Connect the Dots

April 14, 2022, the day when Los Alamos postdoc Namyoung Ahn made the latest contribution to a discovery that could someday soon shake up the entire computing industry, was one of those rare spring days in Los Alamos that are clear, cool, and windless. The 12,000-foot peaks of the Sangre de Cristos, snow covered and visible out Ahn’s office window, looked close enough to touch. Not that he ever saw them. Ahn was where he has spent the bulk of his time so far at Los Alamos: in a dark and windowless lab abuzz with the electrical hum of nitrogen-filled gloveboxes and machines customized to manipulate nanoparticles. He was staring at a monitor connected to a Wheat Thin-sized chip of glass and feeling glum. “Like I’d done all I could to contribute to the project and had nothing left to offer,” he says. Regardless, Ahn did that afternoon as he had done thousands of times before and tried again. He sent electricity coursing through a device he and a team of colleagues had built over the past 20-plus years that could one day turn quantum dots, synthetic nanoparticles about 1/10,000th the thickness of a human hair, into a laser capable of communicating information between microelectronic devices. This time, Ahn saw a tiny spot on the glass wafer glow a brilliant red.

By that point, Ahn had just started year three of a three-year postdoc appointment under Victor Klimov, a Los Alamos scientist who has worked on quantum dots since the late 1980s. Klimov wanted to harness the dots’ unique quantum properties to make electrically driven lasers that could be fabricated cheaply using simple benchtop chemistry and could sit on any substrate, including silicon. If realized, such devices could send and manipulate information using light alone. The idea could be a boost to microelectronics and the 515 billion-dollar annual silicon chip industry. Phones, computers, weapons, manufacturing, transportation, communication—silicon chips are the beating heart of very nearly all modern technologies, and the industry is forecast to crest a trillion dollars by decade’s end. Electrically driven lasing quantum dots could revolutionize all of it. 

If all that sounds too good to be true, by April, Ahn, who came to Los Alamos from his native South Korea, had come to agree. At his feet sat two cardboard boxes filled with several thousand spent glass shards he and three colleagues had imprinted with quantum dots and layers of various other materials over the years. Each shard took at least a day to build, and not one had produced the desired results. Ahn watched a monitor as the line indicating the photon flux emitted from one edge of the tiny device climbed steeply before falling, tracing a shape like a shark’s tooth which he had seen many times before. But then the line did something new. It climbed much higher, saturating his detector. Light amplification. Light amplification! Inside the device, persnickety and meticulously assembled layers of atomically thin materials had taken electrical pulses and converted them into photons that were now multiplying. Ahn’s device was lasing!

Lasing quantum dots could revolutionize the silicon chip industry

He printed the results and walk-ran the two blocks to Klimov’s second-floor office. Ahn gave a cursory knock on Klimov’s door but didn’t wait for an answer. “It lased,” he said. Klimov said nothing but grabbed the paper and used a ruler to measure the difference between the two peaks. “Calm down, Namyoung,” Klimov said, his Russian accent thinned but still apparent after decades in the U.S. Typically quiet and mild mannered, Ahn took a deep breath.

“Now,” Klimov asked, “Where do you want to publish this: Nature or Science?”

The promise

Quantum dots are nanoparticles made up of somewhere between 100 and a few hundred thousand atoms of a semiconducting material, such as cadmium selenide. Their allure to science lies in their size. They’re essentially zero dimensional—so tiny that they exhibit quantum behavior. Electrons are tightly confined in the material such that they can only occupy atomic-like discrete energy levels. When electrons in a quantum dot are stimulated either by light or electricity, they move from the valence band, where the electrons sit immobile when the material is unexcited, into the conduction band, where they can move freely through the material. The vacancy the electron leaves behind in the valence band is called a hole. The electron and the hole together form an electron-hole pair—an essential component of quantum dots’ great potential. When the electrons eventually snap back to their holes in the valence band, they release energy as either heat or light.

