I didn’t really want or need to write this article, because after my test of the AMeCh SGT-4 in October 2024 and the cases of damage documented later, which I addressed in the follow-up article, it seemed that everything had been said. However, it is now clear that the issue cannot simply be put to rest with technical facts or scientific analysis. In almost all of the cases that I have seen or reported in the meantime, the period of purchase and use coincide exactly with the paste that I tested at the time. This is proven by numerous entries and image documents in the relevant forums such as Quasarzone, which I have already referred to in the follow-up. Such public experience reports are common in the hardware community and are part of a transparent exchange between users and testers.

Important preliminary remark on social networks

However, what happened next goes beyond any objective discussion. Due to repeated derogatory statements by the provider, who publicly disparaged me as an incompetent “German YouTuber” and “influencer” (which is in stark contrast to the email correspondence below), a downright smear campaign arose in South Korean social networks, which a Korean reader then drew my attention to in an email. In the course of this campaign, hundreds of users visited my YouTube channel to leave racist and nationalist insults (also against other community members), including openly anti-Semitic and National Socialist comparisons. I have never actively written anything in these Korean forums, but right at the beginning, users who had been harmed by the paste had spoken up and then referenced my test as proof because it exactly matched their own experiences. Incidentally, I would never have written the second article or made the video if I hadn’t been so badly insulted. Because until that day, I had complied with AMeCh’s personal request from 2024 and didn’t comment any further.

Normally, I am above personal criticism in such networks, it is part of my work and I now find it very difficult to be triggered. In the past, I have had to learn a lot myself, such as the fake fan tests of a former freelancer and now legally convicted forger, under whose results I had put my name because I believed the manufacturer’s words from the hospital and didn’t check it again. All of this was dealt with very transparently with the involvement of third parties. But when it comes to Nazi and Hitler insults, a limit has been reached that is neither morally nor legally acceptable, especially as the slowly forming nationalist mob also targeted the Korean customers concerned, i.e. even their own countrymen. These events show an alarming degree of brutalization and lack of responsibility, especially when they are indirectly fueled or at least tacitly tolerated by a company. To date, I have deliberately not entered into any discussions, because there is no point anyway. However, I have now reported the most serious failures. It remains to be seen what YouTube and Google will make of this. One example of over 600 (!) posts that I deleted and banned the users concerned from the channel. The following was still relatively mild, I don’t even want to quote the others here:

(I’ve never seen such human scum as you. I hope you suck Hitler’s ass and then shoot yourself in the head with your own hand, you damn dog.)

I very much hope that the readers of Quasarzone will read today’s article carefully, because it concerns a South Korean company that has primarily deceived its own countrymen. Instead of facing up to this responsibility, they are still looking for a convenient scapegoat and, one year after the first tests, have publicly declared me to be the main culprit. Not because of the facts, but out of a need to cover up their own failures. What happened on social networks afterwards crossed every moral boundary. When nationalism is misused as a shield and people start attacking critics with racist or even National Socialist attributions, then this is no longer defense, but deliberate incitement. And no, I didn’t drive the company to ruin, it had already done that itself. With an announcement.

Why I am writing this article today

The paste in question was proven to contain ingredients that were mixed in with a certain amount of euphoria, but definitely did not belong there and ultimately led to the damage I predicted back in 2024. This criticism has therefore been valid from the outset. This in turn was even confirmed to me personally and in writing by the provider in an email after my test article, even though he unfortunately repeatedly denied this fact to the outside world. The reason for this (perhaps to his credit) is that he didn’t really understand the email from the German manufacturer BYK about the additive used and its odor. My only, but for me annoying, mistake in the whole thing was that at the moment I became aware of these true circumstances, I did not subsequently change the contents of the test from the suspected acetoxy silicone to the amino groups confirmed by BYK, so as not to further harm AMeCh, as requested. Similar odor and completely identical effect including consequences. This is a negligence that I blame myself for, but which does not change the actual problem of the destruction of copper. Of course, I want to set the record straight in advance, because that is the right thing to do.

