MicroLEDs entering the display market
Today, the (flat panel) display market is mainly comprised of liquid crystal displays (LCDs) and organic light emission diode (OLED) displays. They target a broad range of end products, for example, large displays for televisions and large information displays, smaller displays for tablets and smartphones, and microdisplays for AR/VR applications. LCDs use (inorganic) LED backlight to emit light through a matrix of liquid crystals to generate the images. OLED displays, on the contrary, are self-emissive, using organic compounds to emit light in response to an electric current.
In addition to LCDs and OLEDs, the microLED display is making inroads, promising more color and higher brightness at lower power.
More recently, a third player, the microLED display is making inroads, which promises more color and higher brightness at lower power than the traditional displays. The first products using this technology have recently made an impressive entrance into the market. Think about microLED televisions and video walls that have been introduced by major display players, promising the ultimate visual experience for home, entertainment as well as business environments.
MicroLEDs are microscopic versions of traditional LEDs (>1mm), measuring less than 50µm. Whereas traditional LEDs are packaged individually, microLEDs are used in bare die form, and a multitude of them need to be incorporated into a display. They are organized in sub-pixels, each containing red, green and blue microLEDs to form a color display. The emitted color is set by the (bandgap of the) inorganic material used to make the LED, e.g., AlGaInP for red and InGaN for green.
Large, modular displays: an opportunity for microLEDs on thin-film transistors
Active matrix microLED displays are controlled by an array of backplane transistors, responsible for switching and driving the individual display pixels. When looking at the material used for making the backplane, microLEDs fall into two categories. One uses Si-based transistors, made with standard CMOS fabrication process flows. These transistors can be made very small, resulting in backplanes with a very high transistor pitch – ideal for e.g. high-resolution VR/AR applications or projectors. Si backplanes are, however, relatively expensive and limited in size and non-transparent.
A second category uses thin-film transistors (TFTs), made of either amorphous Si, low-temperature polycrystalline Si (LTPS) or indium-gallium-zinc-oxide (IGZO). TFTs can be processed on larger substrates than Si, and have the potential to achieve a lower cost per unit area when processed in larger volumes. MicroLEDs on TFT substrates are the focus of this article.
MicroLEDs can be used in large, modular displays intended for televisions or video walls for home, cinema and advertising, or for conference and meeting rooms.
One of the envisioned applications are large, modular displays intended for televisions or video walls for home, cinema and advertising, or for conference and meeting rooms. Depending on the size and number of the modules, the final display can measure 100 to 200 inches, and even larger. In this application, microLEDs on TFTs are expected to outperform OLEDs on TFTs. MicroLEDs are far more efficient, offering a higher luminance for the same drive current. Also, as there are no organic layers, the microLED displays do not require encapsulation, making a seamless transition from one module to another easier. With OLEDs, on the contrary, each module needs to be encapsulated separately.
Unlike OLEDs, microLEDs intended for larger displays can not be monolithically processed on one and the same substrate. Therefore, for mid- to large sized displays, other approaches such as pick-and-place are needed for their fabrication. Following this approach, the microLEDs are fabricated using three different epiwafers (for red, blue and green), diced, and transferred onto a TFT backplane using a high-speed pick-and-place system.
Design considerations for microLED TFT backplanes
Enabling high-performant microLED displays introduces new challenges to the way the backplane is designed. Today, different electronics design approaches are being pursued, one starting from active-matrix OLED (AMOLED) design, the other starting from a passive, PCB-based design approach.
Each of these approaches comes with its own pros and cons in terms of e.g. gray level, flickering, pixel pitch, heat dissipation or power consumption. Imec, drawing on its many years of experience in TFT circuit design, designs alternative pixel circuits that can take microLED displays into the next level.
Imec designs alternative pixel circuits that can take microLED displays into the next level.
Below is an example of a new TFT circuit design for modular microLED displays – co-developed with Barco – that combines the best parts of different existing driving approaches.
Example: A new, 6T2C pixel circuit for driving microLED displays
When developing a backplane circuit for a specific type of display, designers need to make several decisions. They have to choose the optimal matrix architecture (active matrix driving vs. passive matrix driving), the way the gray level is set (analog driving vs. digital driving), and the LED programming mode (voltage vs. current programming). Imec has compared and assessed various state-of-the-art approaches. As a result, researchers have come up with a new, hybrid approach that combines the best part of the different approaches in a new 6T2C pixel circuit – able to tackle the numerous emerging challenges for microLED displays.
Active vs passive matrix architectures
Both passive matrix and active matrix displays use horizontal and vertical lines to respectively select a line and apply the corresponding image data to the columns to drive the pixels in that line. The lines are activated in high-frequency order, such that the human eye regards it as a planar frame rather than a ‘line scan’. In passive matrix driving, used for example in today’s microLED walls, pixels in the unselected lines are off. They turn on only for a short period, when the line is chosen. In other words, only one line is emitting light at the time. In active matrix driving (such as in AMOLED designs) all pixels contain a storage element. This allows them to remain active throughout the switching cycle until their values are updated again. Hence, they also emit light when other lines are being programmed; reducing the pixel brightness requirement and therefore the current levels. That’s the fundamental difference between active and passive matrix driving.
Passive (left) vs. active (right) matrix driving schemes.
