Why monolithic integration wins the race to build microLED displays
Sep 23, 2019
We’ve become used to very high resolution displays on our smartphones, with some offering pixel densities of up to 400dpi – arguably greater than most human eyes can resolve. These displays are almost always built using LCDs, and represent the current state of the art for a technology that has been in development for over half a century.
As we move to a more intimate interaction with technology and the virtual world it brings us, through body-worn electronics such as smart watches, fitness trackers, virtual-reality helmets, and augmented-reality glasses, the display technology we use will have to change. LCDs, which use changing crystal alignment to block or allow the passage of photons from a backlight, must give way to emissive technologies that turn electrons into photons directly.
There are a couple of key contenders for this replacement technology, including OLEDs and microLEDs, that is arrays of small LEDs mounted at tight pitches on some sort of backplane. OLEDs have had some traction in the market, but are between 100 and 1000 times less bright than LEDs for the same electrical input. This really matters in applications such as head-up displays, where, because of the difficulty of coupling the light emitter to an optical waveguide, very little of the light produced actually reaches the user’s eye.
MicroLEDs rely on the electro-optical characteristics of III-V materials such as gallium nitride (GaN) to directly emit photons. Some companies create microLED arrays by mounting discrete GaN LEDs on sapphire substrates, at pitches that begin at 50um and may have to shrink down to as little as 10um for very small or very high-resolution displays. To build a full colour HD panel this way, with a resolution of 1920 x 1080 pixels in three colours, would mean accurately placing more than 6.2 million LEDs with positional accuracies of at least +/-10um.
This kind of individual device placement will never make economic sense, if it means tying up an expensive pick and place machine for days or weeks simply to produce a single display for a $500 pair of AR glasses. To tackle this, manufacturers have experimented with techniques such as transferring multiple microLEDs from the wafer they were made on using an adhesive film as a transport. Work is also going on to build massively parallel pick and place machines, which can pick up an array of LEDs from the source wafer and then place them simultaneously on their target substrate. In both cases, there are multiple issues with maintaining positional and alignment accuracy across all the placement sites.
In a sense, the development of microLED arrays is a classic Moore’s Law story, in which monolithic integration is obviously the right ultimate answer, but it takes many intermediate steps to achieve it. However, Plessey has developed their proprietary monolithic approach to building GaN microLEDs directly on a CMOS silicon substrate, so that placement and alignment accuracy are now controlled by lithography, a very well understood process in the semiconductor industry.
Plessey showcased a World’s first prototype earlier this year at Display Week in San Jose. The microLED display features an array of 1920×1080 (FHD) current-driven monochrome pixels on a pitch of just 8 microns. Each display requires more than two million individual electrical bonds to connect the microLED pixels to the controlling backplane. The Jasper Display Corp eSP70 silicon patented backplane provides independent 10-bit single colour control of each pixel.
Silicon substrates have greater thermal conductivity than sapphire, allowing the microLEDs to be driven harder before their output becomes thermally limited. A CMOS silicon substrate also makes it very easy to build the addressing and switching matrix for controlling the microLEDs, and to integrate other supporting circuitry, perhaps even graphics processing units, on the same wafer. The Plessey GaN-on-silicon process already scales to 200mm wafers, and can be scaled further, enabling economies of scale in manufacturing that will not apply to working on smaller sapphire substrates.
Most current microLED arrays use three devices to produce the R, G, B outputs necessary for a full colour display. But recent research suggests that it may be possible to build more sophisticated microLEDs that can produce R, G, B and intermediate colours (through mixing) from one device. Once again, the monolithic integration predicted and championed by Moore’s Law wins.