Introduction

SID Display Week (DW) 2025 is coming to San Jose, CA, May 11–16. My last article discussed the LCOS technology I saw at Display Week 2024, and this article will cover many MicroLED technologies. Of the conferences I go to regularly, Display Week has the most information I have found from MicroLED companies in one place.

For example, I first heard of Jade Bird Display (JBD) at DW 2019. I have already written JBD demonstrations and technology at DW 2024 in Jade Bird Display’s MicroLED Compensation. Still, there were many other MicroLED companies at DW that I haven’t covered. But before addressing these other companies, I thought I would share a picture I took from my first encounter with JBD nearly six years ago. In 2019, JBD had a small one-table booth (above) with their MicroLED demos on the table, and by DW 2024, they had a large booth (right).

In terms of the order of companies, I grouped them by the way they implement colors. The categories are native single color, single diode emitter with multiple wavelengths, and quantum dot converted. Within a category, I tried to put companies I had not discussed before first.

Partnering with SID and Discount Code

I’ve partnered with SID to share my insights from Display Week (DW) — past, present, and future. If you’re planning to attend Display Week, SID has provided the code DW25KARL for a free exhibit hall pass — don’t miss the chance to explore the latest innovations shaping the future of displays.

Single Color (Green) Versus Full-Color Nits

When you hear specs like native single-color green having millions of nits while full-color green and full-color QD color converted typically have “only” 150K to 300K nits, you might wonder what is going on. There are many reasons for this order-of-magnitude difference. I’m going to touch on a few of the key reasons in this section and move some of the details to the appendix. Suffice it to say the conversion efficiency of the QD is not the main reason.

Native LED Full Color

Native (non-color converted) MicroLED makers talk of impressive multi-millions of nits outputs, and they talk of single colors, often green, devices. Green MicroLEDs typically have (by far) the most nits per watt of electricity. Blue LEDs, while having higher “wall plug efficiency” (WPE = Light power/electrical power) due to the human visual response, contribute a few nits. Native red LEDs have been proven to drop dramatically in efficiency as the LEDs are made smaller, typically making it the most problematic color for MicroLEDs (somewhat ironically since with big LED red was the easiest to make). So, even ignoring all the complexities of getting colors combined either on a single device or with optics, blue and red add significantly to the power consumption while adding relatively few nits. Power consumption and heat (and the effects of heat) will limit the light output.

Then, there is the issue of efficiency losses when combining colors. Using an X-cube to combine three single-color MicroLEDs is expensive and has optical losses. It also will grow 3-dimensionally as resolution increases, as the pixel sizes can’t be scaled much further. Spatial (r, g, b side-by-side) color with native MicroLEDs has massive technical issues. Spatial color makes the light-emitting area of a given pixel larger, which means that microlenses used to partially collimate the light to produce more nits from the same lumen output of a pixel will be less effective due to etendue. Stacking red, green, and blue LED epitaxial layers on top of each other causes significant issues of light loss as the layers block each other and cause thermal issues on top of the process complexity.

Quantum Dot Full Color

Quantum dots (QDs) absorb light of a shorter wavelength and emit light of a longer wavelength, blue to green and red, or from UV to red, green, and blue (and perhaps other colors, including white). QD-MicroLEDs then have a single LED crystal layer (blue or UV) and print the QD in a spatial pattern to produce color. As discussed above, Blue/UV MicroLEDs have much greater WPE than green or red, and even with some losses in the QD conversion, they can have higher net efficiency in terms of power-in to lumens/nits-out. ]

QD conversion of a single color, blue or UV, results in a simple LED structure on a single layer of a single crystal. Blue and UV have much better WPE than green and red. Even after the losses in conversion, it can be more efficient to generate red and green than with native red or green MicroLEDs.

The QD color subpixels are typically laid out side by side or often in a 2×2 rectangle with two green subpixels (or sometimes a white or other color subpixel), known as “spatial color.” As discussed earlier, spatial color pixels are bigger and thus have worse etendue. Thus, microlens will be less effective in collimating/collecting the light to increase nits (see Collimation, Etendue, Nits (Background for Understanding Brightness)). Any method for producing full color has significant losses.

