Reactivity of Anode Raw Materials and Anodes for Production of Aluminium

• Main Author: Engvoll, Marianne Aanvik
• Format: Doctoral Thesis
• Language: English
• Published: Norges teknisk-naturvitenskapelige universitet 2002
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:no:ntnu:diva-34; http://nbn-resolving.de/urn:isbn:82-471-5388-2

In the Hall-Héroult process for primary production of aluminium, a considerable amount of anode carbon is lost through unwanted gasification in air and CO2. The carbon gasification reactions are catalyzed by a number of inorganic impurities normally present in the anodes. Some of these impurities follow the anode raw materials while others are introduced during the anode manufacturing process. The aim of this work is to obtain a fundamental knowledge of how the bath compounds: AlF3, Al2O3, NaF, Na3AlF6 and CaF2, which may be introduced in various amounts to prebaked anodes through the addition of recycled anode butts, influence the air and CO2 reactivity of anode carbon. In order to avoid the disturbing and possibly masking effect of other impurities normally present in industrial anode materials, this work uses cokes made by carbonization of high purity carbon precursors (tar oil/petroleum pitch) in a laboratory scale coke reactor. Known amounts of fluoride salts or the corresponding metal acetylacetonates are added to the liquid precursor prior to carbonization. “High”-sulfur cokes are prepared by also adding 4.5 weight percent dibenzothiophene to the precursor (corresponds to an addition of 1 wt% elemental sulfur). The calcined coke samples are characterized in terms of reactivity towards air and CO2 gasification, size and shape of optical texture units and degree of turbostratic order. Scanning electron microscopy and surface area measurements are used to study the surface textural changes resulting from the catalyzed gasification. The extent of gasification inside industrial prebaked and Søderberg anodes is investigated by characterizing anode core samples in terms of air permeabilities, contamination profiles and reactivities towards air and CO2 gasification. The characteristics are related to findings from electron microscopy examinations. COKES DOPED WITH SODIUM ACETYLACETONATE, SODIUM FLUORIDE AND CRYOLITE Sodium acetylacetonate decomposes completely to sodium carbonate during carbonization and calcination of the coke samples. At the initial stages of gasification, the sodium carbonate particles decompose to a sodium oxide phase, which catalyzes the air and CO2 gasification reactions strongly. Sodium fluoride and cryolite also act as strong gasification catalysts. Due to formation of higher amounts of inhibiting fluorine gases (COF2, AlOF2), the cryolite doped cokes are less reactive than the corresponding sodium fluoride cokes. The difference between the two coke series is especially pronounced during air gasification. The air reactivity of the sodium-doped cokes is markedly reduced when 4.5 weight percent dibenzothiophene is added to the coke precursors prior to carbonization. During carbonization, the sulfur is stabilized in large aromatic molecules and it is only liberated when the carbon matrix is gasified. The free sulfur adsorbs on the active sites of the sodium particles and lowers their catalytic activity. The Na-S adsorption complexes are thermally unstable at the CO2 gasification temperature (960 °C) and the CO2 gasification rates are therefore not affected by the dibenzothiophene additions. The additions of various amounts of sodium acetylacetonate, sodium fluoride or cryolite (< 0.7 wt%) do not affect the optical texture and turbostratic structure development processes during carbonization and calcination of the coke samples. COKES DOPED WITH CALCIUM ACETYLACETONATE AND CALCIUM FLUORIDE Both calcium oxide (from decomposition of calcium acetylacetonate) and calcium fluoride act as very strong catalysts towards the carbon-CO2 gasification reaction. Presumably because of the low reaction temperature (525 °C), the air gasification rate is only slightly enhanced by the calcium additions. Due to sulfur poisoning of the catalytically active calcium oxide particles, the CO2 reactivities of the calcium acetylacetonate doped cokes are markedly reduced when 4.5 weight percent dibenzothiophene is added to the coke precursors prior to carbonization. Because of the low reactivity of the calcium acetylacetonate doped cokes towards air gasification, the amount of sulfur released from the coke matrix is too low to provide a measurable air reactivity reduction. Sulfur is unable to adsorb on and deactivate the catalytic sites of calcium fluoride. Neither the optical texture nor the turbostratic coke structure