Intelligent metasurface design and detection method for optical part

Abstract

The invention relates to an intelligent metasurface design and detection method for an optical part, and the method comprises the steps: generating a superposed signal containing a high-frequency detection signal sequence and a low-frequency driving signal sequence, and loading the superposed signal to a liquid crystal metasurface unit; a feedback current signal generated by the liquid crystal metasurface unit is collected in real time, and a capacitive reactive current component having a phase difference with the high-frequency detection signal sequence is separated from the feedback current signal; determining an actual optical phase according to the signal change trend of the low-frequency driving signal sequence and the capacitive reactive current component or a preset step-down mapping logic; extracting a direct-current bias component in the feedback current signal, performing time domain accumulation calculation on the direct-current bias component to determine the ion migration degree in the liquid crystal, and generating a compensation bias voltage for offsetting an internal residual polarization field according to the ion migration degree; and dynamically adjusting the amplitude of the low-frequency driving signal sequence according to the deviation value of the actual optical phase and the target phase in combination with the compensation bias voltage.

Description

Technical Field

[0001] This application relates to the fields of optical modulation and micro / nano photonics technology, and more specifically, to a method for designing and detecting intelligent metasurfaces for optical components. Background Technology

[0002] Optical metasurfaces, through the precise arrangement of microstructural units on a subwavelength scale, enable flexible control of the phase, amplitude, and polarization state of light waves. By integrating electrically controlled media such as liquid crystals into metasurfaces, dynamic beam deflection, adaptive optics correction, and holographic dynamic displays can be achieved using electric field driving.

[0003] Existing liquid crystal metasurface control schemes typically employ preset control strategies, such as open-loop driving based on a factory-calibrated voltage-phase (PV) mapping table; or introducing external optical sensors (such as wavefront analyzers) to construct a feedback loop and correct the driving command by monitoring the quality of the far-field light spot.

[0004] However, a core problem faced by the aforementioned existing technologies in practical applications is that the lack of effective acquisition of the real-time physical state of the devices by the control system makes it difficult for the system to maintain subwavelength-level control accuracy in complex dynamic tasks. Summary of the Invention

[0005] Based on this, it is necessary to address the aforementioned technical problems and, in order to obtain the real-time physical state of the device, this application provides a method for intelligent metasurface design and detection of optical components.

[0006] A method for designing and detecting intelligent metasurfaces of optical components, the method comprising: A superimposed signal containing a high-frequency detection signal sequence and a low-frequency driving signal sequence is generated and loaded onto the liquid crystal metasurface unit; wherein, the amplitude of the high-frequency detection signal sequence is lower than the activation threshold of the liquid crystal molecules and the frequency is higher than the mechanical response frequency of the liquid crystal molecules; the low-frequency driving signal sequence is determined based on the target phase; The feedback current signal generated by the liquid crystal metasurface unit is acquired in real time, and the capacitive reactive current component that has a phase difference with the high-frequency detection signal sequence is separated from the feedback current signal. Based on the signal variation trend of the low-frequency drive signal sequence, and based on the capacitive reactive current component or the preset step-down mapping logic, the actual optical phase is determined. The DC bias component in the feedback current signal is extracted, and the DC bias component is accumulated in the time domain to determine the degree of ion migration inside the liquid crystal. Based on the degree of ion migration, a compensation bias voltage is generated to counteract the internal residual polarization field. Based on the deviation between the actual optical phase and the target phase, and in conjunction with the compensation bias voltage, the amplitude of the low-frequency drive signal sequence is dynamically adjusted.

[0007] The beneficial effects of this application are as follows: By introducing specific high-frequency characteristic signals into the driving excitation and extracting the corresponding components, this application can establish a correlation mapping between electrical feedback and the microscopic physical state of the liquid crystal metasurface unit, which significantly reduces the degree of nonlinear mismatch between the driving command and the actual optical phase response. Since this application can quantify and offset the negative impact of internal field strength evolution and physical hysteresis to a certain extent, it is more conducive to suppressing the performance drift phenomenon under long-cycle operation. Thus, without the need to introduce external redundant optical sensing, it effectively enhances the control consistency and output accuracy of the metasurface system under dynamic conditions, and improves the reliability of integrated optoelectronic devices. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the accompanying drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below.

[0009] Figure 1 This is a schematic diagram of the structure of an optical metasurface unit in one embodiment of this application; Figure 2 This is a flowchart illustrating a method for designing and detecting intelligent metasurfaces for optical components in one embodiment of this application. Detailed Implementation

[0010] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0011] This application provides an optical metasurface unit whose physical configuration strictly corresponds to the hierarchical relationship in the claims. For example... Figure 1 As shown, the optical metasurface units are constructed sequentially from bottom to top (in the opposite direction to the incident light direction): First electrode layer: First, a metal thin film (preferably gold (Au), aluminum (Al), or silver (Ag) with a thickness of 100nm-300nm is prepared on a substrate (such as a high resistivity silicon wafer or glass). In this embodiment, the first electrode layer has a dual function: electrically, it serves as the common ground (GND) for electrical excitation, and optically, it serves as a reflector for light waves.

[0012] Subwavelength superatomic array: Located above the first electrode layer. The subwavelength superatomic array is composed of periodically arranged micro- and nano-structures of high refractive index media (such as polycrystalline silicon, titanium dioxide, etc.). Its lateral dimensions and height are both on the subwavelength scale, and it is used to achieve initial modulation of the phase of the incident light wave.

[0013] Nematic liquid crystal layer: Selected with high birefringence ( Nematic liquid crystal materials (such as E7 type) are filled into the gaps of the subwavelength superatomic array and completely cover the array through a perfusion process.

[0014] The second electrode layer, located above the nematic liquid crystal layer, is a transparent conductive film made of indium tin oxide (ITO). This second electrode layer possesses excellent light transmittance, allowing incident light to enter the space below and working in conjunction with the first electrode layer to establish a driving electric field.

[0015] In this embodiment, the nematic liquid crystal layer and the subwavelength metaatomic array are both located within the "electric field driven space" formed between the first electrode layer and the second electrode layer. To ensure physical consistency within this space, this embodiment further includes: Orientation structure: Polyimide (PI) orientation layers are spin-coated on the inner surfaces of the first electrode layer and the second electrode layer (i.e., the side facing the electric field driving space), and the initial pretilt angle of the liquid crystal molecules is determined by rubbing treatment.

[0016] Gap control: By doping support spheres (spacers) between the first electrode layer and the second electrode layer, the height of the electric field driven space is precisely controlled between 3μm and 10μm.

[0017] The specific physical structure described above fulfills the hardware prerequisites for the "in-situ sensing" of this application, and the specific coordination relationship is as follows: Construction of dynamic capacitor channel: The first electrode layer and the second electrode layer not only provide the driving field, but also utilize their parallel configuration to form a dynamic parallel plate capacitor by using the nematic liquid crystal layer and the subwavelength superatomic array located between them as a composite dielectric.

[0018] Physical realization of impedance mapping: When liquid crystal molecules in a nematic liquid crystal layer are deflected in an electric field-driven space, their equivalent dielectric constant ( The change in impedance directly causes a change in the complex impedance between the first and second electrode layers. Because the subwavelength superatomic array fixes the local microscopic electric field distribution, the impedance change and the optical phase change are physically locked together.

[0019] The signal extraction node support: The first electrode layer (bottom metal electrode) serves as the signal output terminal, allowing external circuits to extract the true direction vector of the liquid crystal molecules in real time without interfering with the optical modulation state. The capacitive current component of the information.

[0020] Based on the above functional support, in this physical configuration, the direction vector of the liquid crystal molecules ( It exhibits unique "optical-electrical dual-domain coupling" characteristics: Optical domain: Pointer ( The effective refractive index around the superatomic array is determined. ), thereby controlling the phase of the reflected light ( ); Electric domain: Pointer vector ( The equivalent dielectric constant of the interlayer medium is determined by ( ). ), thereby controlling the dynamic capacitance of the control unit ( ).

