“…Actually, we verified that in our experimental setup, the excitation beam profile could not be described as a fundamental Gaussian beam; moreover, the thin sample approximation was not fully verified. Therefore, we used the known value of the CS2 sample nonlinear refractive index for hundred-femtosecond pulses, i.e., γref = 2 × 10 −6 cm 2 /GW [23,[25][26][27], as a reference value to calibrate our system and to obtain the Kerr nonlinearity of other samples, using the method described in [28], essentially exploiting the linear dependence of the Z-scan peak-valley amplitude ΔTpv on the nonlinear refractive index for small nonlinearities. Here, the peak-valley amplitude of the CS2 reference sample was ∆𝑇 = 13%, and thus, given the measured value ∆𝑇 of an unknown sample in the same excitation conditions, the associated refractive index is given by As evidenced by an attempt to fit the data with the commonly used simple theoretical model [20], assuming a TEM 00 Gaussian beam, thin sample and small nonlinearities, the experimental traces deviated substantially from the model.…”