Image Credits: Maxx-Studio/shutterstock.com
There are several kinds of computer memory and has its drawbacks, which is why modern computer technology uses several kinds of memory.
Computer researchers around the world are working to create better data storage devices, with the gold standard being a ‘universal’ memory device with the quickness of RAM, the storage space of a hard drive and the stability of a flash drive.
Non-volatile memory based on ferroelectric thin films is one promising development that is currently in the research stage. A memory device based on ferroelectric thin-film ‘remembers’ by storing the direction of an externally-applied electric field based on a residual polarization charge.
Thin-film ferroelectrics have been used to make memory devices for some time. About a decade ago, such a concept for computer memory was proposed after ferroelectric properties were demonstrated in ultra-thin single-crystal films of perovskites. The primary challenge in implementing the technology is making a system with sufficiently high data storage density, and therefore a high storage capacity on a tiny area. The adoption of ferroelectrics is also hampered by the fact that they are fashioned from materials that are ‘incompatible’ with production processes for standard microelectronics.
Ferroelectric materials are primarily insulators that don't conduct electricity. However, if the material layer is extremely thin, electrons can ‘pass’ through with a set probability, due to a phenomenon known as quantum tunnelling. The likelihood of tunnelling is dependent upon the size and shape of the potential barrier (the energy qualities of the structure), with the electrons that happen to ‘pass through’ establishing a tunnel current.
Storing Information with Ferroelectric Thin Films
In ferroelectric tunnel junctions, data can be written by sending a voltage to electrodes alongside an ultra-thin ferroelectric, and it can be read by determining the tunnelling current. Theoretically, this kind of memory ought to have an incredibly high density, quick reading-and-writing speeds, and a low degree of power usage. It could become a non-volatile replacement for conventional DRAM (dynamic random-access memory) technology.
Modern DRAM has a relatively short retention time, after which the information must be either lost or overwritten, and because of this, DRAM technology requires a relatively large quantity of power. Memory founded on ferroelectric tunnel junctions will be capable of saving much of this power, which is valuable for portable electronic devices. However, until recently, ferroelectrics have not been compatible with silicon-based technology due to production processes.
In a 2016 study, scientists described the creation of polycrystalline alloyed films of hafnium oxide that were capable of retaining ferroelectric qualities on a silicon substrate. The films were developed by atomic layer deposition (ALD), which is widely used today to fabricate microprocessors.
One of the explicit benefits of ALD is that it permits functional layers to be created in three-dimensional structures. Since structures of the material used in the study are suitable for silicon technology, new non-volatile memory devices with ferroelectric polycrystalline sheets of hafnium oxide can be placed directly onto silicon, scientists behind the study said.
In a 2017 study, researchers announced another major step forward for ferroelectric thin-film devices: the creation of a flexible memory device.
The study described a novel process to create ferroelectric thin films at low temperature ranges and incorporate them with carbon-based organic semiconductors to fashion highly-flexible memory devices. The new device platform is able to be integrated with other flexible electronic technology.
We made a low-voltage, non-volatile, vertical organic transistor using a hafnia thin film. That level of detail may only be exciting for those in the electrical engineering field. For everyone else, that means this is a practical discovery with very real applications.
Jacob Jones, Professor, Materials Science and Engineering, North Carolina State University
A fully-functional prototype developed by the team retained its functionality even when bent up to 1,000 times, the study said.