[0021] Therefore, under an ideal physical model, the optical phase ( ) and unit capacitor ( There exists a one-to-one single-value mapping relationship, which constitutes the physical premise for real-time monitoring of optical phase using electrical feedback.

[0022] However, in actual physical operation, due to the influence of the intrinsic properties of materials, the above ideal mapping relationship will degenerate in the following ways: Mechanical domain delay (hysteresis effect): Influenced by rotational viscous resistance, the mechanical rotation of liquid crystal molecules (…) Lagging behind the drive voltage ( Changes in the driving voltage () result in different phases corresponding to the same voltage. It is no longer a reliable characterization of phase.

[0023] Electrochemical interference (ion migration): The accumulation of impurity ions at the electrode interface generates a built-in electric field. This neutralizes part of the external driving electric field. This results in even the external voltage ( While the electric field is constant, the actual electric field experienced by the molecules also drifts, causing phase misalignment.

[0024] Given the driving voltage ( Phase with target ( The deterministic logic between ( ) has been broken due to the aforementioned physical defects. This application recognizes that it is necessary to bypass the unreliable driving voltage and directly extract the true pointing vector of the molecule from the electrical domain. ( ) electrical signals.

[0025] Due to the unit capacitance ( The phase "true value" is a real-time mirror image of the molecular director. By extracting the capacitive component in the feedback current, a phase "true value" independent of the driving trajectory can be obtained. To extract this weak feature without interfering with the modulation state and simultaneously counteract the electric field deviation caused by ion migration, this application provides a method for intelligent metasurface design and detection of optical components, applicable to detection equipment, such as... Figure 2 As shown, the specific steps are as follows: steps S201 to S205: Step S201: Generate a superimposed signal containing a high-frequency detection signal sequence and a low-frequency driving signal sequence, and load the superimposed signal onto the liquid crystal metasurface unit; As a dielectric material with dielectric anisotropy, liquid crystal molecules exhibit two distinct physical dimensions in response to an external electric field: mechanical rotation and polarization. This is the physical basis for achieving in-situ sensing. Mechanical response dimension (phase modulation domain): Liquid crystal molecules possess intrinsic physical mass (generating rotational inertia) and rotational viscosity (generating internal frictional resistance). When the applied electric field frequency is in the low-frequency range, the electric field torque can overcome the viscous resistance, driving a macroscopic spatial deflection of the molecular long axis. This mechanical rotation directly alters the effective refractive index of the metasurface unit for incident light waves, thereby achieving optical phase modulation.

[0026] Electropolarization response dimension (state-aware domain): Within a molecule, the electron cloud and dipoles can undergo instantaneous displacement polarization in response to an alternating electric field. Even when the electric field frequency exceeds the mechanical response limit of the molecule's long axis, this microscopic polarization capability persists, and its equivalent dielectric constant is strictly dependent on the molecule's instantaneous deflection angle. This means that a high-frequency electric field can provide feedback on the molecule's spatial orientation through charge response without driving molecular rotation.

[0027] Traditional driving schemes use only a single low-frequency driving signal. However, due to the significant viscoelastic hysteresis effect of liquid crystal materials within their nanoscale confinement, the optical phase and driving voltage are not simply linearly mapped, limiting open-loop control accuracy. To establish in-situ closed-loop feedback, a "sensing carrier" that does not interfere with the optical modulation state but can extract molecular angle information in real time must be embedded in the driving signal. Therefore, this step solves the problem of coexistence of "control" and "feedback" within the same physical channel by superimposing a high-frequency detection signal and a low-frequency driving signal in the time domain.

[0028] In this embodiment, the low-frequency drive signal sequence Defined as the voltage envelope for executing phase modulation commands. Its discrete-time form is expressed as: in, This is the sampling sequence number. This is the preset target optical phase. The specific generation steps are as follows: First, based on the target phase The corresponding static reference voltage is retrieved from the pre-stored "phase-voltage mapping table (P-VTable)". This voltage represents the nominal field strength required to achieve the target phase under standard conditions.

[0029] In order to overcome the response hysteresis of liquid crystal molecules, It exhibits a controlled evolution trajectory. Real-time calculation of the current driving voltage and... The difference is used to plan a slope with a specific gradient. The sequence. The slope The positive and negative polarities indicate whether the liquid crystal molecules are in a "forced flip" or "elastic reset" state, providing a logical reference for the subsequent step S203 to resolve hysteresis ambiguity.

[0030] High-frequency detection signal sequence Defined as a "dielectric state-sensing carrier", its mathematical expression is as follows: in, For amplitude, To detect frequency, Sampling frequency, This is the initial phase. Its parameter settings are subject to the following physical boundary constraints: Frequency constraint (physical inertial locking): Set to a cutoff frequency much higher than the mechanical response frequency of liquid crystal molecules (typically) Due to the rotational inertia and viscous resistance of molecules, their long axis cannot undergo physical displacement with the high-frequency alternating electric field, thus exhibiting an inertial locked state in spatial orientation (i.e., optical phase), eliminating phase interference caused by the sensing signal.

[0031] Amplitude constraint (small signal non-intrusive): Strictly limited to the initiation threshold voltage of liquid crystal molecules The following (usually) This micro-scale design ensures that the Coulomb torque generated by the probe electric field is insufficient to alter the macroscopic equilibrium position of the molecules, thus maintaining the control precision of the optical wavefront.

[0032] In the digital domain, the two sequences are linearly superimposed to synthesize the final composite drive instruction sequence: The specific actions to be performed are as follows: The two sequences described above are generated using a digital frequency synthesizer (DDS). To ensure that the extremely weak characteristic current can be extracted via phase-sensitive detection in the subsequent step S202, the... The generation clock and the sampling clock of the detection device are synchronized to ensure that the high-frequency signal of each cycle has a definite phase starting point.

[0033] Synthesized The voltage is converted into an analog voltage by a high-speed digital-to-analog converter (DAC) and applied between the top transparent electrode and the bottom metal electrode of the metasurface unit to establish a composite electric field within the liquid crystal layer.

[0034] Through the above path, this step constructs a dual-frequency coexisting excitation environment at the physical level: the low-frequency driving signal guides the liquid crystal molecules to move along the command trajectory, while the high-frequency detection signal utilizes the "mechanical high-pass filtering" characteristics of the molecules to excite a capacitive feedback current that reflects the instantaneous pointing information of the molecules in real time without interfering with the molecules. This provides the purest underlying data sequence for the high signal-to-noise ratio feature extraction and phase closed-loop calibration in subsequent steps.

[0035] Step S202: Real-time acquisition of feedback current signal generated by liquid crystal metasurface unit, and separation of capacitive reactive current component that has a phase difference with high frequency detection signal sequence from feedback current signal; Since the excitation signal applied to the liquid crystal cell is a composite signal, the resulting feedback current contains a variety of complex physical components. This step aims to accurately extract characteristic electrical quantities reflecting the instantaneous deflection angle of liquid crystal molecules from a noisy background using orthogonal demodulation techniques in the electrical domain.

[0036] Superimposed signals Driven by the total feedback current flowing through the liquid crystal metasurface unit It exhibits a complex evolution of impedance characteristics, which mainly includes the following physical components: Low-frequency drive component: by The generated strong current is used to maintain the macroscopic deflection of molecules, but it does not carry microsecond-level angle sensing information.

[0037] Resistive active component (loss current): Under actual working conditions (such as high temperature or long-term pressure), liquid crystal materials will generate weak ionic conductivity or dielectric loss. This part of the current is in phase with the excitation voltage and is greatly affected by environmental interference, which can easily lead to artifacts in angle determination.

[0038] Capacitive reactive component (sensing core): As a dielectric, the current generated by the displacement polarization of liquid crystal molecules leads the high-frequency detection voltage in phase. Exactly 90° (i.e., orthogonal phase). Due to this current amplitude... The instantaneous equivalent capacitance of the liquid crystal layer alone Decision, and It is the molecular direction vector A direct physical mirror image, therefore It is the purest and most reliable underlying physical indicator for characterizing the optical phase state.

