Chief Product Officer at Dispelix, Jussi Rahomäki is an industry visionary in diffractive and nanoscale optics, pursuing technology and business initiatives at the forefront of the augmented reality industry. He has built his strong expertise in augmented reality and nanoscale optics technologies at various business and academic positions over the last decade. Dr. Rahomäki received his B.S., M.S. and Ph.D. in physics from the University of Eastern Finland and holds 14 Publications and 7 Patents.
After Jussi’s presentation at Laser World, we sat down with him for a deeper dive into the technology and design that power Dispelix waveguides.
Q: What is your role at Dispelix and where were you before you joined the team?
A: I'm the Chief Product Officer at Dispelix. I have spent almost 20 years in micro-optics. I started my university studies for micro-optics in early 2000, then moved into academia, and then plastics at Dispelix.
Q: You’re working on some pretty impressive technology at Dispelix. Can you give us more detail?
A: Dispelix is an advanced waveguide designer and manufacturer, delivering next-generation visual solutions for both consumer as well as enterprise AR and MR wearables. At Dispelix, we are building the technology for AR, specifically the near-eye display. If you think about eyeglass lenses, we bring the virtual image functionality in the lens. Our core focus is developing the technology, how to manufacture and get it in a real application.
Q: With so much in the world to see and for us to process, how are you bringing a near-eye display to lenses and how does that work?
A: It all starts with miniaturization, which allows you to fit more transistor nodes onto a smaller integrated circuit. This enables AR devices to display complex images that require a lot of data processing for visualization while also giving the consumer a form factor that they’d want to wear. Without miniaturization you’d have glasses that are quite big and not comfortable for any application – consumer or industrial – because of the weight and clunkiness.
Q: In your opinion, what is the holy grail when it comes to waveguides?
A: The holy grail for us is to make waveguides that help to replicate the look and feel of eyeglasses. For us this means maintaining form while also evolving functionality. How we get there, now that's the big thing. The industry is struggling to solve this problem. It's not only about miniaturization, you want to have excellent image quality at the same time in addition to squeezing down the size.
Q: So how is that accomplished?
A: There are a couple of different approaches to achieving this. Let's start from the display side, where the initial image is created by the reflective panel, such as the Elco, TLP, or micro-LED panel. Then there are the laser beam scanning systems. They have different limitations and different impacts on the end solution.
From our point of view, we are mostly limited by the efficiency with the defective optic system. And that's contributing to what you cannot see from the device or in other ways, what the power consumption of the device is. Then depending on how many pixels we can create from the panel, kind of limited, what kind of field of view we can create, so that the user is still in the angle, you can have enough pixels to create a self-centered user experience. So ultimately, we currently realize 25 degree FOV.
Then how to fit this? How these different types of displays all fit with defractive optics displays. There are different optical architectural challenges. Finally, of course, artifact control, based on things like coherence, manufacturing artifacts, and overcoming other challenges.
Q: How is Dispelix approaching this challenge?
A: First of all, at the heart of the company is the optical design of this component. So how do you create optical architecture to support the display solution? The core, something which makes this possible is really how to make the design. For example, if you use traditional ray tracing, the modeling of light transport, you are limited to simple solutions because you quickly run out of the calculation on power.
What we have been developing is an approach, which can be 20,000 times faster than ray tracing. We are going further because there are still challenges around calculating complicated things. We are also focusing on developing waveguides that can be mass manufactured and are currently working with a number of manufacturers to ensure that we can deliver millions of waveguides to satisfy customer demand.
Q: Dispelix is focused on creating the best waveguides and technology on the market. But it doesn’t end there, correct? Who else are you working with to make Dispelix designs a marketable reality?
A: This is an ongoing industry effort. Dispelix was one of the founding members of the non-profit LaSAR Alliance. We currently have roughly 20 members who are contributing their combined knowledge to this laser beam scanning solution for AR. Either core components, materials, manufacturing, defective optics or solutions.
Q: Can you explain in more detail about how Dispelix waveguides work?
A: Light passes through the waveguide, and then it propagates with the total internal reflection, and we couple it out. If we do this right, the image going in is the same image that comes out.
In this form we are basically multiplying this incoupling pupil. We are building it inside what is called the Eye Pupil Expander, EPE, and outcoupler. At any point you can see the full image, which provides a 20, 30, 40, 50-degree field of view, depending on the source. The key conditions are defined by this nice Donut up here. You are thinking in the angular world, what kind of conditions you can couple the light inside the waveguide so that it will stay there until you want to take it out. Things like wavelength, incident angle, refractive index, pupil size, and waveguide tunnels are affected. We have to think about what these things mean for the user experience.
Q: How is Dispelix able to achieve faster optimization?
A: First we have a waveguide solver, which uses rigorous physical models to define how the light interacts with our surface relief gratings. Then we have analyzer functions, which analyze what is the desired output of the system. We do the optimization, which is the secret here. So, looping this around, we should come up with the ideal solution faster than with other means.
Here are a couple of examples to demonstrate. This is again, because of the laser. A couple of artifacts we are introducing in some cases, for example, coherence, which is creating Newton rings on the surface. We have some what we call the EP interference. The propagating light is interfering with itself inside the waveguides causing some uniformity challenges. In some cases when using the laser, you have a Speckle. We successfully get rid of these challenges with our simulation tool. Also, Speckle is not really the problem with waveguide technology, because you do not have the few surfaces, but you can easily introduce Speckle by adding some optics in the projector which has some small roughness. The uniformity controls the light, because you have a different part from incoupling to outcoupling, how to balance this propagation from incouple to outcouple so that you can keep the colors as desired.
There is a strong color change over the field of view and here is less. What is very important from an application point of view, the display cannot look like a display. We have a grading or waveguide where you cannot see the gratings, which is good here. You cannot recognize that there is a display in these glasses when it's turned off. You make it very lightweight, very thin so that you can really integrate it in a small form factor.
Here are a couple of examples of our technology demonstration devices. So here we are solving the solution for the reflective panel, which is our DPX Selva platform. So, pushing down the combined projector plus waveguide in a really eyeglass format. And here is another one, which is DPX Sade, our laser beam scanning compatible waveguide with all projectors and electronics in this demo device. Still little bit room to squeeze down, but good progress.