Under normal circumstances, I would not publish any content from private emails or conversations. In this case, however, I unfortunately feel compelled to list individual passages in a condensed form and summarized to the essentials as a report in order to classify the character and credibility of this company and to show how far the will to take responsibility apparently (does not) extend here. For data protection reasons, I will not publish the complete mail correspondence here in the original, only some important passages later as excerpts and original quotations. However, all messages, including all attachments, are fully archived and will also be used legally in case of doubt. Even if AMeCh’s request was for me not to add any further criticism in order to ensure its survival, I should have done so at the time. From today’s perspective, the fact that I went so far out of my way to accommodate the company was misunderstood compassion.

Chronological summary of the process including the mail traffic (ascending date)

I had bought around 20 pastes anonymously on Amazon to test whether there were any real gems and price/performance winners away from the established providers. The AAirhut products showed that such tests are definitely worthwhile. What I always do, however, is check the necessary documentation for the marketing of thermal pastes in Germany. And that is precisely why I have personally written to all the manufacturers of these purchased pastes and asked for the safety data sheets (SDS) as well as RoHS and REACH certificates. One of these emails also went to AMeCh, which led to a lengthy conversation, as I am fair and also exchange measurement results and ask questions in advance. As the paste posed problems right from the start and the company presented itself as a start-up, I did even more than I usually do just to provide a little support here. And now one thing at a time:

8. October 2024 at 02:04 am In response to my request, AMeCh first sends the safety data sheet and REACH documents for SGT-4 by email. There are no further statements on the content, the documents are sent formally without comment.

8. October 2024 at 12:56 pm I inform AMeCh about first laboratory measurements. Findings include a strongly perceptible sour-pungent odor when heating up, the assumption of possible acetoxy-crosslinking silicones as a possible reason, a doubling of the bondline thickness between 20 °C and 80 °C at a defined force and a rapid drying and hardening process after less than ten cycles. The result: the paste cannot be used for a long time. These statements are based on our own measurements.

8. October 2024 at 10:31 am AMeCh responds with a reference to available TGA data (Thermogravimetric Analysis). They speak of almost zero weight loss up to about 120 °C, above that slightly by 0.2 %, from which they deduce that the paste does not dry out in the CPU window. The company interprets the linear increase in thermal resistance with the layer thickness as an indication of stable conductivity, sees no evidence of cracks or bubbles and asks for the detailed measurement setup and the data from igor’sLAB to be sent to them. (Note: It is better not to measure something like this openly in a gas jet, but at constant pressure and increasing temperature in order to record practical interactions and changes in the internal structure. The strong increase in bondline thickness under thermal cycling is consistent with several effects. These include a reorganization of the polydimethylsiloxane on the filler and metal surface, partial crosslinking or association of functional groups, an increase in viscosity and possible release of volatile components. A TGA with low mass loss does not exclude these effects, since interfacial reactions and viscosity changes can proceed without significant mass release and TGA conditions are often anhydrous, which underestimates hydrolytic contributions)

8. October 2024 at 18:45 I explain my own measurement methodology, refer to industrial standards and previous cases, again criticize the strong odor, the dramatic BLT increase and the siloxane used with presumably reactive additives and ask for reliable company information, as the address, website and register entries were unclear. I also provide exemplary links to the methodology.

8. October 2024 at 13:38 Now AMeCh replies that it develops and manufactures itself, but does not share internal data comprehensively for competitive reasons, knows the TIMA5 tolerances, therefore considers the igor’sLAB data supplied to be absolutely clean and comprehensible, but contradicts the impression of rapid drying out, referring instead to high viscosity and the fact that reuse of removed paste is generally problematic. Pohang, South Korea, is now generally mentioned as the location. Laser flash results are available, but are not published for consumer protection reasons. There is no transparent information on how and with what the paste is produced.

