Paolo Samorì will chair Europe's leading conference in graphene and related materials next year, Graphene Week 2020. He is a Distinguished Professor and director of the Institut de Science et d'Ingénierie Supramoléculaires at Graphene Flagship partners the Université de Strasbourg and CNRS, where he heads the Nanochemistry Laboratory. Samorì is also the Deputy Leader of the Graphene Flagship's Functional Foams and Coatings Work Package and the leader of the Spearhead project ChemSens.
His research focuses on the chemistry of graphene and layered materials, from their production and modification to applications in the fabrication of real devices. I had the opportunity to speak with Samorì about his work, and he gave me a great look into the role graphene and layered materials might play in the computers of the future.
First and foremost, can you tell me what non-volatile memory is?
It's a type of data storage device capable of retaining information, without a constant power supply, for a long time – typically more than 10 years.
This sets them apart from traditional volatile memories, such as dynamic and static RAM (random access memory), which are used to hold temporary data or program code that a computer processor needs to function. These are considered volatile, which means their memory is lost or reset when power is removed from the system.
In contrast, non-volatile memories remember the information they are currently storing, even after the power is switched off. They are commonly found in computer and smartphone hard drives, USB sticks and flash drives. They will also play a key role in the emerging Internet of Things technologies, which will structure the future of our society.
Why use graphene?
Nowadays, a huge range of materials with outstanding physical and chemical properties, including graphene and related layered materials, are being investigated. We believe they can improve the performance of non-volatile memories in terms of speed, density, power consumption, reliability and cost.
For example, graphene offers excellent flexibility and transparency, which enables non-volatile memories to be integrated into flexible materials, wearable electronics and smart objects. These are properties that conventional semiconductors don't have. Graphene also has extremely promising characteristics to boost the performance of emerging magnetoresistive RAM devices, which showcase a new method of storing data bits using magnetic states, instead of using electrical charges like in dynamic RAM.
The ultimate goal here is to enable the next generation of low-cost, lightweight, portable and wearable data storage devices.
Do you think graphene-enabled non-volatile memory can compete with existing memory?
We're not trying to take on standard memory devices in the usual way, competing with current technologies. We're trying to see what kind of functionality graphene and layered materials can offer. For example, new types of smart and wearable electronics are going to require ultra-thin, flexible, transparent storage. If we can do this with graphene-based non-volatile memories, this would be an absolute game-changer.
If you want to have a revolution from a technological point of view, you need to have something unique, not just a straightforward numerical improvement. Of course, we want to match the performance of the technology that currently exists but, more importantly, we want to add new functions and possibilities. We have to look further and see how this can be integrated with, for example, artificial intelligence and machine learning. I also want to stress that this technology is really promising for wearable electronics, and we want to keep going in that direction.
What challenges do you face using graphene and related materials in non-volatile memory?
There's the persistent problem of obtaining large amounts of high-quality graphene or related layered materials. But other than this, the biggest challenge is integrating these materials into existing conventional devices, without altering their unique properties.
Furthermore, even though we've achieved proof-of-concept and laboratory prototypes, integrating graphene and related materials at the industrial level is still a formidable roadblock. To overcome this, we need large-scale funding initiatives like the Graphene Flagship, which enable synergies with academic and industrial partners alike.
When do you think memories enhanced with graphene and related layered materials could hit the market?
New technologies typically need two or three decades of intense research and development before an exciting proof-of-concept turns into a game-changing technology capable of competing with the state-of-the-art and being sold commercially. But this assumes that sustainable funding support will be provided to leaders in the field, and that the industrial floor will come together with the existing large networks of researchers to complete the value chain for innovation.
Who are the key EU companies helping progress in the field of non-volatile memories?
STMicroelectronics, Wilk Eletronik and Swissbit are well-known European leaders in the field. There are also many smaller companies engaged with new and emerging non-volatile memories, such as Intenso, Angelbird and 3D Plus. These companies produce and develop memory devices based on more traditional materials.
What is the future of graphene and related materials in memory technology?
The future of non-volatile memories enhanced by graphene and related materials will depend on the paradigm shift provided by research into these new materials, as well as investments in new startups.
In light of the results over the last ten years, we envision two types of technology involving graphene and related materials: firstly, low-cost, flexible and transparent information storage devices, to be integrated into wearable systems and smart objects – and secondly, high-speed and high-capacity non-volatile memories.
In addition, the integration of layered materials into various emerging magnetoresistive RAM non-volatile memories could enable both information processing and storage technologies within the same device, which is considered the holy grail in the industry. This could allow for a combination of highly efficient machine learning, with unprecedented storage capability.