[0039] because For small-signal detection, the excited sensing current is extremely weak and easily drowned out by low-frequency driving current background and random thermal noise. Traditional amplitude detection methods cannot distinguish between "changes in capacitive current caused by molecular rotation" and "changes in resistive current caused by material leakage". To resolve this identification ambiguity, this step introduces orthogonal phase-sensitive demodulation logic, which utilizes the orthogonality of the phase dimension to simultaneously implement dual filtering in the frequency and phase domains.

[0040] To ensure that the weak detection features are not drowned out by background noise, this embodiment employs a high signal-to-noise ratio acquisition link: A wideband precision transimpedance amplifier (TIA) is deployed in the feedback loop of the liquid crystal cell. This circuit utilizes the deep negative feedback characteristic of the operational amplifier to control the nanoampere-level feedback current flowing through the liquid crystal layer. It is converted into a voltage signal in real time, while maintaining the equivalent potential distribution at both ends of the unit to prevent the acquisition circuit from having a counteracting effect on the modulation field.

[0041] The converted voltage signal is digitized using a high-bit-depth (e.g., 16 bits or more) analog-to-digital converter (ADC). The sampling process must be performed by the generator. The same clock trigger ensures that the feedback current sequence is strictly phase aligned with the excitation source sequence on the time axis, thereby locking the phase reference origin of the signal in the digital domain.

[0042] The core operation of extracting the capacitive reactive current component from the complex total feedback current signal lies in performing a "characteristic projection of the phase dimension": First, based on the detection frequency defined in step S201... The spectral position of the probe component is located from the feedback signal sequence. Utilizing the orthogonality of the electrical vector space, the acquired high-frequency feedback signal is decomposed into two mutually perpendicular components: one component along the direction of the excitation voltage vector (representing loss), and the other component along a direction perpendicular to the excitation voltage vector (representing capacitive response). By extracting the amplitude of these perpendicular components, the "capacitive reactive current component" is obtained. This component effectively eliminates resistance fluctuations (resistive interference) caused by liquid crystal material aging or environmental impurities, and also removes baseband offset caused by the envelope of the low-frequency drive signal.

[0043] This step implements a "virtual sensor" function at the physical layer. It transforms the abstract, non-intuitive deflection process of liquid crystal molecules into a highly deterministic and physically unique electrical characteristic quantity through complex impedance feature extraction technology. This characteristic quantity serves as the input to step S203, solving the problem of "voltage and phase not being unique due to hysteresis" in traditional control, and laying the data foundation for achieving subwavelength-level optical closed-loop control.

[0044] Step S203: Determine the actual optical phase based on the signal change trend of the low-frequency drive signal sequence and the capacitive reactive current component or the preset step-down mapping logic. After extracting the "capacitive reactive current component" that characterizes the instantaneous arrangement of liquid crystal molecules, this step aims to eliminate the interference of the intrinsic hysteresis effect of the material and accurately restore the electrical characteristic quantity to the optical phase quantity.

[0045] The deflection process of liquid crystal molecules under an electric field is controlled by internal viscoelastic torques. These torques exhibit significant path dependence: Energy injection stage (pressure boost): When the external electric field is enhanced, the molecules need to overcome rotational viscous resistance to flip over, and their arrangement is a hysteretic response to the change in electric field.

[0046] Energy release phase (voltage reduction): When the external electric field weakens, the molecules mainly rely on the intrinsic elastic energy of the material to reset. Due to the different directions of viscous resistance in the two processes, the deflection trajectories of the liquid crystal molecules cannot coincide during the voltage increase and decrease processes.

[0047] Physically, this forms a closed "hysteresis loop." This means that under different motion trajectories, the same electrical characteristic value (capacitive reactive current component) may correspond to two completely different spatial pointing angles. If only a single mapping table is used for conversion, the optical phase determination result will cause a drastic jump when switching between boost and buck voltage.

[0048] Although the current component extracted in step S202 is pure, it is physically a "non-single-valued function" and cannot directly and uniquely determine the optical phase. To resolve this "one cause, multiple effects" identification ambiguity, an additional logical dimension must be introduced—namely, the direction of motion of the liquid crystal molecules. By identifying the evolution trend of the driving signal, the complex bidirectional hysteresis loop can be decomposed into two independent single-valued mapping curves. The purpose of this step is to force a one-to-one correspondence between the current characteristics and the optical phase through "trend prediction + logical splitting".

[0049] The key features involved in this step and their physical definitions are as follows: Signal variation trend: defined as low-frequency drive signal sequence The numerical gradient (i.e., the direction of the first derivative) on the time axis. It represents the dynamic intention of the external excitation source to exert a torque on the liquid crystal molecules.

[0050] Actual optical phase : Defined as the actual phase delay of the incident light wave caused by the superatomic array under the current physical environment.

[0051] Mapping logic branch: defined as two mutually exclusive mathematical transformation models preset for different motion modes of liquid crystal molecules (forced flipping and elastic reset).

[0052] Based on the digital signal processing unit, the following specific logical judgments and calculations are performed: Real-time comparison of the current frame with historical frames Numerical value. If If it is determined to be in the numerical increment phase; The system determines whether the value is decreasing or remaining constant. This determination provides a logical trigger signal for subsequent algorithmic flow splitting.

[0053] During the numerical increment phase, the "boost mapping logic" is activated. This logic uses the capacitive reactive current component output from step S202 as the independent variable and substitutes it into a preset rising edge characteristic function. This function calculates the instantaneous phase of the molecules under forced deflection by correcting the phase lag shift caused by rotational viscosity. .

[0054] During the decreasing or plateauing phase, the system switches to "de-pressure mapping logic." Since molecular motion is primarily driven by endogenous elastic energy during this phase, phase recovery is performed by invoking a pre-defined de-pressure trajectory curve. In specific implementation paths, this logic can perform fine calibration based on the current feedback value or perform trajectory fitting based on the de-pressure initiation point and time constant, thereby outputting the accurate phase for that mode. .

[0055] This step implements a "logic deconstructor" at the physical level. It utilizes the changing trend of the driving signal as a criterion to effectively resolve the mapping ambiguity caused by the hysteresis effect of the liquid crystal material. By multidimensionally fusing electrical feedback (current component) with the time-domain trajectory (trend), it can maintain a precise grasp of the "true optical phase" throughout the rapid dynamic switching of liquid crystal molecules. This solves the phase control inaccuracy problem caused by "path dependence" in traditional open-loop control, providing a high-confidence feedback criterion for the final closed-loop adjustment in step S205.

[0056] Step S204: Extract the DC bias component from the feedback current signal, and perform time-domain accumulation calculation on the DC bias component to determine the degree of ion migration inside the liquid crystal. Generate a compensation bias voltage to counteract the internal residual polarization field based on the degree of ion migration. After solving the "mechanical domain hysteresis" problem of liquid crystal molecules (step S203), the challenge of "electrochemical domain drift" caused by the microscopic properties of the material still exists. This step aims to sense and cancel the residual polarization field inside the liquid crystal layer by real-time mining of the extremely low frequency components in the feedback signal.

[0057] During the synthesis and encapsulation of liquid crystal materials, trace amounts of impurity ions (such as synthesis byproducts or environmentally dissociated ions) inevitably remain. Under the influence of an external driving electric field, these charged ions will undergo directional migration along the electric field lines: When a superimposed signal is applied to both ends of the electrode, not only is there molecular dipoleization within the liquid crystal layer, but there is also forced motion of ions.

[0058] As the working time progresses, the migrating ions reach the "alignment layer interface" on the electrode surface. Because the alignment layer has insulating properties, the ions cannot penetrate the electrode and are blocked at the interface, resulting in physical accumulation and forming the so-called "electric double layer effect".

[0059] The ionic charges accumulated at the interface generate a "built-in polarization field" that is opposite in direction to the external driving electric field. According to the principle of superposition of electric fields in physics, the effective field strength actually experienced by the liquid crystal molecules is the vector sum of the external driving field and the built-in polarization field: As the amount of ion accumulation increases, the built-in polarization field gradually strengthens, causing even the external driving voltage (in step S201) to increase. If the electric field remains constant, the actual electric field experienced by the molecules will also decrease.

