MicroLED (µLED) displays offer the potential of wider color gamut, higher contrast ratio, and deeper blacks than LCD (liquid crystal display) and OLED (organic light-emitting diode) displays. MicroLEDs match OLED technology for response time and view-angle performance, but exceed OLED in brightness and ruggedness, with much lower power consumption. And like OLED, microLEDs can be deposited on a variety of substrates including glass, plastic, and metal to enable flexible, bendable, and foldable displays.
As the demand for higher and higher resolution displays continues to expand, microLEDs are poised to be the next breakthrough technology in the industry. MicroLEDs are particularly appealing for devices such as smart watches, head-up displays (HUD), and augmented reality (AR) devices where high luminance and resolution are particularly important for visual performance in small spaces, at close proximity to the eye, or under variable ambient lighting conditions.
PlayNitride summarized the differences and advantages of mMicroLED display technology compared to OLED and LCD. (Original image © PlayNitride1)
To realize the potential of microLED display technology, however, a number of production challenges must be overcome. Ensuring the quality and uniformity of these displays is challenging, since each diode is its own emitter and can exhibit wide variability in luminance and color. Whether measuring microLEDs at the individual wafer level, after deposition onto a substrate, or assembled into the display panel of a consumer device, microLED makers need a reliable inspection approach to measure and quantify microLED light and color output with precision.
The Big Hurdle: Production
Due to their size, microLEDs have required the display industry to develop entirely new production assembly technologies, as they use different die structure than traditional LEDs or OLEDs and require new manufacturing infrastructure.
Eric Virey, an analyst at Yole Développement explained: “Unlike OLED, inorganic LEDs can’t be deposited and processed over very large areas. LEDs are grown on 4- to 8-inch wafers and the art of making microLED displays therefore consists in singulating individual emitters and transferring and assembling them onto a backplane substrate. For most consumer displays such as TV or smartphones, microLED with die size ranging from 3-10μm are required to ensure cost compatibility with the applications.”2
This means fabricators must find methods that yield high quality with microscopic accuracy while also achieving mass-production speeds. A microLED screen is composed of millions of tiny pixels in chip form, each of which can be monocrhomatic or contain some combination of red, green, blue and/or white subpixels. To fabricate a display, first microLED wafers must be created via high-yield epitaxial growth. Then each individual wafer must be transferred to a substrate or backplane that holds an array of units in place.
A simplified illustration of the microLED production process. (Image © Allos Semiconductors)
The transfer equipment used to place microLED units on a substrate needs a high degree of precision, with placement accurate to within ±1.5 µm. Existing pick-and-place (parallel) LED assembly equipment can only achieve ±34 µm accuracy (multi-chip per transfer). Flip-chip bonders typically feature accuracy of ±1.5 µm—but only for a single unit at a time, thus production speeds are slow.
Neither of these existing LED transfer methods are sufficient for mass production of larger displays such as TV screens where millions of microLEDs are required, (e.g., approximately 6 million for an HDTV). Because traditional chip bonding and wafer bonding processes don’t provide efficient mass transfer for microLEDs, various thin-film-transfer (TFT) technologies are also being explored.
Methods such as elastomer and electrophotographic transfer, roller transfer, and fluid assembly are being tried. Researchers are also working to resolve the challenges associated with integrating compound semiconductor microLEDs with silicon-based integrated circuit devices that have very different material properties and fabrication processes.
Addressing Mass Transfer Challenges
At the recent SPIE AR VR MR Conference, microLEDs were a hot topic, with multiple companies in the industry discussing the technology, its potential use in augmented and virtual reality devices, and presenting their unique methods for overcoming production and mass transfer challenges. Virey described the microLED industry as being in a “Cambrian explosion phase”, characterized by continuing experimentation and innovation with many new startups emerging. The theme of microLEDs carried through into the overlapping SPIE Photonics West conference and expo. Some relevant highlights from the two events:
- PlayNitride, Inc., a leader in microLED production, uses a stamp method for mass transfer. They explained that even with a tiny error rate, a 4K display could have as many as 100,000 defective microLED chips. While it would be possible to build in redundancy for every chip to make up for defects, that would be quite expensive. Instead, Playnitride uses a “re-stamping” process to identify defective pixels add in new diodes to repair the defects, one color at a time.
- Glō, which emerged from Sweden’s Lund University Nanostructure Consortium, is using a patented room-temperature wafer transfer technology based on nanowires. Nanowires are 3D structures that can enable red, green, and blue microLEDs similarly to inorganic materials. The efficiency of nanowire microLEDs improves with decreasing size. Glō reports that millions of its GaN nanowire-based RGB microLEDs can be transferred with high yields and bonded to active backplanes made out of silicon, glass, or flexible substrate materials.
