The relationship of Vbe and temperature in BJTs

[dohzer]

Is it possible to get BJTs that have a very low change in Vbe with respect to temperature, or are they all roughly the same?

 

[LvW]

vbe versus temperature

Due to physical reasons the VBE voltage of ALL BJT´s changes with temperature at app. -2mV/deg if Ic is kept at a constant value.

 

[timof]

vbe vs temperature

Is it possible to get BJTs that have a very low change in Vbe with respect to temperature, or are they all roughly the same?

Temperature dependence of bipolar transistor current gain is described by the following formula:

gamma ~ exp((dEe-dEb)/kT),

where dEe is the emitter band gap narrowing (with respect to the base) due to heavy doping effect, and dEb is the band-gap narrowing in the base due to presence of Ge in SiGe HBTs (SiGe has a narrower band gap than Si), and kT is the thermal energy.

In bipolar (homojunction) transistors current gain increases with temperature, while in typical SiGe HBTs it is decreasing with temperature. I guess one can select such a (low) content of Ge in the base of SiGe HBT that these two factors cancel each other to make current gain independent of temperature (and thus Vbe at constant Ic will also be independent of temperature).

I have not seen any literature on this, and I doubt that someone would go into the complexities of HBT technology (compared to Si BJT) just to make gamma independent of temperature.