[0060] This electric field shielding phenomenon causes the deflection angle of liquid crystal molecules to drift slowly over time. If this internal polarization field is not actively intervened, even if the phase calculated in step S203 is accurate, the control command will fail to drive the molecules to the predetermined position because the "effective electric field is subtracted".

[0061] The key physical characteristics involved in this step are defined as follows: DC bias component: refers to the extremely low-frequency (close to zero-frequency) component in the feedback current signal. Its amplitude directly characterizes the net migration rate of impurity ions within the current liquid crystal layer, and is the only window for sensing the ion movement state.

[0062] Ion migration degree: defined as the net amount of charge accumulated at the electrode interface. This characteristic quantifies the shielding ability of the current built-in polarization field against the external electric field.

[0063] Compensating bias voltage: refers to the additional DC offset level applied to counteract the built-in polarization field. Its purpose is to physically "fill in" the potential loss caused by ion accumulation.

[0064] Based on the logic of electrochemical in-situ sensing, the following specific calculation and generation actions are performed: The full-bandwidth feedback current signal acquired in step S202 is subjected to low-pass processing. By filtering out the high-frequency detection signal component and the low-frequency drive signal AC component, the "DC bias component" reflecting the intensity of ion directional flow is extracted.

[0065] The extracted DC bias component is discretized and integrated in the time domain. By tracking the integral trajectory of the current over time, the scale of net charge accumulation at the interface can be quantified in real time, thereby determining the current "degree of ion migration".

[0066] The calculated ion migration rate (charge) is mapped to the corresponding residual polarization potential. Based on the magnitude of this potential, a "compensation bias voltage" with opposite polarity and equal amplitude is generated in real time. This action aims to establish an external field opposite in direction to the built-in polarization field, causing the two to physically neutralize each other within the liquid crystal layer.

[0067] This step establishes an "electric field adaptive balance mechanism" at the physical level. It not only senses the dynamic migration of ions in situ but also "physically compensates" for the effective electric field loss caused by charge accumulation by actively injecting a compensation field. This fundamentally suppresses the phase drift phenomenon of the liquid crystal metasurface under long-term continuous operation, ensuring that when the final closed-loop adjustment is performed in step S205, the control operator can act on a liquid crystal layer with a physically pure environment and precisely controlled field strength.

[0068] Step S205: Based on the deviation between the actual optical phase and the target phase, and in conjunction with the compensation bias voltage, dynamically adjust the amplitude of the low-frequency drive signal sequence.

[0069] As the endpoint of the control link in this application, this step aims to fuse the "true phase state" obtained in step S203 with the "internal electric field environment patch" obtained in step S204 using multi-dimensional parameters, and drive the liquid crystal metasurface unit to accurately lock the target optical state through closed-loop iteration.

[0070] Precise control of liquid crystal molecules is essentially a "field-state coordination" process. In the physical model, the goal is to make the molecules reach the desired target phase. Not only is it necessary to apply a corresponding nominal driving electric field, but it is also necessary to counteract external disturbances and changes in the internal electrochemical environment in real time. Phase Feedback: The actual optical phase provided in step S203 Phase with target The deviation between them represents "position misalignment" caused by temperature fluctuations, mechanical response hysteresis, or model nonlinearity.

[0071] Field Compensation: The compensation bias voltage provided in step S204 This represents the "field strength premium" that must be paid in order to neutralize the built-in polarization field generated by internal ion migration.

[0072] This step involves algebraically superimposing the parameters from the two dimensions mentioned above to construct a composite driving instruction capable of real-time self-healing.

[0073] In open-loop control or single feedback control, one aspect is often sacrificed: focusing solely on phase feedback (step S203) may ignore the potential shift caused by ion migration (step S204), leading to frequent over-adjustment of the controller in the "spurious electric field" and causing phase oscillations; while focusing solely on ion compensation cannot correct the mechanical response error of molecules. This step, through dual-drive adjustment of "phase deviation correction" and "polarization electric field counterbalancing," ensures that the control command can respond quickly to target changes while maintaining static stability over long periods of operation.

[0074] The key physical characteristics involved in this step and their functions are as follows: Phase deviation value Defined as the algebraic difference between the target command and the actual observed value. It is the error source that triggers the "dynamic catch-up" action.

[0075] Dynamically driven increment This refers to the instantaneous voltage correction calculated to eliminate the aforementioned phase deviation. It represents the degree of intervention of the control algorithm in the molecular motion trend.

[0076] Target driving amplitude The envelope height of the final synthesized low-frequency drive signal sequence that directly affects the next control cycle.

[0077] Based on the integrated feedback operator, the following specific adjustment actions are performed: Will and Real-time comparison is performed to obtain the deviation value. Subsequently, using feedback control operators (such as adjustment logic based on deviation integrals or rate prediction), the deviation in the optical domain is mapped to a "dynamically driven increment" in the electrical domain. This increment guides the molecule to converge toward the target position with an optimal trajectory.

[0078] The parameter accumulator is activated to linearly synthesize the initial reference voltage determined in step S201, the dynamic drive increment calculated in this step, and the compensation bias voltage calculated in step S204. This action physically completes the triple tasks of "command issuance," "error correction," and "field strength counterbalancing" simultaneously.

[0079] Using the synthesized target driving amplitude as a new constraint, the low-frequency driving signal sequence is reconstructed in real time. The envelope of the liquid crystal cell. This update action directly changes the equivalent electric field strength applied to both ends of the liquid crystal cell.

[0080] This step implements a "full-environment adaptive controller" at the physical level. It transforms the liquid crystal molecules from a "passive, forced system" constrained by hysteresis and ion interference into an "active servo system" with self-correcting capabilities: real-time deviation correction eliminates phase drift caused by model inaccuracies and environmental interference; the intervention of the compensating bias voltage solves the "fading" or phase deviation phenomena of the liquid crystal device under continuous power. Regardless of whether the liquid crystal molecules are currently in a boost or buck historical trajectory, this closed-loop logic ensures that they converge to the target phase via the shortest path. This provides physical reliability for high-frequency, high-precision dynamic wavefront control (such as beam deflection and dynamic focusing).

[0081] In an exemplary embodiment, the physical constraints of the "superimposing high-frequency detection signal sequence" action in step S201 are refined, detailing how to achieve the technical goal of "perception without interference control" through adaptive selection in the frequency domain and precise suppression in the amplitude domain.

[0082] Optionally, generating a high-frequency detection signal sequence includes: setting the frequency of the high-frequency detection signal sequence above the mechanical response frequency of the liquid crystal molecules and below the dielectric relaxation frequency of the liquid crystal molecules; using a symmetrical waveform with zero mean as the basic element of the high-frequency detection signal sequence, and the time-domain integral of the basic element within a single signal period is zero.

[0083] Within the dual-frequency driving framework of liquid crystal metasurfaces, high-frequency detection signal sequence Although used for sensing, if its electric field torque is too strong, it will produce the following side effects: If the detection frequency happens to be in the mechanical resonance band of the liquid crystal molecules (determined by molecular viscoelasticity and anchoring force), it will cause the molecules to wobble additionally. If the amplitude of the detection signal is too large, the equivalent electric field torque it generates will be superimposed on the driving field, causing the liquid crystal molecules to produce an unexpected secondary deflection, causing the phase sensing in step S203 to lose its reference.

[0084] This step ensures that the detection signal exists only as a "reactive probe" by establishing the dual criteria of "frequency non-resonance" and "amplitude weak constraint".

