LED display manufacturers are perpetually searching to optimize the efficiency of their products without sacrificing performance. While these fields are frequently inversely correlated, occasionally an innovation will arise that allows manufacturers to improve both simultaneously. One such development is “flip-chip” architecture, an advancement that represents an alternative method for mounting LED packages onto package substrates or, in the case of Chip-on-Board packages, directly onto circuit boards. While flip-chip is a catchy phrase, a more accurate term for the concept is probably “wireless bonded chip architecture.” While not a new concept to the world of electronics at large considering it was first commercially deployed by IBM in the 1960’s, wireless bonding has taken on heightened importance in the LED display industry for the potential improvements it can deliver regarding production costs, heat dissipation, reliability, and durability. Government systems have taken advantage of these benefits since at least the 1980’s and will continue to invest in the technology moving forward. With such broad potential benefits and applicability, the LED display manufacturing industry recognized that wireless bonded architecture might be worth adopting in LED chip production for large-format displays if it were feasible. So is it?
How Do Traditional Wire Bonds Work?
Before we forecast the large-scale feasibility of this alternative chip architecture, we’ll need to detail how wireless bonded architecture works. But before we do that, let’s recap how a traditional wire bond is assembled. The architecture of a conventional wire bond LED package has the active area of the semiconductor chip facing upwards as the chip is mounted onto the substrate or board with epoxy, usually with a dielectric layer in between. Wires are then used to interconnect bonding pads on the outer edges of the active area of the chip to the external circuitry of the substrate or board it is mounted on. These bonding pads are located on the outsides of the active area in order to minimize the amount of wiring needed to reach them. In this arrangement, light emits from the top of the chip and heat dissipates through the bottom.
What’s Different About Wireless Bonded Chips?
Wireless bonded architecture flips this design upside down (literally) by rotating the orientation of the emissive elements of the chip, allowing an unobstructed path for light from the chip to the viewer. Though ultimately it delivers a more efficient product and requires fewer materials overall, it does add a new step in manufacturing. Towards the end of the chip manufacturing process, the bonding pads on the active surface of the chip receive a small dot of solder. Unlike with traditional wire bonding, these pads do not necessarily need to be located on or near the outside edges of the surface since rather than connecting to external circuitry via wires looped around the other layers of the chip, they are simply attached to the circuitry directly through thermosonic bonding or reflow soldering. These bonds leave a small sliver of space between the surface of the active area of the chip and the surface of the substrate or board, which is then filled with epoxy to act as a thermal bridge heat can use to escape.
The shorter distance of the interconnections between active area and circuitry, along with the fact that manufacturers are now free to include as many of these touchpoints as possible, reduces inductance, cutting heat production while increasing heat dissipation efficiency, a benefit that is compounded by the fact that most of the heat created can now release through the top of the package, rather than its bottom. This in turn reduces thermal decay and allows the package to handle slightly higher currents, thereby increasing light output, which itself is unimpeded by the presence of wires. The compacted architecture also improves the package’s resistance to shock.Occasionally an innovation arises that allows #LEDdisplay manufacturers to improve both efficiency and performance. Introducing “flip-chip” architecture. Click To Tweet
As various manufacturers direct massive investments into improving their microLED manufacturing, they have found the benefits to size, durability, and light production solve a lot of the problems they’ve faced in producing reliable and high-performing microLEDs, specifically for head-mounted display applications (for virtual reality purposes). The benefits of this investment into wireless bonding architectures, though initiated by manufacturers interested in microLED, can be leveraged by the LED display manufacturing industry at large to improve all packaging efficiencies for diodes of all sizes.
So is it Feasible?
So by rearranging the assembly process, packages with wireless bonded architectures become notably smaller, more efficient, and more productive. When aligned en masse, these smaller packages create slimmer, smoother display surfaces for consumers as well. So why isn’t the display industry rapidly transitioning to this method? Here’s where we return to the feasibility question. Despite the plentiful benefits wireless bonding can introduce, because the manufacturing process is different, many manufacturers are not currently equipped to handle the large-production of these packages. Second, it’s not clear that overhauling their current set-ups would pay off, considering such a dramatic change would likely hinder pursuits of other R&D pursuits already underway. Furthermore, wireless bonded packages are much more difficult to maintenance, though their increased durability and reliability means failures are far less likely. A minor issue is that within wireless casings the shortened interconnections are stiff, so as chip temperatures change, the interconnections can crack. These packages also prevent you from testing parts in advance so issues can arise if the diode binning process is not strict enough. Still though, with durability and reliability increased by roughly 10 times, performance issues are not really much of a worry.
Wireless bonded architecture presents an intriguing path for display manufacturing but due to the drawbacks mentioned here, it is ultimately probably best suited for use in small-scale arrays rather than for the entire composition of a display. That said, we’ll see what the future holds – perhaps another innovation will flip this upside down, too!