QD MicroLED still has the issue that blue light contributes only about 4% to the nits, while red contributes about 30% (see appendix). Thus, red and blue (depending on the various wavelengths) contribute to power while adding only about 34% to the combined nits.

QDs have additional issues with very small LEDs as opposed to larger direct-view displays. Because MicroLED feature sizes are very small, the QD layer needs to be very thin. It is difficult to absorb and convert all the blue (or UV) light in a thin layer. Any blue (or UV) light that is not absorbed and re-radiated in the new color is energy that is wasted and will either desaturate the color or need to be blocked with a color filter that will cause further losses. If the QD is converting UV, there is an additional requirement to block all the UV light so it can’t harm the eye.

Another issue with QD is isolating the light between subpixels so that the blue (or UV) light that illuminates, say, the red subpixel does not also partially illuminate the adjacent green subpixel, thus desaturating the colors. Any blue light leakage will also reduce the overall contrast. Some light-blocking structures have to be between each subpixel to prevent serious cross-contamination of colors. This light block will also result in less area for light emission and absorb some light, contributing to light loss. I have yet to see a QD MicroLED with what I would consider great color control, but then I am looking at prototypes and not final products.

Native (Single Color) MicroLEDs

Native single-color MicroLEDs, primarily from JBD, currently dominate AR glasses-type headsets that are appearing. Most lightweight AR glasses use green only, and a few use X-Cube optics (for example, TCL). In contrast, others use waveguides (for example, Meta’s Orion) to combine three JBD monochrome MicroLEDs to produce full color. Native MicroLEDs are very durable and can produce very high brightness in the millions of nits. However, producing monolithic full color with native LEDs is a very difficult problem and will inevitably come with significant brightness losses, as discussed above.

QubeDot

QubeDot was in the Germany pavilion, demonstrating their devices with GaN red, green, and blue MicroLEDs. Their capabilities include design and foundry services for GaN MicroLEDs.

VueReal

I have met multiple times in the last few years. I wrote about and discussed them in a video about CES and AR/VR/MR 2024. VueReal makes both single-color microdisplay devices (currently with a 5.2-micron pixel pitch) and single- and multi-color singulated and transferred MicroLED displays, what VueReal calls “MicroSolid” displays with pixel pitches from ~50 microns to ~450 Microns or larger.

While I have put VueReal in the “single native color category,” they have also used phosphor-coated blue LEDs to produce white and other colors, as well as Quantum dots, for larger MicroLEDs MicroSolid displays.

VueReal’s MicroSolid “cartridge printing process” takes tested and singulated LEDs and makes relatively small displays that require limited magnification and direct-view displays that can be largely transparent.

Pixel pitches larger than 150 microns are typically used in direct-view displays with no optics or in VR headsets with simple optics. At these larger pixel sizes, the LEDs are spread so far apart that VueReal can make transparent displays (right). These displays can be laminated into glass for various uses. VueReal is also claiming that they can use the gaps between the LEDs for solar cells in a self-powering watch.

VueReal has used phosphor coating to make a transparent white display (below left). They can also use quantum dots with their MicroSolid process. The MicroSolid process can also produce displays with very large pixels to make automotive taillights that can show simple messages (below right). One could see these types of devices in simple HUD-like display uses or with an Uber/Lyft display that could also show the passenger’s name or other information at pickup.

VueReal currently has single-color, single-chip MicroLED microdisplays with a 5.2-micron pixel and over 1 million nits (with green). Similar to Jade Bird Displays, they use GaN for blue and green and (I think) AlInGaP for their red LEDs. While these might be useful for waveguide display headsets, unlike their larger MicroSolid displays, I have not seen them used even in a prototype.

“Tweener” 57-micron Pixel Pitch

Typical AR microdisplays (LCOS, MicroLED, Micro-OLED, and DLP) used with waveguides use pixel pitches from 3 to 12 microns, and 150 microns and above are typically used for direct view and VR. VueReal has also been showing a “tweener” pixel-size microdisplay with a 57-micron pixel pitch, 160×90 pixels, 0.42″ (9.1 x 5.1 mm), and a peak brightness of 10,000 nits. It’s not bright enough for waveguides, and the pixel sizes are too small for a direct-view display. So, the types of optics used with small OLED displays would be required. The device is aimed at relatively low-cost “data snacking” type AR.

