Phu-Quan Pham
Memristor concept
A memristor (short for "memory resistor") is a type of passive electronic component that has the unique property of being able to remember and store its resistance value based on the amount of charge that has passed through it. In essence, it can "remember" the amount of current that has flowed through it over time, which makes it a promising candidate for various memory and computing applications. The memristor is a specialized electronic component that exhibits a unique property known as "memristance." Memristance refers to the ability of the component to change its electrical resistance in response to the amount of electric charge that has previously flowed through it.
The memristor was theoretically proposed by Leon Chua in 1971 as the fourth fundamental two-terminal passive circuit element (alongside the resistor, capacitor, and inductor). However, it wasn't until 2008 that researchers at Hewlett-Packard (HP) successfully demonstrated a physical memristor device, which sparked significant interest and research in the field of memristor-based electronics
Figure 1. The missing electronic element
Figure 2. Memristor, a two-terminal device
The "pinched hysteresis loop" is a fundamental characteristic of memristors that helps define and distinguish them from other electronic components. To understand the concept better, let's break it down:
Hysteresis: Hysteresis in memristors refers to a phenomenon where the relationship between the charge and the flux in the device depends on the past history of the device's operation. In other words, the behavior of the memristor is not only determined by the present input, but it also depends on its previous state and the path it has taken to reach the current state. The term "hysteretic" is used because this behavior is reminiscent of hysteresis loops seen in magnetic materials, where the magnetic flux density depends not only on the present magnetic field but also on the material's past magnetic history.
Pinched Hysteresis Loop: When we apply a varying voltage across a memristor and measure the current flowing through it, we can plot a graph of current (I) against voltage (V), known as the I-V curve. In some cases, if we gradually increase the voltage from zero, we might observe that the current does not increase linearly with voltage. Instead, it might follow a particular curve that shows a gradual increase in current until a critical voltage is reached. Beyond this voltage, the current might suddenly jump to a higher value, forming a "pinched" or "butterfly" shape in the I-V curve.
Hysteresis in memristors is essential for certain applications, such as memory and computing, as it allows for the storage of information in the form of resistance states. This property has garnered significant interest in the field of electronics and computing and has the potential to revolutionize memory and computing architectures.
Figure 3. The I – V characteristic of memristor, hysteresis loop.
Resistive Switching Effect/ Behavior
The resistive switching effect in memristors refers to the ability of these devices to change their resistance in response to applied electrical stimuli. A memristor is a non-volatile memory device whose resistance can be altered by passing an electrical current through it. The resistive switching phenomenon is at the core of how memristors function and is responsible for their potential applications in memory and computing technologies. When an electrical voltage is applied across a memristor, it can undergo a change in its resistance state. This change can be reversible, meaning the memristor can switch back and forth between different resistance levels multiple times. The two primary resistance states are often referred to as the "low-resistance state" (LRS) and the "high-resistance state" (HRS).
The resistive switching effect is based on the redistribution of mobile ions or defects within the memristor's material. Depending on the direction of the applied voltage and the current flow, these ions or defects can migrate and accumulate at certain interfaces or regions within the memristor. This migration causes a change in the effective cross-sectional area available for electron conduction, leading to a change in resistance. The ability to precisely control and manipulate these resistance states makes memristors promising candidates for various applications:
Non-Volatile Memory: The ability to retain the resistance state even after the power is turned off makes memristors suitable for non-volatile memory applications, potentially leading to faster and more energy-efficient storage solutions.
Neuromorphic Computing: Memristors can emulate certain aspects of biological synapses, enabling them to be used in neuromorphic computing systems for machine learning and artificial intelligence.
Analog Computing: The continuous resistance modulation in memristors can be exploited for analog computing tasks, which could lead to more efficient and specialized computing architectures.
Resistive Switching Random Access Memory (RRAM): Memristors are being explored as a basis for a new type of non-volatile memory technology known as RRAM. Overall, the resistive switching effect in memristors plays a pivotal role in their unique behavior and potential applications across various fields of electronics and computing.
