Hello,

I am using Keithley 2450 Sourcemeter to attain an I-V characteristic of a thin insulating film grown on a hard substrate. At low voltage, I-V curve is linear and displays a correct resistance of ~4k ohm (film dimension = 1cm x 1cm x 100nm). However, when the program runs sweep up to large voltage, there are strange "jumps" that I do not fully understand whether it is intrinsic to the sample or my probe station setup/software is faulty.

I understand a large voltage bias can lead to heating of a sample or capacitive buildup that can bend the curve to make it look like a log curve, but this "jump" I am observing does not really make sense to me. On the graph attached below, 0 to 20 V is applied and back down to 0V. And this "jump" seems to be observed at different voltage input for different direction of voltage sweep.