Fingerprint-resistant scratch-resistant touch film photoelectric signal conversion system

Abstract

The application discloses a kind of anti-fingerprint scratch-resistant touch film photoelectric signal conversion systems, the present application relates to touch display and photoelectric signal conversion technical field, including: transparent conductive functional layer;Anti-fingerprint scratch-resistant hard coating is arranged on the transparent conductive functional layer upper;And photoelectric conversion unit for converting optical signal into electrical signal;It further includes: stress buffering and optical compensation structure, is arranged between anti-fingerprint scratch-resistant hard coating and the transparent conductive functional layer.The anti-fingerprint scratch-resistant touch film photoelectric signal conversion system, by configuring only responding to thermal-mechanical stress stress simulation unit and double slope reference dark current offset loop, realizes the independent extraction and reference comparison to parasitic dark current component, avoids the damage of conventional filtering method to effective touch signal edge steepness.

Description

Technical Field

[0001] This invention relates to the field of touch display and photoelectric signal conversion technology, specifically to a fingerprint-resistant and scratch-resistant touch film photoelectric signal conversion system. Background Technology

[0002] In the technological evolution of touchscreen photoelectric signal conversion systems, to improve the visual cleanliness and operational durability of the terminal, the industry generally covers the transparent conductive sensing layer with a surface-hardened treatment layer that has anti-fingerprint and scratch-resistant functions. This treatment layer is usually composed of highly cross-linked fluorinated organic polymers or silicone-modified hard resins, forming a dense oleophobic and hydrophobic interface through UV curing or thermosetting processes to prevent the adhesion of sebum and sweat and resist scratches from hard objects. However, while the introduction of this physical protective structure improves the mechanical and chemical resistance of the touchscreen surface, it brings systemic risks to the accuracy of the underlying photoelectric signal conversion. Specifically, there is an elastic modulus mismatch between the anti-fingerprint and scratch-resistant hard coating and the flexible transparent conductive functional layer. When the touchscreen is bent during assembly or subjected to repeated pressure, the surface hard coating, due to its high rigidity, is difficult to deform synchronously with the underlying flexible substrate. Under long-term mechanical stress, residual internal stress at the microscale gradually accumulates inside the hard coating and is released along the interface in the form of submicron-level crack propagation, forming a non-uniform strain field distribution between the layers. This strain field directly causes a local gradient change in the refractive index of the optical path region of the photoelectric conversion unit, resulting in spatial phase distortion of the reference optical signal. The distorted optical signal is converted into a parasitic electrical signal output by the photodetector, which is unrelated to the touch press event, manifested as continuous drift and fluctuating amplitude of background dark current noise. For the touch sensing stage, patent publication CN117234363A describes a self-capacitance sensing system, which uses a ramp-controlled current source to discharge the self-capacitor and generate a ramp transition curve to reduce the impact of clock jitter on the sensing voltage, thereby improving the signal-to-noise ratio at the capacitance detection level. However, this solution focuses on the accuracy of signal processing in the circuit domain, and the sensing front end still relies on a conventional transparent conductive stacked structure, failing to address the problems of optical path refraction distortion and photodetector dark current drift caused by stress release within the surface-hardened protective layer. During prolonged continuous touch operation, the rise of the noise substrate erodes the dynamic range of the effective sensing signal, leading to a decrease in touch coordinate calculation accuracy and jitter breaks in the continuity of the scribing line, ultimately creating a dilemma where anti-fingerprint and scratch-resistant performance and photoelectric conversion fidelity cannot be simultaneously achieved.

[0003] Therefore, the central technical problem to be solved by a fingerprint-resistant and scratch-resistant touch film photoelectric signal conversion system is to suppress the uncontrolled drift of dark current in the photoelectric conversion unit caused by the release of internal stress in the hard coating while maintaining the surface scratch-resistant physical protection performance and oleophobic and fingerprint-resistant properties. Summary of the Invention

[0004] The purpose of this invention is to provide a photoelectric signal conversion system for fingerprint-resistant and scratch-resistant touch film to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film, comprising:

[0006] Transparent conductive functional layer;

[0007] A fingerprint-resistant and scratch-resistant hard coating disposed above the transparent conductive functional layer; and a photoelectric conversion unit for converting optical signals into electrical signals;

[0008] Also includes:

[0009] A stress buffer and optical compensation structure is disposed between the anti-fingerprint and scratch-resistant hard coating and the transparent conductive functional layer. The stress buffer and optical compensation structure includes a composite layer composed of an elastic matrix and refractive index adjusting particles dispersed therein. The elastic modulus inside the composite layer has a non-uniform gradient distribution in the direction perpendicular to the layer plane, and the spatial concentration distribution of the refractive index adjusting particles has a preset non-linear correspondence with the gradient distribution of the elastic modulus.

[0010] The stress buffer and optical compensation structure is configured to: respond to the interfacial stress generated by the deformation of the anti-fingerprint and scratch-resistant hard coating, guide the stress to dissipate to the low modulus region through the internal modulus gradient, and simultaneously drive the refractive index adjustment particles to migrate in a directional manner to compensate for the local refractive index change caused by residual stress, thereby suppressing the dark current drift of the photoelectric conversion unit caused by optical path distortion.

[0011] Preferably, the elastic modulus gradient distribution pattern inside the stress buffer and optical compensation structure corresponds to the biomimetic fractal stress flow channel pattern obtained by iterative solution through a cross-scale simulation model. This pattern breaks down the input external concentrated stress into a diffusely scattered stress distribution with spatial incoherence.

[0012] Preferably, the system also includes a process morphology verification subsystem, which includes a laser interferometer and a local infrared pulse array. This subsystem is used to compare the deviation between the actual leveling morphology and the biomimetic fractal stress flow channel pattern in real time during the wet film curing process of the stress buffer and optical compensation structure, and to correct the local prepolymerization degree of the material by adjusting the local infrared pulse parameters, so as to force the execution of at least two microchannel morphology reconstruction cycles.

[0013] Preferably, the photoelectric conversion unit is coupled to a dual-slope reference dark current cancellation circuit, the dual-slope reference dark current cancellation circuit comprising:

[0014] The first ramp generator is used to drive the main sensing channel to discharge.

[0015] A second ramp generator is used to drive the dark current reference channel to discharge; and

[0016] The stress simulation unit, located at the input of the dark current reference channel and without touch sensing function, is used to generate a parasitic dark current signal that only reflects the effects of thermo-mechanical stress.

[0017] Preferably, the dual-slope reference dark current cancellation loop further includes an adaptive filter configured to perform at least two cyclic input-output cross-validation operations, including: in the first round, synchronously acquiring the noise floor waveforms of the main sensing channel and the dark current reference channel; and in the second round, using a recursive least squares algorithm to convolve the time-domain waveform of the dark current reference channel with the sensing signal of the main sensing channel to eliminate noise.

[0018] Preferably, it also includes a lightweight digital twin model set at the edge computing end, which is constructed based on Gaussian process regression and has embedded a nonlinear mapping relationship of "surface morphology-refractive index distribution-dark current spectral density";

[0019] The digital twin lightweight model is used to receive the real-time noise waveform acquired by the dark current reference channel as the initial disturbance variable, and to perform no less than three forward prediction and backward correction iterations to generate nonlinear compensation current control commands for canceling the higher-order harmonic components of the dark current.

[0020] Preferably, it further includes a high-frequency pulse width modulation current injector, which is coupled to the virtual ground terminal of the preamplifier of the photoelectric conversion unit, and is used to inject a nonlinear compensation current that is out of phase with the predicted dark current waveform and whose amplitude is synthesized by probability weighting into the virtual ground terminal according to the control command output by the digital twin lightweight model.

[0021] Preferably, it also includes a surface energy state closed-loop processing device, which includes a spectral ellipticity monitoring unit and an atmospheric pressure plasma jet unit, used to perform at least three surface molecular reconstructions and uniformity verifications after the anti-fingerprint and scratch-resistant hard coating has been cured, until the surface energy dispersion is lower than a preset anti-fingerprint function maintenance threshold.

[0022] Preferably, the refractive index adjusting particles are nano-zirconia particles, and their spatial concentration distribution in the stress buffer and optical compensation structure is determined by the compensation path function output by the joint simulation of the ray tracing algorithm and the finite element mechanical model.

[0023] Preferably, the anti-fingerprint and scratch-resistant hard coating is a fluorinated polyether modified acrylate UV-cured layer or an organosilicon modified thermosetting hard coating.

