Why are the heatproof composites used for semiconductor packages never used for coil formers, connectors and switches?

It is very hard to destroy the material normally used for "plastic" IC and transistor packages with IR or hot air soldering tools, unless you apply a measure of heat that would reliably destroy the semiconductor in it first. From what datasheets say, the material is more akin to polymer concrete than a common plastic.

It is very easy to destroy the thermoplastic materials used for tactile switches, smd coil formers, and connectors the same way.

If the first material is cheap (which one would assume, given there are very cheap semiconductors, whereas connectors often are not that cheap), why is it not used commonly to make parts in the second group?

• Good question +1 Oct 29, 2020 at 16:44
• The common IC housing material is quite brittle. The IC package is massive enough and not subject to mechanical stress. Oct 29, 2020 at 16:47
• I'm not so sure the Epoxy Molding Compounds and Liquid Crystal Polymers and what have you are as cheap to make as the plastics used for connectors etc. They are higher-specced when it comes to thermal expansion, thermal stability, resistance to melting, moisture absorption, etc., so I could well imagine other plastics are cheaper to produce and machine. Oct 29, 2020 at 17:03
• Thermoplastics are easy to mold into complex shapes; their low melting point allows the material to flow into small features in a mold. High melting point materials require higher temperatures and high pressure for molding and make the molding process more complex and time-consuming to keep the mold temperature constant. Mold wear is a problem, requiring new tooling periodically and resulting in flashing as the mold wears. Otherwise, machining is an option; both approaches are more expensive. Oct 29, 2020 at 18:54
• Mold wear sounds like a valid point to consider indeed when it comes to mineral filled plastics (which AFAIK the IC materials are, that's why I likened them to polymer concrete).... Oct 30, 2020 at 7:46

2 Answers

The material that most ICs are encapsulated in (the ubiquitous hard black plastic) is epoxy, which is a thermosetting polymer. This is opposed to most materials referred to as plastics, which are thermoplastic polymers.

Thermoplastic polymers are made up of many separate polymer chains of various lengths (molecular weight) that don't actually have any bonds to other nearby chains. Instead, the material is an amorphous, glassy tangle of chains jumbled together, or a more orderly crystalized structure of chains arranged in a repeating pattern, depending on the specific material. As a result, with enough heat, these chains can freely slide past each other, and the more heat, the easier it becomes. Eventually, the heat makes the plastic behave like a liquid more than a solid, which allows for things like injection molding or 3D printing.

There is a temperature where this transition from solid to deformable semi-solid to viscous liquid occurs called the glass transition temperature, or $$\ T_{g} \$$. Here in lies their weakness. All of the common and cost-effective thermoplastics also have relatively low glass transition temperatures which are well below the melting point of even leaded solder. Add in that a soldering iron's tip is usually a fair bit hotter than the melting point of solder (for good heat flow to occur), these materials don't stand a chance against a soldering iron. They're really only designed to withstand one to a few brief reflow heating cycles in a reflow oven, and sometimes not even that for through hole connectors. There are some high performance plastics like polyetheretherketone (PEEK) or thermoplastic polyimide (Kapton) that have glass transition temperatures far in excess of soldering temperatures, but these materials are also extremely expensive - PEEK is on the order of 20 times the cost of more common plastics. Yet other thermplastics like PTFE (Teflon) can certainly withstand the temperatures but have other properties (usually mechanical) that make them unsuitable for the use in connectors etc. You can buy teflon insulated wire (at some added cost), and a soldering iron won't hurt the insulation on those wires at all. They're also extremely frustrating to strip because they're so slippery.

Thermosetting polymers, on the other hand, are still polymer chains like thermoplastic polymers, but these chains have undergone a chemical reaction that has caused these chains to cross-link, or form ionic or covalent bonds with other chains. In other words, the chains are linked and locked in place. Heat doesn't cause them to soften. Sure, if hot, their chains would indeed slide past each other easily like with thermoplastic polymers, but this doesn't matter because the chains can't move. This is the same mechanism that separates soft and meltable natural or synthetic rubber from the much harder and more durable vulcanized rubber used in tires. It is the same material, but the vulcanization process induces cross-linking between the chains of rubber.

Thermosetting polymers, like any 2-part epoxy you might buy at the hardware store to the epoxy used to encapsulate power transistors to ICs alike, usually exist as liquids in their uncured (no cross-links and short chains) state. They are mixed with a hardener that forms part of the cross-links as well as initiates them (sometimes heat and/or pressure is also needed) and the material will set into a hard, rigid material. This process is generally irreversible, and for this reason, thermoset polymers don't melt at all. Turning into a liquid would require breaking the same chemical bonds that make up the epoxy itself, so any heat capable of melting it simply decomposes (destroys by ripping the chemical bonds of the molecules apart) into different chemicals instead. This also means that you need temperatures hot enough to thermally decompose the material, or break the covalent bonds forming the very molecules themselves.

