I need to attach a heat sink to a PCB. The PCB is a 560 mm x 9 mm strip of LEDs. The LEDs vary from mid-power to high power. I have a pure copper solution for the high power. I am investigating the lower power end where a 25 mm wide strip would be sufficient for thermal management. The lower end, mid power, is budget sensitive and am experimenting with aluminum extruded heat sinks.
My concern about an aluminum heat sink is galvanic corrosion where aluminum touches copper. In the first prototype for the pure copper the thermal pad on the PCB were unplated bare copper.
This image shows the pure copper thermal path from the LED's thermal pad to 1/2" copper water pipe.
This is the footprint for another prototype recently fabed. It shows the copper thermal path from a Cree XP or OSRAM high power LED. The center pad of the footprint is the LED's thermal pad. It goes to a copper strip along the edge which gets screwed to either an aluminum bar or copper bar.
With a thinner copper bar I can get the copper bar closer to the LED reducing thermal resistance.
I have new strips with ENIG plating.
HASL plating is not an option
This is using a footprint for an Oslon SSL similar to the KiCad image above.
The next experiment is to try three different aluminum extrusions.
I used the ENIG assuming I can screw either aluminum or copper to ENIG with very little chance of galvanic corrosion but have no idea why I believe that may be true or not. If not, it's still a nice planar surface.
I have purchased 0.062" thick $2.75 6061 T4 aluminum (1.35" wide) and $5.50 C110 copper (1" wide) bars 24" in length.
I am going to assume for this issue that the aluminum will do as well as the copper.
The bars screw to the PCB and the heat sinks screw to the bar.
I assume there will be a problem with a copper bar mating with the aluminum heat sink. I could use a thermal tape TIM like 3M's 5 mil Thermally Conductive Adhesive Transfer Tapes. Or screw the parts together with a thermal compound.
I could use an aluminum bar screwed to the PCB and there will be no problem with the mating of the bar and heat sink.
The questions:
Will there be a galvanic corrosion issue with the ENIG and copper or aluminum?
Will there be an issue with the copper bar and aluminum heat sink?
Best Answer
Aluminum is is a class of it's own and does not like copper. There is some risk of putting aluminum directly on copper, and it is smart to start questioning this.
Corrosion occurs when there is a ionic solution (most likely moisture) with two metals that are not in the same galvanic series. The galvanic series that your materials might be most exposed to is something like tap water, but it can be different depending on the environment. The idea is to match the potentials closely, the further apart the metals are, the greater risk for corrosion.
Source: The effects of electroless nickel coatings on the galvanic corrosion of other metals
With Aluminum on the top at around -1100mV, nickel in the middle at ~-300mV and copper at the bottom +80mV. So sandwiching nickel between the two lowers the potential between the two and risk for corrosion.
If the aluminum comes into contact with copper there won't be corrosion necessarily, there needs to be water or some other solution to kick the ions around, so try and keep moisture out as a first rule, but there is always some moisture in the air.
The second thing would be to plate the aluminum with nickel (I've heard chrome isn't as good for heatsinks). An enig coating on the PCB side might be ok also, as long as it doesn't get too scratched. Or even better, an ENIG coating on both.
This article shows what can happen on a micro scale with an ENIG coating on aluminum, it needs to be continuous or there can be problems.
Yes copper on aluminum is bad, with a coating of ENIG there will be less corrosion.