Transistors - die pictures

[Noopy]

TIC236, a Triac built by Texas Instruments.
There are eight indices with voltage ratings from 100V to 800V. M stands for 600V.
The TIC236 allows a continuous current of 12A.






In the middle of the die is the gate contact. In the lower area of the die there are two big bondwires contacting the upper main terminal (surge current is 100A!).




A etched trench guarantees clean edges for high voltage ratings.




At the gate contact there are two etched hemispheres.
You can spot the different doping...




Removing the metal layer doesn´t help very much but you can guess the construction.






There are two antiparallel connected thyristors. In the middle the gate with the p- and the n-doped areas can generate free charges in every operating point and fire up the thyristor.
The hemispheres at the gate contact were neccessary so that the n-doping of the gate doesn´t short to the n-doping of the main contact.


https://richis-lab.de/Bipolar62.htm

 

[mawyatt]

We had an on-site Flash Xray lab to emulate a nuclear event to help with these developments.

I´m sure that was an interesting lab! 

Thanks!

Yes indeed!! They had an audible alarm system to let everyone know they were about to fire off the Flash Xray or Explosive Decompression test, you didn't want to be around for either

Keep those great semiconductor images coming, these are fascinating to view, especially for those of us that worked in this field 

Best,

[Noopy]

Keep those great semiconductor images coming, these are fascinating to view, especially for those of us that worked in this field 

I will! 
I still have quite a lot semiconductors: big and small, old and new, complex and simple, known by everyone and very special... 

I wonder how long it will take till the shutter of my camera quits. 
I bought a new hard disk for my NAS so memory won´t be a problem the next months.

[exe]

I still have quite a lot semiconductors: big and small, old and new, complex and simple, known by everyone and very special... 

I vote for FRED/FERD diodes

[Noopy]

I still have quite a lot semiconductors: big and small, old and new, complex and simple, known by everyone and very special... 

I vote for FRED/FERD diodes

Acknowledged! 

[Noopy]

Some more pictures of the Motorola MJ3001:




Yeah, base emitter breakdown! 
The driver transistor is glowing first because at lower currents (0,03A) the base emitter resistor of the power transistor (50 ) doesn´t  provide enough voltage to drive the base emitter junction of the power transistor into avalanche breakdown.




0,2A




At 0,5A there is a first light in the power transistor. Can you locate it? 
Here we have a voltage drop of 26V. Assuming the voltage across the driver transistor is still around 13V the voltage drop across the power transistor is 13V too. It´s plausible that both have the same breakdown voltage. To get 13V with a current of 0,5A the resistor across the base emitter junction has to be around 26 (datasheet says 50 ).
With a voltage drop of 10V across the whole MJ3001 you can estimate a base emitter resistance across the driver transistor of 7k (datasheet says 2k ).
Well the exact resistances aren´t important and resistors in semiconductors often own a high tolerance.




0,6A, there are quite some "highlights" and darker areas.




0,9A, still the light is not very uniform.
Through the base emitter resistor of the power transistor there has to flow a current of 0,5A. For the base emitter junction there are only 0,4A left. 0,4A over a larger area gives a dimmer light than 0,9A through the smaller area of the driver transistor.




That´s interesting: The light in the driver transistor is quite wide.
You can see it´s no focus problem because in the corners there are sharp structures.






The light extends wider into the inner base area than into the outer emitter area. Perhaps that is due to the doping concentrations? The higher doped emitter occupies less of the junction than the lower doped base.








Here we see the small defect that shows first light and consumes current from the nearby areas that stay dark at higher currents.
We have seen a similar defect / effect in one of the BUX22: https://www.richis-lab.de/Bipolar07.htm




The defect is a point with a diameter of round about 4µm.


Some more pictures here:

https://www.richis-lab.de/Bipolar64.htm

 

[RoGeorge]

That´s interesting: The light in the driver transistor is quite wide.
You can see it´s no focus problem because in the corners there are sharp structures.

Why does the light at the 4 corners located inside of the enclosed area (the 4 corners right near the bonding wire) appears as thick as in the straight lines, while in the rest of the corners the light appears to be thinner?

[Noopy]

That´s interesting: The light in the driver transistor is quite wide.
You can see it´s no focus problem because in the corners there are sharp structures.

Why does the light at the 4 corners located inside of the enclosed area (the 4 corners right near the bonding wire) appears as thick as in the straight lines, while in the rest of the corners the light appears to be thinner?

I assume it´s all about current distribution. The concave corners get a lot of current while the convex corners have to glow with less current since a lot of the current disappears on the way across the edges.

[Noopy]

TIC263, a Triac a little more powerful than the TIC236: 25A / 175A.






Yes, there is some damage. 
The edge length of the die is 4,3mm. They used three bondwires to connect the upper main terminal.




A cleanly etched edge.






We know these structures. 


https://www.richis-lab.de/Bipolar63.htm

 

[Miyuki]

Just wondering, do beasts for 100 Amps and more still have the same structure just huge, or are many small in parallel?

[Noopy]

Just wondering, do beasts for 100 Amps and more still have the same structure just huge, or are many small in parallel?

"Normal" big Thyristors are just huge single Thyristors.
GTOs have a special structure at one of the main terminals that prevents current crowding while switching off (the triggered switch-off, not the zero-current switch-off).

[SeanB]

How about unijunction transistors, those are hard to emulate these days as a drop in arrangement.

[Noopy]

Here we had the Programmable Unijunction Transistor 2N6027: https://www.richis-lab.de/Bipolar14.htm
It´s a built like a thyristor.

