Removing dies from processors without breaking them

[snoman701]

Is it possible?

I also don't want to change the "image" on the die, and I believe that if I leach the solder away, the image is dulled some, but I may be wrong.

Is it possible to easily remove the die? Even heating from the bottom to release the solder?

I'd like to use them in an assemblage.

 

[goldsilverpro]

Is it possible?

I also don't want to change the "image" on the die, and I believe that if I leach the solder away, the image is dulled some, but I may be wrong.

Is it possible to easily remove the die? Even heating from the bottom to release the solder?

I'd like to use them in an assemblage.

Certainly you can remove them without damaging them.

I once developed a process to recondition the packages from reject IC's. At that time, the most prevalent type packages were white and purple ceramic side-braze, with gold plated lids and legs - the 40 lead ones were the most common. A successful business was built around my process and our main customers were Intel, AMD, and AMI (now defunct).

In general, we first heated the parts to the melting point of the gold/silicon braze and then plucked off the lids and the chip with a small vacuum probe. Next we stripped off the gold brazes and the gold plating with my secret, sooperdooper, extremely selective stripper (took a year of everyday experiments to develop) without causing any damage to the other metals - Ag/Cu braze on legs, Kovar legs and seal ring, Mo/Mn traces, internal bonding fingers and chip pad. Finally we gold plated the parts. Intel, etc., then brazed in a new chip, re-lidded them, and sold them as a new part. The fall-out rate for our re-conditioned packages was lower than that of brand-new packages.

The melting point (MP) of the 80Au/20Sn eutectic brazing alloy, used to attach the lid, is about 533 F and the MP of the 97.15Au/2.85Si eutectic braze, used to attach the silicon chip to the gold chip pad, is about 686 F. We ultimately built a machine that heated the parts on rails and there were about 30 vacuum probes on a manifold. There were hold-downs to keep the parts in place. The temperature the parts were heated to was a little over the Au/Si eutectic of 686 F. The probes were automatically positioned to pluck the lids off 30 parts simultaneously and then the probes moved to an area where the vacuum was released and the lids were dropped into a container. This was repeated to remove the chips.

In the experimental stage, we had rails (about 4" long) made of some metal that would conduct the heat fairly well and take the heat with little or no corrosion. They were made of various widths to fit the widths of the parts (mainly 3/8" and 1/2"). A horizontal hole was drilled in the rail to accept a rod shaped heating element, which was attached to a good electronic temperature controller. At first, we had no vacuum probe so, when the braze melted, we removed the chips by sliding a very thin lab spatula underneath them. When you get good at this, you can do it without damaging the chip. Keep in mind that the molten braze will hold the chip down pretty tight, due to surface tension. Later on, I got a vacuum probe with a push button (to release the vacuum) on the handle. The hole in the probe was small, maybe 1/32" or less.

In the very beginning, I had no fancy heated rails, so I just used a short piece of square steel stock placed on the coil of an electric hotplate and held the part down with long lab tweezers spread apart. Depending on the part size and shape, I'm sure you can figure out a proper piece of steel that will sit flat against the bottom of the part, especially in the area under the chip. You only have to heat it right under the chip.

If you overheat the chip, I would guess that you could damage it. Ideally, you just want to barely, but thoroughly, melt the braze.You might want to remove the bonding wires from the chip before heating it. A #11 Xacto blade under a lighted microscope works well. Use a gentle precise flick with the tip of the blade and you shouldn't damage the chip. Part of the Au or Al wire ball or stitch bonds will remain on the chip but they shouldn't affect the chip's properties. You could remove the wires after you remove the chip but it maybe more difficult than doing it before. Also, instead of using a spatula to lift the chip, try the #11 XActo blade on a handle.

 

[snoman701]

Thank you! Off to the shop!

 

[snoman701]

Next we stripped off the gold brazes and the gold plating with my secret, superdooper, extremely selective stripper (took a year of everyday experiments to develop) without causing any damage to the other metals - Ag/Cu braze on legs, Kovar legs and seal ring, Mo/Mn traces, internal bonding fingers and chip pad.

So...gotta take that one to the grave?

