Review of SPL Processing Technology 1997

There are two general categories of technology: (1) technology that treats and stabilizes the waste (rendering the waste harmless) and enables landfilling of the stabilized-waste, and (2) technology that focuses on making products out of the waste. There are hydrometallurgical and pyrometallurgical options for both categories.

Treatment/Stabilization Technology:  This Technology has the disadvantage that a long-term liability for the land-filled material exists into perpetuity. Examples of such technology are Reynolds Process, the Pechiney Lime Suspension Process, the Alcan Inertization Process, the Pechiney SPLIT process, the Imco Process, one option of the Comalco Process, the IGT "Cyclone" Process, the Alcoa Flow-Through Leaching Process. This type of technology promises to make the waste harmless by today's standards before it is landfilled.

Legitimate Recyclers:  As discussed, product-oriented technologies can avoid the long-term liabilities associated with the stabilization technologies. However, the product-oriented technologies tend to be more complicated and hence, more capital intensive. As such, they are more sensitive to economies of scale and require more sophisticated management. Examples of this technology include the Alcoa/Ausmelt technology used in Portland, Australia, one option of the COMTOR process developed by Comalco, and the Alcan Hydrometallurgical Process. The first two of these processes involve a pyro process followed by a hydro, or chemical process. (It is significant to note that Alcoa also has a hydrometallurgical process that is completely a "hydro" process that they have chosen not to commercialize to date.) The COMTOR process originally concentrated on destroying the cyanide by calcination, and, as an apparent afterthought, considered alternatives for recovering the fluorine from the calcined residue.

Preferred Technology

Of the product-oriented technologies, the Alcoa/Ausmelt and the Alcan Process are clearly more viable than the other processes of which the author is aware. The Alcoa/Ausmelt approach is probably more capital intensive, but, less risky, than the more complex process from Alcan. In the Alcoa/Ausmelt process, the SPL is ground, melted in a furnace with fluxes, and the HF driven off by air lancing of the slag. The process that Alcoa uses to capture and convert the HF to AlF is unknown, but, could be similar to that used in their Port Comfort plant in Texas that manufactures AlF from fluorspar. Materials of construction and refractory selection are challenging, but, Ausmelt and Alcoa has piloted this technology with SPL and several companies sell technology for making AlF including Buss Chemical Company and Lurgi.

Alcan Process:  One of the more promising technologies is a process developed by Alcan.In the Alcan Low-Caustic Leach and Lime Process (LCLL), the finely ground SPL is leached with caustic to remove the fluorine, free and complexed cyanide, alumina, and some silica into the leach liquor. After washing, the carbon rich residue is a potential fuel/reductant for metallurgical operations. However, it may contain excessive amounts of sodium for most applications.

The cyanide is then destroyed by hydrolysis in a high pressure autoclave. After evaporation of the remaining liquor, a 50% caustic solution is produced and separated by filtration from sodium fluoride crystals through a a difficult filtration process. This sodium fluoride can be converted into aluminum fluoride or calcium fluoride by aqueous processes. (Alcan operates a 40,000 tpy aluminum fluoride plant in Jonquiere, Quebec.) The main advantage of the Alcan process is that it produces more products than the Ausmelt/Alcoa process, probably at a competitive capital and operating cost for a given through-put. The placement of the lower valued caustic product and carbon product in many localities will be challenging. The caustic liquor is extensively recycled in the process leading to the possibility of impurity build-up in the circuit.The Alcan process will also require more sophisticated technical management (hydrometallurgists) to operate effectively. (DEPCO also has a hydrometallurgical process, however, it is at the early stages of development and not ready for commercialization.)

The Ausmelt/Alcoa Process:  This process uses the proven furnace technology of Ausmelt to melt previously groundSPL into a slag.Cyanides in the SPL are destroyed during the thermal processing. An iron source is used to produce a commercial fayalite slag commonly used in non-ferrous metallurgy smelting. After the slag is formed, air and natural gas are blown through the melt via a lance to liberate the fluorine. The fluorine reports to the gas stream as HF along with other volatile fluorine species.

The volatile HF in the off-gases from the furnace could be treated by a process developed by Buss Chemical Company Technology, now Kvaerner Technology. The fluorine is recovered from the gas stream by scrubbing with concentrated sulfuric or oleum to remove the fluorine. The fluorine is then distilled from the sulfuric acid and condensed. The AHF collected may then be sold or passed through a fluid-bed reactor containing alumina to produce alumina fluoride.

