Charge-Trapping Memory Refresh Programming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Charge loss in charge-trapping memory cells leads to inaccurate data storage due to charge detrapping from shallow traps in tunnel ONO layers, causing threshold voltage (Vth) to decrease, which complicates reliable long-term data retention.
Innovation Solution
Implementing an in-place refresh programming method that identifies and classifies memory cells by Vth downshift, applying tailored refresh programming to subsets based on the amount of downshift, and customizing subsets for different data states to minimize complexity and avoid program disturb, allowing for frequent data retention checks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full reprogramming is performed frequently to correct charge loss, then data retention accuracy is improved, but device endurance and time consumption deteriorate
Solution Approach 1:
The patent segments memory cells into different subsets based on their charge loss characteristics and applies differentiated refresh programming strategies to each subset. This allows targeted correction of charge loss in cells that need it while avoiding unnecessary reprogramming of cells that maintain their charge, thereby improving data retention accuracy without proportionally increasing device wear and time consumption.
Solution Approach 2:
The patent changes operational parameters by adjusting refresh programming frequencies and voltages based on measured charge loss levels. By monitoring threshold voltage shifts and adapting refresh operations accordingly, the system maintains data retention accuracy while optimizing between refresh frequency and device endurance, avoiding both excessive full reprogramming and insufficient charge correction.
2Reliability
If refresh programming is applied to all memory cells, then data retention is improved, but device complexity and time consumption increase
Solution Approach 1:
The patent implements local quality by applying different refresh programming operations to different subsets of memory cells based on their individual charge retention characteristics. Instead of uniform treatment, each subset receives customized refresh operations matched to its specific needs, improving overall data retention while reducing unnecessary operations and simplifying the control logic compared to blanket refresh approaches.
Solution Approach 2:
The patent applies partial action by performing refresh programming only on memory cell subsets that exhibit charge loss beyond acceptable thresholds. By identifying and targeting only the necessary subsets for refresh operations rather than all cells, the system maintains data retention reliability while reducing operational complexity and time consumption associated with comprehensive refresh cycles.
3Ease of manufacture
If uniform refresh programming is applied to all data states, then manufacturing simplicity is maintained, but measurement precision and adaptability deteriorate
Solution Approach 1:
The patent changes parameters by implementing data state-specific refresh programming voltages and thresholds tailored to each data state's charge loss characteristics. This allows precise correction of threshold voltage downshift for each data state while maintaining a relatively simple manufacturing process, as the different parameters are implemented through software control rather than requiring different hardware configurations.
Solution Approach 2:
The patent achieves universality by using a single refresh programming mechanism that can adapt to multiple data states through parameter adjustment. The same hardware infrastructure performs refresh operations across all data states, with the flexibility to modify operating parameters based on the specific data state being refreshed, thereby maintaining manufacturing simplicity while improving measurement precision and adaptability.
Data Source
Figure 1A~1B
Figure 1C~1D
Figure 1E
AI summary
Techniques are provided for periodically monitoring and adjusting the threshold voltage levels of memory cells in a charge-trapping memory device. When a criterion is met, such as based on the passage of a specified time period, the memory cells are read to classify them into different subsets according to an amount of downshift in threshold voltage (Vth). Two or more subsets can be used per data state. A subset can also comprise cells which are corrected using Error Correction Code (ECC) decoding. The subsets of memory cells are refresh programmed, without being erased, in which a Vth upshift is provided in proportion to the Vth downshift. The refresh programming can use a fixed or adaptive number of program pulses per subset. Some cells will have no detectable Vth downshift or a minor amount of Vth downshift which can be ignored. These cells need not be refresh programmed.