[0024] This invention provides a photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film. It has the following beneficial effects:

[0025] By incorporating a stress buffer and optical compensation structure with a non-uniform gradient distribution of internal elastic modulus and dispersed refractive index-modulating particles between an anti-fingerprint and scratch-resistant hard coating and a transparent conductive functional layer, active intervention is achieved in the stress release process caused by forced deformation of the hard coating. This structure utilizes the modulus gradient to guide the interface stress dissipation along a pre-defined low-modulus path, reducing stress concentration at the interface of the transparent conductive functional layer and thus suppressing optical path refractive index gradient distortion caused by the formation of a regular strain field. Simultaneously, residual stress drives the refractive index-modulating particles to migrate directionally. The resulting local refractive index change spatially matches and cancels out the optical path shortening caused by stress densification, ensuring that the wavefront phase of the beam reaching the photoelectric conversion unit through this composite structure remains essentially stable. This in-situ optical self-compensation mechanism at the physical level breaks the causal chain between the mechanical deformation of the anti-fingerprint and scratch-resistant hard coating and the dark current drift of the photodetector, enabling the system to maintain a stable background noise substrate for photoelectric conversion while preserving the high hardness and oleophobic / hydrophobic properties of the surface layer.

[0026] At the level of electrical signal processing, this invention achieves independent extraction and reference comparison of parasitic dark current components by configuring a stress simulation unit that only responds to thermo-mechanical stress and a double-slope reference dark current cancellation circuit, thus avoiding the damage to the edge steepness of the effective touch signal caused by conventional filtering methods.

[0027] By introducing a lightweight digital twin model to iteratively predict the higher-order harmonic components of dark current noise, and based on this, generating a nonlinear compensation current with inverted phase and probability-weighted amplitude injected into the virtual ground terminal of the preamplifier, the random noise components that may still remain after physical compensation are effectively eliminated. Combined with closed-loop verification of the biomimetic fractal flow channel pattern of the stress buffer structure and multiple reconstructions of the surface energy state of the anti-fingerprint coating during the manufacturing process, this invention forms a three-dimensional defense system against dark current drift at multiple levels, including material preparation, structural topology, electrical processing, and surface engineering. This ensures the accuracy of touch coordinate calculation and the continuity of the drawing trajectory during long-term continuous touch operation. Attached Figure Description

[0028] Figure 1 This is a data flow diagram of a photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film according to the present invention.

[0029] Figure 2 This is a flowchart of the process morphology verification subsystem of the photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film according to the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Please see Figure 1 and Figure 2 This invention provides a technical solution: a photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film, comprising:

[0032] Transparent conductive functional layer;

[0033] A fingerprint-resistant and scratch-resistant hard coating disposed above the transparent conductive functional layer; and a photoelectric conversion unit for converting optical signals into electrical signals;

[0034] Also includes:

[0035] The stress buffer and optical compensation structure is disposed between the anti-fingerprint and scratch-resistant hard coating and the transparent conductive functional layer. The stress buffer and optical compensation structure includes a composite layer composed of an elastic matrix and refractive index adjusting particles dispersed therein. The elastic modulus inside the composite layer has a non-uniform gradient distribution in the direction perpendicular to the layer plane, and the spatial concentration distribution of the refractive index adjusting particles has a preset non-linear correspondence with the gradient distribution of the elastic modulus.

[0036] The stress buffer and optical compensation structure is configured to: respond to the interfacial stress caused by the deformation of the anti-fingerprint and scratch-resistant hard coating, guide the stress to dissipate to the low modulus region through the internal modulus gradient, and at the same time drive the refractive index adjustment particles to generate directional migration to compensate for the local refractive index change caused by residual stress, thereby suppressing the dark current drift of the photoelectric conversion unit caused by optical path distortion.

[0037] It should be further explained that in a fingerprint-resistant and scratch-resistant touch film photoelectric signal conversion system, a stress buffer and optical compensation structure is provided between the fingerprint-resistant and scratch-resistant hard coating and the transparent conductive functional layer. This stress buffer and optical compensation structure uses an organosilicon elastomer as the matrix material, and refractive index adjusting particles are uniformly dispersed within the matrix. The refractive index adjusting particles are nano-zirconia particles.

[0038] In the direction perpendicular to the layer plane, the elastic modulus of the composite layer exhibits a gradient distribution pattern that continuously decreases from the first interface adjacent to the anti-fingerprint and scratch-resistant hard coating to the second interface adjacent to the transparent conductive functional layer. Specifically, the crosslinking density and Young's modulus of the first interface region are higher than those of the second interface region. Matching the elastic modulus gradient distribution, the volume concentration distribution of nano-zirconia particles within the composite layer also exhibits a nonlinear gradient along the direction. Specifically, the particle concentration is relatively sparse in the high-modulus region near the hard coating and relatively dense in the low-modulus region near the conductive functional layer. The elastic modulus gradient of the stress buffer and optical compensation structure continuously decreases from the upper interface to the lower interface. The elastic modulus at the upper interface is controlled between 100 MPa and 300 MPa, and the elastic modulus at the lower interface is controlled between 10 MPa and 50 MPa. The gradient change rate remains stable without abrupt changes. The refractive index adjustment particles are nano-zirconia particles with a diameter of 20 nm to 50 nm, and the particle doping volume fraction is controlled between 5% and 15%. The specific doping amount is adjusted synchronously according to the elastic modulus gradient distribution to ensure a predetermined nonlinear correspondence between the particle concentration distribution and the modulus gradient. This concentration distribution curve is obtained by simultaneously solving the finite element structural mechanics equations and the ray tracing equations. During the solution process, the optical path difference compensation under interface stress input is used as the optimization objective function for iterative convergence determination.

[0039] In the preparation stage, after coating the transparent conductive functional layer with a liquid precursor mixture, the gradient distribution of the above modulus and particle concentration is formed by a segmented ultraviolet irradiation process. Specifically, a low irradiation intensity is first used for deep pre-curing to induce particles to migrate and accumulate to the bottom under the action of viscosity gradient, and then a higher irradiation intensity is used to complete the surface hardening cross-linking.

[0040] During actual system operation, when external touch pressure or bending deformation is applied to the anti-fingerprint and scratch-resistant hard coating, the microcrack propagation stress and interlayer shear stress generated inside the coating are transmitted downwards to the stress buffer and optical compensation structure. Since the elastic modulus of this structure decreases from top to bottom, the stress energy is preferentially guided to the bottom low-modulus region for viscoelastic deformation dissipation, rather than being concentrated at the interface with the transparent conductive functional layer, thus avoiding the generation of a regular strain field at the interface due to stress concentration.

[0041] Meanwhile, the residual stress guided to the low-modulus region causes local compression and densification of the organosilicon molecular chain segments in that region. This densification process drives the dispersed nano-zirconia particles to migrate and rearrange directionally along the stress gradient direction. Since the initial particle concentration distribution has been pre-set to a compensation state corresponding to the modulus gradient, the local equivalent refractive index increase caused by particle migration precisely matches and cancels out the optical path shortening caused by stress compression. This ensures that the wavefront phase of the light beam passing through the composite layer into the transparent conductive functional layer and then to the photoelectric conversion unit remains stable, suppressing stray light interference and scattering effects caused by non-uniform refractive index changes.

[0042] In the photosensitive area at the front end of the photoelectric conversion unit, since the spatial distribution of the received light intensity and the phase information are not distorted or disturbed, the background dark current component in the output photogenerated carrier recombination current does not drift or rise with the accumulation of touch presses. Thus, under the premise of maintaining the original pencil hardness and hydrophobic and oleophobic contact angle parameters of the anti-fingerprint and scratch-resistant hard coating unchanged, effective control of dark current drift is achieved.

[0043] The elastic modulus gradient distribution pattern inside the stress buffer and optical compensation structure corresponds to the biomimetic fractal stress flow channel pattern obtained by iterative solution through a cross-scale simulation model. This pattern breaks down the input external concentrated stress into a diffusely scattered stress distribution with spatial incoherence.

