A substrate’s thermal performance influences die cooling, and hence the lifetime of solid-state lighting.
Legislation and environmental concerns are prompting a mass switch to newer equipment and appliances. The rapid growth of solid-state lighting is just one example. In Europe and North America, eco laws have consigned incandescent bulbs to history. LED lighting has taken over, mainly by realizing high-performing lighting at far lower power levels.
Prices for LED replacement lamps are dropping, but remain higher than traditional incandescent bulbs. The power consumption is lower, of course, and that’s the main point, but their greater longevity is needed to offset the price premium in buyers’ minds. LED bulbs potentially can last the lifetime of the light fitting, eliminating the cost and inconvenience of replacing failed units.
To achieve this requires designers to consider the desired operating lifetime of the chosen LED light engine under the expected operating conditions. This is essentially a thermal management challenge, to extract heat from the LED chip in pursuit of greater reliability. Poor thermal design will permit the die temperature to rise, ultimately shortening its lifetime. Excessively high die temperatures can also affect the brightness and color consistency of the light output.
Reliability vs. cost. In today’s price-sensitive markets, however, over-engineering thermal management adds costs that are simply unsupportable. An optimal solution will perform well enough to ensure the LED die temperature remains below a limit that will permit the bulb to last as long as the blister pack says.
Because the lens occupies most of the upper surface of a packaged LED component, most of the heat from the die must be extracted through the underside and goes into the substrate. The package is engineered to have low thermal resistance between the die and base, but this represents only the first stage of the heat’s journey. The substrate’s thermal performance has a critical influence over cooling the die, and hence the lifetimes of all the LEDs mounted on the substrate, which make up the light engine.
Standard FR-4 materials have relatively poor thermal conductivity. Historically, designers have employed various ways to enhance the thermal performance, but some of these – such as inserting heat slugs or filled and plated vias – can significantly increase board cost. This is undesirable in cost-sensitive applications like lighting or, indeed, many other high-performance computing or power-conversion products, such as power adapters, battery chargers, smart-TV sticks or boxes, smartphones, and automotive electronics like ECUs or autonomous-driving computers.
Economical but high-performing solutions to thermal-management challenges are needed. Where heat generation is high, such as LED lighting panels or inverters for high-power applications like welders or large motor drives, insulated metal substrate (IMS) has proved to be a valuable technology. Materials specialists continue to develop higher-performing dielectrics that combine high thermal conductivity with good electrical insulation, despite the fact that these two parameters are often at odds: good electrical insulators. This, of course, has always been at the root of FR-4’s inadequacies as a thermal substrate.
New choices, new opportunities. In the past, laminates and prepregs have often been heavily loaded with ceramic filler particles to boost thermal conductivity. Unfortunately, these required very high pressures and temperatures during board fabrication. New generations of materials are easier to work with, and are a viable option for those applications that need higher performance than conventional FR-4 can deliver, but where an IMS may be uneconomical or undesirable for any other reason. Typically, there is also more freedom to optimize the weight and trace widths of the encapsulated copper foil. Novel laminate/prepreg formulas have polymer matrix structures that are compatible with a range of epoxies or polyimides, including low-signal-loss materials, yet provide several times better thermal impedance than FR-4.
The arrival of new materials signifies a real broadening of choice for design engineers. This should provide much greater flexibility to design for reliability within cost boundaries that are acceptable to their target markets. When the materials available can make a real difference to the quality and performance of the end-product, product designers can give due consideration to thermal aspects in the early stages of their projects.
This would indeed be a break with the past; surveys have shown engineers typically regard heat dissipation and thermal management as low priorities, and often give thermal issues only cursory consideration, after the design has been completed or during prototyping. With the wider variety of thermal materials now available, spanning a wide price/performance spectrum, more lateral thinking designers can use them to gain an advantage.
Materials suppliers, for their part, must encourage awareness and adoption of the latest materials throughout the industry, including printed circuit fabricators, board designers, circuit designers, and systems designers. We all need to make the effort to qualify new products according to important industry standards, deliver technical support and direct instruction for engineers and fabricators, and provide easy access to process guides and calculators to help choose the optimum approach and ensure the best possible performance.