Silicon dioxide SiO2 is the most abundant mineral on earth. It exists mainly as crystals (quartz, cristobalite), glasses and amorphous artificial silicas. Silicon oxides are used in a wide range of applications, covering five main industrial fields: electronics and jewelry (resonators, oscillators, sensors, semi-precious stones), cosmetics and food (release of oils and perfumes, antibacterial, toothpaste, sun protection...), specialty chemicals (additions for building materials, encapsulation and release of compounds, mineral fillers in polymers and tires...), health (delivery of poorly soluble active ingredients, components of dressings and wound-healers). In their many applications, silicon dioxides can be produced or used in the form of powders with micron, submicron or even nano-sized grains. These powdered forms can also be generated by abrasion during the performance of certain trades (construction and insulation, quarry work, miners, porcelain workers, prosthetists, etc.). We therefore felt it important to take stock of current knowledge on the toxicity of SiO2 silicon dioxides. Crystalline silica powders (quartz, cristobalite) and those derived from industrial glass cause inflammation of the respiratory tract (fibrosis, oedema, lung cancer) and are cytotoxic. This cytotoxicity is due to the fact that the grinding of these crystalline silicas can create very high concentrations of surface radicals (ROS for Radical Oxygen Species), which generate strong oxidative stress. Moreover, this toxicity can be heightened by the fact that some quartz or cristobalite can be contaminated with iron, the presence of which leads to Fenton-type reactions, a further source of hydroxyl free radical formation. Amorphous silicas are often used as nanostructured powders in applications such as mineral fillers to improve rheological control and mechanical behavior, catalysts, desiccants, toothpastes and cosmetic or therapeutic vectors. These amorphous silicas are produced in large quantities (world production of amorphous silica nanoparticles was estimated at 1.3 tonnes per year in 2000), making them the most abundant synthetic nanoparticles on earth.
Amorphous silica nanoparticles are prepared in two main ways: at high temperature (1200 to 1400°C) by flame pyrolysis of chlorinated precursors (SiCl4) followed by rapid thermal quenching to form so-called silica fumes, or by soft chemistry in aqueous solution via polycondensation reactions involving silanol groups (≡Si-OH + HO-Si≡ ⇔ ≡Si-O-Si≡) to form dense colloidal or mesoporous silicas.
Amorphous silicas lack long-range order, and due to a flat energy landscape, their structures are highly dependent on kinetic and environmental factors. These factors manifest themselves mainly in differences in :
- the architecture of the siloxane network, which consists of a combination of rings whose concentration, pattern (number of members making up the ring) and distribution (throughout the material or mainly near the surface) ;
- the extent of the hydrogen bonding network formed by the silanol groups (≡Si-OH).
We discussed structure/toxicity relationships for amorphous silica nanoparticles of the same size synthesized at low or high temperatures. On the basis of hemolytic tests on erythrocytes, assessment of cell viability and ATP levels in epithelial cells and macrophages, it clearly appears that silica fumes exhibit significant toxicity, whether freshly prepared or annealed or even after air aging. On the other hand, colloidal silicas, produced by soft chemistry under identical processing conditions, are non-toxic. For silica fumes in particular, there is a good correlation between toxicity and hydroxyl group concentration, and their potential to generate reactive oxygen species (ROS) that cause hemolysis of red blood cells. By combining spectroscopic characterizations (vibrational Raman and infrared spectrometry, electron paramagnetic resonance, etc.) and physical analyses (X-ray scattering, microscopy), determining the state of silica aggregation, the concentration of hydroxyl groups, the relative proportion of constrained and unconstrained siloxane rings and their potential to generate hydroxyl radicals, the structural origin of the toxicity of certain amorphous silicas has been better identified. The toxicity of silica fumes seems to derive mainly from their large population of constrained three-membered rings created by their mode of synthesis, and their high concentration of hydroxyl groups. Hydrogen bonds and electrostatic interactions of silanols on the surface of silica fume aggregates lead to strong perturbations of the extracellular plasma membrane, detected by inflammasomes whose subsequent activation leads to cytokine secretion.
The hydroxyl radicals generated by the small, tense rings present in silica fumes, but very much in the minority in colloidal silicas, contribute strongly to inflammasome activation.
In conclusion, it seems that crystallinity is not a prerequisite for the toxicity of silicon oxides. Some amorphous silica nanoparticles exhibit toxicities at least equal to, if not greater than, those of crystalline silicas. In amorphous silicas, ROS production is strongly promoted by the presence of strained three-membered siloxane rings formed at high temperature and highly unstable to hydrolysis. These strained rings are easily observed by Raman spectroscopy and are absent in all colloidal silicas (Stöber silica, silica mesoporous, LUDOX silicas, silicalite gels and silica gels) which, based on tests carried out in the literature, do not present any toxicity. The small rings present in high-temperature silica fume and vitreous silica act as a reservoir for ROS, which can be generated continuously via surface hydrolysis or nanoparticle dissolution, thus accounting for much of the toxicity observed in these silicas. The use of polymeric coatings or lipid bilayers masks the interactions between cells and hydroxyl groups, thereby reducing or even eliminating cytotoxicity. The strategy used in recent work on active ingredient vectors is, on the one hand, to select amorphous silicas elaborated by soft chemistry and, on the other, to coat them with a hybridizing organic layer carrying multiple functionalities. This double condition makes the use of these nanomaterials more reliable.