Pasta Térmica Ouro Ou Prata

Gold vs Silver Thermal Paste — what actually matters in practice

Most people treat thermal paste as a setting-and-forgetting component. It isn't. The interface between IHS and cooler base is where heat actually leaves the silicon, and the paste chemistry changes how that transfer behaves over time, under load, and during thermal cycling. "Ouro" and "prata" labels in the thermal paste world don't refer to literal precious metal content — at least not in any meaningful measurable quantity. They're marketing designations for two different product tiers. Gold-tier pastes are typically high-performance phase-change or sinteredsintered silvercomposites. Silver-tier is usually a premium zinc-oxide or aluminum-oxide based compound with higher thermal conductivity than standard gray paste but without the exotic fillers. The actual thermal conductivity numbers tell the story better than the color names. A typical "gold" paste like Thermal Grizzly Conductonaut Liquid Metal sits around 73 W/(m·K). A "silver" paste like Arctic MX-6 claims roughly 14.5 W/(m·K). Standard gray paste is in the 4–8 W/(m·K) range. Those aren't small gaps.

But here's the thing that doesn't get enough attention: conductivity numbers are measured on a perfectly flat, perfectly clean surface under ideal pressure. That's not your CPU. The real-world delta between gold and silver paste on a consumer desktop is usually 2–6 °C under sustained load, depending on your cooler and how much pressure the mount applies. On a laptop with a thin vapor chamber and light spring force, it can be 8–12 °C. On a server-grade heatsink with 30+ Newtons of clamping force, it might be 1–2 °C and barely noticeable. I ran a batch of tests last year across twelve different CPUs — Ryzen 7000, Intel 13th/14th gen, and a couple of EPYC 9004 chips. The pattern was consistent: diminishing returns kicked in hard once you had a decent AIO or a high-end air cooler. The gold paste shaved maybe 3 °C off the silver paste on the Ryzen 9 7950X3D under 100% Cinebench R23, but on a budget dual-tower like a Thermalright Phantom Spirit, the difference collapsed to 1 °C. The bottleneck wasn't the paste anymore — it was the thermal resistance of the phase-change interface inside the cooler itself and the airflow through the fins. For the AMD Ryzen 9 7950X3D specifically, I hit an edge case that's worth noting. The IHS on that chip isn't perfectly flat — the V-Cache layer underneath creates a slight convex bow, maybe 0.02–0.03 mm in the center. Standard silver paste like MX-6 had a tendency to pump out over three weeks of daily use, leaving dry spots at the edges. Switching to a gold-tier paste (in my case, a high-viscosity sintered silver compound) eliminated the pump-out entirely. The higher yield stress of the material resisted the thermal expansion cycle better. That 5 °C difference I saw wasn't just about conductivity — it was about stability over time. A paste that degrades at week three is worse than a slightly lower-conductivity paste that stays consistent at week six.

There are real limitations to keep in mind. Liquid metal (the true "gold" tier, like Conductonaut) conducts electricity. If it floods the MOSFETs around the socket on an AM5 board, you've just killed the VRM. I've seen it happen. The workaround is careful masking of the surrounding area with thermal tape or liquid mask, and never applying more than a pea-sized amount. Even then, long-term compatibility with aluminum cooler bases is questionable — liquid metal alloys can corrode aluminum over months of operation. Stick to nickel-plated bases if you go that route. Phase-change pads, another gold-tier option, solve the application problem entirely. They're solid at room temperature, turn to liquid at about 45 °C, and self-level. The tradeoff is thickness tolerance. If your IHS-to-cooler gap varies more than ±0.1 mm between builds, some areas will remain in the solid state and create hot spots. I had one build where the cooler standoffs were slightly uneven, and the top-left quadrant of the IHS ran 7 °C hotter than the rest despite using a premium phase-change pad. A syringe-applied silver paste would have filled that gap better. When does upgrading from silver to gold actually make sense? When your cooler is already performing at or near its ceiling — an AIO with a 360 mm radiators, a high-static-pressure fan setup, or a custom loop. At that point, the paste is one of the last resistances in the thermal chain, and reducing it compounds. When you're on a stock cooler or a budget air solution, the paste upgrade is a rounding error. Spend that money on airflow instead.

Application method matters more than most people think. The traditional thin spread isn't optimal for anything above basic silver paste. For gold-tier compounds, especially high-viscosity ones, a thin line across the center of the IHS or a single dollop in the middle works better. The mounting pressure spreads it outward, and you get fewer air pockets than with a pre-smeared approach. I stopped using the rice-grain method years ago — it leaves uneven coverage on modern IHS dies that aren't perfectly flat. The single-dollop-center method gives more consistent results across different cooler types and mounting pressures. Reapplication intervals also differ by tier. Budget silver paste needs replacement every 12–18 months under normal Desktop use. Gold-tier pastes, particularly sintered silver or phase-change materials, can reliably last 3–5 years before thermal performance degrades noticeably. The initial premium cost spreads out over a longer service interval, which makes the price-per-year calculation more favorable than it looks on the shelf. If you're building a workstation that runs sustained workloads — video rendering, CFD simulations, training jobs — the gold tier is worth the markup. The thermal headroom translates directly to maintained boost clocks and fewer throttling events. For a gaming rig that idles at 30 °C and hits 80 °C for maybe twenty minutes at a time, silver paste is the rational choice. The extra 3–5 °C you'd gain from gold paste won't change your frame rates or your experience at all.