3.2. Method performance
The method performance was assessed by checking linearity in
a matrix-matched calibration curve, by evaluating LOD and LOQ, as
well as selectivity, accuracy and precision.
Linearity of the analytical method, previously verified by Barp
et al. (2013) for a mixture of paraffin oil and offset printing ink
in n-hexane (in the range 1–50 lg mL1 each), was confirmed also
in the presence of the food matrix. To this purpose a blank pasta
sample (with no detectable mineral oil) was spiked with increasing
amounts of printing ink solvent (9% of MOAH) in the range
between 0.20 and 25.0 mg kg1 food (corresponding to 0.18–
22.75 mg kg1 of MOSH and 0.02–2.25 mg kg1 of MOAH). The
analyses were performed in duplicate at 4 different fortification
levels. The slopes of the calibration curves built using the printing
ink standard in n-hexane and the one built analysing the fortified
pasta sample were compared by running a t-test at the 5% significance
level to confirm the absence of any matrix effect (p > 0.05).
The least squares method was used to estimate the regression line
and linearity and the goodness of the curve was confirmed using
lack-of-fit and Mandel’s fitting tests (Fcalc < Ftab)