8. October 2024 at 15:24 I object to the attribution “influencer”, describe qualification, equipment and precise test procedure, ask again for comprehensible company and address details and transparently present my concrete test sequence for BLT and cycle tests, including renewed questions about the cause of the strong odor.

8. October 2024 at 23:53 Now follows a short post from me. A BLT increase due to outgassing to over 100 % is observed, a striking peak around 90 to 95 °C as a possible indication that corresponding adhesives release corrosive substances during crosslinking. The different viscosity compared to reference pastes is noted.

9. October 2024 at 08:52 AMeCh justifies the slow communication with a small two-person operation, describes manual production activities, lack of resources for website maintenance and announces that it will seek external laboratories for further testing. The submitted TGR material (TGA post-processing) is a pure and non-binding courtesy of third parties, more detailed conditions will be submitted later. AMeCh is impressed by the laboratory equipment at igor’sLAB.

9. October 2024 at 21:49 I transparently and extensively detail the ASTM approach, distinguish interface and bulk parts of thermal resistance, justify the multiple measurements and criticize laser and hot-wire methods as completely unsuitable for this issue. I also point out that sales via Amazon in Germany are not legally compliant without documents and that the siloxane used is problematic. The recommendation is to change the silicone oil. The legal information refers to the German environment applicable at the time, specific contacts with authorities are not mentioned in the email.

10. October 2024 at 15:34 AMeCh writes that they did not know that the sale in Germany was illegal without specific documentation and asks about possible steps to be taken against Amazon in addition to stopping sales and disposing of remaining stock.

10. October 2024 at 11:00 pm I write to AMeCh informally that it is practicable to provide download links for the safety data sheet, REACH and RoHS in the product description, as Amazon rarely checks formally, but again point out that the content is problematic and that changing the silicone oil is the key measure. Another example of a questionable paste offer is mentioned. These references are purely pragmatic assessments, not legal advice.

11. October 2024 at 16:30 This is followed by a very personal AMeCh email with thanks and emotional wording, but no new technical content.

12. October 2024 at 03:46 am I’ll answer briefly, apart from the siloxane, the paste looks fine. This statement differentiates the problem area, but explicitly refers further to the siloxane.

16. October 2024 at 06:00 The test of the paste goes online. I am committed to honesty towards my readers, so the article does not contain colorful benchmark bars, but all laboratory results and available information and assessments, as well as a warning of possible consequential damage (which unfortunately is now visible).

7. November 2024 at 17:12 AMeCh sends a longer clarification after my published test article. They explain the vinegar-like smell with an added PDMS additive from BYK, which causes an odor when heated. BYK had confirmed that the product contained amino groups but was safe, the European MSDS listed “no hazardous ingredients”, some fields were “not provided” or “not applicable”. AMeCh deduces that there is a misunderstanding about acetoxy crosslinking silicones (but does not understand that amino groups can be just as reactive). In addition, a very low BLT below 10 micrometers under typical pressure conditions is claimed (but this is untenable based on the grain sizes used, which I was able to disprove in the calibrated measurement setup). Subsequently, AMeCh sends another, very emotionally worded email. It asks for indulgence, emphasizes the fragile company situation, apologizes for any harsh wording and asks for “a spark of compassion”. This message does not contain any new technical information. (But I had decided not to follow up here and comply with the request)

Classification of the core statements

First of all, I would like to highlight the section that AMeCh itself has written and thus confirmed to me in full. Please pay particular attention to the underlined section. Here is a literal quote, as it is too important to withhold from readers:

Note: In the EU, the safety data sheet SDS according to CLP is regularly used for this, this classification concerns the substance law in handling, not the inertness on electronic surfaces

What do such additives mean and what does “safe” mean?

