LEDs, Smoke and Thermal Resistance

By Giles Humpston | Cambridge Nanotherm

“Red hot ‘n’ smoking” aptly describes what happens to LEDs, when the thermal system designer doesn’t quite get the sums right.  Similar to most other semiconductor devices, if an LED gets too hot then all sorts of bad things start to happen.  These can range from a reduction in efficiency, light quality and reliability through to shorter lifetime and ultimately catastrophic failure.  To prevent this, LEDs are attached firmly to a heat sink.  Between LED and the heat sink will be a MCPCB that provides the electrical connections to the LED.

With only three components, LED, MCPCB and heat sink, the thermal design of an LED system should be straightforward.  All it requires is that the pathway has low thermal resistance.  In theory this can be accomplished by selecting materials with high thermal conductivity and is, indeed one of the reasons why most heat sinks are made of aluminium.  Thermal conductivity, which has units of W/mK, is a simple material constant that refers to the ability to transfer heat by conduction with a high number being preferable.

In a MCPCB of the type commonly used with LEDs, the base plate is typically 1 mm thick aluminium and the circuit layer 1 oz (35 µm) thick copper.  Between them is a dielectric to provide the electrical isolation.  Because most MCPCBs have this basic structure, selecting the best should just be a matter of choosing the one with the highest thermal conductivity dielectric in the appropriate price bracket.  Right?  Wrong!

What actually matters from the perspective of thermal design for LEDs is the thermal resistance. Thermal resistance takes into account the length of the pathway that the heat has to travel and also the area of material available.  To simplify the comparison between different materials and make the units practicable, the area is usually normalized to 1 cm2.  So instead of selecting the dielectric with the highest thermal conductivity, we should select the one with the smallest thermal resistance.  Right?  Wrong!

Putting two thermally conductive materials together does not in itself guarantee the ready transport of heat between them.  Indeed, it is very difficult to predict how good the thermal pathway will be due to what is known as the interface resistance.  Interface resistance is a measure of the inability of heat to travel between two materials in physical contact.  Unfortunately, and as the manufacturers of thermal interface materials are acutely aware, interface resistance is extremely difficult to predict and often has to be determined by experiment.  Almost every physical factor that affects the condition of a surface will have an influence on interface resistance including roughness, porosity, hardness, state of oxidation, reflectivity, adsorbed species, pressure, temperature and even time.  Interface resistance is not a negligible parameter and can therefore have a significant impact on the ability to conduct heat away from LEDs.  Understandably, given the plethora of influencing factors it is the one parameter seldom given on data sheets.

Sometimes manufacturers will attempt to incorporate thermal resistance into data sheet values by quoting a number for absolute thermal resistance.  This is done commonly for heat sinks where the interface resistance between the heat sink and an infinite body of air at 25°C can be measured easily.  The absolute thermal resistance (Rth) of the heat sink is then expressed as a temperature rise for a given heat input (°C/W).

The paucity of information presents the LED thermal designer with a real challenge, as it is often impossible to make direct comparison between competing products.  Real-world comparative tests are then necessary to select the best components for a given application.  If, during testing, the LED glows red and starts to smoke, it’s usually an indication the thermal design is seriously flawed!

CN_HumpstonAbout the Author
Dr. Giles Humpston is a metallurgist by profession and has a doctorate in alloy phase equilibria.  He is a cited inventor on more than 250 patents and has co-authored over 150 papers as well as several text books.  Dr Humpston currently works as the Field Applications Manager for Cambridge Nanotherm on thermal substrate technologies.

2 Responses to LEDs, Smoke and Thermal Resistance

  1. EAlvarezB says:

    We have found that in many systems after achieving the right conductivity, the convection stage then becomes the limiting factor. Can you comment if you have seen this behavior as well ?

  2. Pingback: LED Journal: LEDs, Smoke and Thermal Resistance | Cambridge Nanotherm Ltd

Leave a Reply