Dr. Babak Nikoobakht and a team at the National Institute of Standards and Technology (NIST) have developed a nanoLED with up to 1,000 times the brightness of conventional submicron-sized LEDs. They overcame “efficiency droop” — the drop-off of brightness after current increases past a certain point.

Tech Briefs: How does the light output of your LED compare with others in the same size category?

Dr. Babak Nikoobakht: The base area of our LED is less than 1 mm2 so it falls in the class of sub-micron LEDs. A typical sub-micron LED emits about 22 nanowatts of power but ours can produce up to 20 microwatts.

Tech Briefs: How did you obtain such high brightness?

Dr. Nikoobakht: We used a new architecture for making the LEDs, but with the same materials as in conventional LEDs. The difference in ours is the shape. Unlike the flat, planar design used in conventional LEDs, we built a light source out of long, thin zinc oxide strands we refer to as fins. For typical LEDs, increasing the current increases brightness, but only up to a certain point. After that, the brightness drops off as the current continues to increase. This phenomenon is called “efficiency droop.”

Tech Briefs: How do fins allow you to overcome efficiency droop?

Dr. Nikoobakht: Fins have two distinct large side facets. In trying to figure out how to bring n-metal contact to fins we saw two options: to its top facet or its side facet. Large side facets are appealing because they offer a larger metal-semiconductor interface area, thus a larger area for current injection. It is the larger facet that allows more effective current injection to a nanocrystal.

Both our experimental and theorical results show that a fin shape is much more effective in collecting the holes from the p-GaN side of the heterojunction (relative to a planar LED). This allows a larger fraction of injected electrons and holes to recombine in the fin (and not the p-GaN side). We also show that the fin shape allows a lower non-radiative Auger recombination rate compared to other LED designs. This all means a lower optical loss in the fin cavity.

Typically, when LEDs (both conventional and nanoLEDs) are pumped harder with electrons, the rate of non-radiative decay of e-h pairs also increases. We found that for a fin-shaped geometry, that is not the case. As we inject more electrons, the non-radiative decay saturates and no longer increases. This results in brighter LEDs and more efficiency at high injection currents.

Tech Briefs: What made you think about following this line of research in the first place?

Dr. Nikoobakht: This goes back to 2002, when I began my research at NIST — you know how research is, you never know what you end up with. At that time there was a lot of promise around nanowires, including LEDs and waveguides for photonics. For the following ten years, people demonstrated the capabilities of nanowires. In our lab, we developed a technique that gave us lateral wires. We decided to study them, as we saw potential in their predictable directionality. While working on them, by accident, we found that we could grow them into fins. Later we developed a technique to control their height that could extend to 1 or 2 microns (1000 to 2000 nanometers). It became like a wall, which some people call a nanowall.

A fin LED pixel includes nitrogen (N) doped zinc oxide. (Image Credit: B. Nikoobakht, N. Hanacek/NIST)

Tech Briefs: Would that be 2D?

Dr. Nikoobakht: From a materials science perspective, you can call that 2D, even though it’s a 3D object. But the point is that it has two large dimensions: length and height, but its width is around 100 nanometers, more or less. So, sometimes we call these 2D structures or quasi-1D structures.

We started looking at the different properties of these structures. One of the questions we wanted to address was: how are they emitting light as an LED? we were interested in making lasers for on-chip applications, since currently there are no light sources available in the UV and visible light range for on-chip applications. Everything is pretty much in the telecom industry wavelengths of 950 nm to 1.5 microns — near-infrared.

So, our primary goal was to get our fins to light up for on-chip applications to be able to detect molecules and do chemical spectroscopy. We saw these large side facets, but it was difficult to get access to them because they are so tiny. It took us a couple of years to find a way to get a metal contact exactly to the side of the fin and nowhere else, because we didn’t want to have a random process, we wanted it to be very deterministic. Finally achieving that was our first major breakthrough.

We realized that they were emitting light that was fairly bright. But that was just by using our eyes, looking under a microscope. Then by doing quantitative measurements, we looked at their efficiency as we were pumping them with more electrons. We wanted to see at what point they would start to dim —begin losing their output power, and also, at what point their structure would start to melt. To answer these questions, we did a set of careful measurements and realized that these LEDs were droop-free. As we loaded them with more and more electrons, they kept getting brighter and brighter. After verifying the droop-free behavior by using external efficiency measurements we developed a theory for that behavior. And the theory matched what we observed in our experiment, which was very exciting. So, from a fluke, it started to become something real. That, in a nutshell is the story of the past 15 years.

Then, we made another surprising discovery as we increased the current. While the LED shone in a comparatively broad range of wavelengths at first, it eventually narrowed to two wavelengths of intense violet color. The explanation grew clear: Our LED had become a tiny laser. Converting an LED into a laser usually takes a large effort. It requires coupling a LED to a resonance cavity that lets the light bounce around to make a laser, but our fin design can do the whole job on its own, without needing to add another cavity. We expect that these lasers can be used for on-chip applications.

Tech Briefs: Could you give me an idea of some applications for on-chip lasers?

Dr. Nikoobakht: Laser light from wrist band sensors can examine physical properties of skin, like blood oxygen and heart rate or even the level of skin hydration/dryness. It can measure properties of the skin or right underneath the skin.

But an important application for on-chip lasers is analytical spectroscopy — the interaction of laser light with molecules. Our light source fits well with this kind of platform. Although people have been working on this for many years, most of this kind of chemical interrogation of materials has been done in the near IR because of its development in the telecom industry. But the best region for spectroscopy is in the visible and UV ranges because of the rich electronic transitions that you don’t see in the near IR. The metrology community is also interested in this because, after all, measurement science is a key part of it. There’s a great deal of unknown science here that needs to be developed for these kinds of applications.

Tech Briefs: What are the advantages of developing chip-size optical microscopes? Why are they better than larger ones?

Dr. Nikoobakht: For one thing, larger ones are expensive and bulky. And the throughput, the volume of measurements you can do with an optical microscope, is low. Although these instruments are very accurate, you can’t do 100s of measurements per day. If you want to do a DNA analysis, for instance, it would take hours. So, the idea is to speed up the measurements, increase accuracy and precision, and reduce cost.

Tech Briefs: Are there other possible applications for these LEDs?

Dr. Nikoobakht: Their two-dimensional arrays could make them useful for smart lighting and optical data communication.

Tech Briefs: How close are you to developing a working prototype that can be used outside of the lab?

Dr. Nikoobakht: I think it depends on the application. Generally speaking, there are still some engineering aspects that should be explored and tuned, such as lifetime and stability.

Tech Briefs: What are your next steps?

Dr. Nikoobakht: Exploring other materials for wavelength tunability and integrating fin light sources into photonic circuitries.

An edited version of this interview appeared in the October 2020 issue of Tech Briefs.