Drama, second round: AMECH (AIMAC) SGT-4 or when heat-conducting paste becomes an odor nuisance and source of corrosion

Augrund neuer Erkenntnisse und Umstände möchte ich das offizielle Statement gern voranstellen, welches auch die Hintergründe beleuchtet und Einiges klar stellt:

The AMeCh SGT-4 Case – The Story Behind the Story and Corrosive Amino Groups | Statement

After my original test of SGT-4 here on my website, almost exactly one year ago to the day, there are now more and more user reports from South Korea that confirm my laboratory results 100%. The paste not only exhibits chemically atypical behavior, but apparently also causes visible corrosion damage and harmful fumes. However, there has been no objective response from the manufacturer so far, but this has been compensated for with unobjective and personal attacks, which are definitely worth writing about.

In my first test of the AIMAC SGT-4, I had already pointed out several anomalies that were not consistent with a conventional PDMS-based thermal compound. In addition to a clearly perceptible vinegar smell, the ASTM comparison test showed significant deviations in thermal behavior. The material expansion when heated was not linear and the paste exhibited an unusually high adhesive force after loading. In the same test, the thermal conductivity was measured to be well below the manufacturer’s specification. This already made it clear that chemically it could not be a classic, inert polysiloxane paste. As a reminder, here is the link to the original at the time:

AMeCh STG-4 in test – Korean thermal paste with free vinegar as price/performance winner?

Start-ups and micro-suppliers in the technology sector often suffer from a remarkable form of resistance to learning. They confuse early attention with competence and believe they can outwit market laws through brash self-promotion. History repeats itself with almost tragicomic consistency: those who ignore warnings and prefer to defiantly point to supposed “misunderstandings” are caught up in reality at the latest when their own customers begin to report publicly about odors, corrosion or simply broken hardware. I am a polite person and will not quote from the correspondence with the “company” at the time, but I will let the customers have their say.

What I documented back then in the laboratory is now a practical certainty, backed up by numerous user reports: the problems are real, reproducible and home-made. They could have been avoided if scientific criticism had not been seen as a personal attack. Instead, experiments were carried out on end customers as if they were part of an involuntary field test. But anyone who turns the paying customer into an experimental animal is ultimately playing with their own survival – a lesson that the two self-proclaimed heat conductors have also had to learn painfully in the meantime. I don’t offer such companies any support, because they certainly needed it.

Possible adhesive admixture and its consequences

The analyses at the time and also later revealed evidence of a mixture of polydimethylsiloxane (PDMS) and an amino-based or acetoxy-crosslinking RTV silicone, which was apparently added as an adhesive to hold the paste together even at very low BLT. Such a system utilizes prepolymeric silanes such as methyltriacetoxysilane, which react under moisture to form polysiloxanes, releasing acetic acid. This acetic acid explains both the odor and the observed copper oxidation. At the same time, the partial cross-linking causes the paste to tighten after a certain period of operation, becoming sticky and almost gluing surfaces together. In my test, an immense separating force was required to release the two test surfaces after the heating cycle. A pure thermal paste would never exhibit such behavior. Exactly these symptoms can also be found in the user reports documented below. And now it becomes bitter for those involved…

Damage reports from the community

Several reports on the AIMAC SGT-4 appeared in the Korean forum Quasarzone in October 2025. Affected users described significant material damage and health complaints:

• 귀요밍1221ㅎ reported severe discoloration of the copper surface, erased CPU lettering, a pungent smell of vinegar and headaches lasting several days after prolonged gaming in a closed room. He had used the paste several times over the course of a year.
• 라이젠구팔삼디co nfirmed the copper corrosion on two systems (Ryzen 9 9800X3D and 9700X) and announced that he would switch to Arctic MX-6 in the future.
• 멋있는나 documented the same effect on an RC1900N cooler and found that pure copper coolers are significantly more affected than nickel-plated variants.

The phenomena described correspond precisely with the damage pattern caused by acetic acid release from acetoxy-crosslinking silicones. I have extracted the pictures from the public forum on Quasarzone for documentation purposes:

001

002

003

004

005

006

007

008

010

011

012

013

A technically experienced user in the forum explicitly referred to my original report and drew the correct conclusion: If the paste actually contains acetoxy-crosslinking silicone, acetate compounds would be formed that attack copper through localized carboxylic acid corrosion. He correctly calls this effect “ant-nest corrosion”, a classic example of selective pitting corrosion, which is also described in specialist metallurgical sources. This expert feedback independently confirms that the observed damage is not a coincidence, but can be clearly explained chemically.

Reactions of the manufacturer and forum users

The supplier of the SGT-4 responded to the allegations with a series of derogatory comments. In public responses, he described my criticism as unfounded, claimed that the product was safe and RoHS/REACH-compliant, and even attacked critics in the forum personally. Instead of providing a technical explanation, however, the manufacturer referred to general environmental certificates, which have no relevance whatsoever to the actual problem, namely chemical reactivity with metals. Particularly striking was the repeated reference to alleged misinformation from “a person from Germany”, which obviously meant my analysis. Of course, the content of the counter-argument remained inconsistent: a pure PDMS paste can neither produce an acetic acid smell nor cause copper discoloration. The evasion to personal accusations instead of a laboratory test therefore appears unprofessional, which the forum members also noted several times.

The reactions of the community were therefore more than clear. Several users criticized the condescending tone and called on the manufacturer to disclose the chemical composition or recall the affected batches. Instead, only blanket statements of exoneration continued to be published. Some users publicly recommended switching to established brand pastes such as Thermal Grizzly Kryonaut, Arctic MX-6 or Noctua NT-H2, whose chemical stability has been proven. Others pointed out that AIMAC has neither a verifiable company address nor a professional email domain, a further indication of questionable origins and a lack of quality assurance. Trust in the product is now considered to have been completely lost within the community.

Summary

In the end, a mixed feeling remains. It is not a triumph, but rather a quiet disillusionment when, after months or even years, what you suspected and documented early on is confirmed. In view of the many aggrieved buyers, I find it anything but satisfying to be proved right in the end. Nevertheless, I see it as my job to warn users and buyers in good time about questionable experiments and inadequately tested products before minor anomalies result in real damage. Because defective coolers, discolored CPUs or corroded surfaces are not only annoying, but in many cases also expensive.

My special thanks at this point go to the Korean reader who drew my attention to the discussion at Quasarzone. Without such information, many of these cases would go unnoticed. I’m always open to any form of feedback or mail when it comes to technical irregularities, conspicuous materials or possible product defects. With a bit of luck, proper methodology and a little patience, many things can be cleared up, as this example shows.