This temperature is going to be a lot higher than thermoplastic glass transition temperatures, so it is much harder to damage these materials with heat, and you need much higher temperatures.

Now, the reason they don't simply make connectors or other things out of thermosetting polymers isn't any single silver bullet answer. There are many reasons and the exact ones at play depend on the product and dozens of other factors. But generally:

1. Thermosetting polymers have properties very poorly suited to certain applications. They are very hard and brittle, and shatter with little ability to bend or otherwise deform. They are also much less durable in the thicknesses that thermoplastics tend to be used in, requiring more thickness and bulk for strength. This of course also makes them totally impractical for anything that needs flexibility, like wire insulation. That said, many shielded inductors are manufactured with thermosetting polymers where the coil is fully encapsulated in the shielded grey to black block. Often, the magnetic material is mixed into the thermoset polymer. This is a popular style in SMD power inductors.

2. Thermosetting polymers are impossible to process like thermoplastics can be. You can't use thermosetting polymers in injection molding machines. To reach the final state, thermoset materials fundamentally require a chemical reaction to occur (and it has limits to how fast it can be made to occur), whereas thermoplastic polymers can be processed with heat alone - you heat them to shape them, then cool them to solidify them, and you can do this as many times as needed. This makes the manufacturing processes completely different at every possible stage and imposes all sorts of limitations that a process or manufacturing engineer could probably spend hours discussing.

3. Demand. There simply isn't any widespread market need or demand for soldering iron/hotair rework resistant connectors, wires, etc. and making them so would reduce manufacturability, increase cost, lower durability, and likely increase size. Very old electrical components actually did tend to use thermosetting polymers, but they were also big, heavy, and expensive. Any tube amplifier would be filled with parts encapsulated in Bakelite - the first manmade polymer ever, and also a thermosetting one.

This isn't to say such components don't exist - they very much do. Or at least, high temperature ones. I already mentioned PTFE wire, but there is also silicone insulated wire that can withstand soldering temperatures, and plenty of high performance thermoplastics with melting points even higher than teflon (whose heat resistance is already nothing to sneer at, considering it is used to coat frying pans for cooking). There is magnet wire that is coated with some sort of enamel that can withstand 200°C continuous and is generally impervious to damage from a soldering iron. I actually hate this kind of magnet wire because you can't just crank your iron all the way up and burn the enamel off, you have to use sand paper or something to abrade it off. But it exists, and it isn't even particularly more expensive.

4. Inductors are more often limited by the magnetic core material losing magnetic properties before wire temperature becomes the limiting factor.

5. There are definitely switches (usually very expensive ones) that won't melt - but they're simply made of metal. In this case, there isn't really a compelling advantage to use thermosetting epoxy, as most customers want switches that have high durability and likely are not too concerned with high heat resistance since, presumably, the switch is intended to be operated by a human. Humans are even less tolerant to soldering temperatures than thermoplastics, and have an even narrower operating temperature range. Mine is about 68-72° F and outside of that, most productivity or proper functioning stops and I mostly just complain about the temperature. It's a common malfunction.

There are surely yet more reasons, many probably very specific to a certain application or need behind this. The real, single silver bullet answer is simply this: pretty much everything is made out of the right material as determined by market demand, usage, cost, and various physical properties balanced by what the people buying these parts want but are willing to pay for.

And these days, most parts are made for mass automated PCB assembly and not repair or rework, for better or worse.

Let's take for example the STM32F410RBT6, the cheapest 32-bit MCU on LQFP-64 package from the Mouser. The volume of the plastic pakage approximatly is 140 mm3. The price of 1000 pieces is 2130$. Compare this to the G800W590018EU, a 12-pins connector. The volume of the plastic approximatly is 196 mm3. The price of 1000 pieces is 105$.

Suppose that the plastic cost is a half of the G800W590018EU price. If the price of plastic rise twice then the cost of G800W590018EU increase by 50%. If the price of plastic fall twice then the cost of STM32F410RBT6 decrease only by 2%.

• Also connectors are expensive from the shear amount of metal they need. If you add composite to the mixture price would be higher. As they usually don't go in reflow ovens there's no much point for then to withstand that much heat. Dec 28, 2020 at 15:19
• @OtávioBorges Surely some connectors go into reflow ovens--surface-mount connectors do exist! Dec 29, 2020 at 20:00
• That's why I said usually Jan 1, 2021 at 5:25