This might be a bit of an unanswered question or maybe something obvious, but I do want to ask it anyways.
I've developed a method that when I repaste a die/heatsink for a CPU or GPU, I use isopropyle alchohal and fine surface make up cleaning pads to completely clean and polish the contact surface. Afterwards I set the said subject down with the contact surface upside down(propped up so the surface can't contact anything) to ensure no dust can fall on it. Essentially cleaning the contact surface to be as sterile as possible right up until the moment that I finally apply TIM, and attach the heatsink for good.
Is this excessive, or is there a method to this madness? would this have any affect on the cooling performance or is it just self serving placebo? I do this with both my graphics cards and processors, using my own choice of TIM.
The idea is that using pure water and isoproypel alchohal is that purefied water when used with alchohal has a very hard time clinging and condensing to anything as it evaporates and alcohal obviously evaporates very quickly and provides a very sterilized(dirt and grime, not just microbes) and cleaned surface. That and being very rigorous and maticoulous removes any microparticle contamination, allowing TIM to flow into the micro cracts and surface imperfections that would otherwise be filled up or blocked with dust and insulating particles.
Obviously some large contaminents could compromise the surface purity, but how much affect could micro particles have?
Could actually polishing the contact surface as heavily as possible(or at least till the point where I feel like I don't need to) create an adverse affect?
Is being super maticulous about it a good idea if you really into getting that perfect contact, or is it just fluff/placebo that doesn't actually do anything?
Well I suppose if thermal compound has fallen over the TIM then the make up pads will definitely make it easier to clean up in the corners of the PCB and TIM. Other than that simply using 95% to 99% Isopropyl alcohol along with some coffee filters (since they will leave no fibers behind) will suffice for removing the current thermal compound off of the TIM and contact plate of the heat sink.
EDIT : Have you have heard of the process of lapping, that of which many people are convinced it will greatly reduce temperatures. Basically its starting from a low grate of sand paper and keep grinding it down in increments until you end up with a mirror finish. It looks quite nice but does not always produce amazing results. You can also compromise the effectiveness of the copper heat pipes since you may come to a point where you grind down the heat pipe and create small hoes allowing the gas to escape.
TIM is another name for Thermal Compound, TIM stands for Thermal Interface Material. AKA Thermal Grease. TIM is any material that is utilized to enhance or conduct heat from a Die or heatsource to the cooling apparatus/device. Example, Arctic Silver, or MX-4, ect.
What you meant I assume is the Heat spreader.
And this has little to do with cleaning up previously used TIM.
This is a question of surface flawlessness/purity, and the performance affect it would have on a set up.
Basicially the post is asking: Would a polished flawless surface produce a better heat conduction, over a casually cleaned surface? Or is the effort wasted as it does nothing that can be actually measured.
Ah yes, sorry about that. The heat spreader, not TIM. I can only assume there will be a point of diminishing return where even if there is still some impurities on the surface there should be a negligible difference.
looking at this from en engineering standpoint there are a few things to consider:
there are various levels that effect your thermals:
there is the micro level / a meso level / and a macro level - takling about computers i would say you could view at it like this: - macro level is the room your pc is in - meso level is the inside of the pc / the hardware component as a whole - micro level the contact surface
changes on the micro level have an influence, i will not deny that, especially micro fractures and such can have huge impacts on the structural integrity sometimes. influences/improvement go like this macro>meso>micro lowering your room temperature for example will have bigger influences overall
theoretically and i mean THEORETICALLY you don't want your surface to be completely polished and 100% flat but rather have imperfections(in which the thermal grease can crawl into) leading to a bigger surface area and thus have better thermal conductivity - but i wound not go over it with sandpaper ;)
unless you can control and maintain thermals or general conditions in the macro/meso level like say in a cleanroom laboratory - you would not notice any differences from micro particles or contaminations or surface imperfections unless the contamination is huge. it is good to be meticulous and exact when cleaning - pc components are expensive after all. i clean mine with those medical wipes they use to sterilize the skin before sticking a needle in(they are pretty neat an are already soaked with alkohol)
final words: good to be exact and careful and if you are have confidence in your skills and don't worry about whats going on at a microscopical level - as simple as possible as exact as necessary
If your main concern is temperature choose more efficient way to cool the cpu. If you are using air cooling you options ale very limited. Best one is lapping as Theonewhoisdrunk said but removes id marks on cpu which mean you loose warranty,is time consuming, will reduce or at least even up temperature of cores. If you decide to polish cpu do the same with cooler. Better tim like gelid extreme can shave off 5C or even more. Otherwise better solution is to reduce ambient temperature by getting an air conditioner .