problem about the base current (Ib) of pnp transistor

when i do some simulation on my circuit, i find the base-current of pnp go into the base electrode other than go out of the base. i couldn't find the reason. who can help me? what has happened?

thanks!

What kind of simulator do you use?

Post the diagram and simulation output of your circuit here,

the schematic is below, which is a bandgap reference circuit. when i sweep the temperature, i find when the temperature exceed 100 centigrade, the base current of q4 go into the base electrode other than go out of it. i think it is impossible to happen. i want to know the reason. please give me some advice. thanks a lot!

the simulation is below. the Ie of q3 & q4 is equal. but when the temperature is raising, the Ib of the q4 is gradually reducing. at last it become positive.(i display the -Ib, not +Ib)

If leakage of collector-base junction at 100 C exceed necessary base current p-n-p transistor or Ie/(Bp+1) the total base terminal current can be positive.

Ie of q3 and q4 equal. Base curents should be equal as well. But area of q4 should be larger. So leakage of the q4 c-b junction also larger. That's why you can see positive base current only for q4.

It's due to definitions of the currents of a transistor. The currents are always through into the transistor. For PNP the all current directions are through into the transistor with negative values except emitter current.

All simulators assume these definitions.

thank you for all the help!

I know the definition of the currents of a transistor. The most problem confused me is why the base current goes in to the base eletrode.
Fom, I understand the leakage of the q4 c-b junction is larger. But I cannnot catch exactly why it is the reason for the abnormality. Can you talk more about it?
thanks!

Base terminal current includes 2 currents:
One is negative base current of pnp transistor. This current is defined by external resistors value and not change to much with temperature and depend on resistor temperature coefficient. Moreover if resistor has positive temperature coefficient that's mean emitter and negative base current will be decreased with temperature (Imean absolute value). Beta also will change with temperature (the higher temperature the higher Beta). It also will cause decreasing base current with increasing temperature.
Second current is positive c-b junction leakage current. This current increased with temperature much more than resistor change.
When you monitor transistor base terminal current in really you monitor sum of these two currents. This sum became positive when temperature exceeded 100 C.
You can check how much c-b leakage just simulating single p-n-p transistor with shorted base-emitter terminals (connected to plus) and collector terminal (connected to minus).
Another thing you must check validity temperature range of your model. Some foundries has a limited temperature range for models.

Fom, thanks for your explanation and suggestion!

I simulate the leakage current of the pnp transistor. And I find that when the temperature exceed 100, the leakage current is increasing obviously.
But the simulating temperature range is inside the mode temperature. could I do some change to prevent the abnormaility?

From your schematic I can suppouse 2 things:
1. You ned to get some negative temperature coefficient for your voltage reference.
2. Or you want to get temperature independent voltage about 2.5V instead of 1.25V.
I can propose to you to make the following change:
- remove q1, q2 and m2 from your schematic.
- connect bases of q3 and q4 to GND
- connect inp+ to emitter of q3
- connect inp- to emitter of q4
- add additional q5 with following connections:
base - to common node of R1, R2 and drain of m11
collector - to the GND
emitter - to the drain of m1 and to output of your voltage reference.
I think result will be the same like for your current schematic but you avoid these problems.
Good luck.