High bandwidth flash chip esd protection circuit optimization method and system

Abstract

This invention relates to the field of electrostatic discharge (ESD) protection for chips, and more particularly to an optimization method and system for ESD protection circuits of high-bandwidth flash memory chips. The method includes the following steps: performing multi-point synchronous observation using a charge imaging module to extract multi-step ESD observation information; identifying high-density charge regions based on the multi-step ESD observation information; defining a safe charge threshold, assessing the charge safety of the high-density charge regions, and entering a dynamic control mode; based on the dynamic control mode, dynamically planning the charge discharge path and calculating the real-time residual charge density; detecting that the real-time residual charge density is lower than the safe charge threshold, and completing the local discharge circuit protection operation. This invention provides rapid response and efficient handling of ESD risk events exceeding the safe threshold, improving chip circuit safety and chip lifespan.

Description

Technical Field

[0001] This invention relates to the field of electrostatic discharge (ESD) protection for chips, and more particularly to an optimization method and system for ESD protection circuits of high-bandwidth flash memory chips. Background Technology

[0002] During the long-term operation of high-bandwidth flash memory chips, electrostatic discharge (ESD) can damage or degrade their performance due to high-speed data transmission, high-frequency operation, and complex electrical characteristics. The internal circuitry of flash memory chips is susceptible to external or internal ESD during high-frequency, high-speed data exchange, leading to malfunctions in storage and control units. The risks posed by ESD are particularly significant in high-bandwidth operating environments; if ESD is not effectively suppressed, it can result in data loss, corruption, or system crashes. Therefore, ensuring the anti-static capabilities of high-bandwidth flash memory chips in high-voltage and high-speed data transmission environments has become a pressing technical challenge.

[0003] Traditional ESD protection circuit designs primarily rely on simple protective devices such as TVS diodes, capacitors, and resistors. While these measures can address a certain level of electrostatic discharge (ESD) threat, their limitations are becoming increasingly apparent as flash memory chip operating frequencies and data transfer rates continue to rise. First, traditional solutions often rely on single ESD suppression devices, failing to provide comprehensive optimization for complex, high-bandwidth operating environments, resulting in limited protection effectiveness. Second, existing ESD protection circuits often fail to consider electromagnetic interference and electrical noise in high-speed signal transmission. These factors can easily cause protection circuit failure under high-frequency operation and may even generate additional circuit losses, affecting the overall performance of the chip. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention proposes a method and system for optimizing ESD protection circuits in high-bandwidth flash memory chips, thereby resolving at least one of the aforementioned technical issues.

[0005] To achieve the above objectives, the present invention provides an optimization method for ESD protection circuits of high-bandwidth flash memory chips, comprising the following steps: Multi-point synchronous observation is performed using a charge imaging module to extract multi-step ESD observation information. High-density charge regions are identified based on the aforementioned ESD multi-step observation information; Define a safe charge threshold, conduct charge safety assessment of high-density charge regions, and enter a dynamic control mode; Based on the dynamic control mode, dynamic planning of charge discharge path is performed, and real-time residual charge density is calculated; if the real-time residual charge density is lower than the safe charge threshold, the local discharge circuit protection operation is completed.

[0006] This specification provides a high-bandwidth flash memory chip ESD protection circuit optimization system for performing the high-bandwidth flash memory chip ESD protection circuit optimization method described above, including: The synchronous observation module is used to perform multi-point synchronous observations through the charge imaging module and extract multi-step observation information of ESD. The charge region identification module is used to identify high-density charge regions based on the ESD multi-step observation information; The safety assessment module is used to define the safe charge threshold, perform charge safety assessment on high-density charge regions, and enter dynamic control mode. The discharge circuit protection module is used to dynamically plan the charge discharge path based on the dynamic control mode and calculate the real-time residual charge density; detect that the real-time residual charge density is lower than the safe charge threshold, and complete the local discharge circuit protection operation.

[0007] The beneficial effects of this invention are specifically as follows: Global and local synchronous monitoring: Multi-point synchronous observation avoids the risk of missed detection that may occur with single-point monitoring, improving the ability to capture transient ESD events. Multi-step observation information can capture the dynamic process of charge accumulation and discharge, providing an accurate data foundation for subsequent analysis. By detecting charge accumulation trends in advance, control measures can be taken before the charge becomes too high, reducing damage to flash memory cells. Protective measures are taken only for high-risk areas, avoiding global overprotection and improving ESD protection efficiency. By identifying high-density areas, charge discharge paths can be designed specifically, reducing unnecessary energy consumption. Local high-density charge is a major source of ESD damage; early identification can significantly reduce the probability of cell damage. A safety threshold serves as a quantitative standard, making ESD protection actions intelligent and automated, requiring no manual intervention. The dynamic control mode can adjust the discharge strategy in real time according to actual charge density changes, adapting to the rapid charge accumulation characteristics of high-bandwidth chips. Intervention is only applied to areas exceeding the threshold, avoiding unnecessary operations on safe areas and improving chip performance. Dynamic planning ensures that charge is released quickly and safely along the optimal path, reducing the risk of local overload. By calculating the residual charge density, the adequacy of the discharge can be assessed in real time, enabling closed-loop control. This avoids the cumulative damage to flash memory cells caused by long-term high-density charge accumulation, improving overall reliability. Activating the discharge circuit only when necessary protects the chip and avoids excessive energy waste. Partial discharge avoids full-chip discharge, significantly reducing dynamic power consumption. This enables high-bandwidth chips to respond quickly and efficiently to transient ESD events, ensuring chip stability even during high-speed operation. Attached Figure Description

[0008] Figure 1 This is a flowchart illustrating the steps of an optimization method for a high-bandwidth flash memory chip ESD protection circuit according to the present invention. Figure 2 This is a detailed flowchart illustrating the implementation steps of step S1. Figure 3 This is a flowchart illustrating the detailed implementation steps of step S2. Detailed Implementation

[0009] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0010] This application provides a method and system for optimizing the ESD protection circuit of a high-bandwidth flash memory chip. The execution entities of the method and system include, but are not limited to, mechanical equipment, data processing platforms, cloud server nodes, and network upload devices that can be considered as general computing nodes in this application. The data processing platform includes, but is not limited to, at least one of an audio / image management system, an information management system, and a cloud-based data management system.

[0011] Please see Figures 1 to 3 This invention provides an optimization method for ESD protection circuits of high-bandwidth flash memory chips, comprising the following steps: Multi-point synchronous observation is performed using a charge imaging module to extract multi-step ESD observation information. High-density charge regions are identified based on the aforementioned ESD multi-step observation information; Define a safe charge threshold, conduct charge safety assessment of high-density charge regions, and enter a dynamic control mode; Based on the dynamic control mode, dynamic planning of charge discharge path is performed, and real-time residual charge density is calculated; if the real-time residual charge density is lower than the safe charge threshold, the local discharge circuit protection operation is completed.