It turns out that active matrix designs are more beneficial for power consumption, cost and visual image quality. In passive matrix displays, pixels are only activated during a short period. Hence, significantly higher peak brightness and thus currents through the LEDs are required to achieve the same overall brightness, which drives up power consumption and heat dissipation. For large, modular displays containing millions of microLEDs, an active matrix-driving scheme is therefore preferred over a passive matrix design.
For large, modular displays containing millions of microLEDs, an active matrix-driving scheme is preferred over a passive matrix design.
Gray level: digital or analog driving?
The gray level (or brightness) of an individual LED is determined by the amount of current flowing through the LED. The gray level of each individual LED that makes up a pixel contributes to the overall brightness of the (micro)LED display.
Panel-based designs used for AMOLED displays typically implement an analog-type of driving: the gray level is determined by the analog voltage (or current) applied to the pixel and current through the OLED. In this approach, a higher voltage (or current) level results in a higher light emission, and thus a brighter pixel. But for inorganic (micro)LEDs, this approach comes with a drawback: changing the current through the LED to vary the gray level also affects the wavelength of the emitted light, causing an undesirable color shift.
Digital driving is the preferred option for a microLED display.
Therefore, digital driving is the preferred option for a microLED display. A digital driving approach uses pulse width modulation (PWM) to determine the amount of current flowing through the microLED. Following this approach, a fixed current level is now applied to every LED (so no color shift). But the average time that the LED is on (also called the duty cycle) can be varied to determine the average light emission – and thus the gray level of the pixel.
A 12-bit coding table
To implement digital driving with an active matrix backplane, different coding schemes exist for programming the PWM. These coding tables contain the details of the exact time periods that the microLEDs are on or off. Our team has proposed a unique 12-bit coding table to achieve minimal dark time and optimized optical appearance, resulting in the lowest possible flickering of the display.
The amplitude of the first harmonic in its Fourier spectrum is reduced for all gray levels by implementing the proposed 12-bit coding table, resulting in reduced flickering in the display.
A novel 6T2C pixel circuit
When designing the pixel circuit, again, several choices need to be made. For example, the designer can opt for a more traditional 2-transistor-1-capacitor (2T1C) design, in which one transistor selects the pixel, and another transistor drives the current through the LED by setting a data voltage. However, any variations in transistor characteristics can result in current variations (and hence, color shifts), and therefore, a voltage driven approach to set the current is not a good choice.
Imec researchers have therefore developed a hybrid approach where a fixed current through the microLED is accurately set by a so-called current mirror (defined by two transistors), and the PWM data are applied by voltage levels (enabled by two switching transistors). These switching transistors turn the fixed mirrored current on or off, following the coding table. The two remaining transistors of the 6T2C circuit select the pixel where the input current is updated.
The proposed 6T2C pixel circuit for driving microLED displays.
By using this pixel circuit, the team was able to implement the proposed hybrid design approach to drive the microLED display with the best performance. They have also proposed variations of the 6T2C pixel circuit to broaden its applicability. For example, a variation of the design was proposed to reduce the overall area consumed by the pixel circuit, based on the concept of current mirror sharing. In addition, a global shutter design was proposed to improve the synchronization with which the different sub-modules of a modular display are being refreshed.
The new 6T2C pixel circuit allows microLED displays to be driven with the best performance.
Towards high-volume manufacturing of TFT backplanes
Innovative pixel circuit design is crucial for enabling high-performance large-area microLED displays. But the design choice can also impact the fabrication cost of the overall microLED display. For example, the active TFT-based matrix design as proposed in this article will lead to lower fabrication costs when compared to active-matrix Si CMOS or hybrid Si CMOS/TFT backplane technologies. The lower cost potential of the TFT backplane design is linked to the use of larger substrates (compared to Si) and to the possibility of producing the TFTs in large volumes. The latter can be enabled through a unique foundry model, set up in analogy with the multi-project wafer services that have existed for years for Si CMOS-based technology.
TFTs can be produced in large volumes, enabled through a unique foundry model.
To illustrate these manufacturing capabilities, the team has fabricated a microLED prototype display with LTPS backplane, based on the pixel circuit design that was discussed above. LTPS was preferred over IGZO as the material for making the TFTs since LTPS can provide higher currents for driving the display. Characterization results of the LTPS panel show, for example, very good latch retention data which indicate how long the data can be maintained at the current mirror and on the capacitor after switching.
Latch retention data (oscilloscope).
Conclusion
We have proposed a hybrid approach for designing microLED TFT backplanes that combines the best of existing approaches (i.e., LED wall designs and AMOLED designs): a current programmed, PWM-driven active matrix design. A novel 6T2C pixel circuit allows to implement the hybrid design approach to drive microLEDs with the best possible visual quality. When manufactured in high volumes, the use of TFT backplanes for driving the microLEDs offers a lower-cost alternative while maintaining high-performance operation as compared to traditional passive PCB-based design or to the approach to embed Si CMOS chips in pixels.
This article was published in LEDs Magazine.
Want to know more?
- The results presented in this article were presented by Kris Myny at the 21st International Meeting on Information Display (IMID 2021). Interested in receiving the paper ‘Design considerations for µLED displays’ by L. Verschueren et al? Fill in our contact form.
- Would you like more information on imec’s foundry model for TFT manufacturing? Fill in our contact form.
