First, i think the SETPOINT of your tunneling current while taking the I-V curves should be lower, as the tunneling current has overloaded from -3V to -1V. Second, I-V curve is usually used for interpreting the electrical properties of carriers (eletroons or holes), while the differentials of I-V curve are more commonly usful for interpreting LDOS (local density of states) of surfaces. So ,you should measure the dI/dV curve directly on you samples.