[0012] In the embodiments of the present invention, see Figure 1 This is a flowchart illustrating the steps of an optimization method for a high-bandwidth flash memory chip ESD protection circuit according to the present invention. In this example, the steps of the optimization method for the high-bandwidth flash memory chip ESD protection circuit include: Multi-point synchronous observation is performed using a charge imaging module to extract multi-step ESD observation information. In this embodiment, to accurately capture the accumulation, migration, and discharge behavior of charge during electrostatic discharge (ESD) in a high-bandwidth flash memory chip, a high-resolution charge imaging module is required to simultaneously detect multiple observation points on the chip surface and internally. The charge imaging module typically consists of a nanoscale charge detection array and an ultra-high-speed imaging control unit, with an observation frame rate of up to 10 frames per second. 9 This allows for recording charge behavior on a nanosecond timescale, creating frames.

[0013] During the observation process, the chip was placed in a controlled ESD impact environment, and external electrostatic interference was simulated by gradually applying electrostatic pulses (e.g., within the range of ±2kV to ±8kV). The charge imaging module performed multi-point synchronous sampling on key areas of the chip, and each observation point output charge density, potential distribution, and transient change information in real time.

[0014] All observation data are aligned and integrated using time-series labels to form a multi-step ESD observation dataset, covering the entire process from charge accumulation and dissipation to recovery. The core objective of this step is to capture the real dynamic behavior of charges inside high-bandwidth flash memory chips, providing accurate foundational data for subsequent high-density charge region identification and security threshold analysis.

[0015] High-density charge regions are identified based on the aforementioned ESD multi-step observation information; In this embodiment, after obtaining complete ESD multi-step observation information, spatial analysis of the charge distribution is required to identify high-density charge accumulation regions. First, the charge intensity data at each observation point is normalized to convert the charge density values ​​at different time frames into a unified comparison scale.

[0016] Subsequently, spatial clustering analysis was used to group adjacent regions with charge densities exceeding twice the global average into the same high-density charge region. Through multi-frame data overlay analysis, the migration trajectory and aggregation trend of charges over time could be observed.

[0017] If a region maintains a high charge density (e.g., consistently exceeding 150% of the average density) across multiple consecutive time frames, it is designated as a "stable high-density region." These regions often correspond to dielectric layer interfaces, electrode nodes, or metal interconnect intersections, and are potential ESD energy concentration risk points.

[0018] This identification step can clearly identify the location and dynamic characteristics of charge accumulation on the chip under ESD impact, providing directional guidance for subsequent charge safety assessment and discharge path planning.

[0019] Define a safe charge threshold, conduct charge safety assessment of high-density charge regions, and enter a dynamic control mode; In this embodiment, after identifying the high-density charge region, a safe charge threshold needs to be defined based on the chip material characteristics and structural parameters. This threshold is determined based on parameters such as dielectric layer breakdown strength, metal interconnect withstand voltage limit, and package thermal conductivity, and typically corresponds to the maximum allowable charge density that the chip can maintain stable conduction under ESD stress.

[0020] By comparing with measured charge distribution data, the maximum peak charge and average charge intensity of each high-density region are calculated. When any parameter exceeds the safety threshold, it indicates a potential breakdown risk in that region, and dynamic control mode should be activated immediately.

[0021] In dynamic control mode, the control module will monitor charge migration behavior in real time, evaluate the distribution of discharge energy inside the chip, and provide feedback benchmark for dynamic planning of subsequent discharge paths.

[0022] Based on the dynamic control mode, dynamic planning of charge discharge path is performed, and real-time residual charge density is calculated; if the real-time residual charge density is lower than the safe charge threshold, the local discharge circuit protection operation is completed.

[0023] In this embodiment, after the dynamic control mode is activated, the shortest charge conduction path to the main discharge terminal is first calculated based on the spatial location of the high-density charge region. This path planning considers the conductive layer thickness, resistance gradient, and the uniformity of current distribution between paths to ensure efficient charge discharge and heat dissipation. During the discharge process, the charge migration rate and local residual charge density are monitored in real time. By using time series differential calculation, it can be determined whether the discharge is complete. When the residual charge density is detected to be below a safety threshold (usually less than 10% of the initial peak value), the local discharge task is considered complete. In addition, the thermal response characteristics of the discharge path are recorded simultaneously for subsequent thermal anomaly detection and secondary optimization. This step not only achieves safe charge export but also avoids thermal runaway caused by energy accumulation through a real-time feedback mechanism, ensuring the stability and reliability of the high-bandwidth flash memory chip under ESD impact.

[0024] In this embodiment, see Figure 2 The specific steps for extracting multi-step ESD observation information through multi-point synchronous observation using the charge imaging module are as follows: The charge accumulation and emission paths during the ESD impact process are observed simultaneously at multiple points using the charge imaging module, and multi-step observation information of ESD is extracted. Determine the time slice length, perform dynamic data slicing on ESD multi-step observation information, and obtain multi-time series information; The local charge density gradient and transient electric field intensity are calculated based on multi-time-series information, and the intensity distribution at multiple locations is identified to obtain a charge energy distribution map. Multidimensional coupling feature clustering is performed based on the charge energy distribution map to obtain the charge coupling characteristic distribution of the dielectric layer, metal interconnect layer and node contact region; Real-time ESD characteristic events are generated based on the charge coupling characteristic distribution.

[0025] In this embodiment, during the optimization of the ESD protection circuit of the high-bandwidth flash memory chip, in order to deeply study the dynamic laws of charge accumulation and dissipation during electrostatic discharge events, it is necessary to achieve visual observation of the entire ESD impact process through a charge imaging module. Specifically, a charge imaging device with nanosecond-level temporal resolution is used, and multiple synchronous observation points are deployed on the chip surface, dielectric layer, and metal interconnect layer. These observation points can capture the migration trajectory of charges and changes in local energy distribution during ESD impact.

[0026] During the observation process, the charge imaging module records the charge transfer path at the instant of discharge in a frame sequence, forming multi-step observation data with a time resolution within 1 ns. To ensure synchronization, signal sampling at all observation points is coordinated through a unified trigger clock, thereby ensuring that charge responses at different locations can be compared within the same temporal framework. By spatiotemporally aligning the multi-point observation results, key process information such as charge accumulation, dissipation, and redistribution at each stage of ESD impact can be extracted, forming an ESD multi-step observation information dataset, providing basic input for subsequent time-series slicing and energy distribution analysis.

[0027] After completing multi-point synchronous observations, in order to capture the temporal evolution characteristics of charge behavior during ESD shocks, the observation data needs to be dynamically sliced ​​along the time dimension. First, the time slice length is determined based on the duration of the ESD event and the signal change rate. Typically, when the charge migration rate reaches its peak, the slice interval can be set to 1 to 5 nanoseconds to ensure sufficient time resolution.

