Packaging engineering of LED semiconductors, is a key contributor to producing better discrete component designs that perform more efficiently in a wide variety of operational and environmental conditions, than current conventional formats. Packaging engineering will be of increased importance as demand for LEDs to fulfill new, higher performance, higher brightness applications continues to manifest and gain momentum. Current packaging performance efficiencies, compared to LED die performance attributes, clearly shows that most conventional packages existing to date, are inadequate for the demands of many current and future applications.

This holds true for axial leaded as well as surface mount leaded and leadless packaging technologies (SMT). Most existing LED lamp designs are an outgrowth of traditional formats that support cube type LEDs, with length and width dimensions ranging from 8 to 12 mils and thicknesses of approximately 8 to 10 mils. Most common varieties of packages and the “open” tools that are used to produce them, including the T-1 3/4 (5 mm) lamp, are designed with this size die in mind. Most LED package styles are also designed for conducting substrate LEDs (example AlInGaP on GaAs which have ohmic contacts on the top and bottom of the LED, require one wire bond and conductive epoxy contact for the bottom), as opposed to planar circuit LEDs on insulating substrates (such as InGaN on Al₂O₃ LEDs having two contacts for wire bonding on the top side). Much work also goes into ensuring that existing packaging designs are practical to manufacture and can be produced at low cost. However, LED packages that were originally designed and perform exceptionally well for such applications as indicator lights, and cost pennies to produce, need to be reviewed for more demanding, reliable and cost-effective service.

The T-1 3/4 (5 mm) lamp for example, costs pennies to produce via highly automated assembly procedures, yet has very poor optical performance characteristics. It is estimated that the T-1 3/4 LED lamp has an optical efficiency of <30% with conducting substrate style LEDs and this efficiency is even less with new style InGaN/Al₂O₃ LEDs that are generally 13 mil square with heights running <5 mils in thickness. SMT device packages are generally known to have optical efficiencies even less than this (<20%). Additionally,other factors such as thermal and chemical factors are also associated with traditional style package inadequacies for many new applications. For example, T-1 3/4 packaged Blue LEDs that improperly use encapsulant materials that will prematurely age, will lose an additional 40-50% of their output when subject to ~5 – 10 thousand hour aging. It can be clearly seen, that if the goal is higher performance (for example YAG-coated Blue InGaN Al₂O₃ “White” T-1 3/4 LED lamps currently achieve ~10 lm/W and the goal is to quickly achieve 120 lm/Watt per Japan’s MITI), then to achieve it, this gain will not only have to come from the LED semiconductor, but also the package that it is into which it is assembled.

New package engineering must consider better optical and electrical performance. Reliability and continued high performance after thousands of hours of use must minimize the collective effect of environmental and operational stresses on the resultant package design. Better thermal management needs to be incorporated. Good designs must also consider the variety of new LED die types and wavelengths that they produce and the effects these wavelengths have on the packaging die attach and encapsulating materials and processes selected. A review of all of these parameters is needed in order to achieve optimum LED component product results.

For example, a review of the conventional T-1 3/4 LED lamp, for modification to better accommodate an InGaN/Al₂O₃ 470 nm LED with POWER-Ga(i)N™ (Ref.UNIROYAL Application Engineering Note: “POWER-Ga(i)N ™High Brightness InGaN LEDs”) die geometry, can be used to illustrate how even this style package can be improved to yield better results:

For illustration purposes the sketches and accompanying notes do not present all of the optical and material improvements that may be achieved through the redesign factors cited herein. Combined with the other factors mentioned in the table above, the existing T-1 3/4 could be readily redesigned to better accommodate an InGaN/Al₂O₃ 470 nm LED with POWER-Ga(i)N™ LEDs. With the successful incorporation of new design features, product performance improvements on the order of >50% optical, >100% in power efficiency, better life and reliability, can be realized yielding, even with the T-1 3/4 package, not 10 lm/W but significantly better results with current and future LED die.