[0085] Based on the signal synthesis engine and physical parameter preset module, actions are executed according to the following specific constraint logic: First, the elastic constant and viscosity coefficient of the current liquid crystal material are retrieved, and the forced vibration resonance points of the molecules are calculated. The center frequency of the high-frequency detection signal Configured in a far away The high-frequency flat region (e.g., if the resonant point is around 2kHz, then configure) This action ensures that the electric field alternation rate is much higher than the mechanical response rate of the molecules, so that the molecules can only feel the root-mean-square effect of the high-frequency field and cannot generate synchronous physical oscillations. The non-resonant frequency band is defined as the frequency region that avoids the intrinsic mechanical modes and acoustic branches of liquid crystal molecules (usually between 10kHz and 100kHz).

[0086] Amplitude ratio constraint (1 / 10 principle): requires the peak-to-peak value of the high-frequency probe signal to be within the specified range. Strictly smaller than the real-time amplitude of the low-frequency drive signal One-tenth of it.

[0087] Pointer stability: refers to the angular displacement of the major axis of a liquid crystal molecule under the action of a probe electric field, which is less than 0.1° and is negligible in terms of phase fluctuation in the optical dimension.

[0088] Real-time acquisition of the low-frequency drive signal amplitude planned in step S201 The signal synthesis unit dynamically calculates the maximum allowable amplitude of the high-frequency detection signal based on an amplitude ratio threshold of 1 / 10. Through this "small signal approximation" control, the electric field torque generated by the detection signal is limited to less than 1% of the driving torque (the torque is proportional to the square of the voltage), ensuring that the molecular state regulated in step S205 will not undergo physical drift due to the addition of the detection signal.

[0089] The digital sequence that meets the above frequency and amplitude constraints is injected into the DDS (Direct Digital Frequency Synthesis) module. In the final generated composite driving waveform, the high-frequency carrier is superimposed on the low-frequency envelope in the form of extremely weak ripples, which provides sufficient signal-to-noise ratio for impedance extraction in step S202, while maintaining the purity of the potential energy field inside the liquid crystal layer.

[0090] This embodiment achieves "physical transparency" of the sensing link by refining the parameter constraints in step S201: through weak amplitude constraints, it resolves the paradox of "the measurement itself changing the measurand" in traditional closed-loop control, ensuring that the phase value extracted in step S203 is the pure drive response value. Avoiding the resonant frequency band effectively prevents spurious modulation of the feedback current caused by mechanical jitter, enabling step S202 to obtain a more stable baseband component when performing quadrature demodulation.

[0091] In an exemplary embodiment, the logic of the "generating a low-frequency drive signal sequence" action in step S201 is refined, detailing the calculation process of how to transform static mapping values ​​into dynamic evolution sequences.

[0092] Optionally, generating a low-frequency driving signal sequence includes: determining the reference voltage value required to reach the target phase by searching a preset phase-voltage mapping table based on a preset target phase; monitoring the current real-time voltage of the liquid crystal metasurface unit, and generating a voltage evolution sequence based on the difference between the reference voltage value and the current real-time voltage; and using the generated voltage evolution sequence as the time-varying waveform envelope of the low-frequency driving signal sequence. Specifically, if the reference voltage value is greater than the current real-time voltage, a monotonically increasing voltage evolution sequence is planned and generated; if the reference voltage value is less than the current real-time voltage, a monotonically decreasing voltage evolution sequence is planned and generated.

[0093] In the physical model of liquid crystal superatoms, if a step voltage is directly applied (i.e., the voltage jumps instantaneously from the current value to the target value), the liquid crystal molecules will experience instantaneous phase overshoot or pointer jitter due to rotational inertia and the backflow effect.

[0094] To ensure that the macroscopic deflection process of molecules is continuous and predictable, the change in voltage must be decomposed into a series of discrete gradient steps. By planning a sequence of "monotonically increasing" or "monotonically decreasing" gradients, a clear "energy injection" or "energy release" label can be applied to the system at the driving source.

[0095] Based on the digital signal processing unit, the following specific trajectory planning steps are executed: After receiving the target phase command, the physical endpoint of this adjustment action—the reference voltage value—is locked by searching the pre-stored phase-voltage mapping table. This action determines the final energy level that the driving field strength needs to reach. (Reference voltage value) This refers to achieving the target phase of liquid crystal superatoms under steady-state conditions. The required standard potential.

[0096] Monitor the current real-time voltage through internal registers. And calculate the target residual in real time. Current real-time voltage Defined as the actual driving voltage amplitude applied to both ends of the liquid crystal cell at the end of the previous control cycle. Boost determination: If The current demand is determined to be "energy injection". Pressure reduction determination: If... The current demand is determined to be "energy release".

[0097] Based on the above determination results, the monotonic constraint operator is activated to generate the voltage evolution sequence: Monotonically increasing path: If there is a boost requirement, a path is planned based on the preset voltage slope (Slew Rate) from... Gradually rise to The sequence is a monotonically increasing sequence. This sequence ensures that the Coulomb torque acting on the molecule increases smoothly, avoiding molecular arrangement disorder caused by abrupt changes in field strength.

[0098] Monotonically decreasing path: If there is a pressure reduction requirement, plan a path from... Gradually decrease to The monotonically decreasing sequence allows molecules to smoothly reset using the intrinsic elastic energy of the material, minimizing phase deviations caused by rotational viscous forces.

[0099] Voltage Evolution Sequence It refers to a dynamic wavefront envelope composed of a series of tiny step voltages used to guide molecular motion. The generated... As a modulation envelope, the low-frequency driving signal sequence is updated in real time. The electric field ultimately applied to the metasurface unit exhibits a "controlled evolution" pattern, with its waveform envelope strictly following a planned monotonic trajectory.

[0100] This embodiment achieves deep decoupling between perception and control through refined management of the driving source: since the sequence is forced to be monotonically increasing or decreasing, the sign (directional trend) of its first derivative is uniquely determined throughout the entire evolution cycle. This provides a physical gold standard for accurately invoking the "boost mapping logic" or "buck mapping logic" in step S203. Eliminating the physical overshoot caused by voltage jumps shortens the settling time required for molecules to reach a steady state, resulting in better linearity of the metasurface when performing continuous phase changes (such as frequency sweeps or beam scans).

[0101] In an exemplary embodiment, the specific implementation of step S202 is detailed, showing how to use orthogonal reference signals and delay compensation techniques to achieve "purification" of the signal at the physical layer.

[0102] In the circuit physical model, liquid crystal cells exhibit typical complex impedance characteristics (including the parallel effect of resistance and capacitance).

[0103] In-phase component (active): The current component that is in phase with the external excitation voltage, mainly reflecting the leakage current or energy dissipation of the liquid crystal material.

[0104] Orthogonal component (reactive): Due to the displacement polarization (capacitive effect) of the dipoles of liquid crystal molecules under an electric field, the current phase leads the voltage by 90°.

[0105] To monitor the arrangement of liquid crystal molecules in real time, it is physically necessary to extract only the "orthogonal component" that is 90° ahead, because it is directly proportional to the instantaneous capacitance value of the liquid crystal layer (i.e., the tilt angle of the molecules).

[0106] In actual hardware circuits, signals undergo significant electrical propagation delays after passing through stages such as digital-to-analog converter (DAC), driver amplification, liquid crystal cell, transimpedance amplification, and analog-to-digital converter (ADC). If the original clock generated in the digital domain is directly used as a 90° reference, the phase difference between the feedback signal and the reference signal will no longer be a standard 90° due to the propagation delay. This phase shift causes the resistive component to be mixed with the capacitive component, resulting in the observed "phase" containing noise from fluctuations in the material's conductivity, severely interfering with the accuracy of step S203.

[0107] Optionally, the capacitive reactive current component that has a phase difference with the high-frequency detection signal sequence is separated from the feedback current signal. Specifically, this includes: obtaining the initial phase difference of the feedback current signal relative to the high-frequency detection signal sequence under a preset initial state; synchronously generating an initial orthogonal sequence with the same frequency and a phase difference of 90° in the digital domain based on the synchronous clock that generates the high-frequency detection signal sequence; using the initial phase difference of the system, controllingly shifting the starting phase of the initial orthogonal sequence to obtain an orthogonal reference sequence; performing point-by-point multiplication of the acquired feedback current signal and the orthogonal reference sequence to obtain a mixed digital sequence containing baseband components and harmonic components; and separating the baseband component from the mixed digital sequence as the capacitive reactive current component characterizing the displacement polarization ability of liquid crystal molecules.

