In order to study dust emission from bubble burst in liquid steel, we set up an original experimental device (figure 13) using a vacuum induction melting furnace (Leybold) modified in order to operate at atmospheric pressure under an argon atmosphere. The aims of the experiments are to clarify the way the bubbles burst at a liquid steel bath surface and to quantify the resulting emissions, i.e. film drops and jet drops.
The steel charge (750 g of a commercial steel grade XC38) is melted in an alumina crucible (45 mm inside diameter, 70 mm height), fitted in a graphite susceptor. This configuration reduces electromagnetic convection in the metal bath. The temperature of the liquid steel is controlled by a bichromatic pyrometer; during an experiment, the temperature of the bath is maintained constant, usually at a value between 1600 and 1650 °C.
The gas injection device consists of an alumina tube (7 mm outside diameter, 4 mm inside diameter, 300 mm length), fed with gas through a stainless steel tube, which is connected, outside the furnace, to a mass flow controller and the argon cylinder. The bubbles form at the mouth of an alumina capillary inserted into the injection tube. In order to change the bubble size, we use three different sizes of alumina capillaries (outer diameter: 0.5, 1.2 or 3 mm). Moreover, for a given capillary, the gas flowrate can be modified (between 1 cm3 min-1 and 15 cm3 min-1) as well as the pressure drop, which enables us to vary the bubble size in a wide range, between 4 and 13 mm (all bubble sizes indicated in the present paper are equivalent volume diameters).
Figure 13. Schematic representation of the experimental device The bursting of bubbles at the surface and the formation of the jet drops are observed by means of a high-speed video camera (Kodak Motion Corder) which makes it possible to film the bath surface at a rate up to 10000 frames per second. Actually, good-quality images could not be obtained at such a rate because increasing the shooting frequency entails a reduction in image resolution. We therefore selected rates of 5000 frames s-1 to record the film break and 1000 frames s-1 to observe the formation of the jet drops and to determine the frequency of emergence of the gas bubbles at the surface. The latter frequency is equal to the frequency of the bubble formation at the capillary mouth and thus permits to calculate the bubble size knowing the gas flowrate.
In order to study the film drops, the aerosol formed is exhausted through a rack-mounted tube. The airborne particles are collected on filters inserted in an in-line stainless steel filter holder connected to a flowmeter and a vacuum pump. Two types of filters were used: glass fiber filter (Millipore) for the gravimetric analysis of the particles and PVC membranes (Millipore) for the granulometric analysis and the SEM observation. In order to prevent the saturation of the filters, the exhaust period is limited to 15 seconds for one filter. The flowrate is 4.3910-4 Nm3 s-1, which corresponds to a gas velocity of 0.4 m s-1 at 500 K (typical gas temperature above the bath). According to the Stokes law, it enables to carry particles up to 60 µm in diameter, a size which is larger than that of most of particles contained in the EAF dust. Almost all the particles collected on the filters can be regarded as coming from the steel bath. Indeed, it is possible to remove most of the parasitic particles present in the atmosphere of the furnace by sweeping it with filtered gas. At the beginning of each experiment, the furnace is pumped out and then fed with filtered argon. After one hour of sweeping, there remains in the furnace less than 100 particles with diameters larger than 0.3 μm for 28.3 L of gas and none of these particles have a diameter above 1 μm. Thanks to these experimental precautions, it is possible to obtain a sufficient cleanness of the furnace in order to ensure an accurate determination of the emissions coming from the steel bath.