Scientists realized soon after quantum dots were discovered in Russia in the early 1980s that the nanoparticles could easily be tuned to emit any color of light by changing their material composition and size. Smaller dots emit blue light, larger dots shift red, and modern chemical techniques allow for scientists to adjust the color of the light emitted with nanometer precision. Over the past two decades, scientists have harnessed quantum dots’ potential for everything from solar windows—transparent solar cells that can be applied to glass—to biomarkers that, when paired with antibodies, can tag and illuminate individual cancer cells in vivo. Major electronics manufacturers now layer all different sizes of quantum dots over LED (light-emitting diode) panel screens to create displays with unmatched color purity. “Quantum dots are bridging the gap between virtual reality and reality,” says Ahn. They are making the digital world look deceptively like the real world.

But for a certain group of quantum dot researchers, of which Victor Klimov is a leading voice, the dots’ greatest allure lies in their potential to lase. Quantum dots have long presented an ideal medium for exploiting Einstein’s law of stimulated emission: when an inverted medium, wherein there are more excited atoms than unexcited atoms, is stimulated by an external photon, it will duplicate that photon. If scientists could build a device around quantum dots that would capture and circulate the photons emitted by the inverted medium, they could produce laser oscillations, which in turn could produce an intense, highly directional, narrow-band beam of light.

Klimov first fell for quantum dots during his time at Moscow State University. Back around 1990, he and his colleagues started posing a series of questions, simple at first glance, which over the coming decades would plunge them into a world of enormous complexity: How does lasing work in these tiny particles? What other processes compete with lasing? Is it even possible to make a lasing device? During that first project, the scientists achieved the lasing effect with an early generation of quantum dots: nanocrystals dispersed in a glass matrix. (These samples, pioneered by Alexey Ekimov, were the first example of human-made materials used to gain initial insights into the properties of quantum-dot matter). But that leap in understanding was made during perestroika, the restructuring of society and economy that followed the end of communism in the Soviet Union, when science funding was extremely scarce. To be able to continue his research, Klimov moved to Germany, before landing in Los Alamos in 1995 to study fullerenes and polymers during work hours and quantum dots “after work and on the weekends,” he says. By the end of the 1990s, he had returned to the idea of a quantum dot laser, only this time, he wanted to achieve lasing with colloidal quantum dots, a new-at-the-time kind of dot whose chemical synthesis made it possible to brew the dots quickly, cheaply, and by the bucketload. The new dots had been recently developed by Louis Brus, from Bell Laboratories at the time, and later perfected by Massachusetts Institute of Technology (MIT) professor Moungi Bawendi—advances that earned them and Ekimov this year’s Nobel Prize in Chemistry for the discovery and synthesis of quantum dots.

“Quantum dots are bridging the gap between virtual reality and reality.”

“Here’s a piece of the material our first quantum dot laser was built from,” said Klimov one August day in his office, a west-facing room with views of the Jemez Mountains. He pulled from his bookshelf a small cardboard box. Carefully nested within was a piece of glass graded from yellow to red and shaped like an oversized candy corn. It contained the early generation quantum dots that Klimov first worked with in Russia in the 1990s. “I was immediately fascinated with how these particles behave,” Klimov said, shining a flashlight on the glass. He watched it with the attention of an archaeologist studying a pot shard from a bygone culture. It glowed faintly. “Isn’t that cool?” he asked.

It wasn’t until the 2010s that the knowledge and technology surrounding quantum dots advanced enough that Klimov began to prioritize getting them to lase through electrical excitation, as opposed to optical excitation, which uses light to drive lasing. By then, the semiconductor chip industry had matured from 139 billion dollars in 2001, to 239 billion dollars and was still growing fast. Klimov, who had spent the decade prior optimizing quantum dots for displays and solar cells, came to realize that electrically stimulated dots were a technology more powerful that almost all others in his field. These tiny lasers could crack open the potential of optical computing, a branch of computing distinct from either quantum or traditional computing. First theorized in the 1960s, it works by replacing existing computer components with optical equivalents that, until recently, didn’t exist in broadly usable forms. The possible applications are both large and small. Dots can be produced chemically in batches of trillions for relatively cheap. Optical networks powered by quantum dot lasers wouldn’t produce heat, unlike traditional electric networks, so a data center powered by them wouldn’t need elaborate and expensive cooling equipment to regulate temperatures in server banks. The heat that large data centers generate is detectable from satellites. And because they transmit information in a near-infinite array of colors, as opposed to the 1s and 0s of conventional transistors, dots hold the potential to expand into almost unfathomable complexity the types of questions computers can answer.

