Chunk Redundancy Architecture for 3D Memory Defect Management

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Three-dimensional memory technologies exhibit higher defect rates due to increased effective area, leading to lower yields in manufacturing, as traditional redundancy architectures often require replacing entire blocks or tiles for a single defect, which is inefficient.

Innovation Solution

The chunk redundancy architecture provides redundant blocks in addition to addressable blocks, allowing individual chunks to be replaced by redundant chunks, thereby leaving other redundant chunks available to replace defective addressable chunks, utilizing chunk control circuitry to manage this remapping process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If three-dimensional memory structures are used to increase bit density, then storage capacity is improved, but defect rate increases due to larger total area

Engineering Contradiction:
Improvebit densityVSAvoiddefect rate
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The memory array is divided into multiple independently replaceable chunks rather than treating the entire array as a single unit. Each chunk can be individually identified as defective and replaced using the redundancy architecture, allowing the system to tolerate more defects while maintaining overall functionality.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extends traditional two-dimensional row/column redundancy into three dimensions by adding chunk-level redundancy. This creates a hierarchical replacement structure where chunks, rows, and columns can all be independently replaced, providing multiple dimensions of fault tolerance for the increased defect rates in 3D memory.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Reliability

If traditional redundancy architectures replace entire blocks for single defects, then reliability is improved, but redundancy efficiency deteriorates

Engineering Contradiction:
ImproveyieldVSAvoidredundancy efficiency
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The memory block is segmented into multiple smaller chunks, allowing granular replacement of only the defective portion rather than the entire block. This segmentation enables more efficient use of redundant structures by matching the replacement scope to the actual defect size.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The redundancy architecture dynamically adapts the replacement granularity based on the defect location and size. The system can replace individual chunks, rows, or columns depending on what is needed, making the redundancy mechanism more flexible and efficient compared to static block-level replacement.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS9996438B2Chunk redundancy architecture for memory
Publication Date: 2018.06.12 INTEL NDTM US LLC
  • US9996438B2 patent drawing
  • US9996438B2 patent drawing
  • US9996438B2 patent drawing

AI summary

An integrated circuit (IC) includes addressable blocks of memory, and at least one redundant block of memory. A block of memory includes two or more chunks of memory. The IC also includes redundancy control cells. Control circuitry is included to access a first chunk of a redundant block of memory in place of a first remapped chunk one of the addressable blocks of memory, and a second chunk of a redundant block of memory in place of a second remapped chunk one of the addressable blocks of memory, based on the redundancy control cells.