[0108] In this embodiment, the initial orthogonal sequence refers to a discrete sine / cosine sequence generated by clock synchronization, which is 90° ahead of the excitation source in phase.

[0109] Orthogonal reference sequence: after The calibrated reference sequence used for the final multiplication operation.

[0110] The digital signal processing unit executes actions according to the following specific algorithm sequence: During system startup or static calibration, the processor obtains the initial phase difference of the system through closed-loop backtesting. The action is as follows: the system compares the phase characteristics (such as the zero-crossing difference) of the original high-frequency sequence generated by the digital terminal with the sampled sequence, quantizes the delay of the physical link into a fixed phase offset angle, and stores it as a calibration parameter in the compensation register.

[0111] Based on the same synchronous clock used to generate the high-frequency probe signal, an initial orthogonal sequence is generated in real time in the digital domain. Then, the sequence obtained in the first step is called... A controlled offset is performed on the starting phase of the sequence. Through this offset operation, the "reference ruler" at the digital end is precisely shifted on the time axis, compensating for the transmission loss of the hardware link, thereby generating an orthogonal reference sequence that can achieve "phase alignment" with the signal returned from the analog end.

[0112] Then, the feedback current signal sequence acquired by the ADC is multiplied point by point with the above-mentioned calibrated orthogonal reference sequence. According to the principle of product-sum, this multiplication operation performs "phase domain projection". It shifts the components in the feedback signal that are in phase and have the same frequency as the reference sequence (i.e., the physical capacitive components) to 0Hz (baseband), while shifting the unrelated components to the second harmonic (2f).

[0113] Finally, a low-pass filter with a specific bandwidth is applied to the multiplied mixed digital sequence. The filter removes all high-frequency fluctuations (harmonic components) and random noise, retaining only the stable baseband component. The DC level amplitude of this component is then output as the capacitive reactive current component.

[0114] This embodiment refines step S202 to achieve hardware independence of sensing accuracy: by compensating for the initial phase difference of the system, the problem of "resistive components being misjudged as capacitive components" caused by hardware delay is solved, ensuring the purity of the sensed quantity. Point-by-point multiplication constructs an extremely narrowband dynamic filter at the algorithm level, enabling the system to stably extract nanoampere-level characteristic currents from strong background noise.

[0115] In one exemplary embodiment, the action of "determining the actual optical phase" in step S203 is further refined, specifically describing how to forcibly establish a single-value mapping between electrical feedback and optical state at the algorithm level by identifying the evolution direction of the excitation source.

[0116] Optionally, the actual optical phase is determined based on the signal change trend of the low-frequency drive signal sequence, the capacitive reactive current component, or a preset buck mapping logic. Specifically, if the signal change trend of the low-frequency drive signal sequence is in an increasing phase, the preset boost mapping logic is invoked to determine the actual optical phase based on the capacitive reactive current component; if the low-frequency drive signal sequence is in a decreasing or flat phase, the preset buck mapping logic is invoked to determine the actual optical phase.

[0117] In an exemplary embodiment, the detailed description of step S203 focuses on defining how to precisely tune different physical calibration models at the algorithm's underlying layer by utilizing the positive and negative polarities of numerical fluctuations.

[0118] Under the hysteresis effect revealed in step S203, the trajectory of liquid crystal molecules is divided into a forced boosting segment and an elastic depressurization segment in physical space. Since the boosting segment is propelled by external electric field energy, while the depressurization segment is primarily pulled back by endogenous elastic energy, these two paths are mathematically non-overlapping. Without explicit triggering conditions, automatic addressing within these two parallel physical paths is impossible. This step establishes a "gradient threshold" to achieve real-time classification of molecular motion modes.

[0119] This embodiment defines the key determination features as follows: Rising State: In the digital control domain, this is defined as the current sampling period. low-frequency drive voltage value Strictly greater than the previous sampling period The value indicates that the molecule is acquiring additional electric field energy and is in a dynamic process of being forced to deflect in the direction of the electric field lines.

[0120] The decreasing or holding state is defined as follows: Less than or equal to The state of "decreasing" and "leveling up" are grouped into the same logical branch because in the liquid crystal physical model, once the electric field stops increasing (leveling up) or begins to decrease (decreasing), the forced flipping inertia of the molecules disappears, and they enter a relaxation or steady-state equilibrium process constrained by viscous forces.

[0121] The following determined logical switching actions are executed through the real-time comparison unit: Action 1, Gradient Polarity Detection (Trend Determination): Perform first-order difference operation on the input low-frequency driving sequence to extract the trend factor. :like The current state is determined to be energy injection mode. If It determines whether the current state is either energy retreat or locked.

[0122] Action 2: Targeted invocation of boost mapping logic: When determined to be... At that instant, the boost mapping logic is forcibly mounted. The action is manifested as: reading the output of step S202. The value is then mapped to a pre-defined "rising edge-specific feature table." This mapping table has a pre-set dynamic compensation coefficient for molecular flip hysteresis during the offline calibration phase, thus directly calculating the actual optical phase under forced conditions. .

[0123] Action 3: Targeted invocation of the buck mapping logic: When determined to be... At that instant, the boost mapping is automatically canceled, and the buck mapping logic is invoked instead. This action manifests as follows: using the same... The input is processed, but a numerical search is performed in the "Falling Edge / Steady State Dedicated Feature Table". This logic branch optimizes the model for the elastic recovery characteristics of molecules, ensuring that the output phase value will not be misjudged due to "historical deflection inertia" when the electric field weakens or remains constant.

[0124] This embodiment, through the concretization of step S203, achieves "hardcore cutting" of the physical state: by incorporating the "flat" state into the step-down branch, it ensures that the algorithm has one and only one corresponding mapping logic at any given time, avoiding logical conflicts or idle operations. It solves the problem that a single curve cannot simultaneously handle "rapid flipping" and "slow reset," enabling the metasurface to output highly consistent phase feedback even under complex and irregular phase transformation commands (such as non-periodic random phase adjustments). By using simple discrete numerical comparisons instead of complex curve fitting, the real-time response frequency of the detection device is significantly improved, laying the foundation for sub-millisecond closed-loop control (step S205).

[0125] In an exemplary embodiment, the algorithm for the "extracting DC bias component" action in step S204 is refined to detail how to achieve physical isolation between the ion sensing channel and the driving channel by setting a specific frequency threshold.

[0126] Optionally, the DC bias component in the feedback current signal is extracted, specifically by using a digital low-pass filtering algorithm to separate the DC bias component characterizing the intensity of ion directional migration from the feedback current signal; wherein the cutoff frequency of the digital low-pass filtering algorithm is lower than the fundamental frequency of the low-frequency driving signal sequence.

[0127] Within the ion sensing framework revealed in step S204, the total feedback current is interwoven with a large-amplitude AC drive component and a weak ion migration current (DC). Low-frequency drive signal sequence. The fundamental frequency typically operates in the tens of hertz range (e.g., 10Hz-100Hz). The ion migration process of liquid crystal molecules is extremely slow, and its characteristic frequency is usually below 1 Hz. In order to prevent the AC swing of the driving signal from being misjudged as DC offset, this step completely decouples the two types of physical processes at the algorithm level by setting a "low-pass threshold" much lower than the driving fundamental frequency.

[0128] In this embodiment, the cutoff frequency : Defined as the frequency point where the amplitude of the transfer function of the digital filter drops to -3dB. In this embodiment, this frequency is forcibly constrained to be . (Typical value) to ).

[0129] Digital low-pass filtering algorithm: This embodiment uses a high-order infinite impulse response (IIR) filter or a Butterworth low-pass filter with an extremely steep transition band to ensure sufficient attenuation of the driving fundamental frequency components.