[0044] It should be further explained that the elastic modulus gradient distribution within the stress buffer and optical compensation structure is not a simple linear decrease or a concentric ring distribution, but rather presents a biomimetic fractal stress flow channel pattern. The specific construction process of this pattern is as follows:

[0045] First, based on the actual geometric dimensions, material constitutive parameters, and expected compressive load spectrum of the anti-fingerprint and scratch-resistant hard coating and the transparent conductive functional layer, a cross-scale simulation model was established, incorporating the coupling effect of microscopic molecular chain segment slippage and macroscopic continuous medium deformation. This model describes the conformational entropy changes of chain segments between crosslinking points of the organosilicon elastomer at the molecular dynamics level, describes the overall displacement and stress field distribution of the composite layer at the finite element level, and achieves inter-scale data interaction through a bridging function with local strain energy density as the transfer variable.

[0046] The cross-scale simulation model combines the microscale of molecular dynamics simulation with the macroscale of finite element mechanical analysis. Local strain energy density is used as the transfer variable between scales to establish the correlation between the conformational changes of microscopic molecular chain segments and the deformation of macroscopic structures. The simulation boundary conditions are set according to the actual assembly and stress state of the touch film. The edge of the film layer is set as a fixed constraint boundary, and a uniformly distributed load constraint corresponding to the finger touch pressure is applied to the upper surface. The parameter settings adopt the material constitutive parameters of silicone elastomer and nano-zirconia particles commonly used in touch films. The solution adopts an iterative convergence method, and the solution completion criterion is that the stress field distribution deviation between two adjacent iterations is less than a preset threshold. The simulation process is completed by first performing microscopic molecular dynamics simulation, then macroscopic finite element calculation, and finally cross-scale data interaction calibration.

[0047] Secondly, with the optimization objectives of reducing the maximum principal stress at the interlayer interface and improving the uniformity of stress distribution under concentrated load conditions, and using the local elastic modulus of each discrete element within the composite layer as the design variable, an iterative solution process is initiated. In each iteration step, the model calculates the direction and branching state of the stress streamlines under the current modulus distribution, and uses the topology of the stress streamlines as feedback input to the next modulus adjustment algorithm, until the stress streamline network converges into a stable configuration with self-similar branching characteristics. This stable configuration is the biomimetic fractal stress flow channel pattern.

[0048] The fractal dimension of the biomimetic fractal stress flow channel pattern is set between 1.2 and 1.8. The iterative convergence criterion is that the stress flow line topology does not change significantly after three consecutive iterations and the maximum principal stress between layers decreases to less than 15% of the initial value. When generating the pattern, based on the size and stress distribution characteristics of the touch film, the main framework of the principal stress flow channel is first constructed, and then secondary and terminal stress flow channels are generated step by step. By adjusting the bifurcation angle and width ratio of the flow channel, the stress flow channel forms a self-similar fractal structure. The pattern generation can be completed by controlling the fractal dimension and iterative calibration of the stress field.

[0049] The biomimetic fractal stress flow channel pattern consists of a primary stress flow channel, a secondary stress flow channel, and a finer terminal flow channel. The width of each flow channel gradually decreases, and the bifurcation angle of each flow channel follows the preset geometric fractal dimension constraint.

[0050] Mapping this fractal stress flow channel pattern onto the fabrication process of the stress buffer and optical compensation structure is achieved by introducing a grayscale photolithography mask during the ultraviolet irradiation curing stage. The spatial distribution of the mask's transmittance corresponds to the fractal flow channel pattern. The flow channel region corresponds to higher transmittance, resulting in higher crosslinking density and higher elastic modulus, while the non-flow channel region corresponds to lower transmittance, resulting in lower crosslinking density and lower elastic modulus. This solidifies the gradient distribution of elastic modulus into a complex spatial distribution consistent with the fractal stress flow channel pattern.

[0051] When concentrated external pressing or bending forces are applied to the structure, the stress first enters the main stress channel, propagates along its extension direction within the main stress channel, and is distributed to the secondary channels at the bifurcation nodes. The secondary channels further divert the stress to the terminal channels. Due to the non-periodic characteristics of the orientation and branching angles of each channel, the propagation path length of the stress wave varies in different channels. This causes the originally in-phase concentrated stress components to develop path differences during multiple bifurcation and merging processes, resulting in randomly discrete stress components at each output port. These randomly phased stress components superimpose at the interface of the transparent conductive functional layer to form a spatially incoherent diffuse scattering stress distribution. Its spatial autocorrelation length is much smaller than the size of the photosensitive region of the photoelectric conversion unit, making it impossible to construct a uniform refractive index gradient field within the photosensitive region. This eliminates the conditions for the formation of a regular strain field that causes phase distortion of the optical path from the source of mechanical excitation.

[0052] By selecting a coupled platform of molecular dynamics simulation software and finite element analysis software, setting appropriate fractal dimension constraints and iterative convergence criteria, and generating corresponding lithographic patterns through photolithography mask design tools, the biomimetic fractal stress flow channel pattern is solidified and transferred in a standard UV curing device, thereby reproducing the stress cracking and incoherent diffuse scattering functions.

[0053] It also includes a process morphology verification subsystem, which includes a laser interferometer and a local infrared pulse array. This subsystem is used to compare the deviation between the actual leveling morphology and the biomimetic fractal stress flow channel pattern in real time during the wet film curing process of the stress buffer and optical compensation structure. It also corrects the local prepolymerization degree of the material by adjusting the local infrared pulse parameters, thereby forcing at least two microchannel morphology reconstruction cycles.

[0054] It should be further explained that the process topography verification subsystem consists of a laser interferometer, a local infrared pulse array, and a feedback control unit that is linked to it for data transmission.

[0055] A laser interferometer is installed after the wet film coating station of the stress buffer and optical compensation structure and before the UV curing irradiation area. Its probe beam covers the wet film surface in an oblique incidence manner. By analyzing the phase difference of the interference fringes between the reference beam and the measurement beam, the height values ​​of each discrete point on the wet film surface relative to the reference plane are calculated, and a three-dimensional point cloud map characterizing the current actual leveling morphology is generated. This three-dimensional point cloud map is transmitted to the feedback control unit in real time. The feedback control unit pre-stores a target morphology dataset corresponding to the biomimetic fractal stress channel pattern. This target morphology dataset describes the spatial concave or convex contours that each level of stress channel should present after curing.

[0056] The feedback control unit performs point-by-point registration and comparison between the actual topography point cloud map and the target topography dataset, calculates the topography deviation vector at each spatial location, and marks the region to be corrected when the magnitude of the deviation vector in a certain region exceeds the preset allowable range. For each region to be corrected, the feedback control unit calculates the required infrared pulse energy and pulse duration based on the direction and amplitude of the deviation vector, combined with the reaction kinetic parameters of the photosensitive monomer of the material, and issues control commands to the local infrared pulse array.

[0057] The localized infrared pulse array consists of multiple independently addressed infrared emitting units arranged in a two-dimensional matrix. The emission spot of each infrared emitting unit covers a discrete region on the wet film surface that matches the minimum characteristic size of the stress flow channel. The selected infrared emitting unit radiates infrared energy of a specific wavelength to the corresponding region to be corrected in a pulsed manner. This wavelength is within the non-photochemical absorption band of the organosilicon precursor solvent and the photosensitive monomer, increasing the local temperature of the material only through thermal effects without triggering additional photocrosslinking reactions. The increase in local temperature causes a momentary decrease in the material viscosity, driving the wet film material to fill in depressions from the periphery or spread outwards from convex areas under the action of the surface tension gradient, thereby changing the local leveling morphology of the region.

[0058] After one infrared pulse irradiation, the laser interferometer scans the same area again and recalculates the deviation vector. If the magnitude of the deviation vector still exceeds the allowable range, the feedback control unit adjusts the infrared pulse parameters according to the new deviation value and triggers irradiation again. This closed-loop process of "morphology measurement to deviation calculation to infrared correction to re-measurement" forces at least two complete cycles within each defined area until the magnitude of the deviation vector between the actual morphology and the target morphology converges to within the allowable range before proceeding to the correction process for the next area.

[0059] The grayscale photolithography mask for the biomimetic fractal stress flow channel is designed according to the fractal stress flow channel pattern. The flow channel region corresponds to the high transmittance region, and the non-flow channel region corresponds to the low transmittance region. The transmittance gradient is matched with the elastic modulus gradient. The infrared pulse parameters are set so that the pulse wavelength is between 3μm and 5μm and the pulse energy is controlled at 50mJ / cm². 2 Up to 200mJ / cm 2 The pulse duration is 10ms to 50ms. During wet film curing, low-intensity ultraviolet light is used for deep pre-curing, with an irradiation intensity of 50mW / cm². 2 Up to 100mW / cm 2 The surface layer is then cured using high-intensity ultraviolet light, with an irradiation intensity of 200 mW / cm². 2 Up to 500mW / cm 2 The total irradiation energy was controlled at 1500 mJ / cm². 2 Up to 2500mJ / cm2 According to these parameters, the stress buffer and optical compensation structure can be fabricated.

