Making Industrial Processes More Cost Efficient
Toxicology studies have indicated that stone wools with different compositions have different levels of biopersistence. It is however not easy to carry out a large number of in vivo experiments. It is also known that in vitro dissolution rates at two pH values correlate with two dissolution phenomena in lungs. In this work, nonlinear models of in vitro dissolution rates at pH 4.5 and at pH 7.4 have been developed for stone wools of widely different compositions from a limited amount of experimental data. These in vitro dissolution rates, in combination with some other variables, have then been correlated with in vivo retention half times. It is expected that once the models predicting half times are reliable enough, testing on animals could be reduced to a negligible fraction of its amount today.
The dissolution rates of stone wools depend on their composition in a complicated manner. Not only are the effects not linear, there are strong cross effects of combinations of variables. Therefore, the conventional linear techniques are not effective. Phenomenological modelling is hardly possible since almost nothing is known about the kinetics of the potential surface reactions taking place at different pH values.
New techniques of nonlinear modelling have made this kind of model development feasible, as is illustrated in this paper. These techniques have opened up new possibilities in materials science, including fibre technology and ceramics.
Man-made vitreous fibers, synthetic mineral fibers and synthetic vitreous fibers are generic terms indicating amorphous fibers including glass fiber, mineral wools and ceramic fiber products. Mineral wool includes stone wools, glass wools and slag wools. Stone wools are often the best choice as insulation materials, because of low thermal conductivity and good high temperature resistance. If the composition of stone wools is right, they have very little bio-persistence implying a low carcinogenicity and a generally lower pathogenicity. Crystalline silicate fibers are known to be capable of causing pathologies including pulmonary fibrosis, lung cancer, mesothelioma and pleural plaques. Stone wools like MMVF 21 are typically made from basalt or dolomite. Stone wools with lower bio-persistence can be produced by modifying the composition in some ways. Alumina content, for example, is known to significantly influence bio-persistence. Bio-persistence can be reduced by additions or reductions of some other oxides also, and alumina is not the main determining factor in bio-solubility.
For about twenty years, a lot of vitreous fibers of various compositions have been tested in vitro for their solubility at acidity levels found in the extracellular environment in the lungs (pH 7.4), and in the alveolar macrophages (pH 4.5) [1-6]. These are supposed to reflect the bio-solubility in the lungs. The Directive 97/69/EC on the Classification, Packaging and Labeling of Dangerous Substances put mineral wools in carcinogenic category 3 but included exoneration criteria based on composition, bio-persistence and dimensions. This has further activated the research on bio-persistence. A smaller number of in vivo tests have also been carried out. The in vivo tests are expensive, time consuming and involve ethical questions. There is a method in which macrophages and fibers are placed on a membrane through which culture medium is allowed to flow, and dissolution rates of silica, alumina and iron oxides are measured.
There are however no models which can reliably predict the in vitro solubility of the stone wool fibers from the composition of the fibers. There are no models which predict the bio-persistence of the fibers from the composition of the fibers, for a good reason. The dissolution rates of stone wools depend on their composition in a complicated manner. The effects are not only not linear, there are strong cross effects of combinations of variables. Conventional linear statistical techniques are not effective at describing these effects, while the new techniques of nonlinear modelling are still not common. These new techniques have made development of such models feasible. An attempt has also been made to correlate these results with the small amount of in vivo data. Once we have sufficiently good nonlinear models correlating the bio-solubility with composition, or models correlating bio-solubility with in vitro solubility, it is envisaged that the in vivo testing will be redundant to a large extent.