4th Press Release – Final REEScue Project 2023

Press Release –REEScue Project

Tuesday, 31 October 2023

Place: Tulcea – Romania


The REEScue project under ERANET-ERAMIN 3- REEScue “Integrated process for the recovery of Rare Earth Elements and Scandium from Bauxite Residueslaunched in 1st of October 2020, under coordination of  National Technical University of Athens (NTUA), in cooperation with its academic partner Necmettin Erbakan Üniversitesi (Turkey) and three industrial partners: Mytilineos SA (Greece), ALUM SA (Romania), ETI Aluminyum (Turkey), was finalised at time.

The aim of the project REEScue is the efficient exploitation of European bauxite residues resulting from alumina production, containing appreciable concentrations of scandium and REEs, through the development of a number of innovative extraction and separation technologies that can efficiently address the drawbacks of the existing solutions. The outcome of the project constitutes a process that allows (i) the recovery of the majority of residual Al produced during the hydrothermal processing of BR in an alkaline solution that can be fed to the Bayer process, and (ii) the recovery of Fe after the implementation of magnetic separation to the hydrothermal treatment residue, allowing recovery of over 90% on >40% Fe grade. Alkaline leaching to the nonmagnetic fraction allows REEs dilution, however their recovery remains challenging. A more detailed analysis of the project conclusions follows.

Three alumina refineries participate in REEScue project: Aluminum of Greece (AoG) (Greece), ETI Seydisehir Aluminyum S.A. (ETI) (Turkey), and ALUM (Alum) (Romania). After characterization of raw materials, small differences were observed, which can be attributed to the different origins of the ores used, as well as to the different processing conditions applied in the Bayer process. More significant deviations were identified for sodium and iron content. Major REEs and Sc scandium were detected in all cases. In the Romanian sample (ALUM), iron was mainly found in the form of goethite, while in the Greek (AoG) and the Turkish (ETI) samples, iron was structured in hematite. Despite the tiny size of BR particles, usually more than one phase coexists on a single grain, thus burdening the purity of samples that can be produced after the implementation of a physical separation procedure.

The hydrothermal treatment of the BR under alkaline conditions with a Fe source allowed the conversion of hematite to magnetite by up to 80%, leading to a considerable increase of the magnetic susceptibility of the material (from 5 emu/g to approx. 40 emu/gr). Hydrothermal treatment takes place at 200-250 °C for 1.5 to 3 h. In addition, the liquor produced after the hydrothermal treatment of BR contained over 5000 mg/L of Al, corresponding to a dilution ratio of 75% of the Al of BR, and it is suitable for reprocessing through Bayer process achieving overall increase of the efficiency of  Al-production process. The product was subjected to magnetic separation using wet magnetic separation technique. In all materials tested, better separation efficiency was achieved for hydrothermally treated samples and the use of ultrasounds, proving the beneficial impact of hematite reduction to the separation efficiency. For each sample of different origin, optimum separation efficiency was achieved under the following conditions:

  • Sample AoG→ Fe grade: 40%, Fe recovery 90% [treated sample, ultrasounds ON, magnetic field intensity 25.3 G]
  • Sample ETI→ Fe grade: 35%, Fe recovery 90% [treated sample, ultrasounds ON, magnetic field intensity 13.2 G]
  • Sample ALUM→ Fe grade: 42%, Fe recovery 88% [treated sample, ultrasounds ON, magnetic field intensity 13.2 G]

The implementation of magnetic susceptibility measurements to raw, hydrothermally treated and magnetic and non-magnetic fractions showed that the hydrothermal treatment increased radically the magnetic susceptibility of the material because of the transformation of hematite to magnetite. However, the difference in measured magnetic susceptibility among magnetic and non-magnetic fractions was not considerable. It appears that, the major limitations of wet magnetic separation are related to the properties of the material:

  • The material fineness is not beneficial to the separation because fine particles are vulnerable to entrainment by hydrodynamic forces
  • As revealed by the mineral liberation analysis and elemental mapping, most particles consist of more than one mineral phase, thus even if all Fe-rich grains are obtained, the coexistence of other minerals rich in Al, Si and Ti is inevitable

Acid leaching agents, namely sulfuric and hydrochloric acid solutions were tested for the leaching of REE from the nonmagnetic fraction of the three BR treated samples. Higher REE dissolution occurred at strong acid solutions. Acid strength at 4M, Temperature 90 oC, reaction time 3h and 25% S/L were the conditions that produced the best leaching REE results. However, the strong acid leaching led to elevated Fe, Al and Ca solutions. The solid to liquid separation was very difficult, since the solid was voluminous and maintained a large part of the leach liquor. Another reason for the poor filtration was the polymerization of the dissolved silica content. A washing step for the solid residue was necessary to obtain a part of the leach solution.

The recovery of REE from such solutions using Ion Exchange (IX) is challenging since Fe (30g/l for HCl leaching and 60-100g/l for H2SO4 leaching) and Al (9-15g/l for HCl leaching and 8-11g/l for H2SO4 leaching) must be eliminated before the IX stage. Chemical precipitation of these metals is difficult due to their high concentration in the leach liquor and will result in increased REE losses. The complexity of the stages prior to the IX stage, with the very dilute REE solutions to be treated appear to make the IX process unattractive.

