiloxane mold, and then convert it into a semi-cylindrical ionic gel by exposing it to ultraviolet light (365 nm, 10 W) for 2 min. By radially stretching the ionic gel, we can transfer it to a polyethylene terephthalate film and adhere it there by van der Waals interactions. Next, place a mask plate with a precise 1 x 1 mm hole array (250-unit device) on the ionic gel, form the sensing element column by injecting the dopant ink containing polypyrrole nanoparticles into each hole, and then perform a 5-minute ultraviolet exposure. After removing the mask, the array of sensing structures is exposed on the top of the ionic gel.

[0035] By regulating the pulse intensity and frequency, the dynamic light response mechanism was studied in depth, revealing the pulse intensity-dependent plasticity, highlighting its potential application in neuromorphic optical synapse function Figure 2 a). This mechanism is based on the principle of thermally activated ion migration, which endows the receptor with thermal sensitivity to achieve temperature perception. Under the stimulation of a heat source, the receptor exhibits typical stimulus intensity-dependent plasticity: when the temperature rises from 25 °C to 60 °C, the induced excitatory postsynaptic current increases linearly from 63.6 nA to 1024.7 nA Figure 2 b).

[0036] The tactile sensing performance of the artificial skin was evaluated through systematic pressure response tests. In the low pressure range (<50 kPa), the receptor based on the ion redistribution mechanism induced by mechanical deformation exhibited a high sensitivity of 0.07 nA / kPa Figure 2 c).

[0037] In addition to touch, heat, and light perception, the humidity synapse perception function of the artificial skin is crucial for dynamic wet environment adaptive response. The gel network infused with ion conductors becomes an ideal carrier for constructing humidity synapse interfaces due to its adjustable ionic conductivity and inherent humidity sensitivity. When the relative humidity rises from 50% to 80%, the current generated by the heterogeneous gel receptor jumps from 0.03 nA to 0.75 nA, with a sensitivity of 0.024 nA / %RH Figure 2 d).

[0038] To evaluate the multi-modal fusion decoupling ability, first, the raw signals of the simultaneously deployed optical, thermal, humidity, and pressure sensors were monitored in real-time. As Figure 3 shown, when the electronic skin was exposed to 0-8.64 μW / mm 2The receptor exhibits a highly coupled response under the conditions of UV light, 22-55 °C heating, 32-95% humidity, and 0-10.0 N pressure. These heterogeneous data streams are fused to form a unified perceptual representation that can capture the complex environmental states and their interactions. The core decoupling strategy separates the independent contributions of each sensory modality by extracting latent features such as response amplitude, learning time, and memory duration. Typical examples show that when light causes a sudden increase in current and heat drives a significant increase in amplitude, under the condition of multiple stimuli superimposed, the signal changes caused by humidity and pressure cannot reverse the overall downward trend, which is mainly dominated by the dominant temperature response.

[0039] Figure 4 a highlights the response amplitude differences of the four stimulus sources: the amplitude change induced by humidity is limited to a narrow interval of 0.37-0.76 nA, while the amplitude value triggered by pressure spans a wider range of 2.47-5.47 nA, indicating the distinguishability based on amplitude. In contrast, there is an overlapping region in amplitude between heat stimulation (21.15-279.72 nA) and light stimulation (20.06-72.62 nA), which cannot be reliably distinguished by amplitude alone. However, Figure 4 b reveals significant differences in the time dynamic characteristics of the two, including learning time and memory duration. Due to the transient nature of the light response, its learning time stabilizes at 1.1 seconds; while the gradual heating process extends the learning time to about 12.0 seconds. More importantly, the significantly higher average amplitude of the heat response (279.72 nA vs 72.62 nA) is strongly correlated with its longer memory duration (123.4 seconds vs 64.4 seconds for light), providing clear discriminators.

[0040] Based on the above characteristics, Figure 4 c shows a classification decision tree model integrating response amplitude and learning time for receptor signal decoupling analysis. The model uses the least squares method to accurately determine the optimal split threshold, and is verified by 138 experimental data sets containing current amplitude / learning time / memory time (divided into 70% training set and 30% validation set), finally achieving a decoupling accuracy of >97.5%.

