Attorney Ref. No.: 169593.119972/WO SYSTEMS AND METHODS FOR USING ALUMINUM DROSS IN ALUMINUM ELECTROLYSIS CELLS BACKGROUND [0001] In traditional aluminum smelting operations, such as the Hall-Heroult process, aluminum metal is produced from alumina in an electrolytic cryolite-based bath. The metal produced is prepared for sale to customers in the casthouse. During operations in the casthouse, a waste product commonly known as “dross” is produced. Aluminum dross typically contains aluminum, other metals, salts, and other non-metallic substances, like oxides, carbides, sulfides and traces of pollutants (e.g., polychlorinated dibenzo-p-dioxins and dibenzofurans). When it contacts water, dross may emit flammable gases such as acetylene or toxic gases such as ammonia. Thus, dross poses various potential safety hazards, is potentially harmful to the environmental, and its disposal is expensive. SUMMARY OF THE DISCLOSURE [0002] Broadly, the present disclosure relates to systems and methods for producing aluminum metal in aluminum electrolysis cells using recycled dross as a feed component. In one embodiment, aluminum from an aluminum electrolysis cell is obtained and processed. In one embodiment, a casthouse (or similar) processes the aluminum from the aluminum electrolysis cell. In one embodiment, the processing of the aluminum from the aluminum electrolysis cell results in the production of dross. In one embodiment, the dross is further processed, resulting in the production of a high metal fraction and a low metal fraction. The low metal fraction (herein “processed dross”) may be re-introduced into an aluminum electrolysis cell (or many aluminum electrolysis cells) to at least partially assist in forming metallic aluminum and with little or no material impact on aluminum electrolysis cell operations. [0003] In one embodiment, a method comprises (a) adding processed dross to an aluminum electrolysis cell, wherein the processed dross contains not greater than 40 wt. % metal, (b) monitoring a magnesium fluoride content of the aluminum electrolysis cell, (c) monitoring an electrolytic bath temperature of the aluminum electrolysis cell, and (d) monitoring an excess aluminum fluoride concentration of the aluminum electrolysis cell. In one embodiment, a method comprises a step (e) of determining, based on steps (b)-(d), whether to change a feedrate of the processed dross to the aluminum electrolysis cell. In one embodiment, a feedrate of the processed dross is lowered when excess magnesium fluoride content is realized in an electrolytic bath of the aluminum electrolysis cell. In another embodiment, the feedrate
Attorney Ref. No.: 169593.119972/WO of the processed dross is increased when a depleted magnesium fluoride content is realized in an electrolytic bath of the aluminum electrolysis cell. [0004] In one embodiment, a method comprises, based on steps (b)-(d), maintaining a constant liquidus temperature of the electrolytic bath of the aluminum electrolysis cell, wherein the temperature of the electrolytic bath varies by not greater than 10ºC due to the adding step (a). [0005] In one embodiment, a method comprises adding processed dross to an alumina source of the aluminum electrolysis cell. In one embodiment, the method may further include adding both alumina and the processed dross to the aluminum electrolysis cell via a feeder in communication with the alumina source. The alumina and processed dross may therefore be added concomitantly or simultaneously to the electrolysis cell. [0006] In one embodiment, a method comprises adding processed dross to a solid bath source of the aluminum electrolysis cell. In one embodiment, the method may further include adding both solid bath and the processed dross to the aluminum electrolysis cell via a feeder in communication with the solid bath source. The solid bath and processed dross may therefore be added concomitantly or simultaneously to the electrolysis cell. [0007] In one embodiment, a method comprises adding processed dross to a fluoride source of the aluminum electrolysis cell. In one embodiment, a method further comprises adding both fluoride material and the processed dross to the aluminum electrolysis cell via a feeder in communication with the fluoride source. The fluoride source and processed dross may therefore be added concomitantly or simultaneously to the electrolysis cell. [0008] In one embodiment, a method includes removing molten aluminum metal from the aluminum electrolysis cell, wherein the molten aluminum metal comprises magnesium. The method may further comprise determining an amount of magnesium in the molten aluminum metal, and, based on the determining step, providing aluminum fluoride to the molten aluminum metal in a quantity sufficient to react with the magnesium (e.g., a majority of the magnesium) to produce a magnesium end product. The magnesium end product may be any end product that results from the reaction of fluorine and magnesium, including, by way of example, solid magnesium fluoride, neighborite, and weberite. In one embodiment, a method includes removing the magnesium end product from the molten aluminum metal. In one embodiment, after the removing step, a method includes transferring the molten aluminum metal to the casthouse for treatment. In one embodiment, the providing step comprises adding aluminum fluoride to the molten aluminum metal in a quantity sufficient to both (a) react with the majority of the magnesium to produce the magnesium end product, and (b) react with one or more of sodium, calcium and lithium to create one or more of solid sodium fluoride, solid
Attorney Ref. No.: 169593.119972/WO calcium fluoride, and solid lithium fluoride, respectively. Accordingly, the removing step may also include removing one or more of the solid sodium fluoride, the solid calcium fluoride, and the solid lithium fluoride from the molten aluminum metal. [0009] These and other inventive aspects of the disclosure are provided herein. BRIEF DESCRIPTION OF DRAWINGS [0010] FIG. 1 is a flow diagram illustrating one embodiment of a method for separating processed dross. [0011] FIG.2 is a graph illustrating an estimated increase in the metal pad silicon content due to implementation of D2P recycling using the hypothetical values in Tables 1 – 3. [0012] FIG. 3 is a graph showing a linear prediction model using metal pad magnesium analysis to estimate MgF2 concentration in the electrolytic bath. [0013] FIG.4 is a flow chart describing one embodiment of a sample process for determining when to make an xsAlF3 target adjustment. [0014] FIG.5 is a graph showing cathode voltage drop (CVD) measurements collected during testing of alloyed dross recycling. DETAILED DESCRIPTION [0015] The present disclosure relates to systems and methods for using processed dross in aluminum smelting operations in a safe and cost-effective manner. Indeed, the new dross-to- pot (D2P) processes described herein enables the recycling of low-metal content dross (e.g., casthouse dross) and other residues back into the aluminum smelting process. The D2P process implements systems and methods in the casthouse, bath treatment center, and/or potroom of an aluminum smelter to reduce or eliminate the landfilling of dross and to capture its metallic value. Dross used in the D2P process can be received from a compatible external processor. Pre-processing can also be done onsite in the casthouse to produce a stream of dross of an appropriate size (e.g., no bigger than two-inches (5.08 cm)) that can be used in D2P processing. The D2P process can be used for a range of aluminum alloys, including pure aluminum, foundry alloys, and wrought aluminum alloys (e.g., 1xxx, 3xxx, 4xxx, 5xxx, and 6xxx alloys). i. Safety [0016] During the trial stages of D2P recycling testing, it may be necessary to manually add material into the molten electrolytic bath (sometimes called “bath” herein) of one or more aluminum electrolysis cells (sometimes called “a pot” or “pots” herein). Any solids added to a molten bath should be kept free of moisture and should be properly preheated prior to addition
Attorney Ref. No.: 169593.119972/WO to the molten bath. Personal Protective Equipment (PPE) requirements should be evaluated to ensure protection from any potential bath splashing. Additionally, any of the manual measurements done during the trial period or during deployment of D2P should be done in accordance with a location’s established safety protocols. ii. Casthouse Operations for D2P [0017] The D2P process essentially starts in the casthouse with the formation and handling of dross during furnace tending. For D2P to be deployed, the use of bone ash or any other phosphorous containing whiting agents should be discontinued as such materials may be a source of phosphorous contamination. Any replacement materials should be evaluated for their phosphorous content. In general, non-dross casthouse residues should be kept out of dross containers, such as filters or combo bags. In particular, used phosphate bonded ceramic foam filters must be segregated and not introduced into the dross stream used for D2P. [0018] When D2P is deployed at a location, attention should also be paid to the amount of salt flux being used. While salt flux may be necessary for improving metal quality, effort should be made to avoid excess use of salts. When the dross is returned to the potrooms, chlorides present in the dross from salt additions may form HCl emissions. Additionally, if MgCl or KCl salts are used, the metallic cations may accumulate in the bath and contribute to the overall modification of the bath chemistry. [0019] Nitride formation should also be avoided. While the formation of aluminum nitride may be non-problematic for the D2P process, minimizing the amount of nitrides (e.g., AlN) has the benefit of reducing the amount of nitrogen oxide (NOx) emissions that may eventually form during the introduction of dross to the potrooms. [0020] The use of dross pressing may be incompatible with the D2P process because it results in a pressed dross skull that would be difficult to then feed into later, downstream steps of the D2P process. Instead, the remaining metal fraction retained in the dross skull is typically recovered via a rotary salt furnace. [0021] As a matter of logistics, one straightforward way for dross to be introduced is through a bath crusher. Generally, bath crushers are not designed to handle large pieces of dross with high metal content. To help with this, the casthouse should emphasize good skimming techniques to minimize the amount of metal removed with dross. Once cool, dross can be filtered through a screen to produce a metal rich fraction material of a predetermined size (e.g., larger than two inches (>5.08 cm)) and a smaller, low metal fraction material (e.g., two inches or less (≤5.08 cm)). Other filter sizes may be used. A vibratory screen or rotary screen are
Attorney Ref. No.: 169593.119972/WO generally suitable for filtering. While this screening step is not required, screening does reduce the amount of material that needs to be processed. Specifically, the small, oxide rich pieces in the low metal fraction can be sent directly to the bath crusher, while the larger, metal rich pieces can be processed downstream to recover the metal. Ideally, the metal would be recovered using a method which also produces a dross stream during metal recovery that can be returned to the smelter and safely sent through the bath crusher as well. This process flow is illustrated in FIG.1. [0022] When the D2P process is implemented, it is important to analyze and record the masses of the material at each step of the process as well as the composition. Knowing both the mass and compositions of the material that eventually enters the potrooms is critical for determining the metal returned to the pots through D2P as well as estimating any potential impacts to metal purity. iii. Method of Dross Additions [0023] Prior to re-use in an aluminum electrolysis cell, the dross from an aluminum casthouse furnace may be processed (e.g., to provide for a metal-rich fraction and a low metal fraction). The low metal fraction may include, for instance, a high volumetric or mass content of oxides. Methods of processing dross to achieve such fractions are described in International Patent Application Publication No. WO2019/157589, which is incorporated herein by reference in its entirety. For instance, a batch of dross (e.g., dross resulting from skimming of a metal holding furnace in an aluminum plant, or of aluminum-rich-plant-residue) may be charged into a nonconductive crucible of an induction furnace, e.g., to ensure that the charge is thereby heated above the melting point of the metal to be recovered by electromagnetic induction of electrical eddy currents in the partially metallic charge. An inert atmosphere may be provided to the furnace by filling the furnace with inert gas, such as Ar or N2, for example to prevent oxidation of the metal during the process. The induction coils of the furnace may be energized by an electrical current, for example in order to induce a heating electrical eddy current in the charge to bring it to a temperature above the melting point of the recoverable metal. For example, once melted, eddy currents may cause vigorous stirring of the melt, assuring good mixing. Next, the recoverable free metal may be removed from the furnace crucible, for example by tilting the crucible or by means of a tap hole at a bottom portion of a furnace crucible, where it may solidify. Using this method, a high metal fraction (e.g., 60-80 wt. metal %, or more) may be achieved. The remaining fraction is a low-metal fraction and generally includes not greater than 40 wt. % metal, i.e., “processed dross.” In one embodiment, processed dross may have
