Ilmenite smelting is conducted in both AC and DC electric furnaces. The ilmenite is reduced with a carbonaceous reductant (typically anthracite) to produce a high-titania slag (containing approximately 86% TiO₂) and a low-manganese pig iron. The slag is processed further for the production of white pigments, while the pig iron is used mainly as feedstock in the foundry industry.

Smelting Process

Enlarge this picture Figure 1. Typical ilmenite-smelting furnace.4

The complexity of large pyrometallurgical furnaces often makes it difficult to reason intuitively about their dynamic behavior. Ilmenite-smelting furnaces, in particular, must be operated with a “freeze lining” to prevent attack of the sidewall refractory material by the corrosive liquid high-titiania slag.¹ Because the freeze lining consists of slag solidified from the slag bath, it interacts both chemically and thermally with the slag bath.

This interaction represents a delicate balance in the process and must be maintained by carefully controlling the relative amounts of the process inputs (ilmenite, reductant and electrical energy). The result is that the ilmenite smelting process presents a number of interesting process control challenges. A typical ilmenite smelting DC furnace is shown in Figure 1.

The ilmenite smelting process is conducted within a crucible of solidified, high-titanium-content slags (known as the freeze lining) contained within the furnace refractory walls. This freeze lining protects the magnesia refractory from chemical slag attack.² Apart from increasing the MgO content in the slag, and thereby exceeding the maximum impurity specification, chemical attack by the slag on the magnesia bricks reduces the life of the refractory lining, leading to side-wall breakthrough.

Protecting the integrity of the freeze lining, through firm control over mass and energy balance within the process, is one of the primary objectives of ilmenite smelting. Another potential problem in ilmenite smelters is hydration damage to the magnesia bricks in the lining caused by water leaks.

A Better Lining

Taking all of these refractory problems into account, a new lining configuration was recently designed that can potentially replace conventional magnesia refractories.* The new lining contains carbon and graphite, which are significantly better conductors of heat compared to conventional magnesia linings. This article describes the evaluation of such a lining for ilmenite smelting using a combination of modeling and test work in a 500 kW DC pilot furnace.

The new lining concept combines wall cooling and thermally conductive carbon and graphite refractories to “chill” the refractories by transferring heat away from the furnace lining.³ Effective water sidewall cooling, along with the efficiency of the heat-dissipating conductive refractories, lowers the temperature of the lining below that of the molten materials. This causes a layer of slag and process metal to solidify, or “freeze,” and forms a protective “skull” that completely coats the refractory hot face. Once formed, the slag skull insulates the refractories, reducing heat loss and protecting the lining from erosion, chemical attack, thermal shock, and other stresses. The result is extended life and improved refractory performance.

The new lining concept also enables significant reductions in lining thickness and mass. As a result, the working volume and capacity of the furnace is increased, installation and commissioning time is shortened, and profit-robbing downtime is reduced. In addition, capital costs are lower.

The new linings combine low-thermal-resistant carbon, graphite, and semi-graphite materials with various ceramic refractories specially selected for insulation, electrical insulation, and steel shell temperature control. Within the lining, each component provides the properties required for its application and works together in a cohesive system to enhance lining performance. An added advantage is that, should the protective skull be lost during upset conditions, self-healing and the replacement of the skull will occur quickly due to the low refractory hot-face temperature.

*ChillKote^TM, designed by GrafTEch. ChillKote is a trademark of GrafTech International Holdings Inc.

Furnace Modeling

Enlarge this picture Figure 2. Schematic representation of the “conventional” freeze lining and wall region of an ilmenite-smelting DC arc furnace.1

Zietsman and Pistorius constructed a one-dimensional wall model to illustrate the dynamic response of the furnace sidewall and freeze lining resulting from interactions with the liquid slag bath and changing conditions in the bath.¹ A schematic representation of the freeze lining and wall region of an ilmenite-smelting DC arc furnace is shown in Figure 2. A furnace model was also constructed to characterize the dynamic behavior of and interactions between the freeze lining and slag bath in response to changes in furnace operating parameters. A schematic representation is shown in Figure 3.