[0041] The above only describes the preferred embodiments of the present application. It should be noted that for those skilled in the art, without departing from the principles of the present application, several improvements and refinements can be made, and these improvements and refinements should also be considered within the scope of the present application.

Claims

1. A self-powered multimodal sensing electronic skin, characterized in that: The electronic skin employs a vertical multilayer composite structure, comprising, from top to bottom: a pure ion gel continuous layer, a nanomaterial-doped ion gel array layer, a silver paste connecting layer, an electrode array layer, and a flexible substrate layer. The pure ion gel continuous layer has a light transmittance greater than 90%, serving as a common negative layer to achieve electrical signal transmission cycling while providing protection against external damage. The nanomaterial-doped ion gel array layer serves as a sensing layer, achieving multimodal response through an ion migration mechanism. The silver paste connecting layer has an ultra-low impedance interface, ensuring good electrical coupling between the electrodes and the device, and enhancing photoresponse sensitivity by inducing local surface plasmon resonance.

2. The electronic skin according to claim 1, characterized in that: The ionogel in the pure ionogel continuous layer is formed by copolymerization of 4-hydroxybutyl acrylate, 2-(2-ethoxyethoxy)ethyl acrylate, crosslinking agent ethoxylated trimethylolpropane trimethacrylate, lithium bis(trifluoromethanesulfonyl)imide and photoinitiator 1-hydroxycyclohexylphenyl ketone.

3. The electronic skin according to claim 1, characterized in that: The nanomaterial-doped ion gel array layer covers the ultraviolet-visible-near-infrared spectral response range, and its nanomaterials include one or more of polypyrrole, polydopamine, nano-iron oxide, transition metal carbides, nano-copper sulfide, carbon black, and bismuth oxyiodide; the nanomaterials can be replaced by materials with plasmon resonance effect, including one or more of silver nanoparticles, silver nanowires, and gold nanoparticles.

4. The electronic skin according to claim 1, characterized in that: The electrode array layer is a metal electrode structure encapsulated with polyimide, and its encapsulation material includes one or more of polydimethylsiloxane, polyvinyl alcohol, polyurethane, silicone, and platinum-catalyzed silicone.

5. The electronic skin according to claim 1, characterized in that: The electronic skin generates anion and cation concentration gradients under the stimulation of light, temperature, pressure, and humidity. Through the difference in ion migration rates, a self-driven potential difference is formed, which can output neural electrical signals without the need for an external power source.

6. The electronic skin according to claim 3, characterized in that: The electronic skin directly induces ion migration under the stimulation of light, temperature, and pressure. The difference in cation and anion rates enables signal conversion. The addition of polypyrrole nanoparticles can enhance light absorption and photothermal conversion efficiency. Under humidity stimulation, the gel swelling effect and the expansion of molecular chain spacing promote ion movement, thereby achieving signal conversion.

7. A signal decoupling method for a self-powered multimodal sensing electronic skin according to claim 1, wherein a classification decision tree model is constructed based on the dynamic range of response intensity and the difference in synaptic learning / delay time to achieve multimodal signal decoupling, characterized in that, (1) Extract the response current amplitude, learning time and memory time as feature vectors; (2) The least squares method is used to optimize the decision tree segmentation threshold; (3) The decoupling accuracy of the four-mode signals of humidity, pressure, light and temperature is greater than or equal to 97.5%.

8. An integrated signal processing system based on the signal decoupling method for a self-powered multimodal sensing electronic skin as described in claim 6, wherein the integrated signal processing system interfaces with the electronic skin, the system comprising three core modules: sensing, data acquisition, and visualization, supporting real-time recording and presentation of electrical signal changes generated under various stimulation modes, characterized in that... include: (1) Custom printed circuit boards are used to transmit the original electrical signals to the transimpedance amplifier; (2) The serial communication data acquisition card performs 24-bit analog-to-digital conversion while acquiring data, and uses a software development environment to realize real-time data acquisition and processing. (3) Infinite impulse response band-stop digital filter to suppress power frequency interference.

9. The application of a self-powered multimodal sensing electronic skin according to claim 1, characterized in that, include: Simultaneous monitoring of ambient temperature, humidity, pressure, and light intensity in wearable devices.

Images