Around $6 billion (€4.8 billion) is the annual US energy bill for running large-scale cloud-computing services that use volatile memory. If the power is turned off in volatile memory, such as DRAM (dynamic random-access memory), information is lost quickly. Non-volatile memory, for example SSD (Solid-State Disks), reduces electricity cost as data is retained, even if the power is turned off. Now, organic chemists have created a crystalline material at Northwestern University, US, that promises a cheaper class of non-volatile computer memory that consumes less power, lasts longer, and captures data faster.
This new organic material is made up of very long crystals from two organic molecules. The strong interaction between these molecules demonstrates a long-sought-after property known as ferroelectricity: a material demonstrates a positive electrical charge at one end and a negative one at the other.
This spontaneous electric polarization state can be reversed by passing an electrical field through the substance, which makes ferroelectricity. Then, the material can be used to create computer memory which stores data in 1's or 0's. The organic molecules are stable at room temperature unlike current ferroelectric ceramic and polymer materials. Previous tests produced ferroelectricity in materials, but only at liquid nitrogen temperatures.
This new work may lead to more energy-efficient non-volatile memory and enable faster results for researchers using distributed computing today.
"Non-volatile memories may enable higher performance scientific computing," said Alok Tayi, the co-first author of their research paper, who is a former graduate at Northwestern University and is now a postdoctoral fellow at Harvard University, US. "One advantage of non-volatile memory is that it can retain information even after the power is turned off. Such a fast recall capability may quicken scientific computing and information access. Non-volatile memories also promise faster recall of information and longer lifetime."
According to Tayi, the next steps are for chemists to identify new pairs of organic molecules for new ferroelectrics to improve the purity of the crystals that produce ferroelectricity.
"At the moment, we use millimeter long crystals as one memory device (one bit); this is impractical. Engineers can find clever ways of organizing our ferroelectric materials into more advanced architectures, making each memory device one thousand times smaller," said Alexander Shveyd, the other co-first author of the research paper and a postdoctoral fellow at the University of Rochester, US.
Yoshinori Tokura, a professor of physics at the University of Tokyo who was not involved in the research said, "The organic molecular ferroelectrics will be of particular importance in further developing organic electronics. The key is the fact that the organic ferroelectricity, including piezoelectric [electricity resulting from pressure] and pyroelectric [becoming electrically charged when heated] functions, may be even superior to the existing oxide ferroelectrics, not only in the non-use of toxic or rare-metal elements but also in the functional merit itself. The present materials developed… enjoy high transition temperatures exceeding room temperature as well as well-insulating behavior, thereby ensuring the validity of their new approach."
But, as this new material has a fragile crystalline state, Tokura said that for practical applications in large-scale computing to become a reality further research is needed into molecules that exhibit ferroelectricity, which can withstand nano-fabrication processes; a new era has just begun.
Their research paper was published inNature on 23 August.