[0060] Once the morphological deviations of all discrete regions on the wet film surface are converged, the UV curing light source is activated and completes the overall cross-linking and shaping of the composite layer, thereby transferring the biomimetic fractal stress flow channel pattern from the target dataset to the cured stress buffer and optical compensation structure.

[0061] Based on the above description, a commercial laser interferometric profile measuring instrument, an addressable infrared light-emitting diode array, and a programmable logic controller can be used to build this subsystem. Appropriate deviation tolerance thresholds and pulse parameter search step sizes can be set, and closed-loop control and multiple reconstruction cycles of morphology can be realized in the coating and curing production line.

[0062] The photoelectric conversion unit is coupled to a dual-slope reference dark current cancellation circuit, which includes:

[0063] The first ramp generator is used to drive the main sensing channel to discharge.

[0064] A second ramp generator is used to drive the dark current reference channel to discharge; and

[0065] The stress simulation unit, located at the input of the dark current reference channel and without touch sensing function, is used to generate parasitic dark current signals that only reflect the effects of thermo-mechanical stress.

[0066] It should be further explained that the photoelectric conversion unit is coupled to a dual-slope reference dark current cancellation circuit, which includes a first slope generator, a second slope generator, and a stress simulation unit.

[0067] The first ramp generator consists of a first ramp bias circuit and a first ramp control current source. Its output is connected to the input of the main sensing channel via a first switching device. The input of the main sensing channel is further coupled to the sensing electrode on the transparent conductive functional layer corresponding to the touch sensing area, providing a discharge current to the sensing electrode during the discharge phase to transition from an initial high level to a final low level ramp. The second ramp generator consists of a second ramp bias circuit and a second ramp control current source. Its circuit topology is symmetrical with that of the first ramp generator. The resistor-capacitor network parameters in the second ramp bias circuit are the same as those in the first ramp bias circuit, and the transistor size of the second ramp control current source is the same as that of the first ramp control current source. The output of the second ramp generator is connected to the input of the dark current reference channel via a second switching device, and the input of the dark current reference channel is coupled to the output electrode of the stress simulation unit.

[0068] The stress simulation unit is located in the non-touch sensing area of ​​the touch film. Its film structure, from bottom to top, consists of a transparent substrate, a transparent conductive functional layer extension, a stress buffer and optical compensation structure extension, and an anti-fingerprint and scratch-resistant hard coating extension. The material composition and thickness of each extension are the same as the corresponding film layer in the main touch sensing area. The upper surface of the stress simulation unit does not have any electrode patterns for sensing touch events, and its transparent conductive functional layer extension is patterned as a continuous conductive plane or a floating conductor without capacitive coupling to the surrounding circuitry. Therefore, when a user's finger approaches or touches the corresponding position above the stress simulation unit, no capacitance change signal related to the touch event is generated at the unit's output electrode. The electrical signal generated at the output electrode of the stress simulation unit is caused by only the following physical process: the extension of the anti-fingerprint and scratch-resistant hard coating undergoes thermal expansion or mechanical deformation under the action of ambient temperature change or film bending stress. This deformation is transmitted to the extension of the transparent conductive functional layer through the stress buffer and optical compensation structure extension, causing local strain in the extension of the transparent conductive functional layer. This local strain causes a slight shift in the band structure of the extension of the transparent conductive functional layer, thereby changing the width of the space charge region at its interface with the underlying substrate. This manifests as a slow-varying component in the parasitic capacitance between the output electrode and the reference ground, which is related to the stress change.

[0069] During system operation, the phase-controlled switch controller generates a first timing signal and a second timing signal. The first timing signal controls the first switching device to turn on, initiating the ramp discharge process of the main sensing channel. The second timing signal is in phase and frequency with the first timing signal and controls the second switching device to turn on synchronously, initiating the ramp discharge process of the dark current reference channel. Since the circuit parameters of the first ramp generator and the second ramp generator are the same and the discharge initiation voltage is the same, the discharge current curves generated by the two under ideal stress-free conditions completely overlap. However, when the anti-fingerprint and scratch-resistant hard coating generates internal stress due to forced deformation, the input terminal of the main sensing channel receives not only the touch capacitance change signal but also the dark current component generated by stress-induced optical path distortion, while the input terminal of the dark current reference channel only receives the stress-related parasitic capacitance change component output by the stress simulation unit and does not include the touch capacitance change component.

[0070] The ramp generator of the dual-ramp reference dark current cancellation loop adopts a symmetrical circuit topology. The ramp generators of the main sensing channel and the dark current reference channel use resistors, capacitors and transistors of the same specifications to ensure that the circuit parameters are completely matched. The timing control logic adopts a synchronous triggering method, and the phase-controlled switch controller outputs the conduction signal at the same time, so that the discharge process of the two channels starts and ends synchronously. The film structure of the stress simulation unit is completely consistent with the main sensing area. Only the touch sensing electrode is removed to ensure that it only responds to the parasitic dark current signal generated by thermomechanical stress, so that the independent extraction and comparison of dark current can be realized.

[0071] The discharge currents of the two channels discharge the total capacitance at their respective input terminals. At the end of the discharge period, the input node voltage of the main sensing channel is determined by both the touch event and stress noise, while the input node voltage of the dark current reference channel is determined only by stress noise. These two voltage values ​​are sampled and transmitted to the non-inverting and inverting input terminals of the subsequent differential amplifier circuit, respectively. The voltage output by the differential amplifier circuit is the touch sensing signal after subtracting the stress-induced dark current component.

[0072] Based on the above description, the first ramp generator and the second ramp generator can be built using operational amplifiers, resistor arrays, capacitor arrays and transistors of the same specifications. Stress simulation units that extend continuously from the main sensing area film layer structure are set in the non-visible area of ​​the edge of the touch film, and corresponding switching timing control logic is configured to realize the dual-ramp reference dark current cancellation function.

[0073] The dual-slope reference dark current cancellation loop also includes an adaptive filter configured to perform at least two cyclic input-output cross-validation operations, including: in the first round, synchronously acquiring the noise floor waveforms of the main sensing channel and the dark current reference channel; and in the second round, using a recursive least squares algorithm to convolve the time-domain waveform of the dark current reference channel with the sensing signal of the main sensing channel to cancel the noise floor.

[0074] It should be further explained that an adaptive filter is internally installed in the dual-slope reference dark current cancellation circuit. This adaptive filter uses a digital signal processor or field-programmable gate array as its hardware carrier and has a recursive least squares algorithm engine embedded inside. The input ports of the adaptive filter are connected to the output terminals of the sample-and-hold circuits of the main sensing channel and the dark current reference channel, respectively, and the output port of the adaptive filter is connected to the input terminal of the subsequent touch coordinate calculation unit.

[0075] Within each touch sensing cycle, the phase-controlled switch controller first synchronously turns on the first and second switching devices for a first preset time period. At the end of the first preset time period, the input node voltage of the main sensing channel is captured by the first sample-and-hold circuit and converted into a first digital value, and the input node voltage of the dark current reference channel is captured by the second sample-and-hold circuit and converted into a second digital value. This process constitutes the first round of operation. The first and second digital values ​​acquired in the first round of operation are simultaneously written into the first-in-first-out buffer memory inside the adaptive filter. Subsequently, the phase-controlled switch controller synchronously turns on the first and second switching devices again for a second preset time period, the duration of which is equal to that of the first preset time period. At the end of the second preset time period, the input node voltage of the main sensing channel is again captured by the first sample-and-hold circuit and converted into a third digital value, and the input node voltage of the dark current reference channel is again captured by the second sample-and-hold circuit and converted into a fourth digital value. This process constitutes the second round of operation. The first, second, third, and fourth digital values ​​constitute a complete set of cross-validation data.