“It’s information sent at the ultimate speed—the speed of light,” Klimov explained. “The potential is tremendous.”

A photon avalanche

On a sweltering August afternoon, far away from sunlight, Ahn heads into a lab where quantum dots are synthesized. Ahn, who is soft spoken with dark hair and stylish, plastic-framed glasses, walks through an unlit maze-like corridor, through a room filled with large humming boxes (“No idea what those are,” he says.), and through a series of sliding doors to a brightly lit lab. “We use less than a dollar’s worth of quantum dots for about 20 devices,” he says, proudly shaking a vial of bright red dots recently brewed by Donghyo Hahm, a chemist in Klimov’s group. Getting to this point has been a long road for a big team, and Ahn sees himself as having added just one rung to a ladder that Los Alamos scientists have been building for decades.

For Klimov, building that ladder at the Lab started in 1997. That year, he began a project to understand the nature of light emission and competing processes—heat emission, for example—in quantum dots, and to investigate an optical-gain mechanism, a requirement to building a quantum dot laser. He focused then on making quantum dots lase through optical excitation. While working on the project, he identified three states of quantum dots that are directly relevant to lasing and would guide his work over the coming decades (see “How Quantum Dots Work” illustration). In the ground state, all electrons are in the valence band. This state does not emit photons but just absorbs them, causing the material they’re contained within to glow, like the candy-corn colored glass shard in Klimov’s office. In the single-exciton state, one electron is promoted to the conduction band. When the promoted electron eventually snaps back to its hole state in the valence band, energy is emitted as a photon. In electronic displays, like TVs or computer monitors, quantum dots in this single-exciton state act as color converters: a blue photon is sent in and a photon of another color comes out.

But it was the third useful state that Klimov identified, the biexciton state, that proved the greatest challenge. In order to amplify light—or get the dots to replicate photons—two electron-hole pairs must exist. Once this biexciton state is created, the quantum dot can start to replicate incident photons through the process of stimulated emission. But this biexciton state is also extremely unstable. Depending on the quantum dot’s size (the number of atoms it comprises), the biexciton state collapses through a nonradiative process called Auger recombination, which occurs on extremely short time scales—from 10 to several hundred picoseconds. One picosecond is a trillionth of a second. That’s far less time than needed for the dot to replicate a photon. Not only that, Klimov found that the biexciton state collapses by emitting heat, not light. “Wasted energy. Exactly what we didn’t want,” says Klimov.

Despite that challenge, by 2000, Klimov and his team had found a workaround to Auger recombination. Not unlike a chain reaction, in the regime of stimulated emission, when a stimulated dot releases a replicated photon, that replicated photon then stimulates a nearby dot, which then releases another replicated photon that stimulates the dot next to it, and so on through the medium. What Klimov’s team came to understand was that the more dots that are packed together, and the closer they are, the faster the photon avalanche. The trick was to get such a high density of dots that the photon avalanche could outpace Auger recombination.

At the time, Klimov got his dots from Bawendi’s group at MIT. The samples arrived in Los Alamos in dilute solutions, where the dots were relatively sparsely distributed. To concentrate the dots, the scientists assembled them into a thin, high-density, solid-state film on top of a glass substrate. They then excited this layer with extremely short pulses of light, just 100 femtoseconds (that’s one tenth of one trillionth of one second), and by doing so, achieved a very quick injection of biexcitons across the quantum dot layer before Auger decay set in. The result was lasing through optical excitation. Again, light—not electricity. “These studies laid the foundation for the entire field of colloidal quantum dot lasing and became part of textbooks on colloidal nanomaterials,” says Klimov.

