Unidirectional Racetrack Memory Domain Wall Shift
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Prior racetrack memory devices require long racetracks and high current densities to move domain walls efficiently, leading to high voltage requirements and complex bi-directional motion, which complicates the design and fabrication of these devices.
Innovation Solution
A magnetic shift register system that allows domain walls to be moved in only one direction along a shorter racetrack, eliminating the need for a reservoir and simplifying the design by using unidirectional current pulses to shift and rewrite domain walls, with the option to store data in a microelectronic memory device once read.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If bi-directional motion of domain walls is used to allow reading and writing at any position, then versatility of data access is improved, but device complexity and racetrack length increase due to the need for reservoirs
Solution Approach 1:
The racetrack is segmented into distinct functional regions: a storage region for domain walls and separate read/write regions. This segmentation allows unidirectional motion while maintaining data access capability, as domain walls move systematically through defined zones rather than requiring bidirectional movement throughout the entire track.
Solution Approach 2:
Instead of allowing domain walls to move bidirectionally throughout the entire racetrack, the invention inverts the approach by using unidirectional motion with dedicated read/write zones. Data access is achieved by moving domain walls in one direction through systematically positioned read/write heads, eliminating the need for reservoirs and complex bidirectional control.
2Speed
If higher current density is used to move domain walls faster, then data retrieval speed is improved, but voltage requirements and energy consumption increase
Solution Approach 1:
Current pulses are applied locally at specific positions along the racetrack rather than uniformly across the entire track. This localized current application reduces overall power requirements while maintaining the ability to move domain walls at required speeds by concentrating energy where needed rather than applying high current density throughout the entire racetrack.
Solution Approach 2:
Domain walls are moved using periodic current pulses applied in sequence along the racetrack. This periodic action allows efficient domain wall propulsion by applying current only when and where needed, rather than requiring continuous high current density, thereby reducing average power consumption while maintaining data retrieval speed.
3Quantity of substance
If longer racetracks are used to accommodate more domain walls, then storage capacity is improved, but resistance and voltage requirements increase
Solution Approach 1:
The invention transitions from linear scaling of storage capacity to a multi-dimensional approach by positioning multiple read/write heads at different locations along the racetrack. This allows parallel access to multiple domain walls simultaneously, effectively increasing storage capacity without proportionally increasing racetrack length or voltage requirements.
Solution Approach 2:
Multiple read/write heads are positioned along the racetrack to perform multiple functions simultaneously: reading, writing, and verifying data at different positions. This multi-functionality allows the system to handle larger storage capacities efficiently without requiring proportional increases in racetrack length or voltage, as the same infrastructure serves multiple operational purposes.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables more reliable and efficient data storage and retrieval in a compact, faster, and easier-to-manufacture racetrack memory device, as domain walls are only shifted in one direction, reducing the racetrack length and simplifying the operational complexity.
Implementation Method 1
The spin polarized current carries spin angular momentum, which is transferred to the domain walls causing them to move in the direction of the electron flow
Data Source
Figure 1
Figure 2A~2C
Figure 3A~3F
AI summary
A racetrack memory storage device moves domain walls along the racetrack in one direction only. The reading element can be positioned at one end of the racetrack (rather than in the middle of the racetrack). The domain walls are annihilated upon moving them across the reading element but their corresponding information is read into one or more memory devices (e.g., built-in CMOS circuits). The information can then be processed in circuits for computational needs and written back into the racetrack either in its original form (as it was read out of the racetrack) or in a different form after some computation, using a writing element positioned at the end of the racetrack opposite to the reading element. Such a racetrack can be built more simply and has greater reliability of operation than previous racetrack memory devices.