Illustration of the integrated laser platform created by the Lipson Nanophotonics Group, where a single chip generates narrow linewidth and tunable visible light covering all colors. (Image: Myles Marshall/Columbia Engineering)
Tech Briefs: What are some applications that would benefit from a very small, chip-level laser?

Mateus Corato Zanarella: The applications depend on the wavelength of the emitted light. We made lasers at visible wavelengths from near-ultraviolet, which is about 400 nanometers, up to near infrared (NIR), which is about 780 nanometers, covering blue, green, and red in between. There are several applications for each of these wavelengths. The most notable is quantum optics, which can leverage the narrow linewidth and precise tunability of the lasers we created.

Tech Briefs: What is linewidth?

Zanarella: An ideal laser would emit a single wavelength with zero-linewidth, basically the purest color possible. In practice, a laser always has some linewidth, a small range of frequencies around the center frequency of the light. That’s due to the fluctuations of the wavelength of the laser. So, narrow linewidth basically means very pure color.

Tech Briefs: Could you tell me something about quantum optics?

Zanarella: Yes, it’s a huge field of research with many fronts, and one of the common approaches to realizing quantum bits is to use atoms or ions. The researchers trap and manipulate the species using light, by targeting very specific atomic transitions. The lasers need to have very narrow linewidths because of the narrow linewidths of the transitions themselves. They typically use multiple lasers of different colors, but each one very pure and precisely controllable.

Atomic clocks are another example of a similar process.

Tech Briefs: What is quantum optics most used for?

Zanarella: A lot of it has been research, but perhaps the most notable application is quantum computing. Recently there have been lots of startups aiming for making optical quantum computing a reality using different approaches, some of which interface light with atoms or ions like I mentioned earlier. The specific requirements for the laser sources are different for each particular system, but this is overall the application that requires the most high-performance lasers.

Tech Briefs: What are other applications of visible lasers?

Zanarella: Laser displays are one example. If you have lasers that are pure and spanning wavelengths around red, green, and blue, you can create displays with a wide color gamut and enhanced image quality. And miniaturization is important to enable very compact devices, say for a laser scanning display for AR or VR goggles.

Bioimaging is another application. Fluorescence imaging uses lasers to excite fluorophores and then collects the light to image tissues and other biological samples.

In neuroscience, there are neural probes that use light to control and read the response of neurons in the brain. They use a technique called optogenetics, in which the neurons are modified to produce proteins called opsins that react to visible light, usually blue. The probes then use laser light to excite and control the response of neurons. So, by having compact sources like the ones we created, one can shrink these systems and make very compact and complex probes.

Another application is LiDAR, which typically uses near infrared wavelengths because we can’t see them with our eyes. But if you want to do LiDAR underwater, you can't use NIR light because the water absorbs it all. Its minimal absorption turns out to be in the green and blue wavelength ranges. So, you want to use lasers at those wavelengths, and if you want to do frequency modulated continuous wave (FMCW) LiDAR, you need to use fast tunable lasers like the ones we created.

One other application is visible light communication. Similar to telecommunications, in which we have infrared light being modulated and manipulated to transfer data, you can do the same with visible light.

Tech Briefs: You say that you can tune these lasers with a ring resonator — could you explain how that works.

Zanarella: The ring resonator is dynamically tunable — it has a micro-heater on top of it. When we apply voltage to it, it heats up locally and changes the refractive index of the material of the ring, which in turn changes the wavelength of the laser.

Tech Briefs: Would that be a fixed setting, or would you use it for modulation?

Zanarella: You can do both. You can use it to tune the wavelength to a specific point that you want and then keep it there, or you can keep modulating it. For example, if you are doing quantum optics and you want a laser that's precisely aligned to an atomic transition, you would use that micro-heater to park the laser at that wavelength. But if you're doing underwater LiDAR and want to modulate the optical frequency of the light, you would dynamically modulate the micro-heater. So, you can do either, it just depends on your application.

One of the interesting things about what we did is that you can tune the resonator at high speed, which allows you to control the laser very quickly. That is important in some applications, such as LiDAR and quantum optics.

Tech Briefs: What is the order of magnitude of your fast modulation?

Zanarella: We obtained modulation rates of up to 267 petahertz per second using the laser diode current and 7 petahertz per second using the micro-heater. The 3 dB bandwidth of the heater is 32 kilohertz, which may not sound like much, but the resulting modulation of the optical frequency of the laser is really fast.

Tech Briefs: Can you give me a quick beginner’s class in integrated photonics. So, you start with the laser and that's your big contribution. Laser on a chip is that it?

Zanarella: In this work, yes.

Tech Briefs: So, the laser generates the light, which gets fed into a waveguide. What exactly is a waveguide?

Zanarella: You can think of it as kind of a micro- or even nano-scale version of an optical fiber. In our case it's made of silicon nitride — a dielectric used in the semiconductor industry. The silicon nitride is a material with a refractive index of about two, and we surround it by silicon oxide, with index of about 1.4. The higher index of the silicon nitride confines the light in the same way that the core of an optical fiber would, but on the sub-micrometer size range. So, you can think of our waveguide as a miniaturized optical fiber — chip scale and different material, but same idea.

Tech Briefs: Could you discuss the ring resonator.

Zanarella: You can think of the ring resonator as a waveguide with the termination connected to the input, making a loop. Using the fiber analogy again, it would be the same as getting the two ends of the fiber and connecting them.

Tech Briefs: How do you couple from the waveguide into the resonator?

Zanarella: It’s through what we call evanescent coupling. Even though the light is confined in the silicon nitride core of the waveguide, it leaks out a little — it has an exponential decay outside of the core. If you put two waveguides close enough together, the leakage is strong enough to couple them. So, you're able to transfer your light from one to another just by bringing them really close together — a few hundreds of nanometers is enough.

Tech Briefs: What’s the order of magnitude of the size of these components?

Zanarella: I change the waveguide width depending on the part of the chip to achieve different functions, but overall, about 300 nanometers in width and 175 nanometers in height.

Tech Briefs: How big is the chip?

Zanarella: The chip is about a millimeter long, and since I have different components for different lasers on the same chip, it is several millimeters wide. But the portion that has just the components for one chip-scale laser system is less than half a millimeter wide. The laser diode itself is sub-millimeter, so the overall system with the laser and the chip is on the order of a millimeter long and sub-millimeter wide.

Tech Briefs: Would your laser be less expensive than already existing lasers?

Zanarella: Yes — standard commercially available tunable and narrow linewidth lasers are benchtop, say about a foot long or more. They are big free-space external cavity lasers — we're replacing all that with a little chip. The typical benchtop lasers cost tens of thousands of dollars, whereas our laser diodes cost 10s of dollars — three orders of magnitude less.

When you go to semiconductors and chips, if you mass produce them, they get really inexpensive and that's a big advantage.

Tech Briefs: So, would your chip have to be mass produced to reach that kind of level?

Zanarella: Not really, we already have that kind of level now. A single laser diode costs tens of dollars from a vendor, likely much less if you can work directly with the manufacturer. Then for the chip itself, we pay to use the tools in the cleanroom to fabricate it, but we make about ten thousand or more of them on a single 4-inch wafer. And this is not even mass production, it’s just academic fabrication, but one of those chips already costs less than the laser diode itself.

Tech Briefs: So where are you at now in terms of this work — what are your next steps?

Zanarella: Our idea for the next step is to optically and electrically package the chip-scale lasers to turn them into portable standalone units that can be easily deployed to different applications.