CPU Overview - Gold in the Chips
- Thread starter nigthmare_2002
- Start date
Help Support Gold Refining Forum:
patnor1011
Well-known member
Lionhead said:
Electroplating in CPU is done with pure gold, 24k.
All yellow, gold colored, plated gold on electronics is very close to 24K - ALL is 23K plus. No exceptions and this has held true for at least 45 years! Anyone suggesting that yellow gold plating of a lesser karat is used (or even possible) has no idea of what they are talking about.
Lionhead
Member
Good stuff, Thanks
joem
Well-known member
goldsilverpro said:
Even on fingers?
A
Anonymous
Guest
Yes even on fingers.
joem
Well-known member
mic said:
So then after removing foils from fingers, can they be melted as is ( obviously I know they can), but what karat value would the buuton be?
jimdoc
Well-known member
That would depend on how well you cleaned them to get rid of all the metal from underneath the gold before you melted them. And it is so easy to get it purer at the foil stage, it makes little sense to melt it then.
Jim
Jim
Gold plating is used mainly for the purpose of preventing corrosion to the nickel plated copper or kovar or whatever it's plated onto. For anything that is inserted into a connector, etc., such as fingers or pins, the gold plating is hardened (and brightened) by adding small amounts of a metal salt (usually cobalt or nickel) to the gold plating bath. When plated, the Co or Ni alloys with the gold.
This hardened, alloyed gold will allow more connector insertion cycles (plug and unplug), without excessive wear, than would a pure gold deposit. The hardened gold plating usually runs from 99.00% to 99.90% pure, not pure enough to be deemed Pure, or Fine gold (needs to be at least 99.95% pure to meet this criteria). The most common is probably 99.7-99.8% pure. If too much of the hardening metal is used, the electrical conductivity of the gold deposit will decrease to a level that is unacceptable.
About the only items that are plated with pure 99.99% gold are those that will be subsequently heated for such things as brazed die attachment and/or brazed lid attachment on certain IC packages. If the gold weren't pure, the alloying metals could oxidize and discolor the gold, when heated. This oxidation could cause many problems, besides appearance, with such things as wire bonding, soldering, and poor contacts when the package leads are inserted into a connector.
The only time I have ever seen electronics gold plating of less than 99% purity was a 60/40, Au/Pd alloy plated on certain reed switches, and that deposit was white. Plated gold alloys are of a very different color than their cast counterparts. They are generally much paler. For example, when you buy a 14K or 18K plating bath for jewelry, the actual gold purity of the deposit is usually in the 22K - 23K range. These baths only produce a 14K or 18K COLOR. A similar situation exists with real gold leaf. I have a 25 sheet book of actual 16K gold leaf that is very pale yellow, almost white, in color.
Even 99.0% plated gold will appear paler than 99.9%. It is difficult to see the true color of gold because it is obscured by its glare and brightness. To see the true color of a plated or cast gold item, place a white Kleenex over it and put a drop or 2 of water on the Kleenex. The wet Kleenex, where it clings to the gold, should be smooth with no wrinkles or tears. This eliminates any glare and you'll see the true color. This works best when comparing 2 items, side by side. This wet Kleenex method is often used, by those in the know, to control and adjust the color produced by jewelry plating baths. They plate polished brass panels (about 1" x 4") with the bath in its present state and compare the color with panels plated when the bath was depositing the exact color they want. Jewelers are very picky when it comes to the color of gold.
In 45 years of dealing with electronics refining, I have only seen solid gold used a mere handful of times and, except for the pure gold bonding wires used on ICs, it was only used in very rare, very obscure applications. The typical gold you see is most always plated. Due to it's costs, gold is never plated thicker than is needed. For large plane areas or for other non-wear surfaces, figure 10 millionths of an inch (about $.18/in2, at an $1800 market), or less. For those areas that could experience wear, such as pins or fingers, the average is about 25 - 30 millionths ($.45 - $.54/in2). For parts that must be heated for die attach, etc., about 45 - 60 millionths ($.81 - $1.08/in2). These are just averages and there are many exceptions to these general guidelines. For example, when I had my last refinery, I saw trimmed fingers that ran from $45/# to $405/#, based on an $1800 market. In general, but not always, the gold plating on military parts will be thicker, as a thicker gold will provide greater reliability and longevity.