Memristor Classification
Memristors can be classified into several categories based on their materials, switching mechanisms, and behavior. Here are some of the main classifications of memristors:
Property-Based Categories:
Binary (Digital) Memristors: These are the simplest form of memristors and operate by switching between two distinct resistance states: a high-resistance state (HRS) and a low-resistance state (LRS). The switching is typically achieved through the movement of defects or ions within the memristor material.
Analog Memristors: Analog memristors exhibit a continuum of resistance states, enabling smooth and gradual changes in resistance. This characteristic makes them suitable for applications that require analog signal processing and emulation of biological synapses.
Floating-Gate Memristors: Similar to floating-gate transistors used in flash memory, floating-gate memristors store charge in a floating gate to control resistance changes. The accumulation or removal of charge affects the conductivity of the memristor.
Figure 4. Digital and Analog Resistive Switching Behavior
Mechanism-related to Material-Based Categories:
Spintronic/Magnetic Memristors: Spintronic memristors leverage the manipulation of electron spin and magnetic moments to control resistance. They are based on materials with strong spin-orbit coupling, and changes in the spin configuration lead to resistance modulation. Spintronic memristors have potential applications in non-volatile memory and neuromorphic computing.
Phase-Change Memristors: Phase-change memristors utilize reversible phase transitions between amorphous and crystalline states in a chalcogenide material. The phase change alters the resistance, allowing for non-volatile memory applications.
Filamentary Memristors: These memristors rely on the formation and rupture of conductive filaments within a dielectric material. The migration of ions or vacancies creates or dissolves these filaments, leading to changes in resistance. Examples include conductive bridge RAM (CBRAM) and electrochemical metallization (ECM) memristors.
Ionic Memristors (Oxide-Based): Oxide-based memristors use the movement of oxygen vacancies or other ions within an oxide material to control resistance. The oxygen vacancies can drift under the influence of an electric field, altering the resistance state. These memristors are known for their stable switching behavior and potential for high-density memory applications.
Polymer Memristors: Polymer memristors use conductive polymers that can change their oxidation or reduction state to modulate resistance. These memristors can exhibit flexible and biocompatible properties, making them suitable for wearable electronics and bioelectronic applications.
Biological Memristors: These memristors are inspired by biological synapses and attempt to mimic the behavior of neural connections. They often utilize materials that respond to changes in ion concentration, similar to how neurotransmitters work in biological synapses.
Quantum Memristors: Quantum memristors are theoretical constructs that incorporate quantum mechanical effects to manipulate resistance at the atomic scale. These memristors could enable ultra-dense memory and computing systems with quantum-level precision.
Memristor Crossbar Array and Discrete
Crossbar arrays and discrete memristors are two essential concepts within the realm of memristive devices, each playing a unique role in advancing the field of electronics and computing. A crossbar array is a specific arrangement of memristive devices that forms a grid-like structure. In this configuration, memristors are placed at the intersections of rows and columns, similar to a tic-tac-toe board or a chessboard. Each memristor in the array represents a programmable connection between a specific row and column. The conductance of each memristor can be adjusted to store and process data. Crossbar arrays are particularly interesting because they can enable dense and highly parallel computing and memory architectures. By exploiting the memristor's ability to change resistance based on applied voltage or current, crossbar arrays can perform tasks such as analog computing, pattern recognition, and machine learning operations. The architecture's parallelism and potential energy efficiency make it attractive for various emerging technologies, including neuromorphic computing and artificial intelligence applications.
Figure 5. Crossbar Array Level of Memristor Devices. The current recorded is equal to the sum of arrays
Discrete memristors refer to individual memristive devices that are not part of an integrated array. In this context, "discrete" means separate or distinct. These memristors are standalone devices that can be used for various purposes, including research, testing, and specialized applications. Discrete memristors have been studied to understand the fundamental behavior of memristive devices and to explore potential applications. Researchers often work with discrete memristors in laboratory settings to investigate their electrical characteristics, switching behavior, and performance under different conditions. These individual devices contribute to the broader understanding of memristor technology and help guide the development of practical applications. In summary, a crossbar array is a specific architecture that uses memristors arranged in a grid for parallel computing and memory operations, while discrete memristors are standalone devices used for research, testing, and exploration of memristor behavior and applications.