In the experimental stage, we had rails (about 4" long) made of some metal that would conduct well and take the heat with little or no corrosion. They were made of various widths to fit the widths of the parts (mainly 3/8" and 1/2"). A horizontal hole was drilled in the rail to accept a rod shaped heating element, which was attached to a good electronic temperature controller.

Silver?

I used to use a zeiss heated microscope stage that was made out of silver due to it's thermal conductivity/corrosion resistance.

 

[Amarnathd]

Certainly you can remove them without damaging them.

I once developed a process to recondition the packages from reject IC's. At that time, the most prevalent type packages were white and purple ceramic side-braze, with gold plated lids and legs - the 40 lead ones were the most common. A successful business was built around my process and our main customers were Intel, AMD, and AMI (now defunct).

In general, we first heated the parts to the melting point of the gold/silicon braze and then plucked off the lids and the chip with a small vacuum probe. Next we stripped off the gold brazes and the gold plating with my secret, sooperdooper, extremely selective stripper (took a year of everyday experiments to develop) without causing any damage to the other metals - Ag/Cu braze on legs, Kovar legs and seal ring, Mo/Mn traces, internal bonding fingers and chip pad. Finally we gold plated the parts. Intel, etc., then brazed in a new chip, re-lidded them, and sold them as a new part. The fall-out rate for our re-conditioned packages was lower than that of brand-new packages.

The melting point (MP) of the 80Au/20Sn eutectic brazing alloy, used to attach the lid, is about 533 F and the MP of the 97.15Au/2.85Si eutectic braze, used to attach the silicon chip to the gold chip pad, is about 686 F. We ultimately built a machine that heated the parts on rails and there were about 30 vacuum probes on a manifold. There were hold-downs to keep the parts in place. The temperature the parts were heated to was a little over the Au/Si eutectic of 686 F. The probes were automatically positioned to pluck the lids off 30 parts simultaneously and then the probes moved to an area where the vacuum was released and the lids were dropped into a container. This was repeated to remove the chips.

In the experimental stage, we had rails (about 4" long) made of some metal that would conduct the heat fairly well and take the heat with little or no corrosion. They were made of various widths to fit the widths of the parts (mainly 3/8" and 1/2"). A horizontal hole was drilled in the rail to accept a rod shaped heating element, which was attached to a good electronic temperature controller. At first, we had no vacuum probe so, when the braze melted, we removed the chips by sliding a very thin lab spatula underneath them. When you get good at this, you can do it without damaging the chip. Keep in mind that the molten braze will hold the chip down pretty tight, due to surface tension. Later on, I got a vacuum probe with a push button (to release the vacuum) on the handle. The hole in the probe was small, maybe 1/32" or less.

In the very beginning, I had no fancy heated rails, so I just used a short piece of square steel stock placed on the coil of an electric hotplate and held the part down with long lab tweezers spread apart. Depending on the part size and shape, I'm sure you can figure out a proper piece of steel that will sit flat against the bottom of the part, especially in the area under the chip. You only have to heat it right under the chip.

If you overheat the chip, I would guess that you could damage it. Ideally, you just want to barely, but thoroughly, melt the braze.You might want to remove the bonding wires from the chip before heating it. A #11 Xacto blade under a lighted microscope works well. Use a gentle precise flick with the tip of the blade and you shouldn't damage the chip. Part of the Au or Al wire ball or stitch bonds will remain on the chip but they shouldn't affect the chip's properties. You could remove the wires after you remove the chip but it maybe more difficult than doing it before. Also, instead of using a spatula to lift the chip, try the #11 XActo blade on a handle.

hello I have a requirement where we have to retrieve the die from a Ceramic package with out damaging the die both physically and functionally will you be able to do it for us and if so please share your contact details on ***[email protected] or call me on +91 ***0276269

Edit by moderator.
It is not wise to put your telephone and email in the open. Use conversations/PM

 

[Shark]

hello I have a requirement where we have to retrieve the die from a Ceramic package with out damaging the die both physically and functionally will you be able to do it for us and if so please share your contact details on ***[email protected] or call me on +91 ***0276269

Sadly goldsilverpro is no longer with us.

 

[kurtak]

Sadly goldsilverpro is no longer with us.

And we sure do miss GSP

Kurt