The memory inside an iPhone or a thin laptop relies on transistors made from silicon semiconductors to store data. Computer engineers run into silicon’s physical limits as they try to shrink these devices to fit more data storage in smaller electronics. This means researchers are on the hunt for semiconductor alternatives to silicon.
Now a team of researchers has demonstrated a new type of ultrathin memory device that’s silicon-free (ACS Nano, DOI: 10.1021/nn3059136). The devices, made from graphene and molybdenum disulfide, could someday store more data in a smaller volume than any made of silicon, the researchers say.
Today, the dominant form of memory is silicon-based flash memory. As engineers make flash memory’s silicon components ever thinner, the devices encounter problems such as current leakage, which saps power and reduces the electronics’ performance.
Andras Kis and his team at the Swiss Federal Institute of Technology, in Lausanne, built new memory cells from a pair of two-dimensional materials, graphene and a single-molecule-thick layer of molybdenum disulfide (MoS2). Researchers previously have investigated graphene’s potential as a replacement for silicon in electronics, because it is an excellent electrical conductor. But graphene alone doesn’t have a band gap, an electronic property necessary in transistors. In recent years, materials scientists also have looked at thin films of naturally abundant MoS2, because it does have a band gap. But the two materials have not yet been used together in memory cells.
Besides their electronic properties, graphene and MoS2 are structurally strong materials and transistors made from them have the potential to only be a few atom layers thick. These physical properties mean memory devices made from the materials could bend more easily than relatively thick silicon-based transistors, possibly leading to flexible and wearable electronics.
Kis says the new devices exploit the strengths of graphene and MoS2: the high conductivity of graphene electrodes and the thin profile of MoS2 semiconductors. In each memory cell, charge flows from electrodes and through the MoS2 layer to a stack of multiple layers of graphene. This stack, called a floating gate, stores the charge; the presence or absence of this charge serves as the 1 or 0 for the memory bit. The device reads this stored information by passing current back through the MoS2 semiconductor.
By using graphene and MoS2 together in this way, “you can make a real two-dimensional component where everything is flat,” Kis says. Moreover, he adds, the MoS2 layer is so thin—0.65 nm—that it gains electronic properties that allow it to store more than a single bit of data per memory cell. In most of today’s flash memory, a cell stores only one bit.
Kang L. Wang, an electrical engineer at the University of California, Los Angeles, says the novel, interesting aspect of the new device is that it relies on MoS2 instead of silicon. “And I think for the first time they are putting graphene to its advantage,” he adds.
Kis calls the graphene-MoS2 device a proof of concept device and says that it needs much more engineering before it could be commercialized. For example, one component of the device, an insulator layer, is too thick: It is 6 nm thick, 10 times larger than the layers of graphene and MoS2. His team also wants to improve electrical properties, such as how long the device can hold a charge, to make the devices perform as well as commercial flash memory.