Sample preparation was based on a modified method as previously
described (Yoshida, Isagawa, Kibune, Nagaoka, & Maitani,
2007). Specifically sample preparation was performed in a cold
room (4 C) to minimize the loss of furan by evaporation. Reagents
were also kept cool prior to use. Heterogeneous samples (i.e. bread)
were homogenized in a chilled homogenizer for a maximum of
2 min. 3e5 g of semi-solid/solid samples or 5e10 g of liquid
samples were weighed into chilled 20 mL headspace vials containing
1 g sodium chloride. The total volume of sample solution
was topped to 10 mL using chilled water. All the vials were
immediately sealed and placed in an ice bath. The internal standard,
40 mL working solution of d4-furan (2.5 mgmL1),was injected
through the septum of each vial. Each sample was analyzed in
triplicate. Considering the relatively higher furan concentrations
found in some samples and high sensitivity of the method, a dilution
factor of 5e10 was applied to samples containing more than
120 ng g1 of furan, and the sample was subsequently re-analyzed.
The dilution factors were applied in order to reside within the
dynamic range of the calibration curve