Driving down HB-LED package costs


Griff Resor,

Resor Associates,

Maynard, MA USA

HB-LED packaging accounts for 40% to 60% of finished LED cost. This makes packaging the #1 target for cost reduction.

By convention, LED package costs do not include the cost of converting a finished LED to a lighting product ??? the "Luminaire". Mark McClear of Cree has repeatedly said an integrated approach is needed to deliver an attractive value for consumers. I'll come back to his point at the end.

A few years ago HB-LED packages were assembled from many discrete parts. These included a metal lead frame, a molded pocket for the LED, a zener diode, a mirror coating for the pocket, a heat sink for the LED, wire bonds for electrical connections, a sprayed on phosphor coating, an encapsulant to protect the assembled parts, and an attached lens.

Today LED packaging is using an integrated thin film approach. A single AlN "panel" about 100 mm square is used. Top side patterns provide a LED heat sink, area for wire bond connections, and electrical traces to vias that connect to the bottom side. Bottom side patterns provide for heat transfer and surface mount connections. Significant savings are realized by doing "panel" scale assembly. For example, lenses are molded 500 at a time on one AlN panel.

Philips-Lumileds has changed their LED structure, to enable flip-chip mounting of the LED on the AlN panel. This eliminates the wire bond cost, and eliminates the area needed for wire bonding. This enables a smaller, lower cost LED package. Expect more flip-chip LED designs.

An analysis of packaged LEDs used for LCD-TV backlighting shows that much of the HB-LED package is a "frame" around the LED. This area does not increase significantly as LED area increases. The real goal is to reduce the cost/lumen, not just the package cost. By increasing lumens emitted from a package, one can reduce the packaging cost per lumen.

This tactic is already being employed in HB-LED packages ??? several small LED chips are being mounted inside one package. This provides an added benefit. The LEDs can be wired in series to increase the operating voltage. Higher voltage reduces the power conversion costs in the Luminaire. Which leads back to Mark McClear's point ??? consider the Luminare cost.

The Resor Associates' COO model was designed to study the tradeoff between package costs, LED area and LED yield. It is clear that larger LED chips significantly improve the cost/lumen for finished LED's. As yield improves, the optimum tradeoff shifts to larger LEs. So add improved LED yield to the roadmap for LED package cost reduction!

From the life cycle of ICs and FPDs it is clear that LED makers will use some of their cost savings to elaborate the product in ways that reduce the finished Luminaire cost. The most promising idea that I've seen is "integrated" LEDs. For example, one could form two strings of 40 LEDs each. These could operate directly at a line voltage of 120volts, eliminating the power conversion parts in the Luminaire. Why don't we see integrated LED products today? Defects/cm2 are too high in LED fabs ??? a significant reduction in defect levels is needed first.

Bottom line, there are many ways to reduce the packaging cost in LEDs. But to realize some of these one has to look beyond the traditional LED package to the Luminaire and to the LED chip.

Solid State Technology, Volume 55, Issue 3, April 2012

More Solid State Technology Current Issue Articles

More Solid State Technology Archives Issue Articles