Without successful thermal management, LEDs’ light output, color and efficacy can quickly degrade.
Effective thermal-management designs call for an efficient, heat-conduction path that minimizes thermal resistance (or increases thermal conductance) along a path from the junction to the die submount — the first heat sink — and continues to the printed circuitboard (PCB), the second heat sink, upon which the die submount is installed.
Such a system also needs to integrate appropriate, thermal-conductive materials (that is, high, thermally-conductive substrates such as alumina and metal-core PCBs, and good thermal epoxy or grease), packaging techniques (good solder joints), and substrate geometry that may include thermal (via holes) and flat mating surfaces.
Thermal resistance is defined by the resistance of a material to heat conduction, which depends on the materials’ inherent conductivity property (k) as well as its length and cross-sectional area. It also relates to the temperature difference between the two, adjacent surfaces, where heat transfers from one to the other along a perpendicular path, thus, dividing the heat-energy flow rate.
Both definitions are analogous to electrical resistance. The unit of thermal resistance is the degree Centigrade per watt (°C/W), where W is thermal wattage or power. In an LED, generated, thermal power equals injected electrical power (i.e., voltage times current) driven into the LED, minus the emitted optical power. This equation shouldn’t be mistaken for electrical power when calculating the thermal resistance in an LED module. In more-efficient LEDs, the generated thermal power is lower than that found in a less-efficient LED, even when the same amount of electrical current is injected.
In the solid-state lighting (SSL) industry, it’s common to use a model similar to that shown (Fig. 1), where material 1 is the LED chip, material 2 the chip submount (the first heat sink) and, sometimes, there’s a material 3 for the PCB (second heat sink). The junction temperature TJ is analogous to T1; T2 is analogous to the temperature of the first heat sink; similarly, add T3 for the PCB, the bottom of which is usually the ambient temperature when a stand-alone LED module is tested.
This model is used along with the measured values of forward voltage, T1, and T2 (and perhaps T3) to determine TJ of an LED. Use caution, however, to not mix thermal resistance with electrical resistance and thermal power with electrical power.
A more rigorous way of determining TJ is to use a full, 3-D, finite element or finite difference method to model the thermal behavior (i.e., solving a heat-wave equation) of an LED when a constant current is injected into it, with all of its surrounding components functioning.
Also, be mindful that thermal resistance is a function of ambient temperature, because k and the geometric parameters all change with temperature. The modeling results should be used in conjunction with as many measured parameters as possible, under different ambient temperatures.
The science behind accurately determining TJ and TJMAX are quite complex and time consuming. It requires an iterative effort between rigorous modeling, or calculations, and extensive measurements of various LED parameters. The modeling results should be used in conjunction with test results, but under different ambient temperatures, to determine TJ. Multiple iterations will ensure the convergence of a convincing TJ value.
Quantifying relationships between TJ, TJMAX and optimal drive current for various LEDs modules to be used in the sign and illumination industries and, ultimately, designing brighter and more efficient LED modules, are some subjects of my upcoming research and development plans.
Did you enjoy this article? Click here to subscribe to the magazine.