[0028] Subsequently, the continuous charge observation sequence was divided into time slices, with each time segment corresponding to a charge distribution state. Each slice contains data such as charge density, potential value, and electric field strength changes at each observation point. To prevent noise amplification due to excessively short time slices, a weighted smoothing algorithm was used to smooth the slice boundaries, thus maintaining data continuity.

[0029] Through this process, the entire ESD shock is decomposed into multiple dynamic time segments, each representing an independent charge response stage. This multi-time-segment information not only reveals the temporal relationship between charge accumulation and dissipation but also exposes the sequence of responses and spatial propagation characteristics in different regions during the discharge process, laying the foundation for subsequent electric field and energy distribution analysis.

[0030] After acquiring information from multiple time segments, it is necessary to quantitatively analyze the changes in charge distribution to reveal the energy transfer patterns during ESD impacts. Specifically, this involves differential calculation of the charge density changes at different observation points within each time segment to obtain the local charge density gradient. The magnitude of the charge density gradient reflects the migration trend and accumulation rate of charge in space.

[0031] Next, the transient electric field intensity is calculated based on the charge density change rate, and the direction and amplitude of the electric field are determined by vector superposition. Electric field data from multiple observation points are integrated using a spatial interpolation algorithm to form a three-dimensional distribution matrix of the electric field intensity. Subsequently, high-intensity regions, low-intensity regions, and gradient abrupt change regions of the electric field are labeled through multi-location intensity distribution identification.

[0032] By fusing the electric field intensity distribution with the charge energy density, a charge energy distribution map can be generated. The map uses color or brightness to reflect the degree of energy accumulation, clearly demonstrating the entire process of charge accumulation and dissipation during an ESD event. By overlaying multiple frames of the map, the energy migration path and local overheating risk can be visually observed, providing data support for optimizing the layout of ESD protection circuits.

[0033] After obtaining the charge energy distribution map, it is necessary to analyze the coupling behavior of charges between different structural layers. To this end, a multidimensional coupling feature clustering method is employed to perform cluster analysis on multidimensional features such as energy, field strength, and density gradient in the map. First, the feature data of different structural layers are normalized to make the data of the dielectric layer, metal interconnect layer, and node contact area comparable on the same feature scale.

[0034] Subsequently, a spatial similarity clustering algorithm was used to divide the charge coupling regions. When multiple locations exhibited energy peak correlation or field strength resonance characteristics within the same time period, they were grouped into the same coupling cluster. The clustering results revealed energy transfer channels between the dielectric and metal layers, as well as charge aggregation phenomena occurring in the node contact region.

[0035] The final charge coupling characteristic distribution reveals the interaction modes of different structural layers during ESD impact. This result can reveal which regions are prone to electric field accumulation or breakdown risk, providing a scientific basis for material selection, contact structure adjustment, and potential shielding optimization in subsequent ESD protection design.

[0036] After obtaining the charge coupling characteristic distribution, this static spatial data needs to be transformed into real-time monitorable characteristic events to guide the chip's ESD protection response mechanism. To this end, the ESD characteristic event triggering conditions are defined by dynamically detecting the charge evolution rate and energy accumulation threshold of each coupling region. When the charge density change rate or transient electric field intensity of a certain region exceeds a preset threshold, it is determined to be an ESD characteristic event.

[0037] This process generates event description information in real time, including the event location, duration, peak electric field, and charge release rate. Multiple feature events are sorted chronologically to form an event sequence, reflecting the energy release trajectory of the entire ESD impact. To improve recognition accuracy, an adaptive judgment mechanism is introduced, automatically adjusting the threshold according to the chip's operating status to prevent false positives.

[0038] The final generated real-time ESD characteristic events can serve as the basis for subsequent protection circuit design and response control, enabling the HBF chip to sense changes in energy distribution in real time before and after an ESD impact, thereby optimizing the protection layout at the physical level and improving the overall ESD immunity performance and structural stability.

[0039] In this embodiment, the specific steps for identifying high-density charge regions based on the ESD multi-step observation information are as follows: Based on the real-time ESD feature events, amplitude normalization and timing synchronization processing are performed to obtain standardized ESD feature information. Based on standardized ESD feature information, the charge gradient and potential energy difference between adjacent nodes are calculated to analyze the charge conduction path and identify transient high charge density paths. The internal spatial distribution of transient high charge density paths is identified to obtain a charge conduction path distribution map. Based on the charge conduction path distribution map, the path intersection density and energy accumulation distribution are analyzed to identify high-density charge accumulation regions. Spatial boundary morphological filtering and equipotential line fitting are performed on the high-density charge accumulation region to extract the high-density charge region.

[0040] In this embodiment, the original characteristic values ​​such as electric field intensity and charge density change rate at each observation point are normalized. By introducing a scaling factor, the amplitudes of each signal are mapped to a uniform scale range, making the response data from different observation points comparable. To preserve the true dynamic characteristics of the signal, the waveform morphology is retained during normalization while only the amplitude range is adjusted.

[0041] After amplitude standardization, the time axis of each feature event needs to be synchronized and corrected. Due to the microsecond-level delay in the response of different regions during an ESD impact, timestamp alignment and dynamic time-window resampling techniques are used to map multi-point event signals onto the same time base to ensure the accuracy of subsequent path analysis. After synchronization, the charge change information of all nodes is arranged at equal time intervals, forming a standardized ESD feature information matrix. This matrix can simultaneously reflect the evolution trend and spatial distribution pattern of charge in the time dimension, providing a stable and balanced input data foundation for subsequent charge gradient calculation and path analysis.

[0042] Based on the charge density changes at each observation node, the charge gradient between adjacent nodes is calculated. The magnitude of the charge gradient reflects the direction and concentration of charge transfer in space, while the potential energy difference reveals the electric field strength driving charge migration. By calculating the gradient and potential difference between each pair of nodes, the flow trend of charge between different layers can be preliminarily outlined.

[0043] Next, using graph correlation analysis, each node is treated as a vertex, and the charge gradient is treated as an edge weight to construct a charge conduction network model. By minimizing the path weights, the paths where charge is preferentially generated and diffused during an ESD impact can be identified, i.e., transient high charge density paths. These paths are often concentrated in interconnect channels, dielectric layer defect regions, or metal trace intersections where potential differences change drastically.

[0044] During path resolution, to avoid misjudging pseudo-gradients caused by noise, a sliding window smoothing and local weighted filtering are used to suppress outliers. The resulting transient high charge density path reveals the true conduction behavior of charge within nanosecond timeframes, providing a crucial reference for optimizing the conductive path distribution in ESD protection structures.

[0045] The charge conduction path data is registered with the chip's 3D structure modeling results to ensure that the path point set corresponds one-to-one with the geometric positions of the metal layers, dielectric layers, and interconnect nodes in the chip.

