Sreetosh Goswami and Thirumalai Venky Venkatesan
In the lab

A memristor, or memory resistor, is a switching device that retains a state of internal resistance based on the history of applied voltage. Unlike other modern-day electronics memories like those made from CMOS devices, memristors are able to “recall” their state (that is, the information stored in them), even if you lose power. They also use much less energy.

Resistance-switching memristors made from transition metal oxides, for example, are promising for electronic components in a variety of future device applications, including high-density non-volatile memory as well as emerging applications such as “stateful” logic and bio-inspired neuromorphic devices. Although they show good performance, one of the drawbacks of such memristors is that they tend to have high switching voltage, large switching energy, low reliability and they degrade after large numbers of read-write cycles. They can also be expensive to make at large storage densities (that is, when a large amount of data is stored in them over a small area).

Transition metal complex made from Ru combined with azo-aromatic ligands

A team of researchers led by Thirumalai Venky Venkatesan of the National University of Singapore in collaboration with Sreebrata Goswami of the Indian Association for the Cultivation of Sciences (IACS) and Victor Batista of Yale University has now made an organic device that can compete in terms of performance with oxide-based memristors, is cheap to make and is stable over a long period of time.

The resistance-switching material employed in the new device is a spin-coated active layer of a transition metal complex made from Ru combined with azo-aromatic ligands [Ru(L)3](PF6)2, with L being a 2-(phenyloazo) pyridine ligand. The devices were fabricated and their properties measured at NUS while the molecule itself was first synthesized by Goswami’s group at IACS and converted to a resistive memory by Goswami. The group at Yale performed theoretical calculations.

First industrially competitive organic memristor

“For the first time an organic device is looking industrially competitive,” says Goswami. “and we have developed a clear picture of the molecular mechanism based on our in-situ studies that organic devices have always been lacking. Another benefit of this material that it can be simply spin-coated onto a surface to make the devices with an impeccable reproducibility.”

“The devices are highly robust,” adds Batista, “and show a large difference in their ON and OFF state current, which means that there is a low bit-read failure rate and low error probability.”

No signs of degradation

“They can also be switched on using a very low voltage of just 0.1 V, which is the lowest ever switching voltage demonstrated in a stable device with such a large ON/OFF ratio,” says Venkatesan. “They also have a rise time of less than 30 ns (as measured with the limits of today’s instruments) and have the potential to switch at sub-nanosecond speeds.

“And that is not all: they show no signs of degradation, even after 1012 read-write cycles and are stable at 350 K for over two months. And, since the devices operate at (nearly) the molecular level, they can be scaled down to incredibly small sizes. Indeed, we have demonstrated them down to 60 nm2,” he adds.

Easy to make and cost-effective

“What is more, the way the device works is well understood,” says Goswami, “so optimising it or enhancing its capability will be clearer. Memory technologies can be expensive in general because they require complex fabrication techniques and often involve rare and expensive heavy metals. Our devices are easy to make, and so will be cost-effective compared to current technologies.”

According to the new researchers, the new device might be used to replace any solid-state memory around today, including those found in smart phones computers and other digital devices. “It is thus not hard to imagine that this type of technology being employed in everything from storage of important data (such as bank records, for example), to that obtained from medical and health-monitoring devices,” Batista tells

Towards a universal memory

“At the moment, computers rely on different memory devices for different stages in a calculation and for performing different operations,” adds Venkatesan. “For example, the random-access memory employed in performing routine work relies on different devices than those used for archival memory. With our new device, there might be a potential to develop a universal memory that could perform all these different types of tasks.”

The team, reporting its work in Nature Materials doi:10.1038/nmat5009, says that it is now busy working on improving the performance of its memristor and overcoming possible obstacles to commercialisation. “Our Raman and ultraviolet–visible spectroscopy experiments together with spectroelectrochemistry and quantum chemical calculations show that it is the redox state of the ligands that determines the switching states of the device, and that it is possible to replace the Ru centre with other more abundant transition metals, which would further reduce the cost of the device,” says team member and lead author of the study Sreetosh Goswami,” Our molecular system is a treasure chest for developing a variety of advanced memristors, many of which have not even been thought of yet."

“In experiments, we are now building devices that have more than two states, memory devices that mimic the brain’s memory (neuromorphic systems) and memory devices with optical read-out capability, and so on,” concludes Venkatesan. “Ultimately, we would like to collaborate with an industrial partner to fully exploit the capabilities of this unique material system.”