Attorney Ref. No.: 169593.119972/WO not greater than 35 wt. % metal. In another embodiment, processed dross may have not greater than 30 wt. % metal. In yet another embodiment, processed dross may have not greater than 25 wt. % metal. In another embodiment, processed dross may have not greater than 20 wt. % metal. In yet another embodiment, processed dross may have not greater than 15 wt. % metal. In another embodiment, processed dross may have not greater than 10 wt. % metal. The remainder of the processed dross may contain a variety of impurities and oxides, as described in further detail below (e.g., 5-25 wt. % AlN; 10-50 wt. % alumina). [0024] Processed dross (i.e., the low metal fraction) can be recycled into an aluminum electrolysis cell in a number of different manners. The use of processed dross in an aluminum electrolysis cell is sometimes referred to herein as “dross recycling” or similar. As initially described relative to FIG.1, one way to implement the D2P process is to feed processed dross having a predetermined size (e.g., ≤ 2 inches (5.08 cm)) into the bath crusher located in the bath treatment center. As may be appreciated, any appropriate predetermined size may be used, and processed dross having a size smaller and/or bigger than two inches (5.08 cm) may be used, depending on application. Also, addition of processed dross into the bath crusher may ensure that the processed dross is generally evenly distributed to all pots in small quantities. As the processed dross goes through the crusher, the high metal fraction of the dross may partially mechanically separate from the low metal fraction and some of this high metal metallic fraction may collect in the crusher and build up over time. Accordingly, routine emptying of the metal in the crusher should be implemented. [0025] The processed dross may be added to the pots at any suitable time. One typical time would be to add the processed dross to the pot(s) following an anode change and as part of the process of covering the anode / cell crust with crushed bath. The processed dross may then dissolve into a cell’s bath through different pathways, e.g., from anode cover material falling into the bath, as the crust interacts with the molten bath, or during crust breaking. [0026] Processed dross can be directly added to pots through bagged additions through the taphole. This can be done like other manual chemical additions to a pot. Bagged processed dross should be stored in a dry environment and pre-heated appropriately before being added to the pot. The bag should be constructed of an appropriate material, such as paper, for additions into the taphole. The bag will burn, causing its contents to empty into the electrolytic bath of the aluminum electrolysis cell where the oxide fraction will dissolve in the bath and the metal fraction will melt, adding to the metal pad. [0027] Alternatively, processed dross additions may be made to a pot through an alumina feeder or other auxiliary feeder such as a fluoride or bath feeder. Processed dross may be
Attorney Ref. No.: 169593.119972/WO introduced into a feeder’s storage bin within a cell’s superstructure and may be gradually dosed into the cell each time the feeder dispenses. [0028] Measuring of the average percentage of the metal from the processed dross that stays behind in the crusher may be used for accounting for the location of the metal. This mass balance can be estimated by Eqn. (1), (1) wherein: maddition to metal layer is the mass of the metal from dross the makes it into the pot’s metal layer; fmetal is the average metal fraction of the recycled dross; mdross added to crusher is the mass of dross being added to the bath crusher; and mmetal in crusher is the average mass of metal that remains in the crusher. [0029] As detailed in Eqn. (1), only a portion of the metal present in the processed dross fed to the bath crusher will end up in a pot’s metal layer. To determine how much metal from the D2P process is being added to the metal layer, it may be necessary to determine the average fraction of metal in the processed dross added to the crusher, fmetal, which can typically be measured using XRD analysis of the processed dross samples. Then, knowing the average mass of metal that accumulates in the crusher, mmetal in crusher, will allow for completing the mass balance for the metal contribution from D2P. This helps to ensure that the potrooms are not overestimating the amount of metal introduced into the pots through dross recycling. iv. Behavior of Impurities [0030] Processed dross contains many impurities from many different sources including alloying elements, raw material impurities, bath carryover, and fluxing agents. Generally, the elements present in the processed dross will go to one of three pathways when recycling is done in an electrolysis cell: (a) accumulation in the bath layer, (b) accumulation in the metal pad, or (c) production of gas emissions. Impurity elements such as Si, Fe, Mn, Ti, and Cr may accumulate in the metallic aluminum pad (sometimes called “pad” or “metal pad” herein) at the bottom of the cell since they are reducible with the voltage drop present in aluminum electrolysis. Elements such as Mg, Ca, P, and K (and Cl to a small extent) may accumulate in the bath. Finally, nitrogen and chloride may mainly go on to form gaseous emissions.
Attorney Ref. No.: 169593.119972/WO [0031] Knowing the average chemical composition of the dross is therefore beneficial in assessing the potential impact to the metal purity in the potrooms. To perform this assessment, the following information should be collected: 1. Average dross composition of each alloy series produced at the casthouse; and 2. Dross generating factor of each alloy series. [0032] The use of this data can be seen in Tables 1 and 2 as hypothetical examples of a casthouse product mix and dross generation. This data may then be used to calculate the mass of impurities that will be introduced to the cells during dross recycling and is an input in the mass balance calculations. Table 1: Example table of composition inputs based on hypothetical production of alloy series.