During WWII, and, in earlier years, the gold was often plated quite thick, especially on military parts. At that time, the gold was mainly plated from alkaline cyanide solutions and the deposits tended to be more porous. This, plus the fact that they hadn't yet compiled adequate data, resulted in the manufacturers applying thick gold deposits, just to make sure the components were protected. I once processed several drums of WWII (I think) pins that ran 1 troy oz. of gold per pound. There was also a very thick layer of silver plated under the gold. According to some posts on this forum, much of the Russian electronics from later years also had very thick gold plated over a thick silver layer.
Another possible gold source in these CPUs is when 96Au/4Si or 80Au/20Sn braze preforms are used for die (the IC chip) and/or lid attachment. Although they cover a much smaller area, they are usually 20 to 40 times thicker than the gold plating on those same parts.
This hardened, alloyed gold will allow more connector insertion cycles (plug and unplug), without excessive wear, than would a pure gold deposit. The hardened gold plating usually runs from 99.00% to 99.90% pure, not pure enough to be deemed Pure, or Fine gold (needs to be at least 99.95% pure to meet this criteria). The most common is probably 99.7-99.8% pure. If too much of the hardening metal is used, the electrical conductivity of the gold deposit will decrease to a level that is unacceptable.
About the only items that are plated with pure 99.99% gold are those that will be subsequently heated for such things as brazed die attachment and/or brazed lid attachment on certain IC packages. If the gold weren't pure, the alloying metals could oxidize and discolor the gold, when heated. This oxidation could cause many problems, besides appearance, with such things as wire bonding, soldering, and poor contacts when the package leads are inserted into a connector.
The only time I have ever seen electronics gold plating of less than 99% purity was a 60/40, Au/Pd alloy plated on certain reed switches, and that deposit was white. Plated gold alloys are of a very different color than their cast counterparts. They are generally much paler. For example, when you buy a 14K or 18K plating bath for jewelry, the actual gold purity of the deposit is usually in the 22K - 23K range. These baths only produce a 14K or 18K COLOR. A similar situation exists with real gold leaf. I have a 25 sheet book of actual 16K gold leaf that is very pale yellow, almost white, in color.
Even 99.0% plated gold will appear paler than 99.9%. It is difficult to see the true color of gold because it is obscured by its glare and brightness. To see the true color of a plated or cast gold item, place a white Kleenex over it and put a drop or 2 of water on the Kleenex. The wet Kleenex, where it clings to the gold, should be smooth with no wrinkles or tears. This eliminates any glare and you'll see the true color. This works best when comparing 2 items, side by side. This wet Kleenex method is often used, by those in the know, to control and adjust the color produced by jewelry plating baths. They plate polished brass panels (about 1" x 4") with the bath in its present state and compare the color with panels plated when the bath was depositing the exact color they want. Jewelers are very picky when it comes to the color of gold.
In 45 years of dealing with electronics refining, I have only seen solid gold used a mere handful of times and, except for the pure gold bonding wires used on ICs, it was only used in very rare, very obscure applications. The typical gold you see is most always plated. Due to it's costs, gold is never plated thicker than is needed. For large plane areas or for other non-wear surfaces, figure 10 millionths of an inch (about $.18/in2, at an $1800 market), or less. For those areas that could experience wear, such as pins or fingers, the average is about 25 - 30 millionths ($.45 - $.54/in2). For parts that must be heated for die attach, etc., about 45 - 60 millionths ($.81 - $1.08/in2). These are just averages and there are many exceptions to these general guidelines. For example, when I had my last refinery, I saw trimmed fingers that ran from $45/# to $405/#, based on an $1800 market. In general, but not always, the gold plating on military parts will be thicker, as a thicker gold will provide greater reliability and longevity.