[0046] Subsequently, connectivity analysis of the conduction paths is performed using a 3D spatial reconstruction algorithm to form a complete charge conduction network. This network can clearly display the direction of charge flow, path density, and location of energy accumulation points in 3D space. In this process, spatial voxel partitioning technology is introduced to divide the chip structure into nanoscale grid units. Each unit records the frequency of charge flow and energy distribution, thereby generating a charge conduction path distribution map.

[0047] The distribution diagram not only shows the migration trajectory of charges in space, but also reveals the energy transfer efficiency of different paths. By comparing it with the actual circuit layout, it is possible to intuitively determine whether charges accumulate at sensitive nodes or cross critical insulation areas, thus providing a reliable basis for optimizing the spatial layout and adjusting the overcurrent shunt path in ESD protection design.

[0048] The spatial overlap of each path in the path distribution map is statistically analyzed, and the number of path intersections per unit volume is calculated, i.e., the path intersection density. A higher intersection density indicates stronger charge accumulation and local current concentration in that region.

[0049] Simultaneously, the charge energy in the intersection region is accumulated and calculated to assess the energy accumulation intensity. A spatial integration method is used to superimpose the energy distribution of multiple paths within the same region, resulting in an energy accumulation distribution map. By comparing the intersection density and energy accumulation intensity in different regions, high-density charge accumulation areas can be clearly identified.

[0050] In practice, to reduce the impact of noise on the results, spatial Gaussian filtering and thresholding were applied to the path overlap data. The final output information on the distribution of aggregation areas can intuitively reveal the risk points of charge concentration, especially in the transition areas of metal interconnect layers, dielectric edges, and near vias. This analysis result is an important foundation for subsequent morphological extraction and optimization of ESD protection structures.

[0051] Further precise extraction of its spatial boundaries and morphology is then performed. First, three-dimensional morphological filtering is applied to the spatial distribution data of the clustered areas to smooth local edges and eliminate isolated noise points. Commonly used methods include the combined application of dilation and erosion operations, which can remove abnormal protrusions while maintaining the structural integrity of the main region.

[0052] Next, energy equipotential surface information is extracted based on the filtered data. By fitting the electric potential energy distribution to equipotential surfaces, a continuous equipotential line model can be generated, reflecting the potential energy boundary of the charge in space. This process utilizes three-dimensional surface reconstruction technology to clearly separate the boundaries of different potential energy levels in space.

[0053] The resulting high-density charge region not only contains location and morphology information but also reflects the intensity and directionality of charge accumulation. This region data can be directly used for subsequent ESD protection circuit layout optimization, such as deploying bypass channels in high-risk areas or adjusting the ground potential distribution, thereby effectively reducing the risk of local breakdown and improving the overall ESD resistance performance of the HBF chip.

[0054] In this embodiment, the specific steps for defining a safe charge threshold, assessing the charge safety of high-density charge regions, and entering the dynamic control mode are as follows: Identify basic chip information, including chip material parameters, dielectric breakdown strength, and metal interconnect withstand voltage limit; Based on the chip's basic information, multi-layer charge carrying capacity is predicted to obtain the maximum allowable charge density; The safe charge threshold is defined based on the maximum permissible charge density; Calculate the maximum peak charge and average charge intensity of the high-density charge region; The maximum charge peak value and average charge intensity are assessed for charge safety based on the safe charge threshold. When the maximum charge peak value or average charge intensity is not less than the safe charge threshold, the local discharge path is activated and the system enters the dynamic control mode.

[0055] In this embodiment, key material parameters, including dielectric layer thickness, relative permittivity, metal interconnect layer conductivity, and gate oxide layer material properties, are extracted from chip manufacturing archives and testing platforms. These parameters directly determine the physical boundaries of charge accumulation and discharge behavior. Subsequently, the breakdown strength of the chip dielectric layer is identified by measuring the breakdown threshold of the dielectric under different temperature and stress conditions through nanosecond-level pulse discharge experiments to determine its maximum withstandable electric field strength range. Simultaneously, the withstand voltage limit of the metal interconnect layer is obtained through step voltage loading experiments, recording its resistance abrupt change points under gradually increasing voltage to determine the critical voltage level of the interconnect under high current surges. By identifying and recording these parameters, a dataset of the chip's physical load-bearing characteristics can be formed, providing basic input conditions for subsequent charge-bearing prediction and safety threshold definition. The accuracy of this stage determines the reliability and engineering feasibility of the entire ESD safety control strategy. After obtaining the basic chip information, it is necessary to predict the charge-bearing capacity of different layers based on its material and structural characteristics to calculate the maximum allowable charge density. This process involves comprehensive modeling at three levels: the dielectric layer, the metal interconnect layer, and the contact interface. At the dielectric layer level, based on its thickness and dielectric constant, combined with breakdown strength parameters, the electric field carrying capacity limit of the dielectric under conditions without insulation failure is calculated. Secondly, in the metal interconnect layer, the upper limit of the current density it can withstand is estimated by analyzing the correspondence between conductor cross-sectional area and withstand voltage limit. For the contact interface region, the interface barrier effect and local charge trapping characteristics also need to be considered. Combining these three factors, a multi-layer charge superposition prediction model is used to derive the charge carrying capacity of the overall structure under short-term high-energy ESD events. To ensure prediction accuracy, experimental calibration parameters are introduced, such as material response curves at three temperature conditions (25°C, 65°C, and 95°C), to correct the theoretical estimation results. The final maximum allowable charge density represents the ultimate charge carrying threshold of the chip under safe operating conditions, providing a core basis for defining the safe charge threshold.

[0056] After determining the maximum allowable charge density, a safe charge threshold needs to be defined based on the engineering safety margin principle. This threshold is used to determine whether there is a potential risk of breakdown or degradation in the actual charge distribution. In practice, a safety reduction factor is first set based on the theoretical value of the maximum allowable charge density, combined with the energy accumulation rate and dissipation capability under different operating modes. This factor is generally between 0.6 and 0.8 to ensure sufficient safety margin is maintained even under conditions of temperature fluctuations, manufacturing deviations, or material aging. To adapt to transient current changes in high-bandwidth flash memory chips under different operating scenarios, the safe charge threshold needs to be dynamically adjusted. For example, during high-speed writing or data erasure, due to the high interconnect current density, the threshold can be lowered accordingly to enhance protection sensitivity. The final safe charge threshold is not only a quantitative standard for judging the risk of charge distribution but also the core basis for subsequent dynamic discharge control logic. When the charge density or electric field strength of a certain area approaches this threshold during subsequent monitoring, the protection mechanism will be triggered to prevent local overload and material breakdown.

[0057] After setting the safe charge threshold, quantitative analysis of the previously identified high-density charge regions is required to assess their true energy levels. This process is achieved through spatial sampling and statistical calculation of charge distribution data. First, the coordinate range of the high-density region is extracted from the three-dimensional charge distribution matrix, and the charge density of each voxel within this region is statistically analyzed to obtain the maximum charge peak value, i.e., the point with the highest charge concentration in the region. The mean charge density of all voxels within the region is calculated to obtain the average charge intensity, which reflects the overall concentration of charge energy. To avoid outliers affecting the results, median filtering and quantile truncation are used to remove extreme noise. Through time-dimensional analysis, the trend of charge peak value changes in the high-density region at different sampling times is tracked to determine whether there is continuous charge accumulation or rapid release. The results of this step directly reflect the dynamic characteristics of the regional charge, providing key input parameters for the final charge safety assessment.

[0058] After obtaining quantitative charge data for the high-density area, the charge safety assessment phase begins. By comparing the maximum charge peak value and average charge intensity with safe charge thresholds, it can be determined whether there is a potential ESD risk in the current area. When either indicator exceeds the threshold, it means that the electric field intensity in the local area is approaching the critical level for dielectric breakdown. At this point, it is necessary to immediately activate the local discharge path to divert the excess charge.

[0059] In dynamic control mode, by controlling the variable resistor structure or bypass diode trigger path in the conductive channel, some charge is rapidly transferred to the low potential region, thereby reducing the local electric field peak. To prevent secondary charge accumulation, a time-delayed release strategy is also introduced, that is, the discharge channel is kept in the conducting state briefly after discharge to ensure that the residual charge is completely dissipated.

[0060] This stage is not only the core of charge control logic execution, but also the key to adaptive optimization of ESD protection performance. By responding in real time to situations where the charge threshold exceeds the limit, transient discharge impacts can be effectively suppressed without affecting the normal function of the chip, significantly improving the overall ESD resistance stability and long-term reliability of the HBF chip.

[0061] In this embodiment, the specific steps for dynamically planning the charge discharge path and calculating the real-time residual charge density under the dynamic control mode, and detecting that the real-time residual charge density is lower than the safe charge threshold to complete the local discharge circuit protection operation are as follows: Embedded variable impedance units are deployed in multiple locations within the chip to construct a distributed conduction network; In dynamic control mode, the center point of the region is identified based on the high-density charge region, and a spatial positioning coordinate system is established. Three-dimensional distance calculation is performed on the distributed conduction network based on the spatial positioning coordinate system, and the nearest variable impedance unit is marked. Dynamically plan the charge discharge path based on the nearest variable impedance unit and calculate the real-time residual charge density; If the real-time residual charge density is detected to be lower than the safe charge threshold, the partial discharge circuit protection operation is completed. Real-time temperature sampling and scanning of the charge discharge process and identification of abnormal thermal changes are performed, and abnormal heat sources are marked. Secondary planning of charge discharge path based on abnormal heat source.

[0062] In this embodiment, to achieve efficient ESD discharge and charge shunting control in a high-bandwidth flash memory chip, embedded variable impedance units need to be deployed at multiple key nodes within the chip structure to form a dynamically adjustable distributed conduction network. This process begins at the chip layout design stage, selecting areas prone to charge accumulation as deployment points based on current concentration trends, interconnect layer routing directions, and potential distribution characteristics. Each variable impedance unit consists of a resistance modulation layer, a control gate, and a conduction contact, and can dynamically adjust its equivalent impedance under external control signals.

[0063] The design of a distributed conduction network must balance uniformity and response sensitivity. To avoid uneven local charge discharge or conduction delay, the spacing between network nodes is typically controlled within the range of 10-20 micrometers, and multi-layer interconnection is achieved through a metal interconnect layer. Each impedance unit maintains a high-resistivity state when not in operation, without affecting normal chip operation; when an ESD event is triggered, it quickly modulates into a low-resistivity state to establish a transient charge discharge channel. This distributed structure provides the hardware foundation for discharge path planning in subsequent dynamic control modes, enabling the chip to have flexible adaptive protection capabilities. When the chip enters dynamic control mode, the location of the main charge accumulation area must first be determined to accurately plan the discharge path. By calculating the spatial energy distribution of the previously identified high-density charge areas, the location of the energy center point is determined. This center point represents the area with the most severe local charge accumulation and is the starting point for charge discharge path design.

[0064] To facilitate path analysis in three-dimensional space, a spatial positioning coordinate system needs to be established. The origin of the coordinate system is set at the charge energy center point, and the three-dimensional coordinate axes are defined along the horizontal interconnection direction, the vertical stacking direction, and the main charge flow direction of the chip, respectively. Through this coordinate system, the relative spatial distances between different levels and different conductive units within the chip can be uniformly described.

[0065] Furthermore, during the establishment of the coordinate system, scale normalization is performed based on the chip's packaging structure and metal layer spacing to ensure the accuracy of subsequent distance calculations and energy mapping. The introduction of this positioning coordinate system provides a clear spatial reference framework for the dynamic control process, offering a geometric basis for selecting the direction of charge discharge and matching channels. After establishing the spatial coordinate system, the spatial relationship between the high-density charge region and each variable impedance unit needs to be determined. By reading the coordinate position of each impedance unit in the conduction network and calculating the three-dimensional Euclidean distance between it and the coordinates of the charge energy center point, the spatial distance distribution between each unit and the charge center is obtained.

[0066] After the distance calculation is completed, all cells are sorted in ascending order of distance, and candidate cells in the nearest neighbor region are selected based on a distance threshold. Subsequently, by judging the equivalent resistance state and transient response characteristics of these cells, the nearest variable impedance cell that can be preferentially turned on is marked.

[0067] In this process, to avoid mistakenly selecting ineffective conductive nodes, charge flow direction prediction data is typically used to exclude nodes that are short in distance but located in the opposite electric field region. After screening, the nearest impedance unit remaining becomes the key node for subsequent charge discharge path planning. This step achieves a dual integration of spatial geometric positioning and electrical characteristic screening, providing an accurate physical reference for the dynamic construction of the discharge channel. After marking the nearest variable impedance unit, it is necessary to establish the optimal discharge path from the high-density charge center to this unit. Dynamic path planning is achieved by comprehensively considering charge flow direction, electric field distribution, and impedance change rate. First, based on the electric field gradient direction of the charge energy center point, the preferred direction of charge migration is determined; then, combined with the node connectivity of the conductive network, the candidate paths are traversed and searched to select the discharge channel with the minimum impedance and optimal potential difference.

[0068] After the path is established, the current changes and charge migration rate along the path are monitored, and the residual charge density during the discharge process is calculated in real time. A dynamic decay curve of the charge between the start and end points of the path is obtained through time-series sampling and density integration. If the discharge path is designed reasonably, the curve should show a stable decreasing trend, and the residual charge density should decrease by more than 80% within a nanosecond time.

[0069] The core of this process lies in dynamically adjusting the impedance value of the conducting unit, enabling the charge flow direction to adaptively adjust with real-time potential changes, thereby maximizing discharge efficiency and avoiding local overload. During the continuous charge discharge process, it is necessary to monitor the changes in residual charge density in real time. When this value is detected to be consistently below the preset safe charge threshold, it indicates that the energy in the charge accumulation area has been fully released, and the local discharge operation can be considered complete.

[0070] To avoid premature shutdown causing charge rebound or residual accumulation, the monitoring mechanism sets a stable holding time window, such as maintaining a low charge density state for more than 10 nanoseconds before triggering the impedance recovery operation of the conduction unit. After the conduction unit returns to a high impedance state, the chip returns to a static protection state, ensuring that the charge flow channel is closed and preventing secondary interference.

[0071] This step marks the end of a complete local ESD discharge cycle. Through this dynamic detection and adaptive shutdown mechanism, the chip can automatically resume normal operation after the charge risk is eliminated, maintaining circuit stability and power consumption balance. In long-term operation, this controllable discharge strategy can effectively reduce metal migration and dielectric aging rates, providing durable and intelligent ESD protection for high-bandwidth flash memory chips.

[0072] In this embodiment, the specific steps for real-time temperature sampling and scanning of the charge discharge process, identification of abnormal thermal changes, and marking of abnormal heat sources are as follows: Real-time temperature sampling and scanning were performed during the charge discharge process to obtain temperature information at different locations; The temperature information is subjected to time-series denoising and calibration processing to obtain denoised and optimized temperature information; Based on the noise-reduced and optimized temperature information, temperature rise change analysis is performed to generate a temperature rise change curve; The transient heat flux rate is obtained by calculating the temperature difference between adjacent time frames based on the temperature rise curve. Abnormal thermal changes are identified and abnormal heat sources are marked based on transient heat flux change rate.

[0073] In this embodiment, during the ESD discharge process of a high-bandwidth flash memory chip, the concentrated migration of charge in localized areas can cause a transient temperature rise. If left unmonitored, this could lead to dielectric breakdown or overheating and melting of the metal layer. Therefore, it is necessary to deploy high-sensitivity temperature sampling nodes at critical locations on the chip to achieve real-time temperature scanning of the discharge process. The sampling nodes employ a micro-thermal sensor array structure, distributed across the dielectric layer, metal interconnect layer, and contact nodes.

[0074] The sampling frequency is typically set above 100MHz to ensure the accuracy of capturing nanosecond-level temperature changes. Each sensing node is connected to the data acquisition module via an independent channel to record temperature values ​​and timestamp information in real time. After charge discharge is initiated, temperature sampling scanning starts simultaneously and continues to cover the entire discharge cycle.

[0075] During the scanning process, to avoid misjudgments caused by single-point reading deviations, a multi-node averaging calibration method is adopted to locally fuse temperature data from neighboring sampling points, forming a spatially distributed temperature field dataset. This process yields complete time-series temperature information, providing a reliable data foundation for subsequent thermal trend analysis and anomaly identification. Due to factors such as internal chip electromagnetic disturbances, sampling channel noise, and high-speed signal coupling, the raw temperature data inevitably contains transient interference signals and drift errors. To ensure the accuracy of the analysis results, time-series denoising and calibration processing of the temperature information is required.

[0076] First, a time-series smoothing algorithm is used to perform preliminary filtering on the sampled signal to remove low-amplitude, high-frequency noise components. Based on this, a sliding window weighted average method is used to smooth the temperature changes at adjacent time points, maintaining the continuity and physical plausibility of the temperature curve.

[0077] Subsequently, calibration and compensation processing is performed. By pre-collecting baseline temperature data of the sensing nodes under isothermal conditions, a temperature-time drift model is established, and the measured temperature sequence is offset corrected accordingly. If some nodes exhibit significant abnormal jumps (such as a sudden temperature rise exceeding three standard deviations of the average), they are automatically marked as noise samples and removed.

[0078] The denoised and optimized temperature information obtained after processing can more realistically reflect the thermal changes during charge discharge, providing clean and reliable data input for subsequent temperature rise trend analysis and transient heat flow calculation. After obtaining stable and reliable temperature data, the time series needs to be analyzed to extract the dynamic characteristics of temperature rise during charge discharge. Specifically, by comparing the temperature change trends of each sampling node in different time frames, the correlation between the temperature rise rate and time is calculated, thereby generating a temperature rise change curve.

[0079] The temperature rise curve reflects the entire process of heat accumulation and diffusion in the discharge area. When the ESD discharge path is activated, the temperature in the local area will rise sharply in a short period, followed by a slow decline. Analysis of the temperature rise curve shape can identify whether the charge discharge is uniform and whether the heat diffusion is normal. For example, if the curve shows a sudden spike that lasts for a short time, it indicates a transient current concentration phenomenon; if the curve rises slowly and steadily, it indicates that the discharge is relatively uniform.

[0080] In the experiment, to more intuitively reflect the dynamics of heat distribution, the temperature rise curves of multiple sampling nodes can be superimposed to form a three-dimensional temperature distribution map, presenting the thermal response characteristics of different regions in a spatial dimension. This temperature rise curve is not only used to monitor thermal behavior but also serves as the basis for subsequent calculations of the transient heat flux change rate. Based on the temperature rise curve, differential calculations are performed on the temperature data of adjacent time frames to obtain the temperature change rate per unit time. This rate can be considered an approximate reflection of heat flux changes, representing the speed and direction of heat conduction within the chip.

[0081] In the implementation, the temperature sequence is sliced ​​according to time steps, with each time step corresponding to a sampling period (typically 10 nanoseconds). By calculating the temperature difference between adjacent slices, the distribution data of the transient heat flux change rate is obtained. Subsequently, this data is mapped onto the chip's three-dimensional structural model to form a dynamic heat flux distribution map.

[0082] If the rate of change of heat flux suddenly increases in a certain region, it indicates that there is a rapid charge migration or a sudden increase in current density, which is very likely to correspond to the transient energy peak on the ESD discharge path.

[0083] The key to this step lies in the accuracy of temporal resolution control and spatial interpolation algorithms. By appropriately setting the sampling frequency and spatial resolution, transient thermal behavior at the nanoscale can be captured, providing clear clues for the subsequent identification of abnormal heat sources. After obtaining the transient heat flux change rate, it is necessary to further identify abnormal heat sources that may cause chip damage. The identification process is based on dual feature analysis of time and space: First, regions where the heat flux change rate exceeds the upper limit of the normal discharge range are screened using a threshold determination method. These regions usually correspond to locations where charge flow is obstructed or current is concentrated. Second, the abnormal regions are spatially clustered to determine whether they form continuous high heat flux band structures. If the clustered area exceeds the design threshold (e.g., the local area ratio exceeds 2%), it is identified as a potential dangerous heat source.

[0084] In addition, it is necessary to observe whether the heat source persists across multiple time frames, taking into account its temporal characteristics. If its duration exceeds 50 nanoseconds, it indicates that there may be structural conductive defects or localized material degradation in the region.