Glō RGB microLEDs (Image © Glō, Source).
- New York-based startup Lumiode has developed a microLED method they term monolithic integration that combines established LED and TFT technologies into a patented solution. They start with an LED epi wafer, fabricate a microLED array on it, fabricate silicon thin-film transistors on top of that, then singulate and package the resulting microLED display wafers. They say the method is platform agnostic and can be applied to any LED substrate or wafer size, yielding high-performance, low-power, small-form-factor microLEDs.
- eLux Display, a spin-off from Sharp Laboratories of America, has pioneered a fluid-assembly method. Their microLEDs are fabricated on 4-8” single crystalline substrate, one color at a time. The substrate is made with precise capture points (wells), then a liquid containing microLED disks is sent in an oscillating flow over the substrate until all the wells are filled. Defective diodes can be eliminated before being put into the fluid, and excess disks in the fluid that are left after assembly are then recycled. They report that this method results in 99.5% fill in just 15 minutes—independent of panel size—with 100% correct alignment and excellent uniformity.
Still image taken from a video of the eLux fluid assembly process, showing individual microLED diodes washing over a prepared substrate. Dark circles are 50 µm diameter microLEDs, the light circles are 55µm diameter wells. (Image © eLux3)
The Need for Precision Inspection & Measurement
Ultimately, the appearance of a microLED display is determined at the level of each individual pixel. Because the output of each subpixel (each red, green, or blue microLED) is individually controlled, luminance and color of each pixel is a combination of its subpixel outputs. Due to production discrepancies, there may be variations in luminance for the same electrical signal input throughout the population of same-colored subpixels on the display. Differences in brightness from pixel to pixel can cause variation in display color; this variation demands an inspection system that can perform subpixel measurement to ensure uniformity.
Dead pixels are also an issue, just as they are for traditional displays. For example, to achieve a rate of less than five dead pixels per RGB full-color, full high definition (FHD) display (1920 x 1080 pixels) requires a yield of 99.9999%.4
Solution Example: Wafer-Level Quality Control
The first step in ensuring microLED display performance is inspection and measurement at the LED, chip, and wafer stage to reduce the possibility of dead pixels and ensure appropriate luminance (brightness) and wavelength (color) uniformity. General visual performance standards in the display industry allow for less than 10 dead pixels per display, thus LED yield must be very high. MicroLED manufacturers have to inspect each individual emitter on a wafer to determine uniformity, verify individual distribution of dies, and measure luminance (nits) across red, green, blue, and occasionally white microLEDs.
Active matrix microLED microdisplay chips on a 4-inch wafer, fabricated using JBD’s monolithic hybrid integration technology.
For precise inspection at the individual pixel and subpixel level, microLED manufacturers rely on Radiant’s ProMetric® Imaging Photometers and Colorimeters in various configurations, such as a ProMetric I29 (29-megapixel) Imaging Colorimeter with standard lens or coupled with a Radiant Microscope Lens. To measure microLED luminance only, manufacturers can leverage an imaging photometer, such as our ProMetric Y combined with a Microscope Lens—in Y29 (29-megapixel) or Y43 (43-megapixel) options.
The Radiant Microscope Lens provides objective measurement with 5X or 10X zoom, offering detailed inspection of individual emissive elements of any shape. Microscope lens measurements have been proven for measuring subpixel luminance and color output, registering each pixel using ROI that are hundreds of image sensor pixels in area for ultimate accuracy. Using any high-resolution, low-image-noise ProMetric system (with or without the Microscope Lens add-on) means that every display pixel is captured over several sensor pixels for optimal measurement precision. The system is effective for evaluation of display pixel structures and subpixels of individual microLEDs.
Radiant’s ProMetric imaging system and Microscope Lens (left) can be used to characterize subpixel layouts, shapes, and color patterns (right) with TrueTest™ Software, achieving extreme detail in a very small spatial area.
When used with Radiant’s TrueTest™ visual inspection software, Radiant systems can run multiple tests in sequence, and automatically register pixels using dynamically defined ROI that isolate and cover each pixel precisely to control for issues such as stray light.
- Liu, F., “Next Generation Micro LED Display Technology”, presented at SPIE Photonics West, San Francisco, CA, February 5, 2020
- Virey, E. et al., “Overlooked Challenges for microLED Displays”, SID 2019 Digest p. 129-133, 11-3, DOI:org/10.1002/sdtp.12872
- Lee, J., “Low Cost and Practical Technology to Manufacture Micro-LED Displays: Self-Align Fluidic Assembly, presented at SPIE Photonics West, San Francisco, CA, February 5, 2020
- Ding, K. et al., “Micro-LEDs, a Manufacturability Perspective”, Applied Sciences, 2019, DOI:10.3390/app9061206