Advances in Non-volatile Memory and Storage Technology

Yoshio Nishi, Blanka Magyari-Kope

• Publisher: Woodhead Publishing
• Year: 2019
• eBook ISBN: 9780081025857
• Edition: 2
• Topic: Technology & Engineering
• Subtopic: Physics
• Index: Physics

Advances in Nonvolatile Memory and Storage Technology, Second Edition, addresses recent developments in the non-volatile memory spectrum, from fundamental understanding, to technological aspects. The book provides up-to-date information on the current memory technologies as related by leading experts in both academia and industry. To reflect the rapidly changing field, many new chapters have been included to feature the latest in RRAM technology, STT-RAM, memristors and more. The new edition describes the emerging technologies including oxide-based ferroelectric memories, MRAM technologies, and 3D memory. Finally, to further widen the discussion on the applications space, neuromorphic computing aspects have been included.This book is a key resource for postgraduate students and academic researchers in physics, materials science and electrical engineering. In addition, it will be a valuable tool for research and development managers concerned with electronics, semiconductors, nanotechnology, solid-state memories, magnetic materials, organic materials and portable electronic devices.- Discusses emerging devices and research trends, such as neuromorphic computing and oxide-based ferroelectric memories- Provides an overview on developing nonvolatile memory and storage technologies and explores their strengths and weaknesses- Examines improvements to flash technology, charge trapping and resistive random access memory

Part One

Progress in nonvolatile memory research and application

1

OxRAM technology development and performances

Ludovic Goux Imec, Kapeldreef, Leuven, Belgium

Abstract

This chapter reviews some key aspects of the studies and developments of filamentary-oxide-based resistive memory technology (OxRAM) carried out at imec over the last decade. It describes the switching mechanism and filament properties, details the main advantages of the concept, namely the CMOS-friendly integration, scalability, switching speed, and endurance, however, also the long-standing issues of switching variability and retention tail losses. It finally explains how the tackling of these latter limitations will determine the future application space of the concept, be it in storage-class or embedded memory areas.

Keywords

Nonvolatile Memory; Oxide Resistive RAM (OxRAM); Resistive switching; Conductive filament; Point-contact conduction; Switching mechanism; CMOS integration; Memory scaling; Memory reliability

1.1 Introduction

1.1.1 Nonvolatile memory applications

1.1.1.1 Storage-class memory (SCM)

In today’s computational systems, memories are categorized into volatile memory and nonvolatile memory (NVM) technologies.

Due to the ever increasing demand for more memory capacity, planar NAND Flash has been scaled down to below 20-nm feature size. Concomitantly, three-dimensional (3D) vertical NAND Flash has been developed as a Bit-Cost-Scalable (BiCS) solution and allows today entering the Terabyte era. As a result, the NVM market is by far dominated by NAND Flash technology, and the forecast is that the future of NAND will be NAND [1].

In the volatile memory category, the main technologies are the static RAM (SRAM) and dynamic RAM (DRAM), which are higher-speed and higher-performance technologies, however, exhibit poor scalability. Their role in a central process unit (CPU) is to store data that require immediate access while NAND Flash or hard-disk drive (HDD) store information that is not required immediately but for available future usage [2].

The problem arises when transferring data from DRAM to NAND: the overall performance of the system is limited by the huge latency gap between these two technologies. This gap has been virtually fitted with architectural solutions to increase the data access speed but at the expense of complex system design and increased chip area. In recent years, researchers have started exploring the possibility of novel memory concepts to improve the existing memory hierarchy. The concept of storage class memory (SCM) has been proposed, aiming to fill the access time gap between the “memory-memory” and the “storage-memory.”

As a “bridge” technology between DRAM and Flash, the main requirements for SCM are intermediate between DRAM and Flash and should be cost effective.

In short, SCM should be enabled by a nonvolatile, cheap, and scalable technology having clearly better reliability (write endurance and retention) than Flash. At the lead in this future booming market, Intel-Micron announced the 3D X-point Memory in 2015 [3] and launched products in 2017. Although not officially confirmed by Intel-Micron, it is generally agreed in the memory community that 3D X-point is based on phase-change memory technology.

As will be detailed in the following sections, the filamentary resistive RAM (RRAM) technology has also proven to be a promising candidate in this growing market.

1.1.1.2 Internet of Things (IoT)—Embedded memory

Another booming market is the concept of Internet of Things (IoT), which is defined as “intelligent connectivity of smart devices by which objects can sense one another and communicate” [4].

The advent of billions of connected devices is creating new opportunities and huge markets for the emerging memories. In fact, the market for IoT devices has been projected to more than 3.5 trillion in 2020.

In a typical electronic system, logic and NVM components are fabricated separately due to the incompatibility in integration flow. To accommodate the exponential growth of IoT devices, chip costs have to be reduced.

This drives the need for developing new embedded memory technology where a chip can contain both logic and memory components to lower the cost and save space on the printed circuit boards. This is referred to as system-on-a-chip (SoC). The existing SoC chip uses NOR Flash as embedded memory. However, to integrate embedded NOR Flash in 28 nm node and below, up to 15 extra photomasks need to be added in the overall fabrication process, which makes embedded NOR Flash an extremely expensive technology, all the more for more advanced logic nodes. Therefore, an alternative CMOS-compatible and low-cost embedded memory technology would be highly desired to feed the IoT market as well as other embedded markets.

In this respect too, the RRAM technology is a strong contender, as will be shown next.

1.1.2 Resistive RAM technology

To fulfill the requirements of these new applications, various new memory device concepts have been proposed and studied. The prominent concepts are spin-transfer torque magnetic RAM (STT-MRAM), phase-change RAM (PCRAM), and RRAM or ReRAM. These technologies, due to limited maturity, are categorized as emerging memories.

The RRAM category is the name of a group of memory technologies characterized by an electrically reversible resistive switching functionality between a low resistance state (LRS) and a high resistance state (HRS). One of the most attractive advantages of RRAM is the low-cost integration allowed by the combination of CMOS-friendly materials within a simple two-terminal device structure, which typically consists of a dielectric layer sandwiched between two metal electrodes.

Between the numerous mechanisms potentially at the origin of resistive-switching effects, nano-ionic transport and redox-reaction mechanisms taking place at the nanometer scale [5] have been clearly identified as accounting for the switching functionality of various systems.

In most of the RRAM devices reported [6–8], the resistive switching property originates from the growth and shrinkage of a...

Table of contents

• Cover image
• Title page
• Table of Contents
• Copyright
• Contributors
• Part One: Progress in nonvolatile memory research and application
• Part Two: Emerging opportunities
• Index