Indium from Pentium 4 CPUs

Gold Refining Forum

Help Support Gold Refining Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

ohminator

Member
Joined
Feb 14, 2013
Messages
10
I researched a little bit in Intel's datasheets and found out that some models of the Pentium 4 processors i.e. those with Prescott-cores have the heatspreader soldered directly to the die. This solder is made of pure Indium. It can be easily scraped off after separating the heatspreader from the plastic part.
I managed to obtain 2.8 g out of 15 processors.
 

Attachments

  • DSCF3859.JPG
    DSCF3859.JPG
    153.5 KB · Views: 920
ohminator,
Are you sure that is what you have?
I just did a quick search and what my 10 minute search revealed was a solder made by Indium corporation.

Indium Corp. of America
Alloy 57BI / 42SN / 1AG
with a part number of 52124

57% bismuth, 42% tin, 1% silver.

http://www.6thplanet.com/store/technical/info_processors.html

Patent for a solder with mention of Pentium IV
http://www.google.com/patents/EP1749315B1?cl=en

Other compositions of solder found used in Pentium IV
63% tin 37% lead eutectic solder

Tin solder with 3% silver 0.5% copper


http://www.intel.com/Assets/PDF/designguide/302666.pdf

Can you point me to the documents where you read it was pure indium?
 
Thank you, I've been looking for that information for a while. I too found out that the solder was very soft when ripping apart a lot of P4:s. I still have the gold plated heat spreaders with the indium on them.

So, how to best refine the indium then? Dissolving in HCl and cementing on zinc plate or electrolytic refining, that is the question. Time to start researching. 8)

Butcher raises a good question. I don't think there's any silver in the ones I have... it dissolves in HCl without any white silver chloride forming or any residues. But I've only tested solder from one CPU.
Do you have any reference, Butcher?

But I agree with Butcher, ohminator, do you have any references on the indium claim?

Göran
 
Just one of them: http://www.intel.com/technology/itj/2008/v12i1/1-materials/5-solder.htm

I have no clue how to process big batches the wet way. Since it's already pure the complexity of the procedure should not be very high.
Small batches are best processed mechanically by scraping the metal off I think.
 
Interesting.... the 2 um AuIn2 in the interface seems to contain quite a lot of gold. It would probably end up alloyed to the indium. I see a weekend filled with fun experiments. :mrgreen:

Thanks for sharing!

Göran
 
ohminator,
In the document you posted I did see them speaking of a lead free solder, but what it looked to me like was the indium foil was not the solder but possibly what was being soldered to.

I could be wrong but from looking quickly at the document it looks like the processor does have indium foil probably part of processor composition not the solder, as well as gold, nickel, possibly other metals and other materials, and it also looks like they show where these metals mix when soldered with lead free (solders figure 17), it looks like they are discussing the cracking of the very thin indium foil (250-300Um) when soldered, my guess possibly from oxidation after being solder with lead free solder possibly a problem they had with one of the lead free solders they used at one time, I do not see here were they say they are using pure indium as a heat spreader solder but could be soldering to gold and a thin indium foil.

Although the solder whatever they used could have dissolve some of the other metals into it like gold.
I still am unsure if they are not using one of the common lead free solders to attach the heat sinks.

I will study this document more closely when I have time, to see if I can get more from it.

It seems somewhat vague, and looks like you would have to understand the background information of what they are discussing in the document to figure it out for sure.
As far as I can tell at this point there is at least a little indium in this processor, I am still unsure how much would be in the heat sink solder.

Thank you for posting it, it is informative, I am just a little unsure how to interpret it yet.


Göran
I added some documents in the post above.
 
I think you are over-interpreting something.
It is stated that Indium is used as leadfree solder (thermal interface material) to attach the heatspreader to the silicon chip. To form thermally stable interfaces solder->heatspreader and solder->silicon it is necessary to coat these surfaces with gold. Maybe these coatings are also important for a better wetting behavior of the solder.
That's it.
 
I had a little more time and read this again, I did not realize there were more pages to the article, and it does appear they have stated using indium (my thinking here from fig 17 it is an indium foil with flux)to attach the lids.

From what I can tell it looks like the price of indium ranges somewhere around 2 to 3 dollars US per gram, but I think it should go up as more demand is placed on its use in electronics.

Thanks again.
 
butcher said:
From what I can tell it looks like the price of indium ranges somewhere around 2 to 3 dollars US per gram, but I think it should go up as more demand is placed on its use in electronics.

Thanks again.

According to mineralprices.com the current price for indium is $492.50 per Kg (it was $510.00 Dec.31, 2012 i.e. end of year price - so its down a bit) so thats about 49 cents a gram - not sure were you got the $2 - $3 price from

At $2 -$3 per gram that would make indium worth 2 - 3 times the price of silver --- at 49 cents a gram indium is about half the value of silver

Kurt
 
The use of Bismuth makes sense to me. I have use bismuth in my ancient occupation whenever I needed to solder a component twice. It has a much lower melting point than tin and it allows you to make two reflow passes without unsoldering the first solder contacts. Bismuth is cheap and not toxic.
As for Indium, I have not used it, but it is very common in electronics exsp. for connections between glass or ceramic substrates. It has a very low melting point of around 160C°, so even if it was alloyed together with bismuth and tin you should be able to narrow the search it by finding out the melting point of this alloy.

Btw: Bismuth has a very special smell if molten ( still unsure if that is the metal itself or the flux) It smells good! -To me at least.... :mrgreen:
 
g_axelsson said:
Interesting.... the 2 um AuIn2 in the interface seems to contain quite a lot of gold. It would probably end up alloyed to the indium. I see a weekend filled with fun experiments. :mrgreen:

Thanks for sharing!

Göran
I can't find the text about "2 um AuIn2" and what is meant by "interface" in this context?
 
The linked page was removed, had to dig a bit but found it at the wayback machine...
https://web.archive.org/web/20081011142516/http://www.intel.com/technology/itj/2008/v12i1/1-materials/5-solder.htm

With interface I meant the surface soldered to the heat spreader. In the text they talk about TIM = Thermal Interface Material = the indium foil and the plated metals.

Göran
 
Thank you, Göran!

I reviewed my collection of CPU's and it seems also other Intel CPU's beside Celeron D and Pentium4 have indium soldered heat spreaders: Intel Core 2 Duo and Pentium D

I don't want to look inside my two i5's, my kids will get those for their PC's.

Funny things you can collect, when people know you love that old sh.. :shock: ...stuff :mrgreen:
 
Has anyone even bothered to try to recover the gold on the bottom of the lids? Seem like a lot more work than worth.
 
It is. There's just not enough gold in newer escrap to justify the cost of processing unless you're working on an industrial scale.
 
Back
Top