The mass of recycled dross (mdross recycled) is calculated from the data in Table 1 through the use of Eqn. (2),
where: fdross generation is the dross generation factor for a specific alloy series determined from casthouse experience; malloy production is the planned annual production of a specific alloy series; and xdross returned is the weight percent of the dross produced that gets returned to the potrooms. The value of fdross generation for each alloy can be calculated by Eqn. (3),
Attorney Ref. No.: 169593.119972/WO where mdross generated is the average mass of dross generated for a specific alloy. The value of xdross returned is determined by Eqn. (4),
where: mscreened dross is the mass of any <2 in. fraction of dross screened in the casthouse; mprocessed dross is the mass of dross returned from the external processor, and mdross produced is the total mass of dross produced. In one embodiment, the value of xdross returned is in the range of 0.3 to 0.4. Table 2: Average dross compositions used in total mass calculations.
Table 3: Mass inputs of the alloying elements present in dross from hypothetical annual production numbers.
(“MT” = metric ton) The mass values shown in Table 3 are calculated from the data in Table 1 and 2 using Eqn. (5), (5) where: melement is the annual mass of a specific element returned to the potroom due to D2P recycling;
Attorney Ref. No.: 169593.119972/WO xelement,series is the weight fraction of an element for the dross of a specific alloy series as shown in Table 2; and mdross recycled is the mass of the dross recycled as seen in Table 1 and calculated by Eqn. (1). The data in Tables 1 – 3 may be useful to perform a mass balance calculation to assess the impact of processed dross use on potroom metal purity. While an annual basis was used for example purposes, any periodic basis may be used for the mass balance calculation. Additional useful data for the mass balance model may include: 1. Average of the potroom’s metal pad mass. If not measured recently, it is recommended to perform copper dilution to confirm value. 2. Average of the bath mass in pots. If not measured recently, it is recommended to perform a NaBr dilution to confirm the value. 3. The tap cycle time. 4. Average number of pots in operation during the applicable period (e.g., one year). 5. Average production of one pot (e.g., average daily production). 6. Baseline impurity concentrations for Mg, Si, Fe, etc. [0033] Such data may be used in a mass balance model to estimate the rise in impurities expected from dross recycling, i.e., the use of processed dross in aluminum electrolysis cells. This will allow an aluminum smelting plant to assess trends and any potential impact on its ability to meet process chemistry requirements. The mass balance model performs an estimate of the transient response of the impurities being added through dross recycling. Each step of the calculation represents one tap cycle, because metal tapping is where a small mass of the metallic impurities leaves the electrolysis cell along with the molten aluminum being tapped. Eqn. 6 provides an example of this mass balance for the accumulation of silicon in the metal pad
where the mass of silicon in the metal pad at any point in time, mSi in metal pad, is equal to the silicon brought from raw materials, mSi in, raw materials, and the amount from recycled dross, mSi in, dross, with the amount removed during metal tapping representing the only significant stream out, mSi out, tapped metal. The masses of silicon for each of these values is found from the composition of stream multiplied by the total mass of the stream. This mass balance can then be used to predict the new steady state concentration of silicon. FIG.2 illustrates an estimated
Attorney Ref. No.: 169593.119972/WO increase in the metal pad silicon content due to implementation of D2P recycling using the hypothetical values in Tables 1-3, above. [0034] It may be useful to measure the phosphorous content of the processed dross. Due to the redox behavior of phosphorous when it is present in molten bath, phosphorous contamination may lead to a parasitic loss of aluminum electrolysis current efficiency, wherein, for every 0.01 wt. % P increase in the bath concentration, current efficiency may decrease by 0.68%. Additionally, once phosphorous is present in the bath, it is difficult to isolate and remove. Due to the efficiency of modern potroom emission controls, almost the entirety of any phosphorous emissions will be captured and returned to the pots with the reacted alumina stream. As mentioned previously, in the casthouse the use of bone ash is a large potential source of phosphorous. Any bone ash replacements should be assessed for their phosphorous content. Close attention should made to ensure that casthouse residues, such as filters, are not being introduced into dross bins. The average phosphorous content of the current dross mix should be analyzed to determine if the amount of phosphorous that would enter the potroom during recycling is tolerable. [0035] Magnesium is an element that may accumulate in the bath layer during dross recycling and has a potential to impact the thermal state of the aluminum electrolysis cell. The presence of magnesium in the bath layer as MgF2 may not be considered harmful to the electrolysis process. Nonetheless, MgF2 concentration(s) should be properly managed. This will be discussed further in the section on Bath Chemistry Control. While the majority of magnesium introduced through recycling accumulates in the bath layer, a thermodynamic equilibrium will develop resulting in higher levels of magnesium in the metal pad. The predicted levels of magnesium in the metal pad should be assessed for any potential impact on the ability to produce low magnesium alloys in the casthouse. [0036] The “Treatment of Aluminum in Crucible” (TAC) process to reduce the levels of alkali and alkaline elemental impurities in metal tapped from the potline may also be useful. Specifically, after aluminum metal is removed from a pot, and before it is processed in the casthouse, TAC may be employed. TAC processing equipment may be obtained from, for instance, STAS, 622 Rue des Actionnaires, Chicoutimi (Québec), Canada G7J 5A9. The TAC process involves adding small amounts of AlF3 directly into a crucible and then stirring the molten metal with an impeller, which allows the AlF3 to react with dissolved impurities in the metal (e.g., impurities such as Na, Ca, Li). The result is the formation of solid cryolite phase that can be skimmed off the top of the metal. As it relates to metal produced using D2P, it may be useful to at least partially control magnesium content in the metal using TAC. Specifically,