Jade Bird Display (JBD)

Jade Bird Display (JBD) is currently the dominant MicroLED maker for Augmented Reality. They are designed into over a dozen AR glasses. Every green-only AR glasses I have seen uses JBD’s MicroLEDs. TCL full-color RayNeo glasses use red, green, and blue MicroLED microdisplays from JBD combined with an X-cube) and other companies are expected to also introduce X-Cube combined full-color glasses using JBD MicroLEDs. As discussed in the introduction, I have already written a large article about JBD DW 2024.

MicroLEDs Single Emitter Multiple Colors

Q-Pixel

Startup Q-Pixel was demonstrating their Polychromatic MicroLED Pixels, a single diode that controllably emits different wavelengths. They have demonstrated 6800ppi/~3.7 micron pixels pitches and had a live demonstration of one of their devices working (below). More information on Q-Pixel’s technology can be found in their website article.

The concept of having a single LED emitter for all colors has significant technical advantages in making smaller pixels than with spatial color (separate red, green, and blue subpixels) and for better etendue, leading to better optical efficiency. Multiple companies have demonstrated lab prototypes of a single emitter multiple color technology, including Porotech (see: CES 2023 (Part 2) – Porotech – The Most Advanced MicroLED Technology).  But there are technical barriers, which include:

• The red “electron to photon” efficiency is usually very low with GaN and particularly difficult with a single emitter for variable wavelengths.
• Typically (including Q-Pixel), the wavelength is controlled by the current, and brightness is controlled by pulse width modulation. It is extremely difficult to drive current accurately to a large array of LEDs and thus shifts in color and brightness. Then there is the problem of driving the current accurately while pulse width modulating it (as discussed in my CES 2023 article on Porotech).
• Generating a full-color image would require some form of “field sequential color,” where the various primary colors (there could be more than three) are modulated to form all the colors in between.

Porotech

Porotech (just mentioned above), another single emitter for multiple wavelengths, had a significant booth at DW 2024.  I had recently spent time at several other conferences and met with Porotech at the UK facility. I didn’t spend time with them at DW 2024, but many articles referencing them on this blog, including  CES 2023 (Part 2) – Porotech – The Most Advanced MicroLED Technology).

MicroLEDs with Quantum Dots For Full Color

Raysolve

Raysolve is one of many companies developing quantum dot color conversion of blue MicroLEDs. They are making blue MicroLEDs on 8-inch  GaN-on-Si Epi wafers and then bonding them to 8-inch CMOS substrates. They then apply red and green quantum dot layers, resulting in one red, two green, and one blue subpixel per pixel.

The 2×2 subpixels are ~3.5 microns, resulting in a full-color pixel pitch of ~ 7 microns. Raysolve demonstrated 320×240 and 640×480 devices with the same pixels. Their stated light output was >150K nits.

I was able to take some micro-photographs (using a 1:1 macro lens on an Olympus OM-D E-M5 Mark III) of their 640×480 display directly with my camera (right). Raysolve was also nice enough to give me the source image. Overall, the image looked good, with a much better saturated green color than other quantum QD MicroLEDs I have seen.

If you look carefully at the bowl in the lower right corner of the image, you will see some streaks that are not in the source. I would attribute this to these being prototypes.

As can be seen in the comparison with the source image, there was too much overall blue. I made some quick adjustments, reducing the blue (in Photoshop), and it made a big improvement (lower right). Issues with color balance are common in prototype demonstrations, and I wanted to get a better idea of their potential.

Raysolve was also demonstrating their MicroLED via a (non-Lumus) reflective waveguide. Lumus believes that their geometric (reflective) waveguides are about 9 times more light-efficient with MicroLEDs than typical diffractive waveguides with MicroLEDs (more on this in a future article). This efficiency advantage could explain why RaySolve, with their ~150K MicroLED, is using a reflective waveguide, whereas others using diffractive waveguides use JBD native MicroLED with more than 1 million nits.