Enlarge this picture Figure 3. Schematic of the furnace and process as described by the  furnace model.1

It was concluded from both of these models that independent changes in the operating parameters resulted in changes of the freeze lining thickness. Independent adjustment of operational parameters moves the process away from the current stable point of operation. In some cases, it could result in process instabilities such as foaming and crust formation. In other cases, it could result in loss of freeze lining and associated damage to sidewall refractory material. The models were constructed for a conventional ilmenite-smelting DC furnace lining consisting of a ramming material and mainly magnesia brick in the sidewall.

Enlarge this picture Figure 4. Drawing of the 500 kW pilot-scale furnace showing the new lining.

The furnace used in this campaign was a 500 kW DC arc plasma furnace located in a facility situated at the Kumba Resources R&D site in Pretoria, South Africa (see Figure 4). Due to the size of the pilot-scale furnace, the design of the hearth pad refractories was somewhat restricted. For a full-scale industrial furnace, the hearth pad design would be easier and would look somewhat different than the pilot furnace.

Feed was introduced through a hole in the roof adjacent to the single solid graphite electrode, with metal and slag being tapped from opposing tap holes. An open arc configuration was used because of the conductive nature of the slag. The ilmenite-to-power ratio was maintained as close to 0.56 kg/kW as possible, with a reductant-to-ilmenite ratio of 12%. The furnace was operated using standard feedstock within the operating envelope, as established on previous ilmenite smelting campaigns, to enable direct comparison between the new and conventional linings. The goal for the TiO₂ grade was 86-88%.

Figure 5. The 500 kW furnace setup.

The furnace was started up by heating scrap with a gas burner and arcing to create a molten pool of metal. After the required temperature was reached, feeding commenced, with the initial taps through the lower metal tap hole. The initial low-TiO₂ slag was tapped as soon as possible, with new slag changing the slag composition to the target TiO₂ grade. A photo of the 500 kW furnace is shown in Figure 5.

Figure 6. When viewed from the top, it is clear that no significant wear of the lining has occurred.

Wear of the sacrificial MgO lining was used as the primary indicator that a protective freeze lining was present at all times during the campaign. Possible wear of this sacrificial lining was investigated by performing a mass balance on the MgO. This balance is based on the following assumptions:

• Constant 0.5% MgO grade in ilmenite feedstock, as analyzed at the beginning of the campaign
• Negligible MgO contributed by the ash in the anthracite
• 7% ilmenite feed loss to the off-gas through dusting (based on previous ilmenite smelting campaigns on the furnace)
• No metal tapped with the slag
• No slag tapped with the metal

Based on these assumptions, 97% of the MgO fed into the furnace can be accounted for. Due to the tapping practices, it is known that small amounts of slag are tapped with the metal. Assuming that 3% of the mass metal tapped is slag, all of the MgO can be accounted for.

It was therefore concluded that no wear of the sacrificial lining occurred during the campaign. These findings are backed up by visual inspection of the lining after completion of the campaign. Figure 6 indicates no noticeable wear of the sacrificial lining, even in the high-wear zone at the metal-slag interface.

In this zone, just below the slag tap hole, the metal layer’s rising level removes any protective skull, and, when the metal is subsequently tapped, hot slag comes directly into contact with the unprotected refractory lining. This slag freezes again as a result of the lower hot-face temperature of the lining compared to the liquidus temperature of the slag.

A Viable Option

The successful operation of a pilot-scale furnace equipped with the new lining suggests that it should be a viable option to consider for an industrial-scale furnace. As required, the lining caused the heat flow responses necessary for slag solidification on the refractory brick. Analyzing and comparing the measured results from operating the furnace with the model’s predictions suggest that the model is useful in providing qualitative predictions on the lining performance.

The new refractory technology relies less on operational control than conventional “freeze linings.” Since the hot face of the new lining is easily maintained below the solidification temperature of the slag, skull formation will always occur, as was shown by both modeling and test results.

For more information regarding the new lining, contact GrafTech International Ltd. via e-mail at refractory.systems@graftech.com or visit www.graftech.com.