At this point I must of course make it clear that I was also fooled by the smell and the consequences of the shelf life tests, but I really didn’t expect basic amino groups. Amino groups in siloxanes are clearly more difficult to detect in practice, especially if they are present in very low concentrations or chemically bound. Also, in IR spectroscopy, only acetoxy siloxanes show prominent carbonyl bands (around 1700-1740 cm-¹), whereas amino groups often show overlapping signatures (N-H stretching at around 3300-3500 cm-¹), which are harder to clearly identify in complex polymeric matrices.

From a purely chemical point of view, however, the decisive point is not whether acetoxy or amino functionalities are present, but that both types of additive have reactive groups that act chemically at interfaces with metals and can trigger undesirable subsequent reactions there. It is also known that low molecular weight amines are sometimes interpreted as sour-smelling due to their high volatility and basic irritation, but can be clearly distinguished analytically from acetic acid. But not always with the nose.

Acids and bases can attack copper and nickel surfaces, but the underlying processes are different. Acids destabilize oxide layers by protonation and the formation of soluble copper salts. Basic amines increase the local pH value, complex copper ions and promote the dissolution of oxides via hydrolysis and complex formation. The result in both cases are traces of corrosion, but the cause and kinetics are different, which is relevant for the evaluation of the odor sources and the time dependence of the damage.

Exemplary amino group of a property-altering additive

Polydimethylsiloxanes are basically very stable and inert oils. However, as soon as they are functionalized, whether by amino, epoxy or acetoxy groups, materials with adhesion-promoting (adhesive) properties are created. Under thermal stress and in the presence of moisture or oxide-containing metal surfaces, these inevitably lead to chemical reactions, as I was able to observe in 2024. The paste initially becomes stickier, loses its flowability due to partial cross-linking and microscopic reaction zones can form on metallic contact surfaces. Or more precisely: they react at metal oxide boundaries via coordinative bonds or hydrolysis products and thus lead to adhesive surface changes. These zones show typical corrosion characteristics such as matt, slightly discolored surfaces, localized oxidation spots or efflorescence. In some documented cases, greenish, bluish or brownish deposits appeared on copper or nickel surfaces, indicating reactions between released amines or reaction products and the metal oxides. The greenish to bluish deposits can be consistently explained by copper(II) complexes and basic copper salts, which are formed in the presence of amines, moisture and atmospheric oxygen. Dark, dull zones can occur on nickel due to dissolution and reoxidation of the surface layer.

These observed processes are not classic cross-linking in the sense of condensation (as with RTV silicones), but rather hydrolysis and adhesion reactions on metal oxides. Reactive silicone additives change their structure through thermal cycles and increasingly adhere to the substrates, while at the same time either alkaline (through amines) or acidic (through acetoxy groups) reaction products are formed. These attack the metallic contact surfaces, dissolve ions and promote the formation of thin SiOx-rich films of oxidized siloxanes at the interface as well as traces of corrosion. This results in localized insulation spots, increased thermal resistance and permanent discoloration and even visible material degradation. The “safe” mentioned by BYK generally refers to the classification according to CLP or GHS and does not mean inert to metals or electronic assemblies. An additive can be harmless in regulatory terms and still have a reactive effect on metal surfaces under practical conditions. The decisive factor is the system compatibility of matrix, additive, filler, moisture and temperature profile.

Was AMeCH lying here?

Since I can only speak for the paste from 2024, I will now politely say yes and no. Because I at least give the operators credit for simply not understanding BYK’s statement about the amino groups. You mix something together that you don’t know what reactions can occur when it is heated slowly, exposed to humidity and pressure. This in turn is dangerous and leads to damage that nobody intended. Yes, I always keep an airtight, sealed and cooled reserve sample in the archive for emergencies and as evidence. But I admit that after an already very good four-figure investment of 3 long laboratory days for the ASTM and durability measurements, I simply don’t feel like investing any more money in an extra third-party report to prove the amino groups, which the supplier of these additives has already confirmed. I ask for your understanding, because I don’t want to throw good money after my voluntary services for this company when the facts are so clear. However, what I still find offensive about the whole discussion is the misinterpretation of the REACH and RoHS documents, and we really need to write something about that now.