SAPHLUX (Quantum Dot Color MicroLEDs)

Saphlux, like the previously discussed Raysolve, uses quantum dots (QD) to convert blue LED light into red and green subpixels. It also uses a pattern of two green, one red, and one blue subpixel per pixel.

In addition to the microdisplays for augmented reality, which are the subject of this article, Saphlux makes large chip-on-board direct-view displays. Saphlux, like Porotech, uses a form of porous gallium nitride to make their LEDs.

Saphlux was demonstrating both a 640×480 0.12” green-only (native/not-QD) MicroLED with a stated peak brightness of 3 million nits and a 1024×768 full-color with QD conversion for red and green with 250K nits.

The ~10X difference in brightness between the direct green and the QD full-color MicroLEDs seems typical. Factors that result in this difference include:

• For a given white point, you need approximately (it depends significantly on the wavelengths and white point) 68% Green, 28% Red, and only 4% blue nits. While it takes about the same light energy (in Watts) for the three colors, green provides the vast majority of the nits. Thus, red and blue add power/heat but have relatively few nits. See my 2011 article on the relationship between electrical and optical power (Watts) and lumens and my 2017 article on the relationship between lumens and nits.
• The emission area of spatial color pixels is larger, which, due to etendue, results in less ability to collimate pixels with microlens arrays and thus lower nits. See my 2017 article on the relationship between lumens and nits.
• QD conversion losses
• Any color filtering to block residual blue light after QD conversion to red and green
• Limits on the energy density/heating of the quantum dots.

While I missed getting direct photographs of Saphlux’s display at DW 2024, I made a point to get pictures a few weeks later at AWE. In between conferences, I acquired a 5X macro lens for my Canon R5 to give enough resolution to see the individual sub-pixels (see microphotograph below).

The Saphlux device lacks color saturation/purity of the red and green with a significant amount of residual (non-converted) blue light. One can also see many defects/dead pixels in the high-resolution picture above.  I don’t want to be too judgemental, as this is a prototype, and it should improve. I am pointing it out because it is something I have seen in several QD-based MicroLEDs.

I also got some high-resolution microphotographs of Saphlux’s (native) green MicroLED (below). They look similar to JBD’s green MicroLEDs, which I have seen in terms of uniformity before any demura (pixel-to-pixel uniformity) correction. My usual warning goes out that I don’t know whether the devices (from JBD or Saphlux) have been “cherry-picked,” but they are likely not selected at random. Someone could argue that uniformity is not that critical for “data snacking” applications.

PlayNitride

PlayNitride is one of, if not the largest company, making QD color-converted MicroLEDs. As with Porotech discussed earlier, I was familiar with PlayNitride and had written about them before including my 2023 AR/VR/MR article. See: PlayNitride (Blue with QD Conversion Spatial Color). As a result, I didn’t spend a lot of time with PlayNitride at DW 2024 and get some high-resolution pictures (which takes a lot of time and is hard to set up at an exhibition), but I plan to get an update from them at DW 2025.

At AR/VR/MR 2023, PlayNitride claimed 150,000 nits for their full-color AR microdisplay, but at DW 2024, they increased that value to 300,000 nits (see: https://www.insightmedia.info/playnitride-is-all-microled-at-displayweek/). As pointed out by Chris Chinnock for Insight Media, “100% NTSC” is not a particularly strong statement about the color gamut.

Playnitride showed 30-degree FOV diffractive waveguide-based glasses at DW 2024. The video game content on the glasses didn’t help in evaluating the colors. Not to pick on Playnitride, but this kind of content can hide a lot about color quality.

At AR/VR/MR  2023, I took more time to evaluate their demo, which used Lumus Geometric (reflective) waveguides. I found that their red and green still had significant amounts of residual/unconverted blue light, but I expect this has improved and will continue to improve as they develop the QD process.

Conclusion

SID Display Week continues to be the best single conference I have found to learn about display device technology for Augmented Reality, and this is especially true for MicroLED development. As discussed in the introduction, I learned about Jade Bird Display in 2019 at their small booth, and a few years later, they were in many AR headsets.

