OLED Basics From a Pioneer

Gopalan Rajeswaran (Raj) changed his planned talk from solar panels to discuss AMOLEDs as there was not a specific speech on this topic at the event.

Rajeswaran said that he has been working in the OLED area for 30 years or so since his days at Kodak. OLEDs were discovered in 1987, after four or five years research and the length of times shows that it takes many years for new technologies to come to the market. After his research work at Kodak, Rajeswaran was very involved in the commercialisation of the business with the Kodak/Sanyo venture, SK Displays.

He then explained the core physics of the OLED and how the OLED was discovered, in the process of trying to develop organic solar cells (a quest that continues). In 1989, the concept of using a dye to make colours was discovered and that paved the way for the development of full colour displays. The key to building OLEDs is to get the right molecules in the right place for optimum efficiency. It’s not tricky to make molecules, he said, but it’s very hard to make devices.

In 1995, Rajeswaran started to try to industrialise the technology for Kodak. At that point all the LCD infrastructure was in Japan (LCD was at G3/G3.5 and all the technology was in Japan). Sanyo had some LTPS technology, which was just under development, so a joint venture was formed, SK Displays. LTPS was clearly the backplane of choice because of the higher mobility compared to amorphous silicon.

At the time, the only way of patterning was to use a fine metal mask (FMM). The shadow mask aperture sets the tolerances for the materials and the shapes of the pixels. It takes a long time to master the shadow mask technology – it took three years initially at SK Displays and now it takes around four to six quarters for new entrants to get the process stable and working. The shape of the mask, its alignment and dimensional stability are critical. The masks are expensive, but can usually be re-used some times – Rajeswaran said that 1000 uses of a mask is reasonable. (Making the masks is apparently a profitable business! There are only two suppliers, but Samsung has exclusivity with one.)

In production, around 50% of the processes need an FMM deposition step, not just the patterning of the pixels. For example, the hole transport layers (HTL) need FMMs.

In designing an OLED, you start with material choices that set the final colour gamut and white point. You can then design the electronics to drive the display.

The next step from the design point of view is to choose the final brightness and work that back through polariser and aperture ratio losses to get to the source brightness. At that point you can finalise the grey scaling.

SK Displays made around 2 million displays for cameras and phones before the decision was made to close the business. After around six quarters of production, yield got to around 25%. The problem was mura caused by the combination of the LTPS and the OLED, which was being developed using an RGB structure. There were around 45 different muras, and there seemed no way to fix this problem, so with an eye on the TV application, the white OLED with colour filter structure was developed.

The way to judge an OLED is to look at a white screen at 10% intensity as this shows all the mura and defects

Blue and yellow OLED layers were made to produce white, although the design was short of green and a lot of light was absorbed by the filter. The architecture was developed with multiple white layers and up to three white can be used as layers in the stack. Another invention was to create an RGBW structure with a white subpixel (which works well on general video, but is not so good for saturated colours, for example to create cartoons). Where the white sub-pixel is used, there is no filter, so efficiency improves.

Rajeswaran described the current situation in OLED manufacturing as being two “temporary methods” (RGB patterning using FMM, used by Samsung, and White OLED with filters, used by LG for TV panels) that are limited in their potential. New methods of manufacture are still needed.

Samsung wanted to use amorphous silicon and didn’t want to use LTPS, because of the problems of mura, but it could be made to work with external compensation, that is to say, you check where the mura are and correct for them, both at the time the display is made and later during use on a regular basis.

Moving to current status, although LTPS is the preferred choice for OLED, Oxide TFTs could be a candidate using amorphous-like processes. For mobile displays, LTPS is no barrier. Even now, even after 15 years of development, LTPS is limited to G6 and it’s the combination of this together with the size limitations of FMM, which makes the Samsung technology unsuitable for TV applications.

Top or bottom gate transistor configurations are possible and it’s possible to drive OLEDs using a simple two transistor design. The first OLEDs were bottom emission, where the light is emitted through the substrate that has the transistors on, but this meant low aperture ratios and now top emission is used so that multiple transistors can be used under the pixel to correct muras.

Blue has always been a challenge for OLED and blue material is still fluorescent, but red and green are now phosphorescent materials, which are more efficient. Transport layers to move the electrons to the luminescent layer are important (this is the area that Novaled was good at and the company is now part of Samsung).

The move from fluorescent to phosphorescent materials was critical for getting good efficiency. If you look closely at Samsung’s mobile phone OLED displays, for example on the Galaxy S5, the pixels are relatively small, allowing a black mask area that helps boost contrast.

There is a lot of activity in solution processing of OLED materials, but there is nothing commercial yet, and Rajeswaran said that Samsung has set the performance bar very high for mobile applications, so that displays for that application are unlikely to be solution based.

Encapsulation was an early problem as OLEDs are so sensitive to moisture and oxygen. Most encapsulation is based on glass to glass, but thin film encapsulation is still being tried. Vitex came in with multi-layer encapsulation but is going out again because of cost and speed and there are other candidates – HK Chung (Samsung) is developing organic encapsulation.

In 2007, Kodak developed an OLED that used correction and was made by LG.

Coming forward to today, Rajeswaran said that if you can buy LTPS substrates, you can make an OLED. Korea has put in the effort, but Chinese makers haven’t put in enough effort yet, they keep asking “what is Samsung doing?”, but Samsung has been working hard to control its suppliers and the information on how the process is going.

Samsung has basically failed in moving the FMM process to TV sizes. LG, on the other hand, has decided not to try to attack mobile applications.

Samsung has 17 production lines – each making 17 million displays. There are 300 chambers for deposition.

Later, Rajeswaran believes that the cost of a 55″ OLED will be around $500 eventually, but that means a $1000 TV – will 346 buy it? It’s not certain.

Monitors and notebooks could use OLED, but white displays are difficult for OLED compared to LCD and the power consumption advantages reduce.

OLEDs can be made flexible – you build them on glass, then take them off, but it adds cost

There is a lot of opportunity for OLED to be the second display technology in the market, but it needs to be the first one to gain scale and compete.

At the moment, the OLED manufacturing concept is still the same as it was 15 years ago, but there is a lot of “brute force” involved to make it a commercial business.

Rajeswaran now has a company, Grantwood Technologies, for developing new production technologies to get beyond FMM and WOLED. Why use glass as a substrate, he asked, if you have top emission? It would also be interesting to combine the touch panel with the OLED encapsulation.

He sees the possibility of new techniques that could scale in speed not substrate size – he would like to see tact times in seconds, not minutes.

Display Daily Comments

I remembered seeing Raj speaking at a previous event – it turned out that it was at the DisplaySearch conference in La Jolla in 2005. Looking back at my report of that event, it’s amazing how little has fundamentally changed, although the business is growing and at much higher volumes. Many of the points he made in 2005 remain the same today. (BR)