Advancing memory technologies bit by bit
Melvin M. Vopson
The worldwide demand for digital data storage is increasing exponentially. IBM estimates that 2.5 quintillion of digital data bytes are produced every day on Earth. This is equivalent to 2.5 x 1018 bytes, or 2.5 Exabytes or 2.5 billion Gb produced every day. The increased demand for secondary data storage systems such as HDD, magnetic tape, CD, DVD and solid state FLASH memories is mimicked by similar demand for larger, faster and cheaper primary memory storage, also known as random access memory (RAM).
RAM is a key component in computers and consumer electronics and, depending on the technology used, RAM could be volatile or non-volatile. Volatile RAM means that computer chips store information as long as electric power is supplied to them. Once power is turned off, the information is lost. Usually the information is copied onto a non-volatile secondary data storage memory device.
Computer performance and computational speed would be greatly improved if primary memory storage RAM would be non-volatile. One of the promising candidates for non-volatile RAM memory chips are known as ferroelectric RAM (FRAM or FeRAM) and are based on a special class of dielectric materials called ferroelectrics. FRAM digitally encodes the information using switchable remanent polarization states within a ferroelectric thin film capacitor. The FRAM memories display unique features such as low power consumption, ultra-fast data accesses times and “read / write” endurance of 10 trillion (= 1013 “read / write” cycles) [fujitsu.com]. However, FRAM chips are very limited in terms of their data storage capacity. Indeed, the largest commercially available FRAM memory chips today are 4Mb and they operate typically at 5V or below. FRAM chips are utilized in various applications where the memory size is not a critical requirement, including RF-ID chips, security tags, smart meters, failure analysis in industrial machines, car navigation systems and other consumer electronics.
Alternative non-volatile memory RAM technologies are under development including resistive (RRAM) and magneto-resistive (MRAM) memory technology. However, these have a number of disadvantages including low memory capacity, high power consumption and reduced endurance.
Scientists at the University of Portsmouth and Iowa State University have recently demonstrated a novel non-volatile RAM technology based on anti-ferroelectric materials. Their proposed memory chip has been termed anti-ferroelectric random access memory (AFRAM).
Dr. Vopson and his co-investigator, Dr. Tan, showed that the AFRAM memory cell functions using a similar architecture to FRAM, but it requires a more complex operation protocol. Their initial experimental demonstration of the memory effect in anti-ferroelectric ceramic shows, remarkably, that the proposed AFRAM technology encodes data in both ferroelectric sublattices of an anti-ferroelectric medium. This results in a 4-state non-volatile memory capable of storing two digital bits simultaneously, unlike all other volatile and non-volatile RAM technologies that have 2-memory states and are capable of storing one digital bit per cell. The consequence of this discovery is that existing non-volatile FRAM memory chips could be replaced by non-volatile AFRAM chips using existing production platforms, by only replacing the ferroelectric capacitor with a suitable anti-ferroelectric capacitor. However, the memory capacity per cell is doubled when the newly proposed AFRAM technology is utilized.
In their Electron Device Letters article, the authors presented the device concept of this novel AFRAM technology, while in a parallel article published in Scripta Materialia the authors expanded on the fundamental physics and measurement procedures of these interesting polar dielectric materials.
However, in both articles, the authors acknowledged that there are challenges facing the AFRAM technology, especially related to signal degradation and allocation protocol of the 4-memory states, but it is expected that this research will lead to the possible emergence of future advanced AFRAM memory chips as well as other technologies when anti-ferroelectric materials are used in conjunction with semiconductors and multiferroics.