Hector De Los Santos, Principal Investigator, Research Computing
Now for more than 10 years, he has toiled on the cutting edge of a new way to use the manipulation of plasmons—collective oscillations of free electron density—to do computations. His journey began back in 2009. At that point he despaired on the realization that traditional CMOS logic devices, reliant on current flows, were severely limited. De Los Santos is hopeful that energy-efficient computing will be the standard one day through the use of this new technique.
For 2024, De Los Santos joined forces with their fellow collaborators at the University of South Carolina, Ohio State University, and Georgia Institute of Technology. Through collaboration, they effectively demonstrated a device that exposes the essential features of plasmon-based logic. This announcement is an important breakthrough for his multi-year effort. Ultimately he aims to address the escalating energy loss problems ushered in by the ever diminishing dimensions of electronic devices.
De Los Santos is clear on why it’s urgent to find solutions to these challenges. He points out that with each technological advancement, devices are getting tinier and tinier. This miniaturization creates an associated increase in energy dissipation, which poses significant challenges for conventional computing approaches. The new approach, based on wave flows instead of current flows, aims to go beyond this to change the very physics underlying the principles of computing.
The Challenges of Conventional Logic Devices
Industry analyst De Los Santos points out an emerging key hurdle. Existing technologies fail to fit cleanly into his aspirational vision. He adds, “The biggest hurdle is the technology; it’s just not consistent with where the paradigm is today with logic devices and current flows. This is based on wave flows.” This paradigm change in thinking can be hard for people used to current technologies.
He continues below to explain more of the shortcomings of classic CMOS logic. As he explains, “The idea of plasmon computing came to me around 2009. I saw where the whole field of CMOS logic was heading. Additionally, as devices scale down, quantum mechanical effects as well as short-channel effects increase power dissipation, posing a challenge to conventional approaches.
According to De Los Santos, most specialists do not have the cross-disciplinary experience needed to understand the complexities associated with plasmon-based devices. He notes that no single individual usually possesses all the information required to fully inform users about the device. This points to the astonishing breadth of expertise needed, from metal-oxide-semiconductor physics to cancellations of electromagnetic waves and quantum field theory.
The Breakthrough Device
De Los Santos and his collaborators invented a game-changing prototype. It is a Y junction device with a size of about five square microns. This new device was created to control plasmons using an elaborate structure. In the Y junction, a metal layer sits atop an oxide layer. This arrangement is located on top of a semiconducting wafer and is placed above a ground plane underneath.
De Los Santos produces the relatively static sea of electrons to enable plasmon displacement. He accomplishes this by applying a DC voltage between the metal of the Y junction and the ground plane of his device. This unique structural conformation facilitates the generation of highly localized electron charge density disturbances when optically excited by incoming electromagnetic waves.
“You pattern long, thin wires in a configuration in the shape of the letter Y. At the base of the Y you launch a plasmon. Call this the bias plasmon; this is the bit,” explains De Los Santos.
When the bias plasmon hits the Y-shaped junction, it nonchalantly bifurcates into two identical electric field strengths. That’s the case unless someone hacks the process in a separate way. De Los Santos adds another control plasmon along an angled wire. To encode binary information, he uses this very same technique to redirect the bias plasmon into one leg of the Y junction.
“Now, imagine that at the Y junction you apply another wire at an angle to the incoming wire,” he continues. “You can use the control plasmon to redirect the original bias plasmon into one leg of the Y.”
This new technique opens the door to manipulating plasmons in exciting new ways. De Los Santos points out that this milestone is very important for creating reversible computing devices.
A Vision for Reversible Computing
Ultimately, de Los Santos wants to develop a sleek, intuitive device that only has two controls. This device will additionally show reversible computing abilities. Significant proposals often fall flat, he contends, because the old ways are predestined to fail. He further notes that they are thermodynamically irreversible, resulting in the loss of information and energy throughout the course of a computation.
His vision for plasmon computing is deeply rooted in its energy efficiency potential. In contrast to conventional computing devices, where energy is lost as heat during computation, plasmons can perform operations with almost no energy waste.
“Going back to the analogy of throwing a pebble on the pond: It takes very low energy to create this kind of disturbance. The energy to excite a plasmon is on the order of attoJoules or less,” De Los Santos details. This intrinsic efficiency positions plasmon-based computing as a promising alternative to existing technologies.
As he works towards fabricating a complete device, De Los Santos is focused on demonstrating its potential by concatenating them to create foundational components like full adders integral to computing logic.