The original Ausmelt technology was developed to recycle SPL into products. But what has evolved is a fluorine recycling technology applicable to SPL. The Ausmelt technology reacts moisture from the submerged lance combustion with the fluorides in the molten slag. A more direct approach is to inject steam to remove the fluorides and oxygen to burn the carbon in the SPL. This approach is being modeled by MV, Inc. and will prove to be a competitive method to make anhydrous hydrofluoric acid and the lowest cost method of recycling SPL.

Slags are typically molten mixtures of metal oxides and silicates. Iron blast furnace slags are typically mixtures of SiO2-CaO-Al2O3; this mixture is also the basis for alumina-silicate refractories, glassware and Portland cement. Common fluxes are lime and magnesia in iron and steelmaking, fluorspar in steel refining and silica in copper smelting. The purpose of these fluxes is to give the slage the desired melting point, viscosity, density and/or chemical properties.

Originally, Ausmelt proposed adding limestone and iron ore to the slag for fluxing and then using water from the combustion process to react with the fluorides to produce hydrogen fluoride which leaves in the off gases. While a completely functional approach, it does little to take economic advantage of their own technology. We have proposed substituting fluorspar for limestone and augmenting the water from the combustion process with injected steam.

Limestone which is mostly calcium carbonate dissociates and dissolves as calcium oxide in the slag while releasing carbon dioxide which bubbles through the slag and into the furnace off-gases.

CaCO ---> CaO + CO

In the extreme case, this can lead to slag foaming and entrainment of slag into the off-gases. The lime reacts with silica to form silicates and other metals to form complex metal oxide compounds in the slag. Commercial limestone will have impurities that are characteristic of the given deposit. Magnesia, silica, alumina, and iron oxide are common impurities whose amount vary with the quality of the limestone.

Fluorspar when added to a slag dissolves as a fluoride in the slag. In the slag, the fluorspar dissociates into its ionic components. The fluoride ion becomes indistinguishable from fluoride ions from the SPL or any other source. This fluoride is now free to react with water vapor dissolved in the slag. When the fluospar reacts with the water vapor the overall reaction is given by:

CaF2 + H2O --> CaO + 2 HF

The slag chemistry is not altered by substituting fluorspar for limestone. There may be minor differences in the impurities associated with each raw material, but, not any more so than between two different sources of limestone. In short, the average composition of the slag will not change except for the fluorine content; and the fluorine content of slag already varies from a few percent to twenty percent depending on the batch of SPL being processed. Therefore, the process already has to deal with a variable fluorine content both from the source and from the fact that the fluorine is removed during the course of processing.

The limit of the fluorspar addition is the concentration of CaO that can be tolerated in the slag. There are commercial slag systems with 65 to 70% of CaO. This is different that the slag currently used by Ausmelt in Portland, but, no less doable.The economic incentives to do so are huge.

What does a higher average fluoride concentration do to the remove process. In one word, it will make it much more efficient. As with all chemical reactions, the first part of anything is easier to remove. As the concentration goes lower, the concentration product, or the driving force for the reaction decreases. Does it follow that the reaction will go twice as fast if twice as much fluorine and water vapor is present. Probably not, but, it will go much faster and it will stay faster longer. This impact on the reaction time by adding fluorspar is expected to be minimal. (Technical note: about 1000 ppm of water vapor can dissolve in a slag; equivalent to about 50 ppm of hydrogen.)

Comparison of Promising Technologies

Either the Ausmelt Process, or the Modified Ausmelt Procss, or the Alcan LCLL process are sufficiently developed to be commercialized with acceptable levels of risk. The pyro process, is probably somewhat more capital intensive, but, less risky. Each can produce one or more high value fluorine products, each destroys cyanide, and each produces one or more lower-valued products that may be difficult to market. Operating costs should be similar for each process. However, both offer the promise of eliminating the long-term liability associated with SPL at rates competitive to existing stabilization technology. The rate of implementation of these competitive technologies may be more critical than the relative merits of either in a market with limited SPL.

Only the modified Ausmelt Process offers the possibility of having a competitive process for making anhydrous hydrofluoric from fluorspar whose marginal cost of processing SPL is very low. In fact, in a large regional facility, the fluorine could be recovered profitably from SPL even without tipping fees. Nevertheless, such a facility will never be build in North America.