Transistor leakage

[Newtronics]

Hi there. Is it possible that I could get some help here on the subject of Transistor leakage? Fist of all....Does a transistor have to have leakage? Secondly.....How can it be reduced or even stopped by additional circuitry?
An interesting subject I thought. Hopefully someone among the erudite can shine some light on this phenomina. Walt

 

[nettron1000]

Fist of all....Does a transistor have to have leakage?

To put it bluntly, transistor leakage is a pain in the butt. Manufacturers and designers are still grappling with that problem and have bin since the invention of the transistor.
Its mainly due to imperfections in the manufacturing process, they cannot deal with transistors directly on the atomic scale. Its very difficult to control imperfections in a crystal latice as well as controlling exact amounts of dopants while the transistor is being manufactured. This is why you will never find two transistors that are exactly the same.

>? Secondly.....How can it be reduced or even stopped by additional circuitry? <



As for ways of reducing leakage ive heard of three ways of doing it. One obvious way was to lower the bias voltages, the lower the driving voltages the lower the leakage.

A second was to connect two or more transistors in series, im still not sure about that one.

A third which ive only heard vague info about , was to incorporate a fourth pin ( two bases ? ) to the transistor which would offer an alternate rout for stray electrons, in other words "pull" electrons off the substrate when the transistor is not conducting.

Anyway thats my take on it, but i dont think its possible ( yet ) to completely stop leakage.[/quote]

 

[nettron1000]

Follow up:

I guess i should mention a fourth one; temperature also has an effect on leakage. Lower temperature and lower leakage.

 

[Roff]

Germanium transistor leakage was sometimes a problem. In modern-day bipolar silicon transistors, you can generally ignore it, unless you are running at very high temperatures. JFET leakage shows up as unwanted gate current, but again it is generally so low that it's not a problem, except at high temperatures. Leakage current approximately doubles every 10 degrees C, so, for example, if you have 1 nA of leakage at 25C, you will have about 16nA at 65C.