These displays could play crucial roles in ultra-portable products such as next-generation pico-projectors and in emerging fields such as biophotonics and optogenetics.
Texas Tech University and III-N Technology, Inc., Lubbock, TX,
the US Army RDECOM CERDEC Night Vision and Electronic Sensors Directorate, Ft. Belvoir, VA
In the last 20 years, high-brightness LEDs based on III-nitride semiconductors have achieved dramatic advances alongside developments in indicators and solid-state lighting. For example, InGaN-based white emitters have achieved a luminous efficacy of more than 150 lm/W, which is much higher than those of other self-emissive devices, such as organic LEDs (OLEDs) and electroluminescent emitters. With the incorporation of multiple quantum wells (MQWs) as the active region, LEDs have a narrow emission band of about 25nm, providing a basis for high color purity and chromatic fidelity. With their intrinsic material properties and low-voltage operation characteristics, LEDs have a much longer operational lifetime (>100,000 hours), and can be operated at extreme conditions such as high or low temperatures (-100 to 120 °C) and humidity. All of these intrinsic properties make LEDs an ideal candidate for many applications where performance, reliability, and lifetime are critical.
Since their inception, μLED arrays based on III-nitride semiconductors have emerged as a promising technology for applications including self-emissive microdisplays, and single- chip, high-voltage ACLEDs for solid-state lighting. The InGaN-based μLED array has opened a new avenue for the multi-site photostimulation of neuron cells, and offers the opportunity to probe biological neuron networks at the network level. In particular, III-nitride μLED arrays provide high brightness/contrast/resolution/reliability, long life, compactness, operation under harsh conditions and under bright daylight — properties that cannot be matched by more conventional liquid crystal display (LCD), OLED, and digital light processing (DLP)-based microdisplay technologies.
III-nitride μLED array integration with Si CMOS accomplishes a high-resolution, solid-state, self-emissive microdisplay operating in an active driving scheme. The fabricated blue or green video graphics array (VGA) microdisplays (640 x 480 pixels) have a pixel size of 12 μm, a pitch distance of 15 μm, and are capable of delivering real-time video graphics images. An energy-efficient active driving scheme is accomplished by integrating micro-emitter arrays with CMOS active matrix drivers that are flip-chip-bonded together via indium metal bumps. This success means that InGaN μLED arrays could play crucial roles in emerging fields such as biophotonics and optogenetics, as well as ultra-portable products such as next-generation pico-projectors.
In a monolithic μLED microdisplay, the μLEDs themselves and the interconnection between these μLEDs (the signal transmission paths, including all the metal lines for n- and p-type contacts) are all integrated on the same GaN wafer. This monolithic integration has the merits of easy and quick demonstration and characterization, since the microdisplay itself is an independent package, and the driving circuit can be designed using an off-the-shelf CMOS integrated circuit (IC) chip. However, to achieve a full-scale microdisplay in a monolithic μLED array, connecting the huge amount of control signals from a separate driving circuit to the microdisplay within a limited space is a very difficult, if not impossible, task.
For example, a monochrome VGA format microdisplay with 640 × 480 pixels requires a minimum of 1,120 connections to the rows and columns in order to drive the microdisplay. Given a pixel pitch of 15 μm (the perimeter of the microdisplay is only about 35mm for VGA), wire bonding or tape bonding from the periphery matrix array pads to the driving circuit is highly challenging. A much more serious issue is the driving approach and the achievable performance. Such a monolithic μLED microdisplay connected with a separated driving circuit can only be driven in the passive mode. In this mode, one can only independently access one row at a time. Furthermore, the addressing time of each pixel, or the time one pixel is in the “on” state, is inversely proportional to the number of lines (rows) in the display matrix. For high-information-content (high-resolution) video displays, or for very high-luminance-requirement sunlight-readable displays, the maximum driving current limit is reached, and the light output saturates with further increase in current; hence, the light efficiency and thermal dissipation become serious issues. For high-information-content displays, the desired driving approach is active matrix driving, which cannot easily be implemented in monolithic μLED microdisplays. Until now, microdisplays based on inorganic semiconductors that are capable of delivering video graphics images have not been realized.
These devices will be operated on battery power, and the efficiency of all components in the device package directly affects how long it can operate between charging. The challenge for achieving such a device is that III-nitride microemitters cannot currently be fabricated directly over Si IC circuitry. To overcome the above difficulty, a hybrid μLED microdisplay concept was developed that is similar to the widely deployed scheme of hybrid focal plane array detectors, which utilizes the technique of flip-chip bonding via indium metal bumps. However, the typical size of indium metal bumps used in hybrid focal plane array detectors is around 20μm.
Like conventional III-nitride LEDs, μLED structures were grown on (0001) sapphire substrates by metal-organic chemical vapor deposition (MOCVD). The insulating sapphire substrate provides an ideal platform as the display surface and for isolation between individual μLEDs. All of the emitter structures used are based on InGaN/GaN MQW confined between n-type and p-type GaN barriers. The emission wavelength was tuned by adjusting the In composition in InGaN alloys in the MQW active region. N- and p-GaN layers form the barriers and Ohmic contact layers. Reducing the contact resistance will enhance the hole injection efficiency, reduce threshold current and heat generation, and increase the device operating lifetime.
To obtain a sense of the microdisplay brightness, the luminance of the green μLED pixels were characterized. A 12μm pixel outputs roughly 1 mcd/μA and the luminance increases almost linearly with driving current. Power dissipation within the μLED array is only about 0.8 W for a full VGA (640 × 480 pixels) microdisplay if every pixel within the μLED array is lit up simultaneously. This estimate represents the upper limit of power dissipation since normally only a fraction (~25%) of pixels are lit up for graphical video image displays.
In summary, a prototype full-scale (VGA), high-resolution, self-emissive, blue/green microdisplay is capable of delivering video graphics images for high-brightness pico-projector and head-up/wearable display applications. Compared with microdisplays based on LCD, OLED, and DLP, microdisplays fabricated from III-nitride μLED arrays can potentially provide superior performance. Unlike LCDs that normally require an external light source, III-nitride blue microdisplays are selfemissive, result in both space and power savings, and allow viewing from any angle without color shift and degradation in contrast. On the other hand, OLEDs must be driven at current densities many orders of magnitude lower than semiconductor LEDs to obtain devices with a reasonable lifetime, and hence, are not suitable for high-intensity use.
This work wads done by J. Day, D.Y.C. Lie, J.Y. Lin, and H.X. Jiang of Texas Tech University; J. Li of III-N Technology; and C. Bradford of the US Army RDECOM CERDEC Night Vision and Electronic Sensors Directorate. For more information, visit http://info.hotims.com/40433-305.