But the breakthrough also had another effect. While the advances pioneered during the optical lasing project would prove integral to Klimov’s eventual success building an electrically excited quantum dot laser diode, they also spelled out why it was going to be a much harder feat. Electrical excitation is a slower process. The charge carriers—the electrons and the holes—arrive one by one, typically through a device that first injects an electron and then a hole. Generating a biexciton state requires a sequence of four arrivals, each one competing against the clock with Auger recombination. “Achieving lasing from electrical excitation is virtually impossible unless Auger recombination is tamed,” Klimov concluded after 2000. The team was now in a race against a process that lasted not much longer than a trillionth of a second.

New dots

Klimov picked up a pencil and used it to scratch out on printer paper a series of diagrams that, for the uninitiated, might best be called How Quantum Dots Work and Why Getting Them to Lase Through Electrical Stimulation is Really Hard. “Rectangular potential, right?” he asked, drawing the physical shape of the well that confines electrons in standard quantum dots. It looked like two opposed flat-bottomed Us with their bases separated by a thin gap. The U on the bottom represented the valence band; the U on the top represented the conduction band; open circles represented holes, filled in circles represented electrons; and arrows depicted the particles’ movements. “What happens if you change the shape of this well and use a parabolic shape instead?” asked Klimov, drawing a similar diagram but this time drawing the Us in gentle bowl shapes. Klimov called the curve “smoothed.”

Up until around 2010, the cores of most quantum dots had been made with a single material, cadmium selenide, which creates the hard-edged rectangular well shape that Klimov drew first. But that year, Alexander Efros, a theorist at the Naval Research Laboratory, published a paper that analyzed how the shape of the well—the walls of the quantum dot that confine the electrons—affects the rate of Auger recombination. Efros saw opportunity in particles’ fundamental behavior. Particles don’t like to move between disparate states. By making the wells with different materials, or grading the composition of a quantum dot’s interior, Efros theorized it might be possible to create a starker difference between a particle’s states and thus delay Auger recombination. The question was, according to Klimov, “How to do it chemically?”

Two of Klimov’s other postdocs, first Wan Ki Bae and then Jaehoon Lim, started answering that question in 2013. A few years and many experiments later, they landed on a new design they called alloyed quantum dots. Bae and Lim retained the cadmium selenide core used in most quantum dots but opted to surround it with a cadmium sulfide shell. Compared to cadmium selenide, cadmium sulfide has a wider band gap—it requires more energy to move an electron from its valence band to the bottom of its conduction band. The transition between the two materials would be a continuously graded cadmium-selenide-sulfide alloy, with pure cadmium selenide at the core gradually transitioning to pure cadmium sulfide on the periphery. 

“That reduces the overlap between wave functions,” Klimov said, tapping his page and changing to explanation through quantum mechanics, the only language capable of accurately describing the process. The wave function is related to the likelihood of a particle, in this case an electron, being at a given point in space and time. With continuously graded quantum dots, the idea was that electrons could slide up the flattened curve of the gradually widening band gap, increasing the likelihood that the holes and the electrons would be in a more disparate state, thereby delaying Auger recombination. 

By 2018, the alloyed dots worked, with Bae and Lim showing, as the theory predicted, that they could suppress Auger decay. Over the next few years, work continued on the project, with Lim testing out various materials before landing on a design of a cadmium selenide core that transitioned, through a cadmium-zinc-selenide alloy, to a zinc selenide shell.

“One day in 2018, Jaehoon brought me this,” said Klimov, as his pencil got back to work on a new blank page. “And it was pretty cool.” The new graph he drew showed a simple parabola, but as Klimov drew the shape’s downward sweep, he ticked his pencil upward. “This little, little, little, thing here,” he said, tapping the blip. “And we were jumping—both of us—in my office.” What the blip told Lim and Klimov was that these new continuously graded dots had emitted a photon with energy that was higher than that usually produced by quantum dot LEDs, a proven technology at the time. This meant two things. This photon could only be emitted if the biexciton state, containing two electrons in the lowest energy state, was long-lived enough that an electron could be injected into an even higher energy quantized state. Under the right conditions, the dot could replicate the incident photon. In the new continuously graded quantum dots, Auger recombination was delayed ten-fold. It wasn’t lasing, but it was a critical step toward lasing.

