Graphene for interconnects

Graphene, boasting large carrier mobility and thermal conductivity, coupled with small material volume, emerges as a viable alternative to copper interconnects in electronic circuits.


Copper was an optimal material for interconnects in integrated circuits (ICs) throughout most of 20th century IC history, which was only natural following the great success of copper wires in long-haul electricity networks. Computers, electronic sensors, cameras, television and other microchips relied on the low-cost low-resistivity metal to pass electronic charge from one active layer of a circuit to another, as well as from the circuit to the outside world, with little loss of charges to heat. However, as engineers struggled to keep up with the ruthless Moore's law, doubling the number of transistors on ICs every couple of years, the interconnect size had to keep pace.


With the characteristic length of some interconnect dimensions reaching on the order of 10 nanometers in the last few years, traditional technology is starting to face trouble. At a line width smaller than 100nm, copper surface imperfections and material crystalline structure start to play an increasingly big role in the conduction properties of the metal, and the effective resistivity of traditional interconnects more than doubles as the width decreases from 100nm to 10nm. The rising resisitivity in turn leads to strong heating and increased power consumption of the IC. In fact, the shrinking interconnect is the most likely reason your smartphone heats up with intense use, and it is also why your battery is consumed so quickly. Researchers are of course investigating viable alternatives to copper, and graphene is emerging as one of them.


With unprecedented carrier mobilities, the most natural application for graphene would be exactly in interconnects. However, experimental realization of low resisitivity in graphene took some time to achieve, and it is only recent that graphene is being seriously considered as an alternative replacement to copper.


In 2012, researchers at the University of Illinois used graphene nanoribbons to produce large current densities, surpassing those normally achieved in copper wires by a factor of 1000. The nanoribbons were less than 100nm wide, and only a fraction of a nanometer high. The ribbons were made from CVD graphene, which can be grown uniformly over large areas. At the time, the team revealed that it is already using such graphene interconnects within memory applications, where they serve as the “wiring” for data storage bits.


Along the same research direction, scientists at Korea's Advanced Industrial Science and Technology institute (AIST) have used multilayer graphene intercalated with iron chloride powder. The research, published in September 2013, resulted in resisitivity comparable to that of copper interconnects. CVD graphene was used again, promising high reproducibility and mass production of the material. The iron chloride molecules were embedded between sheets of graphene to lessen its resistivity.


Image: Graphene nanoribbon illustration (courtesy of Georgia Tech/John Hankinson).


A real breakthrough occurred in the field of graphene interconnects with the finding of ballistic transport in epitaxially grown graphene nanoribbons. In the study, published by Georgia Tech and the Oak Ridge National Lab in the Nature magazine last February, ribbons of graphene only 40nm wide were shown to transport electrons ballistically at room temperature, over distances larger than 10 micrometers. Ballistic transport is a mode of operation in which charge carrier, such as electrons, basically do not collide with other electrons nor with the material crystal lattice, and hence experience no loss of energy to heat. The measured conductance surpassed the theoretical predictions for perfect graphene by nearly ten times.


The ballistic transport is enabled by the high quality of the epitaxially grown nanoribbons, which in this particular geometry forces the electrons to behave more like waves in an optical waveguide than like particles in a stream. Other than possibly being used as interconnects in traditional ICs, such wavelike behavior of electrons opens up the possibility to explore options offered by quantum mechanics, with the prospects of radically new (and much faster) types of electronic circuits than the ones we are used to. One challenge for this new technology will be to move the graphene from expensive silicon carbide, on which it is grown, to more comfortable substrates such as silicon and silicon-oxide.


In conclusion, although the concept of using graphene in interconnects comes naturally due to the material's small volume and excellent electrical and thermal properties, real-life research has only recently emerged out of the shadows. Various methods have led to performance that improves on that of copper, and even to radically new technology such as ballistic circuits at room temperature.