Reading time ( words)
Pete Starkey and Eduardo Benmayor, general manager for Aismalibar, discuss ways in which insulated metal substrates have been modified to relieve stresses on the solder joints of high-end LED assemblies during thermal cycling. Eduardo also describes a range of thermally conductive FR-4 laminates that can be processed like standard FR-4 for applications where thermal dissipation can be maximized without changing the design.
Pete Starkey: Eduardo, we are looking at posters describing new materials. Could you give us some idea of the range of new materials that you’re launching? What applications are the materials for, and what are the typical attributes?
Eduardo Benmayor: At productronica, we are showing our new laminates that we are producing. These laminates are trying to follow up on the market demand and the problems that customers are facing with controlling their PCB constructions. And we are following different routes, such as thermal release from any kind of FR-4 construction. A couple of years ago, we developed an FR-4 high-Tg material range with high thermal properties achieving thermal dissipation or thermal conductivity much higher than standard FR-4. These materials are used mainly by PCB makers that need to release temperature from their existing designs.
By using our materials, you can release temperature without changing the design specification. This is very helpful for PCBs because they don’t need to change the design or the overall concept. By only changing the material from standard FR-4s to thermal FR-4s, we are capable of reducing 10–15% of the temperature at the joints on the PCBs. This is very helpful for them because they don’t need to do the re-engineering.
Starkey: How do you achieve high thermal conductivity? Do you incorporate a thermally conductive filler into the resin?
Benmayor: The main development is the amount of mineral content that you are capable of adding into the resin. By adding mineral content into the resin, you gain the thermal properties but may lose many other properties that are needed in the FR-4 lamination, such as peeling strength, expansion, CTE values, MOTs, etc. This package of additional performance that you need on FR-4 is normally reduced with an increase of the thermal mineral content in the resin. We have had to find the balance and the right performance to fulfill all the FR-4 requirements.
Starkey: And in terms of processability, can you process these materials as if they were normal FR-4?
Benmayor: Yes, except in the drilling department, all the rest of the process remains the same.
Starkey: It’s effectively abrasive from the drilling and routing point of view.
Benmayor: Correct. When you are willing to achieve a higher thermal conductivity in a normal FR-4 factory, you will only see the change in the performance in the drilling department, where you will need to significantly reduce the number of holes per drill bit to keep the hole quality. All of the rest of the process—such as copper plating, pressing, laser drilling, and etching—remain with minor variations.
Starkey: Could you also give us some background on stress relief on solder joints?
Benmayor: Another topic that is very widespread in the market today is the solder joint problem that all the LED industry is facing in general. An aluminum base is an excellent material for thermal dissipation because the aluminum is used as a heat sink but has a problem related to thermal expansion.
Starkey: The expansion of the aluminum base.
Benmayor: Exactly. It’s better to use solid copper instead of aluminum for LED MPCB applications. As the coefficient of thermal expansion of aluminum is much bigger than copper, the mechanical stress on the solder joint becomes an issue, especially on the temperature cycle test. When the LEDs are powered-on, they run at a high temperature for a certain amount of time, and when they are powered-off, they cool down. This change in temperature causes stresses which affect the solder joint.
The high-end players have detected this problem, and they are focused on finding the best solution for it. We have been working on reducing the tensile strength of the dielectric layer to minimize the stress on the solder joints during the expansion and contraction of the metal base. By doing this, the young’s modulus is reduced significantly, and we are generating less stress on the solder joints.
Starkey: The dielectric acts as a cushion between the expansion of the metal base and the mechanical strength of the solder joints.
Benmayor: Yes. But, like always, you win something one side, and you lose on the other one. To achieve a lower tensile strength in the dielectric layer, we need to reduce the Tg of the material, and we need to drop the high Tg from 190 down to 100 or 110. That’s where we are working today. We are already releasing samples to our customers; they have to do the thermal cycle test, which is a long test because they are targeting 1,000 to 3,000 cycles. The Japanese are already talking about 3,000 cycles from -50°C to 150°C. And they want to evaluate on real products and see what is the cracking phenomena of the standard IMS compared with a low-modulus IMS.
Starkey: Are we talking about materials still with an aluminum base, a material with a copper base, or both?
Benmayor: The low modulus developments are focused on the aluminum base, where the expansion and mismatch are bigger than on a solid copper base. Other high-end players in the market have already decided to move away from the aluminum and go to the copper, where the expansion is lower, and many of the problems of the cracking are already released.
Starkey: Copper naturally has better thermal conductivity than aluminum, but it’s a more expensive material, and it’s heavier.
Benmayor: You’re right, but customers are more focused on technology than on cost reduction, and many times, they prefer to go to the copper to be on the safe side.
Starkey: They have the choice, and you are able to offer a solution, whatever their choice.
Benmayor: And on our copper-based material, to reduce the tensile strength of the dielectric layer, we add a very thin dielectric special polymer layer underneath the copper. Once again, it’s working very well in relieving the mismatch.
Starkey: Is this special polymer layer placed underneath the copper foil?
Benmayor: Yes. And then we have the thin epoxy layer with high mineral content and the functional copper layer over as usual. By doing this, we also reduce the tensile strength of the dielectric and further compensate for the expansion mismatch between the LED package and copper base.
Starkey: And this doesn’t have any bad effect on the adhesion of the copper foil to the dielectric?
Benmayor: No. We passed all the IPC tests. All of these products are capable of withstanding between four and five minutes in the 288°C solder float test, where normally IPC requires a minimum of 10 seconds. So we have plenty of room to guarantee to our customers that adding this material between the layers is extremely safe.
Starkey: Eduardo, thank you very much. I appreciate it.
Benmayor: Thank you, Pete.