REACH and RoHS

The RoHS and REACH documents submitted are practically worthless in terms of content and form for several reasons, as they do not allow any statement to be made about the actual chemical composition or hazard of the specific product. They merely fulfill a superficial formality without allowing any conclusions to be drawn about the additives used, their reactivity or their origin. I would like to explain this, because not everyone will really be familiar with it.

The RoHS declaration is only a general confirmation of conformity that the product does not contain any of the hazardous substances listed in EU Directive 2011/65/EU above the limit values. These substances are heavy metals, certain flame retardants and PBB/PBDE derivatives. Silicone oils, siloxanes, amines or organic additives are not covered by the RoHS Directive. A product can therefore easily be RoHS compliant, even if it contains chemically highly reactive or corrosive compounds, as long as these are not on the list of restricted substances. In addition, the RoHS document lacks the referenced test as an annex, without which the document is worthless anyway.

The REACH declaration is equally meaningless. It merely confirms that the manufacturer either does not import any substances subject to registration in quantities of more than one ton per year or that the raw materials used have already been registered by upstream suppliers. As a rule, the REACH declaration does not include a complete formulation and does not contain any information on functional additives, their concentration or chemical structure. In addition, manufacturers may exclude “proprietary” substances as trade secrets as long as they do not contain any substances classified as substances of very high concern (SVHC).

A RoHS certificate therefore only proves the absence of prohibited substances, but not the chemical stability or harmlessness in practical use. Neither the RoHS nor the REACH certificate has any chemical-analytical significance. They are used exclusively for regulatory assurance in the EU import process, not for quality or content assessment. They are therefore completely unsuitable for a technical analysis, as would be required for the AMeCh paste, and have no real evidential value in this context.

In practice, this means that REACH or RoHS conformity offers no guarantee whatsoever that a product is chemically inert or compatible with the material. This is particularly crucial in the case of AMeCh paste, as the demonstrable reaction with copper shows that reactive groups must be present that are not covered by either of these two declarations. In addition, the data given in the available documents are not plausibly dated. They are from a period prior to the alleged revision of the formulation, which means that they may refer to an earlier or completely different version of the product. Such a discrepancy is a clear indication that the documents were either created without reference to the actual production batch or were subsequently reused.

Summary and conclusion

I would like to emphasize at this point that I have spent a considerable amount of time trying to help this company free of charge and already warned urgently in 2024 that the composition of the paste, including its additives, could lead to long-term damage in its form at that time. Even then, however, I was told that the formulation could no longer be changed. What I don’t understand now, however, is the following: If the paste was actually produced in-house, it would have been easy to omit or replace the amine-containing components. Why wasn’t this done? After all, these substances were the obvious cause of the problems observed later.

This inevitably raises the question for me as to whether the material was ever produced entirely in-house or whether a larger batch of ready-made paste had possibly already been purchased from a third party (according to their own recipe), then filled and packaged themselves and subsequently marketed under their own name. Wherever, Korea, China, the rest of the world… I don’t want to accuse anyone of anything, but in view of the existing contradictions, I think the question of whether the people involved ever had their own automated mass production facilities is entirely justified. To this day, there are no transparent pictures of their own production and bottling, no information about supply chains or companies involved, although I have asked them several times. It would have been very easy and transparent to prove all of this, as I do.

I even contacted the supplier before my follow-up and video in October 2025 to be fair and asked for a correction on social media, but I still haven’t received a reply. If that had happened, this mob would probably never have happened. In addition, the company has been virtually unreachable for almost 2 weeks, even for its customers who have registered legitimate complaints. Perhaps the community on Quasarzone should first concentrate on investigating the actual problem of the aggrieved customers. I do not work on behalf of a paste manufacturer, nor do I own shares in such a company. I work independently and on request for various institutes, research facilities and paste manufacturers in the field of quality control and measurements. No more, but also no less. And AMeCh had been warned about the damage that has now come to light for over a year.