Somewhere in my warehouse I have a "real" unijunction transistor. Sooner or later we will see one in detail. 

[Noopy]

Tesla GD619, one of the more rare Germanium-NPN-Powertransistors. 16V / 1A / 4W / 1MHz




Some drying agent.




In the package there is some greenish coat.




Now that is interesting. It seems there are whiskers in the package. In some situations these thin crystals grow out of some metals.
Unfortunately I cleaned the package after the first pictures and realized the whiskers later. So no pictures in more detail. 




On the other side of the base contact there are thicker crystals.




There is some red potting on the transistor but it looks quite "dirty".





The complementary GD619 owns nearly the same specifications but is a little slower: 0,6MHz Perhaps because of the slower p-type doping?




No green potting in this package. No whiskers.




That looks a lot cleaner than in the GD609. But at the bottom there is some green goo too.


https://www.richis-lab.de/Bipolar65.htm

 

[Noopy]

Valvo BTY92, a power transistor isolating up to 100V and conducting up to 50A.






Cathode and gate wiring.






The die is protected with some silicone potting.




The gate bondwire was wetched directly on top of the silicon plate.




The General Electric SCR Manual shows how these thyristors were built.




The die is 14,7mm in diameter and 0,2mm thick.




The surface is quite rough and the die seems to get thinner at the edge.






There is a crack around one quarter of the die. I don´t know whether that was a production problem or a damage that occurred later.


https://www.richis-lab.de/Bipolar66.htm

 

[Noopy]

We had these KT808AM built 1984 and 1986. That was the new design of the KT808.




Here we have a KT808AM built 1982.






In this KT808AM we see the old transistor design.






They had some problems with the bondwires. One of the wires got bent accidentally. While this happened the wire was torn off the die.
There seems to be a thin transparent coating on the die.




The edge length of the die is 5,0mm. It´s significant bigger than the die in the newer KT808 (4,3mm).
It´s no surprise older datasheet state a ft of 2MHz while newer datasheets speak of 8MHz.
At the edges there are blue traces probably marking the area where to saw the wafer.
There is a uncommon MESA structure around the transistor.
The bondwires are soldered to the metal structures.




The structures are quite dirty... 




Yeah, lights on!


Some more pictures here:

https://www.richis-lab.de/Bipolar67.htm

 

[Noopy]

The MJ2501 is the PNP brother of the MJ3001 (https://www.richis-lab.de/Bipolar60.htm).






The construction is the same as the construction of the MJ3001.






The die is the same too.





This MJ2501 is a little newer (1979 vs. 1978).






The construction is a little less clean.




And the silicon seems to be a little blurred too. 




Some imperfections.




Base-emitter-breakdown occurs at -18V, a little later than in the MJ3001 (-13V).
When both junctions are in "breakdown mode" we see some brighter dots.






The structures don´t look good in detail. But you don´t find a cause for the bright dot. 


https://www.richis-lab.de/Bipolar68.htm

 

[Noopy]

2N369, a small germanium transistor: Ucb=30V, Ic=50mA, ft=1,3MHz, P=150mW
Manufacturer: ?






Some white paste probably for better heat removal and corrosion protection.






A nice little alloy transistor.
A thin base and a smaller emitter for better specifications.




Quite massive potting in the base of the package.






 


https://www.richis-lab.de/Bipolar69.htm

 

[Noopy]

Fairchild BC550C








The edge length ist 0,33mm.





Fairchild BC560C






0,33mm edge length, same as the BC550C. The structures are the same too.


https://www.richis-lab.de/Bipolar39.htm
https://www.richis-lab.de/Bipolar52.htm

 

[Noopy]

KT808AT, the "round variant" of the KT808.
You can bolt it on a heatspreader with an additional plate.




It looks quite similar to the KT808AM we had last time (https://www.richis-lab.de/Bipolar67.htm).




A nice clean die.




On the lower edge the silicon is splintered.




A alignment aid?






You can distinguish the emitter and the base area. The distances between the electrodes and the pn-junction varies quite a lot.


And now base emitter breakdown (starting at -7,5V):




20mA




100mA




200mA




1A


https://www.richis-lab.de/Bipolar70.htm

 

[David Hess]

I always found something oddly satisfying about those old power transistor packages.

[Noopy]

I always found something oddly satisfying about those old power transistor packages.

 
They look more like engineering than the "quantum physics" you need for modern integrated circuits.
(Of course there is a lot of engineering in modern devices but I think you know what I mean.)

[mawyatt]

Remember the T03, TO39, TO18, the stud type rectifiers, and the Fairchild transistors with circular ceramic base with black epoxy on top. When I was a kid repairing guitar amps, Peavy used these Fairchild transistors. Think Fairchild also used this package with some RTL logic.

Fond memories indeed!!

Best,

[exe]

I like the aesthetics of metal packages.

[David Hess]

... and the Fairchild transistors with circular ceramic base with black epoxy on top. When I was a kid repairing guitar amps, Peavy used these Fairchild transistors.

And they came in two sizes.  The ones shown below had black ceramic but some had white ceramic.  I think the TO-92 plastic package eventually won out because it was compatible with automatic insertion.

Think Fairchild also used this package with some RTL logic.

That was Fairchild "MicroLogic" as seen below in the larger packages from 1968.

http://spingalhistory.blogspot.com/2017/02/fairchild-micrologic-worlds-first-ics.html
http://semiconductormuseum.com/MuseumStore/MuseumStore_Fairchild_923_Index.htm