[0085] Ultimately, the identified abnormal heat sources are marked and their coordinates, duration, and maximum temperature rise are recorded, serving as a reference for subsequent circuit structure optimization and ESD protection design improvements. This identification mechanism enables early warning of hidden thermal anomalies within the chip, thereby improving overall ESD protection efficiency and reliability.

[0086] In this embodiment, the specific steps for secondary planning of the load release path based on the abnormal heat source are as follows: Calculate the spatial location coordinates of the abnormal heat source based on the spatial positioning coordinate system; Based on the spatial location coordinates, adjacent available variable impedance units are detected in the distributed conduction network, and available variable impedance units are extracted. The available variable impedance unit is added to the discharge path planning to perform secondary planning of the charge discharge path and to detect global heat distribution information in real time. The global heat distribution information is analyzed for temperature rise and fall within a predetermined time window to obtain a temperature rise control report; Based on the temperature rise control report, optimize the temperature control operation for local venting.

[0087] In this embodiment, during the ESD protection process of a high-bandwidth flash memory chip, when an abnormal heat source is detected, its precise location within the chip must first be determined. To this end, based on a previously established spatial coordinate system, the temperature sampling node numbers of the heat source are mapped to the chip's three-dimensional geometric layout. This coordinate system, with the chip's geometric center as its origin, uses three-dimensional Cartesian coordinates to describe the chip's physical size distribution in the x, y, and z directions.

[0088] During the location process, the center coordinates of the heat source are determined through spatial interpolation calculations using multi-point temperature data. If the heat source region is diffuse or irregular in shape, the centroid location method is used, with temperature intensity as the weight to calculate the region's center position. This method avoids deviations caused by single-point anomalies, making the location results more consistent with the actual distribution characteristics of the thermal field.

[0089] Simultaneously, by combining time information, the dynamic evolution trajectory of the heat source location can be tracked, and its migration trend or fixed aggregation characteristics can be analyzed. Under experimental conditions, the positioning accuracy can reach the micrometer level (approximately 1~3 μm), meeting the requirements for internal thermal field analysis of the chip. Through this step, the spatial coordinate data of the abnormal heat source can be obtained, providing a coordinate reference for the subsequent detection of variable impedance units in the conductive network and the planning of discharge paths. After obtaining the spatial location of the abnormal heat source, it is necessary to detect the variable impedance units spatially adjacent to the heat source region in the distributed conductive network inside the chip. This conductive network consists of miniature controllable resistor units embedded in the metal interconnect layer, each unit having independent conduction adjustment capability.

[0090] Using a spatial coordinate matching method, a three-dimensional detection radius (typically within the range of 20~50 μm) is set based on the center point of the abnormal heat source. Within this radius, network nodes are scanned to select units that are in operation and whose impedance values ​​have not reached the lower limit. If multiple usable units exist, they are prioritized according to distance, response delay, and impedance adjustment capability.

[0091] To avoid excessive energy concentration during charge discharge, it is also necessary to analyze the spatial coupling effect between candidate units to ensure that no new current convergence region is formed after they are turned on.

[0092] After detection, available variable impedance elements are marked as candidate nodes, and a connection table with the heat source region is established. This step ensures that the control path has sufficient redundancy and current diversion capacity during the secondary planning stage of charge discharge, thereby reducing the risk of local thermal stress accumulation. After obtaining available variable impedance elements, these nodes are incorporated into the charge discharge path planning to form a dynamically adjusted secondary discharge network. The core objective of this stage is to reconstruct the topology of the initial discharge path based on the current charge distribution and thermal state.

[0093] The specific approach involves starting with an abnormal heat source, sequentially connecting available variable impedance units and the main discharge channel node, and determining the optimal path based on the principle of shortest energy dissipation. During path planning, it is necessary to monitor the chip's global thermal distribution information in real time to prevent the formation of secondary hotspots by new paths.

[0094] The monitoring process is achieved through synchronous updates of a temperature sampling array, which collects the global temperature distribution every tens of nanoseconds. If, after the secondary path is activated, the temperature rise rate in a certain area exceeds a preset threshold (e.g., an increase of more than 0.5°C per nanosecond), it indicates that the energy transfer is too concentrated, and the path needs to be optimized again.

[0095] Through this real-time feedback mechanism, the discharge path can adaptively adjust the conduction nodes and impedance values ​​to achieve heat dissipation and energy balance. This step ensures the safe release of ESD energy within the chip while reducing local structural stress accumulation. After the secondary discharge path is stably turned on, the global heat distribution information needs to be dynamically analyzed to verify whether the heat energy has effectively diffused and fallen back within the expected time. The analysis process is carried out in time windows, with the window length typically set to 500 nanoseconds to 1 microsecond, to observe the decay characteristics of the thermal field over time.

[0096] Within each time window, the global average temperature rise and the local peak temperature change trend are calculated to generate a temperature rise and fall curve. If the overall temperature shows a downward trend for three consecutive time windows, and the fall exceeds a set proportion (e.g., falls to below 30% of the peak value), it indicates that the thermal control effect of the venting path is good.

[0097] At the same time, it is also necessary to conduct spatial consistency detection on the temperature distribution in specific areas to determine whether there is residual heat concentration.

[0098] Finally, these data are integrated to generate a temperature rise control report, which includes key parameters such as peak temperature rise location, temperature drop rate, thermal diffusion uniformity index, and discharge energy decay ratio. This report provides a quantitative basis for subsequent temperature control operations, ensuring that charge discharge operates within safe thermal limits. Based on the analysis results in the temperature rise control report, local discharge temperature control optimization is performed. Specifically, when the report indicates that the temperature rise in a certain area is slow to fall or that the local temperature peak remains excessively high, the control module needs to make fine adjustments to the corresponding variable impedance unit.

[0099] By changing the impedance values ​​of these units, local current shunting and conduction depth control can be achieved, thereby reducing local energy density and achieving thermal equilibrium.

[0100] During the optimization process, it is also necessary to coordinate the timing of the control actions to avoid energy distribution oscillations caused by simultaneous adjustments to multiple nodes. To this end, a time-slice-based sequential adjustment method is adopted, allowing only one local region to perform impedance correction operations at a time, with the time slice interval set to 10 to 20 nanoseconds.

[0101] When the global temperature rise rate stabilizes within the set range (e.g., below 0.1℃ / ns), the control process automatically ends, marking the completion of local venting temperature control.

[0102] This step enables closed-loop control of charge discharge and heat distribution, which not only effectively prevents the risk of thermal runaway, but also provides a replicable adaptive optimization model for subsequent chip-level ESD protection strategies.

