How many times do I need to say GOLD? ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
Age gets better with wine!
Regular TTL was about the last large-scale use of gold doping AFAIK.
It drops the storage time by increasing the recombination rate, which trashes the beta.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics
160 North State Road #203
Briarcliff Manor NY 10510
hobbs at electrooptical dot net
http://electrooptical.net
Yep. If I let it all grow long I can play Santa Claus...
A few years ago a ~2 year old kid in a grocery cart, pointed at me and excitedly said, "Ho! Ho! Ho!" ;-) ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Modern power diodes (of the FRED type and such) are doped in such a way that they can achieve 600 or 1200V breakdown, while adding just enough Pt, or beta irradiation, to get carrier lifetime down to the 20ns range.
It's noteworthy that 1200V devices of any power handing seem to require more like 100ns, which should be because of banging into the combined limits of "how much heavy metal doping can we get" versus "how much breakdown can we sacrifice" versus "how lightly can we dope the junction to make it longer". (The ultimate limits being something like 4.6 or
6.6kV for a single Si junction, where recovery times are pretty humble, probably not using any heavy metals or irradiation.)
I wondered the same as you, George. I found this out:
Seems Hoerni developed gold-doping at Fairchild in mid-1959. The effect is to provide localized sites for hole recombination. In practice the downsides are increased forward voltage drop and increased leakage current. A lot of the fast p-n "logic" diodes like 1N914 are Au-doped. It works in germanium too.
and from:
IEEE Transactions on Electron Devices (Impact Factor: 2.36). 01/1977; ED-23(12):1279 - 1283. DOI: 10.1109/T-ED.1976.18650
I got this abstract:
"ABSTRACT Both platinum and gold have been used to reduce lifetimes in fast recovery silicon power devices. There are substantial differences between the energy levels introduced by these impurities. Both impurities introduce acceptor levels which act to reduce hole lifetimes in n-type silicon; however, the gold acceptor is much deeper (E c - 0.54 eV) than the corresponding platinum acceptor (E c - 0.26 eV). In p-type material, on the other hand, the two impurities are quite similar; gold introduces a donor at E v + 0.35 eV, while the platinum donor is at E v
0.32 eV. In terms of basic physics, this paper establishes guidelines to determine, for a given device type, which lifetime killer should be used to provide optimal performance. Platinum offers improved high-temperature properties and turn-on performance when compared to gold and is a better selection for devices which are switched so rapidly that the turn-off is governed mainly by the high injection lifetime. However, when the switching wave form involves low injection recombination tails, gold is a better choice than platinum. "
A transistor contains silicon, phosphorous, boron, oxygen, nitrogen, oddball dopants like gold, and contaminants like sodium. A transistor run in breakdown will occasionally fail due to these moving around.
But, even transistors that are NOT gold-doped will fail this way. Why, oh, why, do you want to blame gold for the failure?
Are you asleep? The subject is the 2N2369... a gold-doped device... to much heat or too much stress will move the gold to places it shouldn't be.
You need to read up on how gold-doped devices are processed... first off, the crudeness will astonish you. ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...
I see about equal hits for diffusion and ion implantation (usually followed by annealing, which will end up with much the same effect, but better dosed).
Well, now, speaking about die-to-die (on same wafer) variations,, once upon a time there were dozens of 2N transistors that sorted out via testing of each transistor on one wafer; that was the norm (then).
Finally someone got the bright idea to do "laydown" (horizontal) diffusion instead of the "stand-up" (vertical) method. Took weeks to discover the new magic.
But many of the 2Ns simply vanished; one could tweak parameters for one part of the population and lose others. Lots of bullet biting and datasheet tweaking and discontinuance of 2Ns.
But transistor parameters became markedly consistent and reliably repeatable (wafer to wafer).
Nice, thanks. I guess that might explain the rather large reverse bias leakage from 1n4148's
BTW thanks for the report on Vbe avalanching in 2n3094's. I assume that the damage depends on local heating (in the base?) And so there is some energy * time type metric. (certainly some grad student must have measured all this.) Such that you can avalanche it forever at low current. But high current for a short time can kill it.
Ion implantation didn't exist when the 2N2369 was born. Gold was "painted" on the back of the wafer and then diffused in, stopped just as it reached the B-C junction... further "annealing" and it'll continue to even out throughout the die and _die_ ;-) ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
2N2369? Are you dumb as a stump or just playing troll? ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
There is none. If you try to keep a transistor just shy of saturation via reducing the base current then it might not work on the next lot because it almost won't turn on at all. Or it fully saturates because of higher gain.
Depending on your "level shifter" requirements you could use a Baker clamp, then no need for a 2N2369. ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
There's this magical new circuit called the 'Baker clamp'. I hear the newfangled TTL uses it. ;)
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics
160 North State Road #203
Briarcliff Manor NY 10510
hobbs at electrooptical dot net
http://electrooptical.net
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.