This select series of anecdotes probably makes it unnecessary to emphasize again the importance of glass in daily and social life stressed above by Bontemps and Figuier. It might in contrast be useful to mention that the antique tradition or ornamental glass was revived at the same period by Georges Frédéric Strass (1701–1773), who became the French King's jeweler, when he invented strass, or rhinestone, a high‐lead crystal glass bearing various metal oxides that is still made today to imitate precious stones.
1.4 The Silica Paradoxes
1.4.1 Biogenic Silica vs. Flint
Historically, glass owes its importance to silicates. But what substances could have replaced silicate glasses in their diversity of uses on a silicon‐free planet? The question would be moot if carbon – the next of kin of silicon in the Periodic Table – and, therefore, life and human beings would have also been lacking. More seriously, however, reflecting on the origin of the silica sources used in glassmaking is not a futile exercise.
It is not widely known that 15 billion tons of biogenic silica glass are yearly produced in seawater by diatoms, sponges, and some other living organisms. Such a biological production has major effects on the Earth's global ecosystem and has now become a biomimetic source of inspiration for designing wholly new materials (Chapter 8.1). Interestingly, biogenic silica also had noteworthy implications for glassmaking because of its recycling into the opal or microcrystalline quartz of flint. Flint, or chert as it is called in geology, is commonly found as abundant nodules horizontally embedded in limestone (Figure 4). Its deposition thus requires carbonate dissolution followed by silica precipitation and, thus, percolating waters undersaturated with respect to calcium carbonates but oversaturated with respect to silica. Without going into the details of the process and of its control by pH and geological context [19, 20], it will suffice here to state that biogenic silica accumulating at the bottom of the sea is the source of the dissolved silica that reprecipitates as flint. And it happens that flint was the raw material used in England from the seventeenth century to remedy the lack of sand pure enough for making optical glass and luxury ware (Chapter 10.10).
Figure 4 The abundant beds of black flint present in a 80‐m high limestone cliff of the English Channel at the Pointe du Chicard in Yport (Normandy). Same beds of the Upper Cretaceous used in the past for making flint glass in England on the other side of the Channel. Height visible on the picture: 10 m.
Source: Photo P. Richet.
In passing, one can also note that silica has been biogenically produced relatively late in evolution compared with calcite and aragonite, the main CaCO3 polymorphs, but then met with immense success especially with diatoms. A major reason was the advantages of an amorphous compared with a crystalline substance in terms of optical or mechanical properties for the materials protecting the living organisms (Chapter 8.1); amorphous calcium carbonates do exist, but they serve instead as intermediate reaction steps, which are short lived and thus end up crystallizing [21], which is not surprising as molten CaCO3 is not itself a good glass‐forming liquid. Interestingly, formation of biogenic silica would have first been a way to evacuate toxic Si at too high concentrations from cells. By a twist of evolutionary history, it would have become a protecting device so efficient for organisms [22] that it has since then played a major role in the global ecosystem, causing, for instance, the Si concentrations to be so low in seawater.
1.4.2 A Quantum‐Chemical Factory: The Production of Silica Sand
Although glassmaking would have been possible without sand, it is unlikely that flint would have led to the invention of glass as it requires thorough grinding to become a reactive raw material. Regardless of grinding costs, it is also doubtful that flint would have been a silica resource widespread and convenient enough for an expanding glass industry. The fundamental importance of silica sand thus remains undisputed. Geologically, sand is produced via the weathering of granite and related SiO2‐rich igneous rocks. The most abundant rock of the Earth's crust, granite is made up of quartz and alkali [(Na,K)AlSi3O8] and plagioclase [(Nax,Ca1 − x)Al2 − xSi2 + xO8] feldspars. Whereas feldspars progressively transform into clay under the action of meteoric waters, quartz resists and accumulates as sand either on the spot or downstream.
The very presence of quartz at the Earth's surface appears to be a clear geochemical anomaly, however, which thus deserves some explanation. With typical 75 wt % SiO2, the melts from which granite crystallizes represent the end products of magma differentiation (Chapter 7.2). Owing to their very high viscosities, they rarely rise up to the Earth's surface to erupt as obsidian flows but crystallize slowly instead at some depth to yield large-grained rocks. These melts are the last produced after partial crystallization of primary magmas, which form themselves deep in the Earth's mantle by partial melting of SiO2‐poor, MgO‐rich rocks (~45 wt % for both oxides, along with ~7 % FeO, 2 % Al2O3, 1 % CaO, and a few ‰ at most alkali oxides). Because oxygen bonds more strongly with silicon than with the other elements (Table A.1), one might think that SiO2‐rich minerals should be the most refractory. As a result, the SiO2 content of primary magmas should be lower than that of their source rock and decrease further through partial crystallization on their way up to the Earth's surface. Such a trend is opposite to the SiO2 increase observed. It is in contrast consistent with the fact that cristobalite, the high‐temperature polymorph of SiO2 at room pressure, is less refractory than lime (CaO), periclase (MgO), and even forsterite (Mg2SiO4) whose melting temperatures are about 600, 800, and 175° higher than the 2000 K of cristobalite, respectively.
The paradox lies in the fact that bond strengths are usually considered within the framework of ionic forces, which are by definition nondirectional. Now, directionality is an inherent feature of Si–O bonding in view of its markedly covalent character. Because electron delocalization through polymerization and creation of Si–O–Si linkages is not large enough to constrain geometrically the arrangements of the SiO4 tetrahedra, the same energy variations are, for instance, caused in H6Si2O7 clusters by a small 0.02 Å change of the Si–O bond length and by a large 20° modification of the O–Si–O inter‐tetrahedral angles (Figure 5). Bending of these linkages is thus so easy that configurational rearrangements take place without involving much energy [24]. The fact is most simply illustrated by the transitions of α‐quartz and α‐cristobalite to their dynamically disordered β‐forms near 573 and 250 °C, respectively. Hence, fusion of these minerals does not require the breaking