is affected by the additions of various amounts of calcium acetylacetonate and calcium fluoride (< 0.5 wt%). COKES DOPED WITH ALUMINIUM ACETYLACETONATE AND ALUMINIUM FLUORIDE When heated to temperatures above 110 °C, aluminium acetylacetonate decomposes completely to aluminium oxide. During carbonization of the tar oil, the small aluminium oxide particles adhere to the surface of the mesophase spheres and prevent them from coalesce on contact. Addition of 1 – 2 wt% aluminium acetylacetonate has the largest textural effects and gives very fine-grained mosaic textures. At higher dopant loadings, the aluminium oxide particles form agglomerates and their influence on the texture development becomes smaller. Since the concentration of active carbon sites is affected by the fineness of the coke texture, the measured coke reactivities are influenced by both the coke texture and the concentration and dispersion of the catalytically active aluminium oxide. Aluminium fluoride reduces the gasification rates if added to cokes with fine optical textures. The air and CO2 reactivities go through a minimum at an aluminium content of approximately 0.4 wt%, which corresponds to an addition of about 1.7 wt% AlF3 to the precursor prior to carbonization. At higher AlF3 additions the reactivities increase again. The observed reactivity behavior probably results from a competition between deactivation of active carbon sites by adsorption of gaseous fluorine compounds (AlOF2, COF2) and the weak catalytic effect of the condensed aluminium species present (AlF3 and Al2O3). At low aluminium fluoride additions, the fluoride deactivation dominates, but at higher additions, the catalytic activity of aluminium oxide and aluminium fluoride takes over. Due to a low concentration of active sites, the addition of aluminium fluoride has only a minor reducing effect on the reactivity of cokes with coarse-grained textures. The reactivities of neither the aluminium acetylacetonate nor the aluminium fluoride doped cokes are influenced by the concurrent addition of 4.5 wt% dibenzothiophene. GENERAL REMARKS ON RELATIVE CATALYTIC STRENGHTS Based on relative catalytic strengths, the contamination species studied may be ordered as follows (strongest catalyst set as 100): “Low”-sulfur cokes: CO2 gasification: CaF2 = CaO >> NaF > Na3AlF6 > Na2O >> Al2O3 100 100 26 21 15 3 Air gasification: NaF = Na2O >> Na3AlF6 >> CaF2 > Al2O3 > Na2O 100 100 57 29 24 19 “High”-sulfur cokes: CO2 gasification: CaF2 >> CaO >> NaF > Na3AlF6 > Na2O >> Al2O3 100 50 25 21 15 3 Air gasification: Na2O >> CaF2 = Al2O3 > NaF > Na3AlF6 = CaO 100 19 19 16 13 13 Aluminium fluoride inhibits the gasification reactions. There are no cocatalytic effects between calcium and sodium i.e. the catalytic activity of calcium is not affected by the concurrent presence of sodium and vice versa. GAS REACTIVITY OF INDUSTRIAL ANODES The carbon dioxide produced at the anode working surface, percolates through the open porosity of the anodes and reacts with accessible carbon in the lower parts. Airburn is mainly of concern at the exposed parts of prebaked anodes (anode tops and the sides near the tapping positions) and under the gas skirts of the Søderberg anodes. From air permeability measurements and electron microscopy examinations, internal CO2 gasification is found to occur in the lower 3 – 5 cm of prebaked anodes. Depending on the anode top surface temperature, air gasification may occur as deep as 4 cm below the anode top surface. In Søderberg anodes the extent of internal CO2 gasification strongly depends on the gas permeability of the anode. In high-permeability Søderberg anodes, CO2 gasification may occur as far as 60 cm above the working surface. Higher up, the temperature is normally too low for gasification. In some particular low-permeability Søderberg anodes (permeability resembling “bad” prebaked anodes), internal CO2 gasification is limited to the lower 15 – 20 cm of the anodes. In both prebaked and Søderberg anodes, the more reactive binder coke is selectively gasified. Since much of the binder phase in the lower parts of the high-permeability Søderberg anodes is consumed, substantial dusting is expected from the working surface of these anodes. In the prebaked anodes, the binder phase is mostly structurally intact. Catalytically active sodium and calcium impurities are mainly introduced to prebaked anodes via the addition of butts. Pot-room dust is an important contamination source in Søderberg anodes. Additionally, the lower parts of especially the high-permeability Søderberg anodes are contaminated by gaseous sodium and aluminium bath species that penetrates into the anode open porosity and condense within the anodes.