Generating high brightness with full color remains an elusive goal for MicroLEDs when working with waveguides, but I continue to see improvement. As pointed out, it is tempting to get sucked in with the multi-million nit specs of single-color native MicroLEDs. Full-color MicroLED devices are going to be much lower in the nit specs. They are going to require more efficient waveguides to provide sufficient brightness, either reflective, which Lumus claims is inherently ~9x more efficient than typical diffractive waveguides, or more efficient diffractive waveguides.

Appendix – More on Single Versus Full Color MicroLEDs

Assuming 630nm Red, 530 Green, and 460 Blue LEDs and a nominal D65 “white” point, red contributes about 26% of the nits, green 70%, and blue only 4%. At these wavelengths, with D65 white, the light power in Watts is about 41% red, 33% green, and 26% blue. Wall plug efficiency (WPE), the ratio of electrical light Watts-out to electrical Watts-in, varies with red, green, and blue.

And then, we have the issue that “nits” (and Lumens) are based on the human response to visible light (done by studies), which varies somewhat dramatically by wavelength (see graph), and why blue, while necessary for color balance, contributes so little to the “nits.” Note also from the graph that the wavelengths of the various colors affect the nits and, at the same time, will affect how much of a given color is needed for white balance.

While blue has a significant effect on the perceived color and blue MicroLEDs are typically much more efficient than green or red, the electrical power spent in the blue LEDs blue’s contributes negligibly to the nits. Native green GaN LEDs are (roughly put) made by adding impurities/crystal strain to what would otherwise be blue LEDs to push the wavelength.

LED makers tend to show graphs with “external quantum efficiency” (EQE) rather than the more easily understandable “Wall Plug efficiency” WPE (light-power-out/electrical-power-in). As it turns out, EQE and WPE are proportional to each other, as explained by the excerpt (right) from Boston Electronics (while it says “UV LEDs,” the equation is generally applicable and spells out all the terms).

The two graphs below are taken from Manuel Gensler’s (from Fraunhofer) presentations at MicroLED and AR/VR Connect in September 2004 and March 2025 (accessing the video presentations requires a MicroLED Connect “pass,” which includes their library of past events). The graph on the left shows the relative peak external quantum efficiency (EQE, which is proportional to WPE) of, for example, native blue, green, and red LEDs as they scale down in size. The graph on the right shows the problem with making green native GaN MicroLEDs and the fact as the wavelength is pushed toward the more desired (for color gamut) ~530nm, the EQE (and thus WPE) drops with what is called “the Green Gap.”

Looking at the Lumens (and Nits) per Watt graph above, you will notice that as the green wavelengths drop from 532nm to, say, 510nm, the nits per (light) Watt drop by a whopping 57% and the color gamut is reduced.

Gensler also discusses the well-known efficiency problems with small red native LEDs made using AlGanInP. The case Fraunhaufer and other QD companies make is that it is more efficient to generate green and red using QD conversion of blue.

There are two types of crystal structures used to make native red LEDs. The more common way is to use Aluminum-indium-galium-phosphate (AlInGaP/AlGaInP), and the other method is to add impurities to InGaN to stretch it to red. Lumileds published an article in Compound Semiconductor in 2022 that shows the WPE as a function of current density for a range of pixel sizes. Consistent with Fraunhaufer’s presentation, the red WPE drops much more dramatically as the LED size is reduced. They also show how InGaN has very low WPE.

Applied Materials in a Display Daily article advocates for using UV LEDs with QD for red, green, blue, and perhaps some other color (or a repeat of green). Generally, the QD conversion efficiency is better with UV or near UV blue light than with visible blue light. Note that they are talking about moderately large “MicroLEDs” in the range of 80-micron pixels for a direct-view watch display where “pick and place” (for example, VueReal) can be used with native MicroLEDs. Even with these larger pixels, the WPE efficiency of red native MicroLEDs is a problem. They claim that a better color gamut can be achieved with UV QD conversion and deliver a higher yield.