[0076] After reading the data set, the recursive least squares algorithm engine inside the adaptive filter uses the sequence formed by the second and fourth digital values ​​as the reference noise sequence and the sequence formed by the first and third digital values ​​as the noise-contaminated main signal sequence, initiating the iterative calculation process. During the initialization phase of the iterative calculation, the algorithm engine sets the first-order transverse filter weight coefficient vector within a predetermined range as the zero vector and sets the inverse of the initial correlation matrix as a diagonally loaded identity matrix. In each iteration, the algorithm engine constructs a reference input vector based on the current value of the reference noise sequence, calculates the prior estimation error, and updates the weight coefficient vector based on the prior estimation error and the gain vector, while recursively updating the inverse of the correlation matrix. The iterative process sequentially traverses all sample points collected in the first and second rounds along the time axis. After traversing all sample points, the weight coefficient vector converges to a steady-state solution, which characterizes the transfer function of the noise component from the reference noise sequence to the main signal sequence.

[0077] When applying the recursive least squares algorithm, a transverse filter order suitable for touch signal processing is selected, and the forgetting factor is set between 0.95 and 0.99. During iterative calculation, signal data from the main sensing channel and the dark current reference channel are collected sequentially according to the time sequence. First, the filter weight coefficients and correlation matrix are initialized, and then the weight coefficients and matrix parameters are updated sequentially according to the collected signal data. After completing two rounds of signal acquisition and cross-validation, the noise waveform of the reference channel is matched and eliminated with the main channel signal in the time domain. The entire iterative process is executed cyclically in the order of signal acquisition, parameter update, and noise matching, thus achieving noise stripping.

[0078] The algorithm engine then subtracts the main signal sequence from the reference noise sequence filtered by the transfer function point by point. The resulting difference sequence is the clean touch sensing signal sequence after adaptive noise cancellation. This clean touch sensing signal sequence is transmitted from the output port of the adaptive filter to the touch coordinate calculation unit to calculate the planar coordinate position of the touch event. By performing at least two rounds of cyclic acquisition, including the first and second rounds, followed by recursive least squares iterative cancellation, the adaptive filter can maintain accurate tracking and removal of noise components even when the statistical characteristics of stress-induced dark current noise change slowly over time.

[0079] Based on the above description, commercially available digital signal processor chips or field-programmable gate array chips can be selected, and corresponding recursive least squares algorithm firmware can be written. Appropriate filter order and forget factor parameters can be set, and the sampling time and data writing timing of the sample-and-hold circuit can be configured to achieve adaptive filtering and cross-validation functions.

[0080] It also includes a lightweight digital twin model set up at the edge computing end, which is built based on Gaussian process regression and has embedded a nonlinear mapping relationship of "surface morphology-refractive index distribution-dark current spectral density";

[0081] The digital twin lightweight model is used to receive the real-time noise waveform acquired by the dark current reference channel as the initial perturbation variable, and performs no less than three forward prediction and backward correction iterations to generate nonlinear compensation current control commands to cancel the higher-order harmonic components of the dark current.

[0082] It should be further explained that the lightweight digital twin model is deployed within the edge computing unit of the touch display system. This edge computing unit is a microcontroller or dedicated neural network acceleration processor located on the same printed circuit board assembly as the touch driver chip. It maintains data communication with the analog-to-digital converter output of the dual-slope reference dark current cancellation loop through a serial peripheral interface bus.

[0083] The construction of the lightweight digital twin model is divided into an offline training phase and an online inference phase. The offline training phase is completed on an external computer with computing power exceeding that of the edge computing unit. First, a batch of touch film samples with the same film structure as the final product and including stress simulation units are prepared. These samples are placed in a controlled temperature and humidity chamber and subjected to cyclic bending loads of different amplitudes and frequencies. Simultaneously, the number of bending cycles, the dark current waveform data output by the stress simulation unit, and the height variation data of discrete points on the surface of the stress buffer and optical compensation structure obtained through offline measurement using a laser interferometer are recorded. The surface morphology height distribution at each sampling moment is expanded using Zernike polynomials, and the polynomial coefficient vector is extracted as the surface morphology feature parameter. The dark current waveform data at the same moment is transformed to the frequency domain using a fast Fourier transform, and the vector composed of the amplitudes of each harmonic is extracted as the dark current spectral density feature parameter. The above morphology feature parameter vector and dark current spectral density feature parameter vector are paired to form a training sample set, which contains thousands of sample pairs collected under various bending conditions and different fatigue accumulation stages. Subsequently, a squared exponential kernel function was selected as the covariance function for Gaussian process regression. This kernel function is expressed as the square of the Euclidean norm of the difference between the two input vectors divided by the square of the feature length scale parameter, then taking the negative exponent. An independent Gaussian white noise term, representing the observation noise variance, is also superimposed on the kernel function. The morphological feature parameter vectors from the training sample set are used as input variables for Gaussian process regression, and the corresponding dark current spectral density feature parameter vectors are used as target output variables. The feature length scale parameter, signal variance parameter, and noise variance parameter are jointly optimized by maximizing the logarithmic marginal likelihood function to obtain the optimized hyperparameter set. The hyperparameter set, along with the training sample set, is stored as a model parameter file.

[0084] The Gaussian process regression digital twin model uses a squared exponential kernel function to construct the covariance relationship. Hyperparameter optimization is achieved by maximizing the logarithmic marginal likelihood of the sample data. The nonlinear mapping relationship between surface morphology, refractive index distribution and dark current spectral density is formed through training and fitting with a large amount of sample data. The collected surface morphology data is used as input and dark current spectral density data is used as output. After multiple trainings, a stable mapping relationship can be established. During online inference, the real-time waveform of the dark current reference channel is used as the perturbation variable. After three forward prediction and backward correction iterations, the prediction results are gradually calibrated, and finally, a matching compensation current control command is generated.

[0085] During the online inference phase, the model parameter file is burned into the non-volatile memory of the edge computing unit. When the touch membrane is in the actual working environment, in each touch sensing frame cycle, the real-time noise waveform acquired by the dark current reference channel is first sent to the input buffer of the edge computing unit after analog-to-digital conversion as the initial perturbation variable. The edge computing unit performs the first round of forward prediction: the dark current waveforms of the current frame and several consecutive frames before it are jointly constructed into an input feature vector, which is substituted into the Gaussian process regression prediction function embedded in the firmware. The prediction function outputs the predicted mean vector and prediction covariance matrix of the dark current spectral density of several frames after the current frame. The predicted mean vector is the first round estimate of the higher-order harmonic components of the dark current. Subsequently, the first round of backward correction is performed: the first round estimate is compared with the dark current spectral density vector actually acquired in the current frame, and the residual vector is calculated. If the norm of the residual vector exceeds the preset convergence threshold, the residual vector is used as feedback to adjust the temporal weighting coefficients in the input feature vector. The adjusted temporal weighting coefficients give higher regression weights to the neighboring frames before the current frame. A second round of forward prediction is performed based on the adjusted time-weighted coefficients to obtain the second round of estimates. The residuals are then recalculated and compared with the convergence threshold. If convergence is still not achieved, the time-weighted coefficients are adjusted again and a third round of forward prediction is performed. This iterative process of forward prediction and backward correction is enforced for no less than three complete loops until the norm of the residual vector falls below the convergence threshold or the preset maximum number of iterations is reached.

[0086] At the end of the iteration, the predicted mean vector is output to the control command generation module. Based on the amplitude and phase information of each harmonic component in the predicted mean vector, the control command generation module synthesizes a pulse width modulation duty cycle sequence corresponding to the nonlinear compensation current waveform. This duty cycle sequence characterizes the time-domain envelope shape of the injected compensation current. Its waveform profile differs from the predicted dark current waveform by half a cycle in phase, and its amplitude scaling factor is determined by a weighted operation of the arithmetic square roots of the diagonal elements of the prediction covariance matrix.

[0087] Based on the above description, Python and the Scikit-learn Gaussian process regression library can be used to complete offline training on a computer. The trained model parameters can be exported as a C language header file. An online inference function for Gaussian process regression can be written on a microcontroller based on an ARM Cortex core. The analog-to-digital converter sampling rate and data buffer depth can be configured, and the convergence threshold and maximum number of iterations can be set to realize the lightweight digital twin model and its iterative prediction function.

[0088] It also includes a high-frequency pulse width modulation current injector, which is coupled to the virtual ground terminal of the preamplifier of the photoelectric conversion unit. It is used to inject a nonlinear compensation current that is out of phase with the predicted dark current waveform and whose amplitude is synthesized by probability weighting into the virtual ground terminal according to the control command output by the digital twin lightweight model.