Attorney Ref. No.: 169593.119972/WO since magnesium falls in the alkaline family of elements, it should readily react with the AlF3 added during the TAC process to form solid magnesium fluoride/cryolite species that can be skimmed off. The solids skimmed from the TAC process can then be returned to the bath treatment center. Higher levels of AlF3 may be used to account for magnesium in the molten aluminum recovered from the aluminum electrolysis cells. In one embodiment, a method includes removing molten aluminum metal from the aluminum electrolysis cell, wherein the molten aluminum metal comprises magnesium, determining an amount of magnesium in the molten aluminum metal, and based on the determining step, providing aluminum fluoride to the molten aluminum metal in a quantity sufficient to react with a majority of the magnesium to produce a magnesium end product. The magnesium end product may be for instance, one or more of solid magnesium fluoride, neighborite, and weberite. The magnesium end product may then be removed from the molten aluminum metal. The molten metal may then be transferred to a casthouse for treatment. In one embodiment, the amount of aluminum fluoride provided to the molten aluminum metal is sufficient to both (a) react with the majority of the magnesium to produce the magnesium end product, and (b) react with sodium, calcium and/or lithium in the molten metal to create one or more of solid sodium fluoride, solid calcium fluoride, and solid lithium fluoride. Accordingly, the step of removing a magnesium end product may also include removing one or more of solid sodium fluoride, solid calcium fluoride, and solid lithium fluoride from the molten aluminum metal. [0037] Depending on the extent of use of potassium salts as a flux in the casthouse (e.g., KCl), a small amount of potassium may be returned to the potroom during dross recycling. Potassium (K) is an element that may accumulate in the bath layer and typically is seen in low concentrations at smelters with conventional bath compositions. During dross recycling, the concentration of potassium should be monitored because high concentrations may pose a risk to the integrity of potlining materials. Like sodium and lithium, potassium may be adsorbed by some aluminum electrolysis cell materials (e.g., graphite) where it may intercalate, potentially leading to swelling of electrodes (e.g., cathodes). It is therefore recommended to limit KF (potassium fluoride) concentrations (e.g., in the range of 2 - 5 wt. %). v. Bath Chemistry Control [0038] In embodiments where processed dross comprising magnesium is recycled to an aluminum electrolysis cell, metallic magnesium and magnesium oxide present in the dross will primarily be converted to magnesium fluoride upon contact with the molten electrolyte. An equilibrium will develop between the bath and metal layers, resulting in the transfer of some
Attorney Ref. No.: 169593.119972/WO magnesium to the metal layer. Due to the lower partitioning of magnesium in the metal layer, however, the magnesium introduced through dross recycling will accumulate in the bath layer. A new, higher, steady-state concentration of MgF2 concentration in the bath layer may eventually develop over the course of the deployment of the D2P process. Changes in dross composition will also result in changes to MgF2 levels. Therefore, proper bath chemistry control may require knowledge of MgF2 concentrations in the bath (e.g., the average MgF2 concentration). [0039] Increasing levels of MgF2 in the electrolyte will modify the liquidus temperature if chemistry adjustments are not made. For every 1 wt. % increase in MgF2, there may be a corresponding 5˚ C decrease in liquidus temperature of the electrolytic bath. Accordingly, it may be useful to determine the average MgF2 concentration in the electrolytic bath over time, including during the start of dross recycling when the MgF2 concentration begins to increase. [0040] Monitoring the MgF2 concentration can be done through bath samples, where XRF or ICP for direct measurement of Mg in the sample can be performed. XRD analysis for quantification of magnesium bearing fluoride phases can also be performed if the magnesium concentration is sufficiently high. An easier alternative method can be to utilize Mg measurement in the potline’s routine metal sample analysis. Since a thermodynamic equilibrium exists between the magnesium present in the bath and metal layer, an empirical model can be developed that allows the use of magnesium in the metal layer to predict the MgF2 concentration in the bath layer. This is done through the collection of simultaneous bath and metal samples. Using the concentration data from the bath and metal samples, a simple linear model relating the MgF2 concentration in the bath to the Mg content of the metal pad can be developed, as shown in FIG.3. If potline operating conditions change drastically, such as operating temperature or line current, then it is recommended to take a round of bath samples to confirm the accuracy of the model. More complex models taking into account bath temperature and the composition of other bath components may also be developed. This model may take the form of Eqn. (7) for example.