Figure 6. The discrete strucutre of memristor device
Figure 7. Memristor structure and potential applications
Memristor for Non-Volatile Random Access Memory applications
Non-Volatile Random Access Memory (NVRAM) refers to a type of computer memory that retains its stored data even when power is removed, making it non-volatile in nature. Unlike traditional volatile RAM (Random Access Memory), which loses its data when the power supply is interrupted or turned off, NVRAM retains data even in the absence of power. This characteristic is crucial for various computing applications, as it allows for persistent storage of information. NVRAM is often used as a general term encompassing various memory technologies that offer non-volatile data storage and random access capabilities such as Flash memory, Ferroelectric RAM, and memristors-based NVRAM.
Flash memory is a widely used NVRAM technology that is commonly found in USB drives, solid-state drives (SSDs), memory cards, and embedded systems. It uses floating-gate transistors to store data in cells, with each cell representing a binary state (0 or 1).
Ferroelectric RAM (FeRAM): FeRAM uses the polarization of ferroelectric materials to store data. It offers fast read and write speeds and has the ability to endure a high number of write cycles.
When considering memristors specifically in the context of NVRAM applications (resistive switching memories), this is refer to three kinds of NVRAM: MRAM, PRAM and RRAM.
Magnetoresistive Random-Access Memory (MRAM): MRAM is another type of non-volatile memory technology that utilizes changes in resistance. It is based on the magnetoresistive effect, where the electrical resistance of a material changes in response to an applied magnetic field. MRAM employs magnetic tunnel junctions (MTJs) that consist of two ferromagnetic layers separated by a thin insulating barrier. The relative orientation of the magnetization in these layers determines the resistance state of the MTJ, allowing data storage. MRAM offers advantages such as high endurance, fast read/write speeds, and non-volatility.
Phase Change Memory (PCM): PCM is an emerging NVRAM technology that utilizes the reversible phase transition between amorphous and crystalline states of a chalcogenide material to store data.
Resistive Random-Access Memory (RRAM): RRAM is a type of non-volatile memory that relies on resistive switching. It involves the change in resistance of a material between high-resistance and low-resistance states in response to an applied voltage. This behavior is somewhat similar to memristive behavior, as it involves the modulation of resistance levels to store and retrieve data. RRAM devices typically use metal-insulator-metal (MIM) or metal-insulator-semiconductor (MIS) structures to achieve resistive switching. RRAM technology has the potential to offer high storage density, low power consumption, and fast switching speeds, making it suitable for various memory applications.
Figure 8. Three kinds of Memristor-based NVRAM
The non-volatile nature of NVRAM makes it valuable for applications where data retention is essential, such as storage drives, embedded systems, and devices that require instant-on capabilities. NVRAM technologies are continually advancing, aiming to provide a balance between fast access times, high endurance, and energy efficiency, making them well-suited for various computing and electronic devices.
References:
- L.O. Chua (1971), “Memristor - The Missing Circuit Element”, IEEE Transactions on Circuit Theory, 18(5), pp.507-519, DOI:10.1109/TCT.1971.1083337.
- W. Jing, et al. (2008), "Solid-state electrochemistry in molecule/TiO2 molecular heterojunctions as the basis of the TiO2 “Memristor”, Journal of the Electrochemical Society, 156(1), p.29.
- D.B. Strukov, et al. (2008), “The missing memristor found”, Nature, 453, pp.80-83.
- N. Nithya and K. Paramasivam, "A Comprehensive Study on the Characteristics, Complex Materials and Applications of Memristor," 2020 6th International Conference on Advanced Computing and Communication Systems (ICACCS), Coimbatore, India, 2020, pp. 171-176, doi: 10.1109/ICACCS48705.2020.9074392.
Post a Comment