[0089] It should be further explained that the high-frequency pulse width modulation current injector is composed of a current-mode digital-to-analog converter, a pulse width modulation waveform generator, a voltage-to-current conversion circuit, and an output coupling capacitor cascaded in sequence. The digital input terminal of the current-mode digital-to-analog converter is connected to the output bus of the edge computing unit, receiving nonlinear compensation current control commands generated by the digital twin lightweight model. These control commands are represented as a set of sequentially arranged multi-bit binary code words, each code word corresponding to a discrete quantized value of the current amplitude within one pulse width modulation cycle.

[0090] The pulse width modulation waveform generator contains a counter and a comparator. The counter increments cyclically at the system master clock frequency. The comparator compares the current count value with the reference threshold output by the current-mode digital-to-analog converter. When the count value is lower than the reference threshold, it outputs a high level. When the count value is equal to or higher than the reference threshold, it outputs a low level. This generates a sequence of square waves with a duty cycle that dynamically changes with the coded word.

[0091] The square wave sequence controls the switching devices in the voltage-to-current conversion circuit to periodically turn on and off. The voltage-to-current conversion circuit adopts a transconductance amplifier topology. Its non-inverting input is connected to a stable reference voltage provided by a bandgap reference voltage source, and its inverting input is connected to the current output terminal of the switching device through a feedback resistor. The output terminal of the transconductance amplifier drives the input branch of a current mirror, and the output branch of the current mirror is the controlled current output node.

[0092] The voltage-to-current conversion circuit converts a pulse-width modulated square wave into a pulse current sequence that is synchronized with it and whose amplitude is proportional to the reference current obtained by dividing the reference voltage by the feedback resistor. The average current value of this pulse current sequence in each pulse-width modulation cycle is proportional to the duty cycle of the square wave. The first plate of the output coupling capacitor is connected to the output branch of the current mirror, and the second plate of the output coupling capacitor is connected to the virtual ground terminal of the photoelectric conversion unit preamplifier.

[0093] This preamplifier employs a transimpedance amplifier structure, with its inverting input serving as a virtual ground. This virtual ground is connected to both the cathode of the photodiode and the first terminal of the feedback resistor, while the second terminal of the feedback resistor is connected to the output of the transimpedance amplifier. Under steady-state operation, the transimpedance amplifier forces the voltage at its inverting input to maintain at the common-mode reference voltage level through the feedback resistor. The virtual ground presents a low impedance to AC signals but a fixed potential to DC bias. When the pulse current sequence output from the high-frequency pulse-width modulation current injector is injected into the virtual ground through the output coupling capacitor, due to the low AC impedance of the virtual ground, almost all of the injected current flows into the feedback resistor, generating a feedback voltage drop proportional to the instantaneous value of the injected current. This voltage drop is directly superimposed on the output voltage of the transimpedance amplifier.

[0094] The control command sequence output by the lightweight digital twin model ensures that the time-domain envelope shape of the injected current corresponds to the predicted dark current waveform with a phase difference of half a cycle. Specifically, when the predicted dark current waveform is in the positive peak region, the injected current waveform is in the negative valley region; when the predicted dark current waveform is in the rising edge stage, the injected current waveform is in the falling edge stage. Simultaneously, the quantization amplitude of each coded word in the control command sequence is determined by a weighted calculation of the arithmetic square roots of the diagonal elements of the prediction covariance matrix. This ensures that the amplitude of the injected current envelope decreases when the uncertainty of the dark current prediction is high and increases when the uncertainty is low.

[0095] Through the DC blocking effect of the output coupling capacitor, the injected current contains only an AC component and will not affect the DC bias operating point of the photodiode. The compensation current injected into the virtual ground terminal is directly added to the photocurrent generated by the photodiode when it receives light and the parasitic dark current at the virtual ground terminal according to Kirchhoff's current law. The AC component of the compensation current cancels out the higher-order harmonic components in the parasitic dark current because they are out of phase and have matched amplitudes. The residual error current after cancellation is then converted into a voltage signal output by the transimpedance amplifier. At this time, the low-frequency fluctuation components and high-frequency harmonic components caused by stress-induced dark current drift in the voltage waveform presented at the output of the transimpedance amplifier are weakened, and only the useful signal component that is linearly related to the capacitance change caused by the touch event is retained.

[0096] Based on the above description, a commercially available current-mode digital-to-analog converter chip, a microcontroller with integrated pulse width modulation peripherals, a general-purpose transconductance amplifier, and discrete transistor components can be used to build this high-frequency pulse width modulation current injector. The period and resolution of the pulse width modulation waveform are set according to the system master clock frequency. The capacitance value of the output coupling capacitor is selected according to the frequency range of the predicted dark current waveform so that the low-end cutoff frequency of the injection circuit is lower than the dark current fundamental frequency. The reference current range is set by adjusting the output voltage of the bandgap reference voltage source and the resistance value of the feedback resistor. This enables the nonlinear compensation current injection and dark current high-order harmonic cancellation functions.

[0097] It also includes a surface energy state closed-loop processing device, which includes a spectral ellipticity monitoring unit and an atmospheric pressure plasma jet unit, used to perform no less than three surface molecular reconstructions and uniformity verifications after the anti-fingerprint and scratch-resistant hard coating has been cured, until the surface energy dispersion is lower than the preset anti-fingerprint function maintenance threshold.

[0098] It should be further explained that the surface energy state closed-loop processing device consists of a spectral ellipticity monitoring unit, an atmospheric pressure plasma jet unit, and a closed-loop feedback controller.

[0099] The spectral ellipticity monitoring unit includes a broadband light source, a polarizer, a sample stage, an analyzer, and a spectrometer. Its measurement optical path projects onto multiple predetermined sampling points on the surface of the anti-fingerprint and scratch-resistant hard coating at a preset incident angle. By analyzing the amplitude ratio and phase difference between the parallel and perpendicular components of the reflected light, the water contact angle, surface energy dispersion component, and polar component at each sampling point are calculated. The atmospheric pressure plasma jet unit includes an RF power supply, a matching network, a gas mass flow controller, and a quartz tube nozzle. The working gas is a mixture of argon and oxygen, with oxygen accounting for between 0.5% and 0.5% by volume. The mixed gas flows through the RF excitation region, generating a low-temperature plasma plume containing active oxygen atoms and metastable argon particles. This plume is ejected from the quartz tube nozzle and acts perpendicularly on the surface of the anti-fingerprint and scratch-resistant hard coating. The input of the closed-loop feedback controller is connected to the data output interface of the spectral ellipticity monitoring unit, and its output is connected to the RF power control terminal and the motion control terminal of the two-dimensional scanning displacement platform of the atmospheric pressure plasma jet unit.

[0100] In the surface treatment process after the anti-fingerprint and scratch-resistant hard coating has been cured by UV or heat, the coating sample is first moved to the area below the spectral ellipticity monitoring unit by a two-dimensional scanning displacement platform. The spectral ellipticity monitoring unit measures the surface energy value of each sampling point in sequence according to the preset grid, and calculates the arithmetic mean and standard deviation of the surface energy of all sampling points. If the standard deviation exceeds the preset anti-fingerprint function maintenance threshold, the closed-loop feedback controller determines that the surface energy uniformity of the coating does not meet the standard and starts the first processing cycle.

[0101] In the first processing cycle, the closed-loop feedback controller calculates the corresponding plasma scanning trajectory, RF power settings, and dwell time for each trajectory point based on the polarity and magnitude of the deviation between the surface energy at each sampling point and the arithmetic mean. Areas with positive deviations exceeding the allowable range receive longer dwell times and moderate RF power, while areas with negative deviations are marked as areas requiring no processing. The two-dimensional scanning displacement platform drives the quartz tube nozzle to scan the area to be processed along the planned trajectory. The active oxygen atoms in the plasma plume undergo oxidation and chain-breaking reactions with the surface polar groups in the organosilicon or fluorine-containing molecular chains on the coating surface, reducing the surface polarity component of some areas with degraded oleophobic properties to near intrinsic levels. Simultaneously, argon metastable particles maintain plasma discharge stability through the Penning ionization effect, thus avoiding etching damage to the carbon skeleton of the coating.