Attorney Ref. No.: 169593.119972/WO where a, b, c, d, e, f, g, h, and i are adjustable parameters that are used to fit the equation to the experimentally measured data, “Mg” is the metal pad content of magnesium in wt. %, xsAlF3 is the measured excess aluminum fluoride of the bath in wt. %, and T is the bath temperature in ˚C. Items T0, Mg0, and xsAlF30 are scaling parameters that will be equal to the midpoint of the range of the measurements used in fitting the equation. For example, if bath temperatures from 940ºC to 1000˚C are used in the data set, T0 will be 970ºC. While potentially requiring more data and analysis, the use of this equation (or similar) should improve the accuracy of the prediction of MgF2 in the bath and may be used in operations where a high level of precision is useful or necessary. [0041] In one embodiment, a near-constant liquidus temperature of the electrolytic bath of the aluminum electrolysis cell is maintained. In one embodiment, a near-constant liquidus temperature is within 10ºC of a target liquidus temperature. In another embodiment, a near- constant liquidus temperature is within 5ºC of a target liquidus temperature. In yet another embodiment, a near-constant liquidus temperature is within 3ºC of a target liquidus temperature. In another embodiment, a near-constant liquidus temperature is within 1.5ºC of a target liquidus temperature. A near-constant liquidus temperature may be achieved by, for instance, adjustments to the xsAlF3 target made in accordance with tracking the predicted MgF2 content in the bath. [0042] In one embodiment, a probe, such as the STARProbe™, may be used to analyze bath samples to determine liquidus temperature. The STARProbe™ is available from Alcoa USA Corp., Pittsburgh, PA or STAS, 622 Rue des Actionnaires, Chicoutimi (Québec), Canada G7J 5A9. Embodiments of the STARProbe™ are described in U.S. Patent Nos. 6,942,381 and 8,088,269, each of which is incorporated herein by reference in its entirety. [0043] While the STARProbeTM cannot independently measure MgF2 concentration, the impact of changes in MgF2 concentrations can be correlated to changes to the liquidus temperature. Theoretically, a 1 wt. % increase in MgF2 would be compensated by a 1.2 wt. % decrease in the xsAlF3 target for a pot’s bath chemistry in order to maintain a constant liquidus temperature. When STARProbeTM is used for bath chemistry analysis and control, it has been found that it is not necessary to fully match the theoretically predicted xsAlF3 target adjustment. As MgF2 concentration increases, the cooling response of the bath sample analyzed during a STARProbeTM measurement changes. In one embodiment, an approximate decrease of 0.48 wt. % xsAlF3 target for every 1 wt. % increase in MgF2 concentration may at least partially help to maintain a constant liquidus temperature. For bath samples having a high MgF2 concentration, the STARProbeTM may no longer match the absolute value of xsAlF3 if
Attorney Ref. No.: 169593.119972/WO comparisons are made against other analytical methods such as XRD or pyrotitration. Importantly though, it should still be considered a reliable indicator of the relative concentration of xsAlF3 that can be successfully used to maintain a constant liquidus temperature when operating with non-conventional bath chemistries. [0044] In one embodiment, the MgF2 and liquidus temperature are measured periodically (e.g., weekly). When a sustained increase of about 0.5 wt. % MgF2 is reached along with a decrease outside the desired liquidus temperature range, a change in xsAlF3 target may be required. Checking both the MgF2 and liquidus temperature before implementing a change in the xsAlF3 target may help to confirm that the change in liquidus temperature is due to a changing MgF2 concentration and not due to a change in another variable (e.g., alumina concentration), for example. This procedure is demonstrated in FIG. 4. For factories with multiple aluminum electrolysis cell potlines, it may be beneficial to check the average of the MgF2 and liquidus temperatures for each potline and adjust them individually according to their respective conditions. [0045] A transition table such as that shown in Table 4 can be used. The liquidus temperature can be, for instance, within 1.5˚ C of the desired liquidus temperature. The pre-recycling values shown in the first row are included as examples; actual average data should be used as the starting point for constructing the table. It is recommended to verify the size of the xsAlF3 target step required for a near-constant liquidus temperature (e.g., ±1.5ºC) during the pre- deployment testing. Table 4: Transition table to help guide xsAlF3 target adjustments during periods of changing MgF2 concentration.
[0046] Direct measurement of the liquidus temperature through the use of the STARProbeTM or another commercially available product may facilitate managing a changing bath chemistry that may result from adding processed dross to an aluminum electrolysis cell. For aluminum
Attorney Ref. No.: 169593.119972/WO electrolysis cells that employ offline bath sample analysis as the means of chemistry measurement and control, calculating the liquidus temperature through the use of published models will be possible as well. [0047] If metal pad sodium (Na) content is tracked (e.g., as an indicator for estimated current efficiency), it should be noted that increasing magnesium content in the metal pad depresses the concentration of sodium in the metal pad. During the transitory phase of dross recycling, as magnesium concentration increases, it is expected that sodium content will decrease. Once MgF2 concentration stabilizes, a relationship between sodium concentration and current efficiency can be established. Readings indicating a decrease in sodium with increasing MgF2 is not indicative of a decrease in current efficiency; it is the result of the thermodynamic interactions between sodium and magnesium present in both the metal pad and bath layers. As MgF2 levels increase, sodium can be more closely associated with this fluoride complex in the bath layer, and partition slightly less into the metal layer. [0048] Throughout the course of dross recycling with the D2P process, value-added-product market conditions may shift. Some periods may see high demand for highly alloyed products, while other periods may see a high demand for pure or lightly alloyed metal. As explained previously, the casthouse product mix will dictate what impurities are returned to the potroom through dross recycling. If a situation arises where there is a switch from mostly high magnesium dross to mostly low magnesium dross, the MgF2 concentration in electrolytic bath will start to decrease. The transition from higher MgF2 levels to lower MgF2 levels can be managed in the same manner as the above-described transition from lower MgF2 levels to higher MgF2 levels. For instance, by following the process laid out in FIG. 4, a generally constant liquidus temperature can be maintained by increasing the xsAlF3 target at the appropriate time(s). [0049] It may also be advantageous in certain situations for the MgF2 concentration to be held at a specific concentration. For instance, if an aluminum alloy end product (e.g., an ingot or billet) requires a specific (target) magnesium content, the amount of magnesium in the metal pad may be controlled such that the ingot/billet cast from the metal of the aluminum electrolysis cell already realizes the target magnesium content, eliminating the need to later alloy the metal with magnesium. Achievement of a specific magnesium fluoride content may be accomplished, for instance, by either increasing or decreasing the feedrate of dross to create a new steady-state MgF2 concentration. At steady-state, the magnesium entering an electrolysis cell needs to be equal to the magnesium leaving the cell. The amount of magnesium leaving the cell (pot) at the new steady state concentration can be calculated by Eqn. (8),
Attorney Ref. No.: 169593.119972/WO (8) where xMgF2, bath is the targeted concentration of MgF2 in weight fraction, xMg,MgF2 is the weight fraction of Mg in MgF2 (i.e., 0.39), m bath tapped is the average mass of bath tapped from a pot over a tap cycle, xMg, metal is the corresponding weight fraction of Mg in the metal pad at the targeted concentration of MgF2 and mmetal tapped is the average metal tapped from a pot over a tap cycle. The parameter xMg, metal may be back calculated from Eqn. (7), which may improve accuracy. The rate of dross additions over a tap cycle can then be calculated by Eqn. (9),
where mMg, out is the mass of magnesium calculated by Eqn. (8) and xMg, dross is the average magnesium concentration in the dross. Due to naturally occurring process variations, it may be necessary to adjust the dross feedrate calculated by Eqn. (9). Adjustments to the feedrate may be calculated, for example, by a proportional control model as described by Eqn. (10).
