Dear Pic'ers,

My reply to Ben,

> Dear Ben,
>
> I am no expert in device physics but I did pull my old college text, M.
> Omar, "Elementary Solid State Physics," 1975.
>
> There are two sections in this book concerning conductivity (electrical,
> that is):  Metal and semiconductor.  The conductivity mechanisms are
> different.  In the semiconductor, the conductivity is proportional to the
> number of n-type carriers.  The constant of proportionality is related to
> the mobility of the carriers.
>
>   conductivity = n*e*(mu_e + mu_h)        (6.35)
>
> The number of carriers is proportional to,
>
>   n = f(T) * exp( -Eg / (2*kb*T) )        (6.36)
>
> and f(T) is a function that depends only weakly on T, e.g., a polynomial.
>
> The end result is stated as, "Thus, conductivity increases exponentially
> with temperature because of the exponential factor in (6.36)." pg. 275.
>
> So, in a nutshell, your intuition is correct in thinking that the electrons
> will be scattered by the increasing lattice vibrations, causing the
> resistance to increase.  Indeed, the mobility of electrons (part of f(T)
> above) does decrease with increasing T.  What it seems your intuition did
> not tell you is that there are more free electrons in the semiconductor when
> temp increases.  It seems to turn out that the increase in the number of
> free electrons happens at a faster rate than the decrease in mobility.
> Overall, it would appear that the resistance of a semiconductor does indeed
> decrease with increasing temperature.
>
> I guess with metals, you don't have this electron-hole carrier effect which
> would explain the difference.
Andy Register