[0130] Based on the digital signal processing unit, the following specific filtering steps are performed: Real-time acquisition of the fundamental frequency parameters of the low-frequency driving signal sequence currently applied to the metasurface unit. For example, if a 50Hz phase modulation command is currently being executed, then the following is identified: .

[0131] Based on the identified It retrieves the corresponding filter parameters from the preset coefficient library and sets the filter's cutoff frequency. The dynamic configuration is set below this fundamental frequency. This manifests as real-time adjustment of the feedback coefficients of the differential equation, compressing the filter's passband to extremely low frequencies (e.g., ...). ).

[0132] The acquired full-bandwidth feedback current signal is input into the configured filter. The filter is suitable for frequencies higher than [specific frequencies not specified in the original text]. All components (including the 50Hz driving fundamental frequency and its higher harmonics, as well as the kHz-level probe carrier) undergo deep attenuation of up to -60dB or more. Ion-directed migration signals with frequencies close to zero pass through the filter without loss.

[0133] The filter outputs a smoothed, single-valued discrete sequence. This sequence eliminates all AC fluctuations, and its value is directly used as the DC bias component characterizing the current ion migration intensity, output to the subsequent accumulation calculation unit.

[0134] This embodiment achieves "high signal-to-noise ratio stripping" in the ion sensing dimension by refining step S204: by forcing the cutoff frequency to be lower than the driving base frequency, the problem of "DC detection zero drift" caused by driving envelope fluctuations is solved, ensuring the purity of the sensing.

[0135] In an exemplary embodiment, the physical level of the actions of "determining the degree of ion migration" and "generating compensation voltage" in step S204 is refined, which refines the mathematical mapping and execution actions from current to charge and from charge to compensation potential.

[0136] Optionally, the DC bias component is accumulated in the time domain to determine the degree of ion migration inside the liquid crystal, and a compensation bias voltage is generated to counteract the internal residual polarization field based on the degree of ion migration. Specifically, this includes: discretizing and integrating the DC bias component in the time domain to quantify the net accumulated charge generated by the liquid crystal metasurface unit at the alignment layer interface; projecting the net accumulated charge into the internal residual polarization potential based on a preset double-layer capacitance model, and generating a compensation bias voltage with the same amplitude and opposite polarity as the internal residual polarization potential.

[0137] Within the electrochemical sensing framework revealed in step S204, ions migrate and accumulate at the interface under the drive of an electric field. The accumulation of ions at the impermeable alignment layer interface is physically an "increase in net charge." According to the definition of current, Therefore, time-domain accumulation of ion current is the only way to quantify the interfacial charge density. The accumulated ion charge and the electrode-induced charge together constitute a nanoscale "electrically double-layer electrode (EDL)". This structure generates a reverse electrostatic barrier, i.e., a built-in polarization field, inside the liquid crystal layer. This step aims to accurately deduce the amplitude of this barrier by simulating the capacitance characteristics of this physical structure.

[0138] Based on the digital signal processing unit, the following specific physical quantity conversion steps are performed: The DC bias component extracted in step S205 As input, a real-time accumulator is started, multiplying the instantaneous current value of each sampling period by the sampling step size, and performing time-domain accumulation. Discretized time-domain integration refers to the accumulation operation performed on the sampled sequence in a digital processor, i.e. ,in The sampling period is specified. This operation converts the dynamically flowing "current" into a statically accumulated "coulomb quantity," thereby outputting the current net accumulated charge in real time. This action reflects the physical process of ion accumulation at the alignment layer interface. Net accumulated charge. This refers to the total effective impurity ion charge retained at the electrode interface during a specific observation period.

[0139] Double-layer capacitor model It is a preset parameter characterizing the physical structure of a liquid crystal cell, which quantifies the voltage drop generated per unit charge accumulation at the electrode interface. The preset double-layer capacitor model is invoked. According to the formula... Quantified charge The projection is the corresponding internal residual polarization potential. . As a preset constant coefficient, it is determined during factory calibration based on the cell area and the dielectric constant of the alignment layer. This projection action accurately estimates the "deduction amount" of the built-in polarization field on the external effective field strength.

[0140] According to the calculation Amplitude, execute polarity conversion logic. Generate an amplitude and... Strictly equal, but with completely opposite signs, compensation bias voltages This action aims to counteract the reverse potential barrier generated by the interface ion charges by actively injecting an equivalent level from the outside, ensuring that the actual field strength inside the liquid crystal layer returns to the nominal state.

[0141] This embodiment refines step S204 to achieve closed-loop quantization from "induced ions" to "physical neutralization": discrete integration transforms the internal field strength, which cannot be directly measured, into a calculable charge quantity, providing a high-precision quantization method. Utilizing the double-layer capacitance model (claim 6), the electrical parameters ( ) and driving parameters ( Coupling ensures that the generation of the compensation voltage is no longer a blind empirical adjustment, but a precise inverse calculation based on physical evidence; opposite polarities It can physically neutralize the built-in polarization field, so that when step S205 is dynamically adjusted, it can act on a completely "neutral" and controlled electric field environment, thus completely solving the common problems of "DC ghosting" and phase shift in liquid crystal metasurfaces.

[0142] In an exemplary embodiment, the control logic for the "dynamically adjusting the amplitude of the driving signal" action in step S205 is refined, and the fusion and output process of the three independent physical components within the same control cycle is refined.

[0143] Optionally, based on the deviation between the actual optical phase and the target phase, and in conjunction with the compensation bias voltage, the amplitude of the low-frequency drive signal sequence is dynamically adjusted. Specifically, this includes: determining the reference mapping voltage value used to achieve the target phase by retrieving a preset phase-voltage mapping relationship based on the target phase; inputting the deviation between the actual optical phase and the target phase into the feedback control operator to calculate the dynamic drive increment used to correct the instantaneous phase misalignment; algebraically summing the reference mapping voltage value, the dynamic drive increment, and the compensation bias voltage to calculate the target drive amplitude for the next control cycle; and using the target drive amplitude to update the waveform envelope of the low-frequency drive signal sequence in real time to change the equivalent drive electric field intensity applied to the liquid crystal metasurface unit.

[0144] Within the closed-loop framework revealed in step S205, the steady-state controlled electric field of the liquid crystal molecules is not a single variable, but a superposition field determined by the needs of multiple physical dimensions: Static energy (benchmark): The nominal potential energy required to maintain a molecule at the target angle.

[0145] Dynamic kinetic energy (increment): used to overcome rotational viscous resistance, accelerate molecular flipping, or correct instantaneous deviations caused by environmental fluctuations (such as temperature).

[0146] Field strength supplement (compensation): Used to counteract the built-in shielding field generated by the accumulation of interface charges.

[0147] This step achieves a three-in-one control of "prediction + feedback + compensation" by performing algebraic summation of the above three factors in the voltage domain.

[0148] Based on the integrated feedback controller, the action is executed according to the following specific parameter synthesis steps: Target phase receiving external commands First, access the phase-voltage mapping table and determine the reference mapped voltage value at the current operating frequency through linear interpolation or direct retrieval. This component provides a preliminary energy estimate for the system to achieve its target. The phase-voltage mapping (P-VTable) refers to an offline calibrated lookup table that defines the nominal effective voltage value required for a liquid crystal cell to reach a specific phase under ideal conditions.

[0149] The actual optical phase obtained in step S203 and Perform a comparison. The resulting deviation values... Input feedback control operator. The feedback control operator is a regulation algorithm (such as a proportional-integral-derivative PID operator) running on the detection device, responsible for mapping the phase error in the optical domain to a correction step size in the electrical domain. This operator calculates the dynamic driving increment used to correct the instantaneous misalignment of the molecule based on the magnitude and rate of change of the deviation. This increment represents the real-time effort the control algorithm makes to catch up with "hysteresis residue" or "model error".

[0150] Start the parameter fusion unit and perform the following algebraic operations: in, The compensation bias voltage generated in step S204. Target drive amplitude. Defined as the peak envelope height of the low-frequency drive signal sequence in the superimposed signal during the next control cycle.

[0151] This action constructs a "composite potential command" within the digital domain, in which... Anchor the target, Eliminate ion interference. Correct dynamic deviations.