where k is an empirically found scaling factor, xMgF2, bath(t) is a recent estimate of the MgF2 concentration from Eqn. (7) or measurements, xMgF2, bath is the targeted MgF2 concentration, and mdross is the initial feedrate of dross over a tap cycle calculated by Eqn. (9). In one embodiment, the amount of magnesium in the metal pad of an electrolysis cell is controlled to be within a certain maximum and minimum (e.g., from 0.5-1.5 wt. % Mg), such as would be required by, for instance, a 5xxx or 6xxx aluminum alloy. In another embodiment, the amount of magnesium in the metal pad of an electrolysis cell is controlled to be at or below a certain maximum magnesium content (e.g., not greater than 0.4 wt. % Mg), such as would be required with a 1xxx, 3xxx, or 4xxx aluminum alloy. vi. Emissions [0050] There are at least two components in processed dross that may convert to gaseous emissions instead of transferring to the electrolytic bath or metal pad. One such component is aluminum nitride (AlN), which may convert to NOx emissions. Another such component is chlorine (e.g., chloride salts), which may convert to HCl and CClF3. During dross recycling testing, it was found that cathode voltage drops (CVD) did not increase, as shown in FIG. 5, indicating nitrogen escape is predominately if not wholly through gaseous emissions. The measurements made during the months shown in FIG. 5 represent the start of the dross recycling trial, three months into the trial, and five months into the trial.
Attorney Ref. No.: 169593.119972/WO [0051] Based on studies of smelting emissions during dross trials, it can be estimated that NOx emissions may be released in the ratio of 1:1:0.05 for N2O:NO:NO2. Accordingly, estimates of the NOx emissions generated from dross recycling may be produced based on the AlN content and total mass of the dross being recycled. An example of this is shown in Table 5, where the assumption is made that all nitrogen present in the AlN is converted to different NOx species, which is likely an overestimation because a non-zero percentage of the nitrogen present in the AlN should convert to N2 gas, resulting in lower NOx emissions than those shown in Table 5. Actual concentrations should be confirmed through actual measurements if there are any regulatory reporting requirements from the environmental agencies with jurisdiction over the smelter location conducting dross recycling. Table 5: NOx emission estimates based on the AlN content of recycled dross.
[0052] Chlorides present in the dross (e.g., from the use of salt flux) may form gaseous emissions. Table 6 provides example emission estimates. In the absence of an anode effect, HCl may form whereas when an anode effect occurs, CClF3 may form. The annual amount of CClF3 produced can be estimated from Eqn. (11). (11) where mmetal production is the total annual metal production for the smelter in MT/yr, the slope factor is the measured Tier 3 slope factor for estimating CF4 emissions from AE rates in (kg CF4/MT Al)/(AE min/pot/day), and CClF3 ratio factor is the ratio of CClF3 emissions to CF4 emissions that occur during an anode effect (typically 0.03). The chlorine consumed during the production of CClF3 is then subtracted from the initial mass of chlorine introduced to the potrooms through recycling. The remaining mass may be assumed to convert to gaseous HCl emissions. Table 6: Chloride emission estimates based on the chloride content of recycled dross
Attorney Ref. No.: 169593.119972/WO
[0053] It should also be emphasized that these example calculations are estimates, and that gas sampling should be performed to obtain actual values. It may also be useful for each location that deploys D2P to determine any regulatory requirements related to gas emissions. vii. Deployment Methodology [0054] Each smelter may have a unique casthouse product mix, which may result in a unique dross composition. Accordingly, it may be useful to perform a small scale D2P trial to confirm the behavior of impurities and aluminum electrolysis cell performance as well as become comfortable operating with a modified bath chemistry. [0055] For instance, a small group of aluminum electrolysis cells can be selected for an initial trial, where processed dross is added (e.g., through the taphole as bagged additions). Due to the need to regularly bring bags of dross to the pots, it is suggested that the logistics of delivering the dross to the pots be considered when selecting the trial aluminum electrolysis cells. Additionally, it may be worthwhile to select aluminum electrolysis cells within a single tapping group in order to manage the change in metal purity in the crucibles being delivered to the casthouse. Another helpful practice is to clearly mark in the cell line the aluminum electrolysis cells intended to receive the processed dross to avoid/minimize any confusion on where the bagged dross should be delivered. [0056] It may be advantageous to coordinate the additions of processed dross with the tap cycle of the aluminum electrolysis cells because the taphole may already be open and the processed dross may be readily added to the aluminum electrolysis cells through the taphole. The size and volume of the processed dross additions should be based on expected amount of processed dross being recycled at full deployment of dross recycling. This can be calculated by Eqn. (12),
where mdross additions is the mass of bagged dross additions per tap cycle, mdross recycled is the total annual mass of dross that would be recycled at deployment of D2P, total pots is the average number of pots in operation for the entire smelter and days in tap cycle is the number days between metal tappings. It is recommended to start at the rate calculated by Eqn. (12) initially.
