“…Among such additives, TiO 2 represents one of the most versatile in aluminosilicate matrices: as a single seed former or in combination with ZrO 2 , it enables the production of a number of technologically relevant materials, including highstrength cordierite and spinel glass-ceramics and low-thermal expansion glass-ceramics based on quartz solid solutions (Qss) [2][3][4]. Despite being investigated since the early days of glass-ceramic development [5], however, the fundamental laws controlling the nucleation efficiency of TiO 2 in different compositional systems are still not fully understood: while 4 mol% are more than sufficient to induce volume crystallization in lithium aluminosilicate glass-ceramics (LAS) [6], higher additions (6-10 mol%) have been shown to be necessary in magnesium aluminosilicate (MAS) [7], sodium aluminosilicate (NAS) [8] and other multicomponent glasses [9][10][11][12][13]; conversely, TiO 2 was found to detrimentally foster surface nucleation in calcium aluminosilicate matrices [14]. The steady improvement of our analytical capabilities has recently shed new light on the complexity of seed formation, particularly within the LAS and MAS systems: (i) incipient liquid-liquid phase separation [15], (ii) a coupled increase in average coordination number of Ti and Al during crystal nucleation [16][17][18], (iii) the resulting pervasive nanostructural heterogeneity of the residual amorphous matrix [19,20] and (iv) the occurrence of TiO 2 (B) as a metastable precursor to the formation of anatase and rutile [21,22] represent only some of the crucial phenomena involved in TiO 2 nucleation which were disentangled in the past few years.…”