odd shape of current-gate voltage curves

[beauyy]

Dear sir,
   I have drawn a current-gate voltage curve for black phosphorus at bias voltage= 0.5 V by using a python script, but the curve is odd. The current are SCF calculated  with LDA.SZ basis set. In most cases, the curves should be monotonic. What's the matter with my calculation? 

[Umberto Martinez]

Are all your calculations properly converged? Anyway, you miss the line device_configuration.update() right after device_configuration.setCalculator(calculator(), initial_state=device_configuration) I would also save the device configurations in case you want to perform further analysis later on. Basically something like this:

Code

    ...
    device_configuration.setCalculator(calculator(),
         initial_state=device_configuration)
    device_configuration.update() 
    nlsave("filename.nc", device_cofiguration)
    ...

[beauyy]

Thanks for your suggestions. I checked my calculation in transmission of black phosphorus. It shows difficult to converge. During calculation, there is a repeated warning
WARNING                                                                     
# The computed multigrid residual is greater than the required accuracy.     
# Computed residual :   2.53311e-007                                           
# Required accuracy :   1.00000e-012     
The device has been optimized with LDA.SZ basis set.
What is the problem with it and how to solve is ?                                     

[Umberto Martinez]

but even with these warnings is the device calculation converged at all?
I doubt that. single zeta basis set is definitely not enough. You should try to increase your basis set.

[beauyy]

I changed the boundary_conditions in the x-direction to PeriodicBoundaryCondition. The computed residual has reduced to the same order of magnitude of required residual, but still larger than it. Without the spatial region, the device is easy to converge. But when a spatial region is applied on the device, it is difficult to converge even if the computed residual is within the tolerable error range.                                     

[Anders Blom]

You electrodes are way too short in Z. See http://quantumwise.com/publications/tutorials/item/115-how-to-set-up-the-electrodes-properly-in-a-two-probe-system