Discontinuity on DC Temperature Sweep for low temperature

Hi all.

I'm simulating a voltage reference temperature behavior using DC Sweep and I'm experiencing a discontinuity on the first derivative for ptat voltage in the first point, it means for -40C to the next temp step. When I check the slope, it looks very linear and smooth.

I'm just asking about the origin of the problem: Could it be a deriv() operator or convergence problem?

Some collegues have also reported this problem. 

Thanks!