“This told us…wow,” said Klimov, putting down his pencil and letting the moment hang. “If you’d asked me in 2000 whether I thought it was possible to get quantum dots to lase through electrical excitation, I would have said absolutely not. This convinced me that we could make it happen.”

Focusing the current

In the years after developing continuously graded quantum dots, a new device began to grow up around the dots. For electrical excitation, Klimov knew they needed quantum dots in a biexciton state, and to excite biexcitons, they needed an LED-like device, with an active layer of quantum dots sandwiched between a layer that injects electrons and a layer that injects holes, which could produce very high current densities. Klimov calculated they needed around 100 ampere per square centimeter, or almost 100 times more than what standard quantum dot LEDs use. Standard LED devices would overheat at around just one ampere per square centimeter. Klimov needed a new electrical design.

“If you had asked me five years ago whether this was possible, i would have said ‘absolutely not.’”

A happy accident provided a lead. Back in 2018, Lim was testing quantum dots in devices he’d built, and did as he’d done many times before and attached two electrodes to the device. But in one test, one of the electrodes was slightly cockeyed, so the injection area was reduced. “It was an accident. He saw a brighter emission because of some kind of misalignment in the device,” said Klimov, indicating on his sketch that electrons had moved uncharacteristically high in the conduction band. At first, nobody understood why. “Everything in steps, you know? You don’t just throw away the data you don’t understand, you try to understand them.”

Soon enough, it became clear that the accidental success owed to how the cockeyed electrode focused and densified the electric current. To mimic the results, Lim added to his device a dielectric that would act as an insulator, a layer of lithium fluoride between a silver hole-injecting layer and the quantum dots. But this time, Lim added a slit to the middle of the insulating layer that would “boost the density” of the electricity flowing into the quantum dots. To make the slit, he plucked a hair from his head and laid it onto the device to create a shadow mask. He then used a special machine that, through evaporative processes, laid down a thin layer of lithium fluoride onto both sides of the plucked hair. “That’s real life,” said Klimov. “I’m not kidding.” When Lim attached the electrodes to it this time, the electricity flowing into the device squeezed through the slit left behind by the shadow mask of hair. Lim measured 20 ampere per square centimeter, enough to produce optical gain (that is, duplicate photons), but still far below the 100 ampere Klimov had calculated they needed to get the dots to lase.

The second generation of current-focusing devices built on Lim’s success, and leaned on ideas developed by Jeongkyun Roh, another of Klimov’s postdocs. This version was designed over two tedious years, between 2019 and 2021, by still another postdoc, Heeyoung Jung. To further enhance current density, Jung prepared his device with the top electrode as a narrow strip that sat at a right angle to the hair-width slit in the lithium fluoride insulator, narrowing the intersectional area and focusing the current into less than two hundredths of a square millimeter. Jung also tweaked the electrical pumping regime. Instead of pumping the electricity continuously, he pulsed it: one-microsecond pulses separated by relatively long millisecond periods that gave the dots a chance to cool. With the addition of what amounted to an aperture that focused the current, the devices were able to sustain, without damage, staggering current densities of more than 1000 ampere per square centimeter—about 10 times more than needed to achieve lasing.

Pumping such high current densities into the dots provided another advantage: it moved additional electrons higher up the conduction band. “You have to put two electrons in the lowest energy state before you can put any electrons in the higher states,” Klimov explained, pointing, in one of the earlier diagrams on his scratch paper, to the line in the conduction band nearest the conduction band’s edge. But once electrons occupy those two places, as many as six electrons can move above them into a higher energy state, leading, in the right conditions, to the replicating of additional photons. “It allowed for strong optical gain across not just one, as before, but multiple quantized optical transitions of a quantum dot,” said Klimov. Put more simply, they were getting optical gain for photons with a wide range of wavelengths—different colors of emitted light produced by the same dot.

