Determination of total Ti in candy

<p>Artificially manufactured TiO2 is a common additive in the food industry. Especially in sweets it is used as the whitest and brightest known pigment. The additive (E171) is permitted in food processing with no maximum level specified. However, recent knowledge about nanoparticles has opened the discussion about the toxicology of nanoparticle TiO2. A risk assessment has not been published yet because of the complexity of the nanoparticles (mineral forms, coatings, size, shape, …), the lack of toxicity data and the analytical difficulties for measuring these particles in relevant food matrices. Measuring total Ti can be used as screening method before further analysis on nanoparticle TiO2. Total Ti concentration can be determined after chemical digestion but dissolving Ti needs the use of concentrated sulfuric acid or hydrofluoric acid (HF). The use of S is not optimal for ICP-MS analysis, because several possible interferences with S. A digestion method was optimized using concentrated HNO3 and diluted HF. The determination by ICP-MS of Ti is not an evidence due to several interferences by different spectral overlaps. Several collision/reaction gasses were tested using an Agilent 8800 ICP-MS-MS instrument. Solutions containing possible interfering ions (Ca, Cr, S, P, Cl or C) were injected to check if interferences could be detected. The signal was monitored in no-gas, He, O2 and NH3 mode at all relevant masses. Both the quadrupoles are used. In no gas and He mode, the same mass is selected on the quadrupoles. The use of O2 or NH3 causes a reaction with Ti, so the element is measured with a mass shift of respectively 16 (O2) or 102 (6 x NH3) on the second quadrupole. The concentration of possible interfering ions was high enough to predict a worst-case scenario. Based on the isotopic abundance of the different masses, the results of this interference test, preliminary tests and literature data, measurements were performed in He mode at mass 47-47. In this mode at m/z 47-47 no interferences were detected caused by Ca, Cr, S, P, Cl or C. The abundance of Ti at this mass is relatively low (7.32 %), but LOQ values were similar compared with more abundant isotopes. The method was validated for different kind of sweets, with a broad application domain (0.16 mg Ti/kg – 390 mg Ti/kg) and resulted in an expanded measurement uncertainty of 20.2%.</p>