6
\$\begingroup\$

I'm trying to make a thermal simulation of a PCB, and one of the parameters I have to put in the simulation software is the component package material. For most standard ICs as far as I know, it's some kind of epoxy, but I'm not sure if it has some reference or material code to investigate. I have not seen any info regarding this but only general descriptions of "plastic" or "epoxy".

Do you know which material it is or a similar material I can use for the thermal simulation?

\$\endgroup\$
7
  • \$\begingroup\$ @periblepsis Do you mean surface impedance? Otherwise it it does not really make sense to talk about sheet resistance, does it? \$\endgroup\$
    – jusaca
    Commented Sep 7, 2023 at 12:33
  • 1
    \$\begingroup\$ A quick search turned up hits like: caplinq.com/semiconductor-epoxy-mold-compounds.html \$\endgroup\$ Commented Sep 7, 2023 at 13:09
  • \$\begingroup\$ Be aware that passive devices will almost always use different materials, so you can't just use the same number for everything if you want that level of precision. Ceramic capacitors are nearly always barium titanate (except for C0G capacitors, which are harder to find information on); inductors are usually various ferrites or occasionally metal powder embedded in epoxy, and resistors are all over the place. \$\endgroup\$
    – Hearth
    Commented Sep 7, 2023 at 19:26
  • \$\begingroup\$ @Hearth Ceramic capacitors don't really dissipate power, though. For thermal simulations, we just need the properties of the main IC's \$\endgroup\$
    – SiHa
    Commented Sep 8, 2023 at 10:53
  • 1
    \$\begingroup\$ @Hearth: Unfortunately even a comprehensive thermal simulation is unlikely to predict someone installing the X7R instead of C0G... \$\endgroup\$
    – Ben Voigt
    Commented Sep 8, 2023 at 16:39

3 Answers 3

22
\$\begingroup\$

You bascially always have an epoxy resin as molding compound ("because of their inherent low viscosity, fast cure properties, low shrinkage during cure, good adhesion to the other components in a chip package, and good overall mechanical stability"), but it is mixed with some silica filler, to match the CTE (coefficient of thermal expansion) of the lead frame. Other additives are carbon black for color and bromine to act as a flame retardant.

The epoxy used changed over time, so it will depend on the manufacturer and age of your chip production.
The first widely used epoxy was bisphenol-A, which was later replaced by Cresol Novolac due to a better heat resistance. In the 1990s this changed again, when larger packages and BGAs became available ("new chemistries and formulations were required to meet their manufacturing and reliability needs"):
Biphenyl resins where used from this time on as one epoxy type, because they could be loaded with up to 90% silica filler, so that bascially no organic material was left to absorb moisture (and by that increase resistance to popcorning of the moulding during reflow). Another type of resin now in widespread use is called mulitfunctional resin, that can impress with a higher glass transition temperature.
To address concerns regarding environmentally friendly molding compounds a new type of epoxy are aromatic resins. These don't require bromine or antimony as flame retardants.

So there are a multitude of epoxy materials, and manufactures can decide to use a variable degree of filler materials like amorphous or crystalline silica. For higher thermal conductivity of the package, sometimes alumina is used as filler.

Source of this information:
Chen, A., & Lo, R.H.-Y. (2012). Semiconductor Packaging: Materials Interaction and Reliability (1st ed.). CRC Press. https://doi.org/10.1201/b11260

The eBook is available as Open Access, so I highly recommend to take a look (chapter 6 is relevant in the context of this question)

\$\endgroup\$
8
\$\begingroup\$

The NE555D is documented to use the following mold compound:

percentage component CAS number
88 % fused silica 60676-86-0
11.15 % epoxy 85954-11-6
0.55 % organic phosphorus 1330-78-5
0.3 % carbon black 1333-86-4

The CAS number allows you to look up information about it, but I have not been able to find the thermal conductivity.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ In particular, FWIW, CAS 85954-11-6 is apparently 4,4'-Bis(2,3-epoxypropoxy)-3,3',5,5'-tetramethylbiphenyl, a.k.a. YX4000 or tetramethyl biphenolepoxy (TBPE). Of course, the thermal conductivity of the mold compound will likely be dominated by the silica filler in any case. The epoxy binder will have some effect, but probably not a major one. \$\endgroup\$ Commented Sep 8, 2023 at 9:17
5
\$\begingroup\$

Plastic is almost irrelevant in your model. Try choosing values from a fairly large range and you’ll see that.

What really matters in thermal performance is the lead frame pattern, and any other bits of metal that transfer heat well. Thickness and material matter. And for that you need to X-ray the chips.

Depending on the resolution you want in the model, it may be perfectly good enough to use lumped parameters for the chips.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.