The delivery

The lab where Ahn assembles the devices is separate from where the dots are synthesized. This one, he keeps dark at all hours of the day because external light can pollute his results. “We build them here,” Ahn says, flipping a switch on a glovebox. The gloves attached to the box pooch out like balloon animals. Ahn’s hands go into the gloves, and he grabs a set of tweezers and starts manipulating a small glass chip within the box. “Between two and five thousand?” Ahn says, guessing at the number of devices he has built in this glovebox over the past two years. The postdocs before him built at least that many, likely more. Building and testing the devices has been only half of Ahn’s job. The other half has been modeling the results on his computer to guide his efforts.

“Where’s the light going? That’s what we were asking the models,” he says. By 2021, as many challenges as Ahn’s predecessors and Klimov had solved, the devices still didn’t lase. So Klimov tasked Ahn and two other researchers, Young-Shin Park and Clément Livache, who specialize in spectroscopic characterization, with figuring out why. While Ahn modeled, Park and Livache characterized how light was amplified and where it was being lost. Early in the process, the disappointing result was that most of the light was going everywhere instead of where they wanted it to go: a narrow section of the device containing quantum dots. Because of this, the light emitted by the quantum dots was being absorbed by the device itself. The reason had to do with the device’s construction. Every layer in the sandwich of materials, except for the quantum dot layer, was a conductor of electricity but also a very strong absorber of light. This meant that many of the photons replicated in the quantum dot layer were, as Klimov put it, “being eaten up” by the device’s architecture.

The solution to this latest problem came from the literature when the team found a paper written in 1976. It introduced the idea of a Bragg reflection waveguide, which was, among other things, a way to direct light by elements external to the device. “If you want to do something cool, go find the old stuff,” said Klimov, of the 1976 paper. “No hype, just business.” 

Normally, light is guided through slabs of high-refraction-index materials, like treated glass, by surrounding it with low-index materials, like air or aluminum oxide. After reading the almost 50-year-old paper, the Los Alamos researchers got the idea that they could concentrate the photons where they wanted them—in the quantum dot layer—by building their device on top of a so-called distributed Bragg reflector: a stack of transparent layers, made of two materials, that each bent light in different ways. The approach allowed them to steer photons away from all the optically absorbent charge-conducting layers. With this idea in hand, instead of using a glass substrate as he’d always done, Ahn assembled his device on a stack of 20 dielectric layers of alternating refractive indices. Together with the silver mirror on top of the device, which doubled as a hole-injecting electrode, these layers formed a reflection waveguide that would concentrate light in the quantum dot layer and steer it out the device’s edge. Once the device was constructed last April, Livache took spectroscopic measurements and confirmed for the first time that the optical gain overwhelmed optical losses. Ahn then took the device to his test station.

On that April afternoon, after decades of steady work and irregular progress, dedication, good luck and bad, for a team of a dozen or more scientists who worked with Klimov over multiple decades, their efforts all came together in that device. Ahn observed all the signatures of light amplification, or lasing: an avalanche growth of light intensity, narrowing of the spectrum of emitted light, and its strong polarization. They had produced a proof of principle for solution-processable laser diodes, a result pursued for decades by researchers worldwide.

It was a proof of principle pursued by researchers worldwide for decades.

Ahn, Livache, and Klimov are now well into the process of patenting their design. And they’ve already moved on to the next steps: working to show that their lasing devices can produce single wavelengths of light and that they can operate on silicon chips, as opposed to glass. “That is what the industry wants to see,” says Ahn. He reports some early promising results on both tasks. With the economic and political prominence of silicon chips rising, the timing couldn’t be better. Over the past year, the Biden administration has put more than 50 billion dollars toward the research, development, and manufacture of silicon chips with the 2022 Chips Act. And a recent U.S. trade ruling restricted international access to American-made semiconducting chips, a move that’s likely to produce a surge of interest in alternatives to conventional microelectronic chips. Quantum dot lasers are one of these alternatives. As Ahn sees it, they are the most promising option. He calls them a leap toward a next generation of computers powered by something far faster, far more efficient, far more abundant, and far more powerful than 1s and 0s: light.

“Pretty exciting,” Ahn says, permitting himself a smile as he quietly shuts the lab door behind him, leaving, as he always does, the lights off.