Attorney Ref. No.: 169593.119972/WO After D2P is implemented, it will likely require about one month for the MgF2 concentration to reach steady-state. This concentration will likely be lower than the predicted concentration for deployment of D2P. In a small-scale trial, the amount of material removed from the cells during anode change out and also the fresh anode cover material (without processed dross) added to the cells serves to dilute the magnesium being added with manual processed dross additions. After one month, it may be useful to slowly increase up to 3 or 4 times the initial rate calculated by Eqn. (12), which may allow for an MgF2 concentration much closer to the expected full deployment value to be reached. This is important for demonstrating stable pot operation at higher levels of MgF2 concentrations. It will then likely require 2-3 months to reach the next MgF2 concentration. After the new MgF2 steady-state concentration has been reached, it is recommended to continue the trial for an additional 2-3 set cycles. [0057] Routine process data should be collected from the test pots as well as properly selected reference pots. Suggested parameters to track during the trial along with the frequency at which to track them are detailed in Table 7. Table 7: Suggested process data to track during a dross recycling trial.
Table 8 details suggested special manual measurements to make during a trial. Measuring the bath and metal mass will help to verify that no significant changes to the ledge thickness/thermal state of the pots has occurred. Measuring the cathode voltage drop (CVD) provides an additional indication if any deposits are forming on the cathode, along with muck surveys. PODFA (porous disk filtration) samples may provide an indication if there are any issues with metal cleanliness due to dross recycling. Near simultaneous bath and metal sampling may be used for establishing the MgF2 prediction model. The bath and metal samples
Attorney Ref. No.: 169593.119972/WO should be analyzed for their respective Mg concentrations. Additionally, it may be helpful to measure the alumina concentration of the bath samples to confirm feed control is operating as expected. This will serve as an additional check since increasing MgF2 levels in bath may cause a bias alumina readings. Table 8: Suggested special manual measurements to take during a dross recycling trial.
viii. Additional Enablers for Dross Recycling [0058] As discussed previously, due to the presence of fluxing salts in the dross, HCl emissions may arise. In one embodiment, processed dross is washed prior to introduction into an aluminum electrolysis cell to at least partially remove fluxing salts from the processed dross. The chloride salts used as fluxing agents in the casthouse are water soluble and should be readily removed by a water wash. To restrict or avoid premature corrosion of machinery (e.g., potroom cranes), an acid filter may be installed on compressors and other components where water may be present. ix. Conclusion [0059] Dross is a difficult to manage and expensive residue with limited options beyond landfilling. With increasing environmental regulations, it may become more even more difficult landfill dross. The D2P process described herein may eliminate the need to landfill dross since the dross may be simply reintroduced into existing aluminum electrolysis cells. As explained herein, even dross from high magnesium alloys can be successfully recycled. x. Miscellaneous [0060] These and other aspects, advantages, and novel features of this new technology are set forth in part in the description and figures herein and will become apparent to those skilled in the art upon examination of the description and figures herein, or may be learned by practicing one or more embodiments of the technology provided for by the present disclosure.
Attorney Ref. No.: 169593.119972/WO [0061] The present disclosure was explained, in part, with reference to the attached drawings, wherein like structures are referred to by like numerals throughout the several views. The drawings shown are not necessarily to scale, with emphasis instead generally being placed upon illustrating the principles of the present disclosure. Further, some features may be exaggerated to show details of particular components. In addition, any measurements, specifications and the like shown in the figures are intended to be illustrative, and not restrictive. Therefore, specific structural and functional details disclosed herein are not to be interpreted as limiting, but merely as a representative basis for teaching one skilled in the art to variously employ the present disclosure. [0062] Among those benefits and improvements that have been disclosed, other objects and advantages of this invention will become apparent from the description figures herein. Detailed embodiments of the present invention are disclosed herein; however, it is to be understood that the disclosed embodiments are merely illustrative of the invention that may be embodied in various forms. In addition, each of the examples given in connection with the various embodiments of the invention is intended to be illustrative, and not restrictive. Therefore, specific structural and functional details disclosed herein are not to be interpreted as limiting, but merely as a representative basis for teaching one skilled in the art to variously employ the present disclosure. [0063] Throughout the specification and claims, the following terms take the meanings explicitly associated herein, unless the context clearly dictates otherwise. The phrases “in one embodiment” and “in some embodiments” as used herein do not necessarily refer to the same embodiment(s), though they may. Furthermore, the phrases “in another embodiment” and “in some other embodiments” as used herein do not necessarily refer to a different embodiment, although they may. Thus, various embodiments of the invention may be readily combined, without departing from the scope or spirit of the invention. [0064] In addition, as used herein, the term “or” is an inclusive “or” operator, and is equivalent to the term “and/or,” unless the context clearly dictates otherwise. The term “based on” is not exclusive and allows for being based on additional factors not described, unless the context clearly dictates otherwise. In addition, throughout the specification, the meaning of “a,” “an,” and “the” include plural references, unless the context clearly dictates otherwise. The meaning of “in” includes “in” and “on”, unless the context clearly dictates otherwise. [0065] While a number of embodiments of the present invention have been described, it is understood that these embodiments are illustrative only, and not restrictive, and that many modifications may become apparent to those of ordinary skill in the art. However, it is to be
Attorney Ref. No.: 169593.119972/WO expressly understood that such modifications and adaptations are within the spirit and scope of the present disclosure. Further still, unless the context clearly requires otherwise, the various steps may be carried out in any desired order, and any applicable steps may be added and/or eliminated.