[0102] After the first plasma scan, the two-dimensional scanning displacement platform moves the coated sample back below the spectral ellipsometry monitoring unit. The spectral ellipsometry monitoring unit measures the surface energy of all sampling points again and calculates the standard deviation. The closed-loop feedback controller compares the second measurement result with the first measurement result. If the standard deviation still exceeds the threshold, the second processing cycle is initiated. In the second processing cycle, the closed-loop feedback controller replans the scanning trajectory and plasma parameters based on the residual deviation distribution after the first processing, and the increment step size of the RF power and the extension coefficient of the dwell time are adaptively adjusted according to the deviation convergence rate after the first processing. After the second processing cycle is completed, the third surface energy measurement and standard deviation evaluation are performed. This closed-loop process of "measurement to plasma processing to re-measurement" is enforced for no less than three complete cycles until the standard deviation of the surface energy of all sampling points drops below the preset threshold before terminating the processing process.

[0103] In the surface energy state closed-loop processing device, the anti-fingerprint function maintenance threshold is set to a surface energy dispersion of less than 3mN / m. This threshold is determined through a large number of anti-fingerprint coating performance tests. The surface energy dispersion is obtained by measuring the surface energy values ​​at multiple points of the coating and calculating its standard deviation. When the standard deviation is less than 3mN / m, the surface energy uniformity of the coating is deemed to meet the requirements for long-term maintenance of the anti-fingerprint function. During the processing, this threshold is used as the calibration standard to complete the surface molecular reconstruction and uniformity verification.

[0104] After the third or subsequent cycles, when the standard deviation converges to within the threshold, the closed-loop feedback controller outputs a processing completion signal and simultaneously records the final surface energy distribution data as a factory quality traceability file. After at least three closed-loop treatments, the anti-fingerprint and scratch-resistant hard coating exhibits a statistically uniform distribution of surface polar groups, ensuring spatial consistency of its oleophobic properties. Furthermore, because the plasma energy density is controlled to act only on a few atomic layers of the surface, the bulk cross-linking structure and pencil hardness parameters of the coating remain essentially unchanged before and after the treatment.

[0105] Based on the above description, a commercially available imaging ellipsometer, atmospheric pressure plasma jet cleaning equipment, and programmable multi-axis motion controller can be used to build the surface energy state closed-loop processing device. By setting a suitable optical constant model and surface energy calculation algorithm, determining the sampling grid density and standard deviation threshold, and writing a program for automatic generation of plasma scanning paths and adaptive adjustment of parameters, the surface molecular reconstruction and multiple closed-loop verification functions can be realized.

[0106] The refractive index adjusting particles are nano-zirconia particles, and their spatial concentration distribution in the stress buffer and optical compensation structure is determined by the compensation path function output by the joint simulation of the ray tracing algorithm and the finite element mechanical model.

[0107] It should be further explained that the refractive index adjusting particles are specifically zirconia nanoparticles, whose crystal phase is mainly tetragonal with a small amount of monoclinic phase. The particles are approximately equiaxed in shape, and their specific surface area and average particle size are controlled by the precursor hydrolysis conditions and calcination temperature during the preparation process. The spatial distribution of the volume concentration of these zirconia nanoparticles within the stress buffer and optical compensation structure is not uniform or a simple linear gradient distribution, but is uniquely determined by the compensation path function obtained by joint simulation solution using ray tracing algorithm and finite element mechanical model.

[0108] The specific process of co-simulation solution includes the following steps: First, in the finite element mechanical model construction stage, based on the measured elastic constants, thermal expansion coefficients, and interface bonding energy parameters of the transparent conductive functional layer and the anti-fingerprint and scratch-resistant hard coating, a two-dimensional axisymmetric or three-dimensional solid finite element mesh model of the three-layer composite structure including the hard coating, stress buffer and optical compensation structure, and transparent conductive functional layer is established. A local normal distributed load simulating finger pressing is applied to the upper surface of the model. The three normal stress components and three shear stress components of each discrete mesh node inside the stress buffer and optical compensation structure under the load are obtained. The strain energy density value of each node is calculated according to the elastic strain energy density formula.

[0109] Secondly, in the ray tracing algorithm construction stage, a geometric ray transmission model is established for the light emitted from the lower surface of the self-stress buffer and optical compensation structure to the photosensitive surface of the photoelectric conversion unit. The initial value of the refractive index of each grid node in the model is given by the linear mixing rule of the refractive index of the organosilicon elastomer matrix and the volume concentration of nano-zirconia particles, and the spatial distribution function of the refractive index inside the stress buffer and optical compensation structure is described in a gridded manner.

[0110] In the initial stress-free state, a collimated incident light beam is transmitted vertically downwards from the surface of the self-fingerprint-resistant and scratch-resistant hard coating. It passes sequentially through the hard coating, the stress buffer and optical compensation structure, and the transparent conductive functional layer. The ray tracing algorithm tracks the refraction angle of each ray at the interface of each layer and the propagation direction within each layer until the light reaches the photosensitive surface of the photoelectric conversion unit. The cumulative optical path value at each position on the photosensitive surface is recorded. This cumulative optical path value is defined as the integral of the product of the geometric path length of the light and the refractive index along the path.

[0111] Under the deformed state with applied load, the strain energy density distribution calculated by the finite element mechanical model is mapped to the ray tracing model. According to the photoelastic effect equation, the refractive index increment of each grid node is expressed as a function of the strain energy density, thereby obtaining the spatial distribution of the refractive index after deformation. Ray tracing calculation is performed again to obtain the cumulative optical path value at each position on the photosensitive surface under the deformed state.

[0112] When ray tracing and finite element method (FEM) are used in a joint simulation, the correlation between stress and refractive index is established based on the physical laws of photoelasticity. The strain energy density data of each region obtained by FEM calculation are directly mapped to the ray tracing model, and the refractive index parameters of each region are adjusted accordingly. The compensation path function is obtained by iteratively optimizing the optical path difference distribution. With the uniformity of the optical path difference in the touch optical path as the optimization target, the spatial distribution of nano-zirconia particles is adjusted one by one until the optical path difference deviation meets the optical transmission requirements, thereby determining the spatial concentration distribution law of the particles.

[0113] By subtracting the cumulative optical path value in the deformed state from the cumulative optical path value in the initial state point by point, the spatial distribution map of the optical path difference caused by the load is obtained.

[0114] In the compensation optimization phase of the co-simulation, the volume concentration of nano-zirconia particles at each grid node is used as the design variable. Upper and lower bounds are set for the concentration distribution, and the optimization objective is to reduce the standard deviation of the optical path difference spatial distribution under deformation to a preset level. A gradient-based sequential quadratic programming algorithm or a population-based non-dominated sorting genetic algorithm is used to iteratively update the design variable. In each iteration, the refractive index of each grid node is recalculated based on the current concentration distribution. Then, a ray tracing algorithm is used to evaluate the optical path difference distribution under deformation, calculate the objective function value, and determine convergence. If convergence fails, the concentration distribution is updated, and the next iteration continues. After tens to hundreds of iterations, the algorithm converges to a set of volume concentration spatial distribution values. The functional relationship between this distribution value and spatial location is the compensation path function.

[0115] The compensation path function is characterized by a relatively concentrated volumetric concentration of zirconia nanoparticles in the low-modulus region near the transparent conductive functional layer. Furthermore, the undulating trend of the concentration distribution along the layer plane spatially corresponds to the principal stress trace direction output by the finite element mechanical model for this region; that is, the concentration change is gradual along the principal stress trace direction, while the concentration gradient perpendicular to the principal stress trace direction is relatively steep. During the preparation stage, the compensation path function is converted into a numerical matrix of concentration distribution in three-dimensional spatial coordinates. This matrix guides the pre-dispersion process of zirconia nanoparticles in the precursor solution and the particle migration kinetics control process during the subsequent segmented UV curing process.

[0116] Based on the above description, a joint simulation platform can be built using a commercial finite element analysis software package and a self-developed ray tracing program. By setting appropriate material parameters and boundary conditions, selecting a suitable optimization solver, a compensation path function for specific product structural parameters can be obtained. This function can then be mapped to the concentration gradient parameters in the manufacturing process, thereby realizing the spatial concentration distribution and optical path difference compensation function of nano-zirconia particles.

[0117] The fingerprint-resistant and scratch-resistant hard coating is a fluorinated polyether modified acrylate UV-cured layer or an organosilicon modified thermosetting hard coating.

