Plasmonic components on integrated circuits—in which energy from light is concentrated into surface plasmon polaritons (SPPs), sub-wavelength electromagnetic oscillations that can propagate along a metal-dielectric interface—have significant promise for enabling large-scale integration in nanoscale optoelectronic chips and devices. That’s because SPPs offer the potential for breaking the diffraction limit imposed by the micrometer-scale wavelength of light in conventional waveguides, and allowing for the nanometer-scale integration common in electronic chips.
But there’s a catch: SPP propagation requires a metal interface, and that means that the electric field attenuates quickly through absorption in the metal—dropping off, according to Fedyanin, a billion times at distances of around a millimeter. And, while it’s possible to compensate for these losses by pumping additional energy into the system, the optical pumping schemes demonstrated thus far to do so have required a large, impractical energy input.
Fedyanin and his team looked at an alternative approach for overcoming surface plasmon energy loss: pumping energy into the system not optically, but electrically. Specifically, the scheme involves inserting a very thin layer of insulating material between the metal-semiconductor interface in the plasmonic waveguide. The insulating layer serves a dual purpose, allowing efficient injection of minority carriers (electrons, in this case) while blocking majority-carrier (hole) current.
The added insulating layer helps to suppress leakage current and ohmic losses in the metal layer. And, when a forward bias voltage is applied, it allows a sufficient concentration of electrons near the semiconductor-insulator-metal interface to create a population inversion in the semiconductor and provide optical gain for the plasmonic mode propagating in the waveguide—amplification that compensates for SPP propagation losses.
In numerical models of the geometry, using a hypothetical system with gold as the metal layer, hafnium dioxide as the insulator, and the p-type semiconductor indium arsenic, the team calculated that the system could fully compensate for SPP propagation losses “at a current density of only 2.6 kA/cm2.” Replacing gold with copper, which significantly increases the minority-carrier injection efficiency, dropped the required current density to 0.8 kA/cm2. “Such an exceptionally low value,” the study concludes, “demonstrates the potential of electrically pumped active plasmonic waveguides and plasmonic nanolasers for future high-density photonic integrated circuits.”
Stewart Wills