[0152] According to the calculation Adjust the low-frequency drive signal sequence in real time The amplitude scaling factor. The DAC receives the updated sequence and changes the equivalent driving electric field intensity applied to the metasurface unit. Under the action of the recombination electric field, the liquid crystal molecules rapidly converge towards the target director, until... Approaching zero.

[0153] This embodiment achieves "high-dimensional self-healing" of the driving instructions by refining step S205: Decoupling of control precision was achieved: by processing "compensation bias" and "dynamic increment" separately, oscillations caused by the mutual coupling of electrochemical drift and mechanical domain hysteresis in the same control loop were avoided. Because The built-in electric field has been offset, and the feedback operator only needs to handle small dynamic deviations, which improves the static wavefront stability of the detection equipment.

[0154] Based on the same inventive concept, this application also provides an apparatus for designing and detecting intelligent metasurfaces of optical components, used to implement the method for intelligent metasurface design and detection of optical components described above. The solution provided by this apparatus is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the apparatus for intelligent metasurface design and detection of optical components provided below can be found in the limitations of the method for intelligent metasurface design and detection of optical components described above, and will not be repeated here.

[0155] The various modules in the device for intelligent metasurface design and detection of optical components can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0156] In one exemplary embodiment, a detection device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method for intelligent metasurface design and detection of optical components.

[0157] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which determines the method for intelligent metasurface design and detection of optical components when the computer program is executed by a processor.

[0158] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the above-described method for intelligent metasurface design and detection of optical components.

[0159] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile deterministic machine-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The components involved in the embodiments provided in this application may be general-purpose, central processing, graphics, digital signal processing, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI), etc., and are not limited to these.

[0160] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0161] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for designing and detecting intelligent metasurfaces of optical components, characterized in that, The optical metasurface unit includes: a first electrode layer, a second electrode layer, a subwavelength metaatomic array, and a nematic liquid crystal layer; wherein the nematic liquid crystal layer and the subwavelength metaatomic array are located together within an electric field-driven space formed between the first electrode layer and the second electrode layer, and the method includes: A superimposed signal containing a high-frequency detection signal sequence and a low-frequency driving signal sequence is generated, and the superimposed signal is loaded onto a liquid crystal metasurface unit; wherein, the amplitude of the high-frequency detection signal sequence is lower than the activation threshold of the liquid crystal molecules, and the frequency is higher than the mechanical response frequency of the liquid crystal molecules; the low-frequency driving signal sequence is determined based on the target phase; The feedback current signal generated by the liquid crystal metasurface unit is acquired in real time, and the capacitive reactive current component that has a phase difference with the high-frequency detection signal sequence is separated from the feedback current signal. Based on the signal variation trend of the low-frequency drive signal sequence, and based on the capacitive reactive current component or the preset step-down mapping logic, the actual optical phase is determined. The DC bias component in the feedback current signal is extracted, and the DC bias component is accumulated in the time domain to determine the degree of ion migration inside the liquid crystal. A compensation bias voltage is generated based on the degree of ion migration to counteract the internal residual polarization field. Based on the deviation between the actual optical phase and the target phase, and in conjunction with the compensation bias voltage, the amplitude of the low-frequency drive signal sequence is dynamically adjusted.

2. The method according to claim 1, characterized in that, The step of determining the actual optical phase based on the signal variation trend of the low-frequency drive signal sequence and based on the capacitive reactive current component or a preset step-down mapping logic specifically includes: If the signal change trend of the low-frequency drive signal sequence is in the numerical increasing stage, then the preset boost mapping logic is invoked to determine the actual optical phase based on the capacitive reactive current component. If the low-frequency drive signal sequence is in a decreasing or flat phase, the preset buck mapping logic is invoked to determine the actual optical phase.

3. The method according to claim 1, characterized in that, The step of separating the capacitive reactive current component from the feedback current signal that has a phase difference with the high-frequency detection signal sequence specifically includes: Based on the synchronous clock that generates the high-frequency detection signal sequence, an orthogonal reference sequence with the same frequency and a phase difference of 90° is synchronously generated in the digital domain. The collected feedback current signal is multiplied point by point with the orthogonal reference sequence to obtain a mixed digital sequence containing baseband components and frequency harmonic components. The baseband component is separated from the mixed digital sequence and used as a capacitive reactive current component characterizing the displacement polarization ability of liquid crystal molecules.

4. The method according to claim 3, characterized in that, The synchronous clock, based on the high-frequency detection signal sequence, synchronously generates an orthogonal reference sequence with the same frequency and a 90° phase difference in the digital domain, specifically including: Based on the synchronous clock that generates the high-frequency detection signal sequence, an initial orthogonal sequence with the same frequency and a phase difference of 90° is synchronously generated in the digital domain. Obtain the initial phase difference between the feedback current signal and the high-frequency detection signal sequence under a preset initial state; Using the initial phase difference, the starting phase of the initial orthogonal sequence is shifted in a controlled manner to obtain an orthogonal reference sequence.

5. The method according to claim 1, characterized in that, The extraction of the DC bias component from the feedback current signal specifically includes: A digital low-pass filtering algorithm is used to separate the DC bias component characterizing the intensity of ion directional migration from the feedback current signal; wherein the cutoff frequency of the digital low-pass filtering algorithm is lower than the fundamental frequency of the low-frequency drive signal sequence.

6. The method according to claim 1, characterized in that, The step of performing time-domain accumulation calculation on the DC bias component to determine the degree of ion migration inside the liquid crystal, and generating a compensation bias voltage to counteract the internal residual polarization field based on the degree of ion migration, specifically includes: Discretize and integrate the DC bias component in the time domain to quantify the net accumulated charge generated by the liquid crystal metasurface unit at the alignment layer interface. Based on a preset double-layer capacitance model, the net accumulated charge is projected as an internal residual polarization potential, and a compensation bias voltage with the same amplitude but opposite polarity as the internal residual polarization potential is generated.

7. The method according to claim 1, characterized in that, The step of dynamically adjusting the amplitude of the low-frequency drive signal sequence based on the deviation between the actual optical phase and the target phase, combined with the compensation bias voltage, specifically includes: Based on the target phase, a reference mapping voltage value for achieving the target phase is determined by retrieving a preset phase-voltage mapping relationship; The deviation between the actual optical phase and the target phase is input into the feedback control operator to calculate the dynamic drive increment used to correct the instantaneous phase misalignment; The target driving amplitude for the next control cycle is calculated by algebraically summing the reference mapping voltage value, the dynamic driving increment, and the compensation bias voltage. The waveform envelope of the low-frequency driving signal sequence is updated in real time using the target driving amplitude to change the equivalent driving electric field intensity applied to the liquid crystal metasurface unit.

8. The method according to claim 1, characterized in that, Generating the low-frequency drive signal sequence includes: Based on the preset target phase, the reference voltage value required to reach the target phase is determined by searching a preset phase-voltage mapping table; Monitor the current real-time voltage of the liquid crystal metasurface unit, and generate a voltage evolution sequence based on the difference between the reference voltage value and the current real-time voltage; The generated voltage evolution sequence is used as the time-varying waveform envelope of the low-frequency drive signal sequence.

9. The method according to claim 8, characterized in that, The step of generating a voltage evolution sequence based on the difference between the reference voltage value and the current real-time voltage specifically includes: If the reference voltage value is greater than the current real-time voltage, then a monotonically increasing voltage evolution sequence is planned and generated. If the reference voltage value is less than the current real-time voltage, a monotonically decreasing voltage evolution sequence is planned and generated.

10. The method according to claim 1, characterized in that, Generating the high-frequency detection signal sequence includes: The frequency of the high-frequency detection signal sequence is set above the mechanical response frequency of the liquid crystal molecules and below the dielectric relaxation frequency of the liquid crystal molecules. A symmetrical waveform with a mean of zero is used as the basic element of the high-frequency detection signal sequence, and the time-domain integral of the basic element within a single signal period is zero.

Images