[0118] It should be further noted that there are two optional chemical composition systems for the anti-fingerprint and scratch-resistant hard coating in its specific implementation:

[0119] The first composition system is a fluorinated polyether modified acrylate UV-curable layer. The curable liquid precursor of this layer is composed of fluorinated polyether diacrylate oligomer, multifunctional aliphatic polyurethane acrylate, reactive diluent monomer, and photoinitiator. The main chain of the fluorinated polyether diacrylate oligomer is a perfluoropolyether segment with acrylate groups at both ends. The number of repeating units of the perfluoropolyether segment is an integer between five and twenty. The oligomer accounts for 10% to 30% of the total resin solid content by mass. The multifunctional aliphatic polyurethane acrylate provides the crosslinking backbone with a functionality of three to six. The reactive diluent monomer is selected from at least one of ethoxylated trimethylolpropane triacrylate or pentaerythritol triacrylate. The photoinitiator is an α-hydroxy ketone compound, and the amount added is 3% to 5% of the total resin mass. After uniformly mixing and vacuum degassing the above components under light-protected conditions, the precursor is coated onto the upper surface of the stress buffer and optical compensation structure using slot coating or microgravure coating methods. The wet film thickness is controlled by the distance between the coating head and the substrate and the flow rate of the liquid supply pump. The coated wet film undergoes surface pre-crosslinking under a first stage of low-pressure ultraviolet irradiation in an inert gas atmosphere, followed by deep curing under a second stage of high-pressure ultraviolet irradiation. The total irradiation energy density is controlled within the range of 1,000 to 3,000 millijoules per square centimeter. The cured coating has a pencil hardness of not less than 3H and a water contact angle of not less than 110 degrees. During film formation, the perfluoropolyether segments in the coating are directionally enriched towards the air interface, forming an oleophobic surface layer enriched with fluorine atoms. Simultaneously, at the interface between the coating and the stress buffer and optical compensation structure, the ester bonds in the acrylate groups undergo ester exchange or hydrogen bonding with the silanol groups on the surface of the organosilicon elastomer, forming a chemically anchored and physically interpenetrating network.

[0120] The second composition system is a silicone-modified thermosetting hard coating. The curable liquid precursor of this coating consists of a silane coupling agent pre-hydrolyzed product, a silicone silsesquioxane resin, a curing catalyst, and an organic solvent. The silane coupling agent pre-hydrolyzed product is obtained by co-hydrolysis and condensation of methyltrimethoxysilane and phenyltrimethoxysilane in an acidic aqueous solution. The silicone silsesquioxane resin is a cage-like or ladder-like polysilsesquioxane, and the curing catalyst is an organotin compound or an organotitanium compound. After coating the above precursor onto the surface of the stress buffer and optical compensation structure, it undergoes leveling and solvent evaporation processes, and then is placed in a constant temperature oven for thermosetting according to a stepped temperature increase program. The final curing temperature is 120°C to 150°C, and the total curing time is 30 to 60 minutes. The cured coating has a pencil hardness of no less than 3H and a water contact angle of no less than 105 degrees. The silicon-oxygen-silicon main chain structure in the coating gives it weather resistance and thermal stability. The stress buffer and optical compensation structure between the coating and the organosilicon elastomer matrix forms a gradual transition interface due to the intrinsic compatibility of the materials. There are no abrupt changes in refractive index or weak adhesion at the interface.

[0121] Under both of the above-mentioned systems, the anti-fingerprint and scratch-resistant hard coatings have optical transparency with a transmittance of no less than 90% in the visible light band, and the difference between their refractive index and the refractive index of the stress buffer and optical compensation structure at the interface is less than 0.05, thereby avoiding the introduction of additional Fresnel reflection loss between the coating and the compensation structure.

[0122] Based on the above description, depending on the specific requirements for production efficiency, equipment compatibility, and environmental tolerance, either the UV curing route of fluorinated polyether modified acrylate or the thermosetting route of organosilicon modified acrylate can be selected. The corresponding commercially available chemical raw materials and coating and curing equipment can be selected, and the preparation can be carried out according to the ratio range and process parameters to achieve the function of fingerprint-resistant and scratch-resistant hard coating.

[0123] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0124] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film, comprising: Transparent conductive functional layer; A fingerprint-resistant and scratch-resistant hard coating is disposed above the transparent conductive functional layer; And a photoelectric conversion unit for converting optical signals into electrical signals; Its characteristic is that it further includes: A stress buffer and optical compensation structure is disposed between the anti-fingerprint and scratch-resistant hard coating and the transparent conductive functional layer. The stress buffer and optical compensation structure includes a composite layer composed of an elastic matrix and refractive index adjusting particles dispersed therein. The elastic modulus inside the composite layer has a non-uniform gradient distribution in the direction perpendicular to the layer plane, and the spatial concentration distribution of the refractive index adjusting particles has a preset non-linear correspondence with the gradient distribution of the elastic modulus. The stress buffer and optical compensation structure is configured to: respond to the interfacial stress generated by the deformation of the anti-fingerprint and scratch-resistant hard coating, guide the stress to dissipate to the low modulus region through the internal modulus gradient, and simultaneously drive the refractive index adjustment particles to migrate in a directional manner to compensate for the local refractive index change caused by residual stress, thereby suppressing the dark current drift of the photoelectric conversion unit caused by optical path distortion.

2. The photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film according to claim 1, characterized in that: The elastic modulus gradient distribution pattern inside the stress buffer and optical compensation structure corresponds to the biomimetic fractal stress flow channel pattern obtained by iterative solution through a cross-scale simulation model. This pattern breaks down the input external concentrated stress into a diffusely scattered stress distribution with spatial incoherence.

3. The photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film according to claim 2, characterized in that: It also includes a process morphology verification subsystem, which includes a laser interferometer and a local infrared pulse array. This subsystem is used to compare the deviation between the actual leveling morphology and the biomimetic fractal stress flow channel pattern in real time during the wet film curing process of the stress buffer and optical compensation structure. It also corrects the local prepolymerization degree of the material by adjusting the local infrared pulse parameters to force the execution of at least two microchannel morphology reconstruction cycles.

4. The photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film according to claim 1, characterized in that: The photoelectric conversion unit is coupled to a dual-slope reference dark current cancellation circuit, which includes: The first ramp generator is used to drive the main sensing channel to discharge. A second ramp generator is used to drive the dark current reference channel to discharge; and: The stress simulation unit, located at the input of the dark current reference channel and without touch sensing function, is used to generate a parasitic dark current signal that only reflects the effects of thermo-mechanical stress.

5. The photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film according to claim 4, characterized in that: The dual-slope reference dark current cancellation loop also includes an adaptive filter configured to perform at least two cyclic input-output cross-validation operations, including: The first round of synchronous acquisition of the noise substrate waveforms of the main sensing channel and the dark current reference channel; In the second round, a recursive least squares algorithm is used to convolve the time-domain waveform of the dark current reference channel with the sensing signal of the main sensing channel to eliminate the dark current.

6. The photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film according to claim 5, characterized in that: It also includes a lightweight digital twin model set up at the edge computing end, which is built based on Gaussian process regression and has embedded a nonlinear mapping relationship of "surface morphology-refractive index distribution-dark current spectral density"; The digital twin lightweight model is used to receive the real-time noise waveform acquired by the dark current reference channel as the initial disturbance variable, and to perform no less than three forward prediction and backward correction iterations to generate nonlinear compensation current control commands for canceling the higher-order harmonic components of the dark current.

7. The photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film according to claim 6, characterized in that: It also includes a high-frequency pulse width modulation current injector, which is coupled to the virtual ground terminal of the preamplifier of the photoelectric conversion unit, and is used to inject a nonlinear compensation current that is out of phase with the predicted dark current waveform and whose amplitude is synthesized by probability weighting into the virtual ground terminal according to the control command output by the digital twin lightweight model.

8. The photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film according to claim 1, characterized in that: It also includes a surface energy state closed-loop processing device, which includes a spectral ellipticity monitoring unit and an atmospheric pressure plasma jet unit, used to perform no less than three surface molecular reconstructions and uniformity verifications after the anti-fingerprint and scratch-resistant hard coating has been cured, until the surface energy dispersion is lower than the preset anti-fingerprint function maintenance threshold.

9. The photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film according to claim 1, characterized in that: The refractive index adjusting particles are nano-zirconia particles, and their spatial concentration distribution in the stress buffer and optical compensation structure is determined by the compensation path function output by the joint simulation of the ray tracing algorithm and the finite element mechanical model.

10. A photoelectric signal conversion system for an anti-fingerprint and scratch-resistant touch film according to claims 1 to 9, characterized in that: The fingerprint-resistant and scratch-resistant hard coating is a fluorinated polyether modified acrylate UV-cured layer or an organosilicon modified thermosetting hard coating.

Images