During WWII, and, in earlier years, the gold was often plated quite thick, especially on military parts. At that time, the gold was mainly plated from alkaline cyanide solutions and the deposits tended to be more porous. This, plus the fact that they hadn't yet compiled adequate data, resulted in the manufacturers applying thick gold deposits, just to make sure the components were protected. I once processed several drums of WWII (I think) pins that ran 1 troy oz. of gold per pound. There was also a very thick layer of silver plated under the gold. According to some posts on this forum, much of the Russian electronics from later years also had very thick gold plated over a thick silver layer.
Another possible gold source in these CPUs is when 96Au/4Si or 80Au/20Sn braze preforms are used for die (the IC chip) and/or lid attachment. Although they cover a much smaller area, they are usually 20 to 40 times thicker than the gold plating on those same parts.
- Joined
- Aug 24, 2011
- Messages
- 88
Patnor great response!
What people don't realize is refining isn't a "perfect" science nor is sampling. No two lots are EVER the same, no matter how identical the material appears with the human eye.
What people don't realize is refining isn't a "perfect" science nor is sampling. No two lots are EVER the same, no matter how identical the material appears with the human eye.
patnor1011
Well-known member
EwasteD said:
You probably meant this one :mrgreen:
:arrow: http://goldrefiningforum.com/phpBB3/viewtopic.php?f=49&t=11954
Barren Realms 007
In Remembrance
Use a hammer and tap on it and the plastic should start craking up so you can remove it.
Somehow that hammer thing did not work for me. Just cracked the die out of the leadframe.
Put the whole on a heater for a while, but that plastic lid wouldnt melt or peel off.
I have cleaned the CPU with hot HCl from tin and other BM now and guess its a case for AR or RE.
The tape seems to hold all the leads in place, so if I remove it they could just fall off and create a big mess (now that the die is away).
Put the whole on a heater for a while, but that plastic lid wouldnt melt or peel off.
I have cleaned the CPU with hot HCl from tin and other BM now and guess its a case for AR or RE.
The tape seems to hold all the leads in place, so if I remove it they could just fall off and create a big mess (now that the die is away).
Barren Realms 007
In Remembrance
Marcel said:
If all else fails then take 2 pair of pliers and bend the chip in half and crack it open where the solution can work on the inside of the chip.
ericrm
Well-known member
why do you want to remove the plastic?
are you hopping to keep the desing intact?
are you hopping to keep the desing intact?
Without the plastic I have a leadframe of pure gold which needs no further processing, but cleaning.
I am not asking for this one single piece but in general, I dont want to manually take every CPU apart, so I though there was a way on how to remove the plastic.
It may burn with a torch, but the gold is thin and lots of it may be gone "with the wind" or encapsuled in smolten plastic.
Best would be, to lay it in something that dissolves the plastic and the gold remains.
As for the die: I will crack that open as you suggested and let AR do the job.
I am not asking for this one single piece but in general, I dont want to manually take every CPU apart, so I though there was a way on how to remove the plastic.
It may burn with a torch, but the gold is thin and lots of it may be gone "with the wind" or encapsuled in smolten plastic.
Best would be, to lay it in something that dissolves the plastic and the gold remains.
As for the die: I will crack that open as you suggested and let AR do the job.
niteliteone
Well-known member
Their are two methods here in the forum that will remove the plastic casing.
Look up pryolsis and wet ashing. Be warned of the dangers with the hot acid.
Tom C.
Look up pryolsis and wet ashing. Be warned of the dangers with the hot acid.
Tom C.
ericrm
Well-known member
i might understand you wrong but if you think this cpu is pure gold, it isnt...
i aways toss those in ap and gold come off ,if it wasent plated it would stay on the plastic...
if i didnt understand what you realy mean ,sorry and forget that i wrote something :roll:
i aways toss those in ap and gold come off ,if it wasent plated it would stay on the plastic...
if i didnt understand what you realy mean ,sorry and forget that i wrote something :roll:
