The pH value 2.0 was selected on the ground of the observation that at a pH value lower than 2.0 MA-DAN derivative is formed more quickly but appeared unstable [12]. On the other hand a pH value higher than 2.0 produces a less stable diazepinium ion, which could be neutralized and hydrolysed [14]. The same effect was observed when the reaction was carried out at a temperature higher than 37 °C. In conclusion, pH 2.0 and 25% ACN were established as optimal conditions for a stable preparation of the MA-DAN derivative at 37 °C within 10 min of treatment. These conditions appear to be suitable to give a reliable value of the free MA content. In comparison, TBA test used for MA determination requires high temperature (at least 90 °C) and long reaction time (at least 30 min) at very low pH value. Treatment of MA with DNPH, although at lower temperature (37–50 °C), is carried out in a concentrated HCl solution.
3.2. Salting-out homogeneous liquid–liquid extraction (SHLLE)
In order to increase the sensitivity of the method, a special attention was addressed to the extraction of the analyte from the reaction mixture that contained 25% ACN. In this view, we experimented that the highly hydrophobic diazepinium ion formed from MA-DAN can be extracted into organic solvents such as diethyl ether or ethyl acetate, provided that an ion-pair is formed with a hydrophobic anion as heptafluorobutyrate. An improvement of the extraction was achieved, using the recently introduced technique, named salting-out homogeneous liquid–liquid extraction (SHLLE) [15] and [16]. Addition of magnesium sulphate to the solution increases its ionic strength and the aprotic ACN forms a distinct upper phase containing the MA-DAN product. Due to the ACN polar feature, the ion-pair of diazepinium ion MA-DAN product with chloride can be extracted, thus avoiding the addition of the hydrophobic heptafluorobutyrate anion.
3.3. Derivative structure and UV response of MA-DAN
The reaction between MA and DAN is expected to afford a 1,5-naphthodiazepinium ion [14], whose mesomeric amino-imino structures 1 (1H form) and 2 (5 H form) are shown in Fig. 2, together with the less probable di-imino structure 3. The UV spectrum of MA-DAN in the solvent mixture selected for the HPLC analysis showed a high absorption at 311 nm, due to the extensive resonance of the naphthodiazepinium ions 1 and 2, that is not possible in a structure like the di-imino cation 3. In order to confirm the attractive feature of high UV response, MA-DAN was compared with MA-DNP. UV spectra of equimolar solutions in their respective HPLC eluents were measured (Fig. 3). MA-DAN response resulted 1.7 higher than that of MA-DNP at their maximum absorption wavelenghts (311 nm and 310 nm respectively) and, therefore, MA-DAN can be considered a suitable derivative for MA measurements by HPLC determination with UV detection.