[0103] In this embodiment, a high-bandwidth flash memory chip ESD protection circuit optimization system is provided, used to execute the high-bandwidth flash memory chip ESD protection circuit optimization method described above, including: The synchronous observation module is used to perform multi-point synchronous observations through the charge imaging module and extract multi-step observation information of ESD. The charge region identification module is used to identify high-density charge regions based on the ESD multi-step observation information; The safety assessment module is used to define the safe charge threshold, perform charge safety assessment on high-density charge regions, and enter dynamic control mode. The discharge circuit protection module is used to dynamically plan the charge discharge path based on the dynamic control mode and calculate the real-time residual charge density; detect that the real-time residual charge density is lower than the safe charge threshold, and complete the local discharge circuit protection operation.

[0104] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0105] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein are implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for optimizing ESD protection circuits for high-bandwidth flash memory chips, characterized in that, Includes the following steps: Multi-point synchronous observation is performed using a charge imaging module to extract multi-step ESD observation information. High-density charge regions are identified based on the aforementioned ESD multi-step observation information; Define a safe charge threshold, conduct charge safety assessment of high-density charge regions, and enter a dynamic control mode; Based on the dynamic control mode, dynamic planning of charge discharge path is performed, and real-time residual charge density is calculated; The real-time residual charge density is detected to be lower than the safe charge threshold, and the partial discharge circuit protection operation is completed.

2. The method for optimizing the ESD protection circuit of a high-bandwidth flash memory chip according to claim 1, characterized in that, The specific steps for extracting multi-step ESD observation information through multi-point synchronous observation using the charge imaging module are as follows: The charge accumulation and emission paths during the ESD impact process are observed simultaneously at multiple points using the charge imaging module, and multi-step observation information of ESD is extracted. Determine the time slice length, perform dynamic data slicing on ESD multi-step observation information, and obtain multi-time series information; The local charge density gradient and transient electric field intensity are calculated based on multi-time-series information, and the intensity distribution at multiple locations is identified to obtain a charge energy distribution map. Multidimensional coupling feature clustering is performed based on the charge energy distribution map to obtain the charge coupling characteristic distribution of the dielectric layer, metal interconnect layer and node contact region; Real-time ESD characteristic events are generated based on the charge coupling characteristic distribution.

3. The method for optimizing the ESD protection circuit of a high-bandwidth flash memory chip according to claim 2, characterized in that, The specific steps for identifying high-density charge regions based on the ESD multi-step observation information are as follows: Based on the real-time ESD feature events, amplitude normalization and timing synchronization processing are performed to obtain standardized ESD feature information. Based on standardized ESD feature information, the charge gradient and potential energy difference between adjacent nodes are calculated to analyze the charge conduction path and identify transient high charge density paths. The internal spatial distribution of transient high charge density paths is identified to obtain a charge conduction path distribution map. Based on the charge conduction path distribution map, the path intersection density and energy accumulation distribution are analyzed to identify high-density charge accumulation regions. Spatial boundary morphological filtering and equipotential line fitting are performed on the high-density charge accumulation region to extract the high-density charge region.

4. The method for optimizing the ESD protection circuit of a high-bandwidth flash memory chip according to claim 1, characterized in that, The specific steps for defining a safe charge threshold, assessing the charge safety of high-density charge regions, and entering dynamic control mode are as follows: Identify basic chip information, including chip material parameters, dielectric breakdown strength, and metal interconnect withstand voltage limit; Based on the chip's basic information, multi-layer charge carrying capacity is predicted to obtain the maximum allowable charge density; The safe charge threshold is defined based on the maximum permissible charge density; Calculate the maximum peak charge and average charge intensity of the high-density charge region; The maximum charge peak value and average charge intensity are assessed for charge safety based on the safe charge threshold. When the maximum charge peak value or average charge intensity is not less than the safe charge threshold, the local discharge path is activated and the system enters the dynamic control mode.

5. The method for optimizing the ESD protection circuit of a high-bandwidth flash memory chip according to claim 1, characterized in that, The specific steps for dynamically planning the charge discharge path and calculating the real-time residual charge density under the dynamic control mode, and detecting that the real-time residual charge density is lower than the safe charge threshold to complete the partial discharge circuit protection operation, are as follows: Embedded variable impedance units are deployed in multiple locations within the chip to construct a distributed conduction network; In dynamic control mode, the center point of the region is identified based on the high-density charge region, and a spatial positioning coordinate system is established. Three-dimensional distance calculation is performed on the distributed conduction network based on the spatial positioning coordinate system, and the nearest variable impedance unit is marked. Dynamically plan the charge discharge path based on the nearest variable impedance unit and calculate the real-time residual charge density; If the real-time residual charge density is detected to be lower than the safe charge threshold, the partial discharge circuit protection operation is completed. Real-time temperature sampling and scanning of the charge discharge process and identification of abnormal thermal changes are performed, and abnormal heat sources are marked. Secondary planning of charge discharge path based on abnormal heat source.

6. The method for optimizing the ESD protection circuit of a high-bandwidth flash memory chip according to claim 5, characterized in that, The specific steps for real-time temperature sampling and scanning of the charge discharge process, identification of abnormal thermal changes, and marking of abnormal heat sources are as follows: Real-time temperature sampling and scanning were performed during the charge discharge process to obtain temperature information at different locations; The temperature information is subjected to time-series denoising and calibration processing to obtain denoised and optimized temperature information; Based on the noise-reduced and optimized temperature information, temperature rise change analysis is performed to generate a temperature rise change curve; The transient heat flux rate is obtained by calculating the temperature difference between adjacent time frames based on the temperature rise curve. Abnormal thermal changes are identified and abnormal heat sources are marked based on transient heat flux change rate.

7. The method for optimizing the ESD protection circuit of a high-bandwidth flash memory chip according to claim 5, characterized in that, The specific steps for the secondary planning of charge discharge path based on abnormal heat sources are as follows: Calculate the spatial location coordinates of the abnormal heat source based on the spatial positioning coordinate system; Based on the spatial location coordinates, adjacent available variable impedance units are detected in the distributed conduction network, and available variable impedance units are extracted. The available variable impedance unit is added to the discharge path planning to perform secondary planning of the charge discharge path and to detect global heat distribution information in real time. The global heat distribution information is analyzed for temperature rise and fall within a predetermined time window to obtain a temperature rise control report; Based on the temperature rise control report, optimize the temperature control operation for local venting.

8. A high-bandwidth flash memory chip ESD protection circuit optimization system, characterized in that, The method for optimizing the high-bandwidth flash memory chip ESD protection circuit as described in claim 1 includes: The synchronous observation module is used to perform multi-point synchronous observations through the charge imaging module and extract multi-step observation information of ESD. The charge region identification module is used to identify high-density charge regions based on the ESD multi-step observation information; The safety assessment module is used to define the safe charge threshold, perform charge safety assessment on high-density charge regions, and enter dynamic control mode. The discharge circuit protection module is used to dynamically plan the charge discharge path based on the dynamic control mode and calculate the real-time residual charge density; detect that the real-time residual charge density is lower than the safe charge threshold, and complete the local discharge circuit protection operation.

Images