The direct leaching of BRs sampled from AoG, ETI and ALUM dams have been done with HCl and H2SO4 using different concentration, temperature, process time and additives. The results obtained under atmospheric conditions have been compared with those obtained hydrothermally in autoclaves. The hydrothermal processes with 10M HCl, 1h duration and 25% S/L ratio at 200oC allow to reach the 85% critical metals (CMs) dissolution which have been expected at the beginning of the project. The H2SO4 process does not allow the highly efficient dissolution for each element, but a strong selectivity of Sc have been observed with 8M H2SO4, 2h duration and 25% S/L ratio at 200oC. The Sc leaching efficiency reached 90% in these conditions. The harsh corrosive conditions are the main drawback for the direct leaching. Moreover, the process facilities must support at least 20 bars under 200oC to realize the hydrothermal leachings.

The recovery of REE from PLS via Solvent Extraction (SX) have been investigated. The extraction efficiencies with using separately DEHPA and Aliquat 336 extractants have been compared. DEHPA in kerosene demonstrate the best efficiency for critical metals. The extractant concentration and the pH are the most critical parameters to optimize the SX. Results demonstrate that the metals present in solution can be successively fixed by the extractant according to the following order: Sc > Ti > Fe ~ Y > La ~ Ce ~ Nd ~ Al. In other words, DEHPA will first fix Sc, then Ti and so on. Thus, the magnetic separation of Fe can be an opportunity to reduce the contamination but not to obtain highly pure samples. The realisation of selective SX on the PLS require a permanent control of CMs concentration which will determine the quantity of DEHPA which have to be introduce in the reactor before each SX. Concerning the influence of pH, fine tuning with less than 0.5 pH value at acidic media is required to separate coarsely some of CMs via SX. Even if the pH control is successfully done, the final samples will not be highly pure and purification process after extraction must be investigated in the future. The SX process allow to control roughly the CMs separation but it is related to very sensitive parameters and its application in industrial conditions seems difficult.

Synthesis of geopolymers is based on the activation of aluminosilicate materials by an alkali metal hydroxide and an alkali metal salt and their transformation into a three-dimensional inorganic amorphous structure. In order to study whether the BR residue could potentially be geopolymerized, is was important to investigate it’s Al and Si content and for this reason the alkaline dissolution of the BR residue in NaOH and KOH 10M solutions was investigated. For all the received samples a series of tests was performed. Regarding their Chemical composition, all the BR samples presented very low content in Si and Al species. Especially for the case of Aluminum (a species necessary for the development of the 3-B aluminosilicate network, since it performs the cross-linking between Si species), all the investigated BR samples presented extremely low  content to almost absence of Al (content below 1%). In the case of silicon, the ARETI (Aluminosilicate Residue of ETI Aluminium) sample appears to have a moderate to low Si content, while the ARAoG (Aluminosilicate Residue of Aluminium of Greece) and ARALUM (Aluminosilicate Residue of VITMECO ALUM) samples present inadequate Si content. Regarding their mineralogical content, all three samples present the same characteristics: there is an absence of amorphous aluminosilicate phases, the LoI is increased, while for the crystalline phases, only iron oxides were identified. Therefore, the XRD diagrams indicate the absence of aluminosilicate phases that could easily provide Si and Al species for the geopolymerisation reaction. Regarding their behavior during the leaching tests, for all three cases the amount of Si and Al that were actually dissolved is low. More specifically, for all cases, the dissolved Al species is around 0.5%, while the dissolved Si is at the best case (ARETI sample) close to 10%. Still, the Si and Al content is considered as not sufficient for the realization of the geopolymerisation reaction, especially due to the almost absence of Al. Therefore, all three BR samples (ARAoG, ARALUM, ARETI) are considered as not suitable candidate materials for geopolymerisation.

The Life Cycle Assessment of the new processes was carried out targeting the identification of environmental challenges associated with the processing of Greek, Romanian and Turkish red mud for the extraction of valuable resources like scandium. The study utilized a Functional Unit of 1,000 kg of bauxite residue treatment and focused on key impact categories such as Abiotic Depletion Potential (ADP), Freshwater Aquatic Ecotoxicity Potential (FAETP), Human Toxicity Potential (HTP), and Terrestrial Ecotoxicity Potential (TETP). Regarding the Global Warming Potential (GWP), ALUM case possesses lowers value of 619.91 kg CO2 eq., followed by AoG (GR) and ETI (TR) with 722.22 kgCO2 eq. and 805.21 kg CO2 eq., respectively.  Overall, ALUM emerged as the most environmentally sustainable choice for scandium extraction, showing the lowest environmental impacts across various categories. Scenario 1 – AoG, also, displayed reduced environmental impacts compared to ETI case, positioning it between the other two cases.


The REEScue project was supported by grants of the:

  • General Secretariat for Research and Technology (GSRT) from Greece for National Technical University of Athens NTUA and Mytilineos SA
  • Scientific and Technological Research Council of Turkey (TUBITAK) for Necmettin Erbakan Üniversitesi and ETI Aluminyum A.S.
  • Executive Unit for the Financing of Higher Education, Research, Development and Innovation (UEFISCDI) from Romania for ALUM, by financing contract no 182/2020


Prof. IOANNIS PASPALIARIS – Project Coordinator

Address: School of Mining and Metallurgical Engineering Iroon Polytechniou 9 str., GR15773 Zografou Campus, Athens, Greece

Email: paspali@metal.ntua.gr

Website: https://reescue.com/