QuantumATK Forum
QuantumATK => General Questions and Answers => Topic started by: kaypu on December 11, 2010, 10:40
-
i calculate the two probe system, in the low bias voltage area,between -0.8V to 0.8V, it is convergence,and i use the temperature 300k.
but from -0.9V to 1.5V, it isn't convergence, so i increase the temperature to 1000k, it converges. will this affact the results much? i mean the transmission spectrum,I-V, conductance.
thank you
-
The temperature has some influence, but probably not very large. You should check it by computing the lower voltages at both higher and lower temperature, and comparing the transmission spectra.
Another option is to use the high-temperature converged state to initialize the lower-temperature calculation, that usually is favorable for convergence. Otherwise it's a general fact that the older ATK versions (2008.10 and earlier) have a harder time converging at higher bias; in ATK 10.8 improved algorithms make the convergence much more stable at high voltage, we have tested up to 10-12 V (not that this is recommended for other reasons) and got nice convergence.
-
thank you very much