Hi friends;

A relatively simple question on Barrier Potential of a p-n junction diode:

The equation for the barrier potential is as follows:

Vb = Vt loge(Na*Nd/square(n)) mV
where Vt=KT/e

This equation shows that the barrier potential is directly proportional to the temperature. So as temperature increases the barrier potential should increase.

But in contradiction to this, I have found the following statement:

" With increase in temperature, more minority charge carriers are produced, leading to their increased drift across the junction. As a result, equilibrium occurs at a lower barrier potential. It is found that the Vb decreases by about 2 mV/deegre celcius. "

Please clarify the above, I will be grateful to you.

Thank you
Vaibhav

 

I'm not sure and have not looked it up but you might be missing a saturation current term (Is) which, is _usually_ very small to what is being considered and hence ignored. However, it is a much stronger function of temperature that overcomes the directly proportional relationship in the KT/e term.

 

Look at it in this way
Law of mass action states that the product of concentrations of electrons and holes is always constant at a fixed temperature
(ni)^2=n*p where i is subscript
Again this law is applicable to extrinsic semiconductors.
The n in the denominator is actually (ni)intrinsic carrer concentration. Since you are increasing the temperature the value of (ni) will increase more as it is also temperature dependent.
Hence reducing the value of the potential barrier Vb

__________________
Artificial Neural Networks are dangerous to mankind

 

btw if you don't mind me sidetracking a moment. would increasing the temperature of a diode cause it to produce electricity ?

__________________
I AM the exeption that disproves the rule in many ways but the rules still apply (unfortunately)

my site:www.simons-photography.com
http://rushdenrotaract.org.uk

 

If you heat one of the junctions and cool the other then yes

__________________
Nothing is impossible.
Once a problem is realised, the rest is just details

 

then i should connect two diodes together and put one in the cold and one in the hot ? how do i connect them ? pn-pn, pn-np or np-pn ?

__________________
I AM the exeption that disproves the rule in many ways but the rules still apply (unfortunately)

my site:www.simons-photography.com
http://rushdenrotaract.org.uk

 

No you cannot heat one junction and cool the other. That can be done in metals ( i.e the thermocouple action)

heating of a semiconductor can cause valence electrons to acquire more energy to break their bonds and enter the conduction band in case of both intrinsic and extrinsic semiconductors. Under open lead condition the direction of flow of electrons or holes is random and hence net current produced is zero. Again no amount of heating can cause a semiconductor to overcome its potential barrier. Too much heat will simply destroy the semiconductor.
But when connected to a proper ckt and proper biasing is done then the current flowing through the diode will depend on the supply and the temperature.

__________________
Artificial Neural Networks are dangerous to mankind

 

Except there is devices out there that work on this priciple. BUT the doping area is bigger.

IF you could practically heat one area and cool the other then current would flow. BUT in hte size of a regular PN junction this isn't really possible

__________________
Nothing is impossible.
Once a problem is realised, the rest is just details