Next Article in Journal
Wrapping a Craton: A Review of Neoproterozoic Fold Belts Surrounding the São Francisco Craton, Eastern Brazil
Previous Article in Journal
Field Simulation Technique to Enhance the Mechanical Strength and Elemental Composition of Soft Clay Soil Using Thermal Treatment
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Minerals
Volume 14
Issue 1
10.3390/min14010042
minerals-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Utilizing X-ray Computed Tomography for Lithium Slag: A Guide to Analyzing Microstructure and Its Potential Influence on Liberation

by
Thu Trang Võ
*,
Thomas Leißner
and
Urs A. Peuker
Institute of Process Engineering and Mineral Processing, Technische Universität Bergakademie Freiberg, Agricolastr. 1, 09599 Freiberg, Germany
*
Author to whom correspondence should be addressed.
Minerals 2024, 14(1), 42; https://doi.org/10.3390/min14010042
Submission received: 9 November 2023 / Revised: 18 December 2023 / Accepted: 23 December 2023 / Published: 29 December 2023
Browse Figures
Versions Notes

Abstract

:
Slag containing lithium aluminate is analyzed for its microstructure. This refers to the mineralogical composition, shape and size of the target and matrix phase, orientation of the target phase, and porosity. To investigate the microstructure of the target phase, a representative sample is taken from the block and measured using the XCT. With the help of the two- and three-dimensional analysis, an insight into the complex structure can be gained. The target phase, in this case, lithium aluminate, has a dendritic structure with different orientations and thus also influences the microstructure of the matrix phase. This is composed of a mix of several minerals and amorphous components. Fine pores with a radius of 10–100 µm are found throughout the sample volume. The fracture behavior of the slag is estimated on the basis of the influencing factors that can initiate fracture in brittle materials or divert the path of the fracture. Since the mineralogical and thus also mechanical properties of the slag can be influenced by its production process, suggestions are given as to how slags should be structured in order to ensure a non-random fracture, which is required for the efficient liberation of the target phase in mechanical processing.

1. Introduction

Slag belongs to the secondary raw materials, i.e., it is understood as a material that can be re-used as a raw material after recycling [1,2]. Usually, slag is found as a waste product from pyrometallurgy, where only the main components of the feed are recovered as molten metal while the remaining components are extracted via the slag. In there, the concept of EnAM foresees that low-concentration critical raw materials are combined in a melt during crystallization to form engineered artificial mineral ores (EnAM-ores) that can be liberated by comminution and further processed. Thermodynamic models, additives, and controlled cooling processes can be used to specifically influence the formation of these minerals.
Particle breakage in comminution processes results from a variety of influencing factors. These can stem directly from the microstructure of the material, such as mineral composition and hardness [3,4], porosity [5,6], comminution conditions such as stress intensity and energy input [7,8,9], or external environmental factors such as moisture content [8,10]. Important in the comminution process is the accessibility of the valuable materials for downstream enrichment processes such as flotation or leaching [11]. At the same time, grinding and milling should become more energy efficient; e.g., the theoretical efficiency of ball mills is about 1.5%–12% [7]. In general, about 21% of the total energy in mining is effectively used (USA), and about 99% of that energy loss goes into heat and noise [12].
To ensure the accessibility of the target phase, non-random fractures are desired, which occur and are influenced by the mineralogical composition and microstructure. The first classification of fracture mechanisms of heterogeneous materials is given by Gaudin [13], in which he distinguishes between liberation by detachment and liberation by comminution. The former occurs purely as a function of material properties and texture, while the latter results in random fracture. A comprehensive compilation on the definition of random and non-random fractures is provided by Mariano [14]. King and Schneider defined four other conditions besides liberation by detachment that do not belong to random fractures: selective, differential, phase-boundary, and boundary-region fracture [15].
These fracture mechanisms require certain material and structural properties in the heterogeneous particle. Microstructures can be described in terms of phases, interfaces, and defects [16]. All have structural parameters, such as chemical composition, and geometrical parameters, such as volume fraction, shape, orientation, arrangement, and grain size. Here, the grains will be represented by the target phase, i.e., the phase that contains the minerals and chemical elements, respectively, that need to be recovered. The remaining phases are summarized as non-target phases in the matrix.
This study focuses on the geometric, i.e., morphological parameters of phases and defects by analyzing lithium-containing slag. Due to the increasing demand, especially in electric vehicles and storage systems [17], the technically difficult extraction, and the limited occurrence of this element, lithium is one of the critical raw materials [18]. As a result, there is a need for the recovery of lithium from the pyrometallurgical waste streams.
The pyrometallurgical method of recovering lithium, e.g., from lithium-ion batteries, allows selective enrichment of the valuable materials without pre-sorting of the batteries [17]. Typically, especially to meet the legislative requirements concerning recovery rates, pyrometallurgy is combined with sophisticated mechanical processing [19,20,21,22,23,24], and only the concentrate of the cathode material containing, e.g., Co, Ni, Mn, and Li undergoes thermal pyrometallurgical processing [19]. Different components, such as Co, Ni, Cu, and Mn, can be separated in the metallic melt or matte phase. By using the pyrometallurgical route, the lithium is potentially enriched as a lithium aluminate compound in the slag phase.
The traditional way to study microstructure in the field of geology is to prepare thin or polished sections of rock samples and analyze them microscopically [25,26]. In the work of Popov et al. [27,28], the microstructure is characterized in three dimensions using mathematical models based on three orthogonal thin sections and allows statistical 3D information about the structure of objects. Another well-established method to give a three-dimensional insight into the microstructure is X-ray computed tomography. XCT is widely used as a non-destructive characterization method in the field of geological, biological, or technical samples [29,30,31,32]. Therefore, the investigated sample can still be used for further studies, e.g., exploring the relationship between breakage experiments with regard to their microstructural preconditions in in-situ setups. An extension of XCT is the coupling with microscopy optics, which allows very high-resolution investigations of fine-scale structures like cracks [5,33,34], pores [35,36], or inclusions [37]. XCT cannot entirely replace the value of thin sections since they offer unique insights into mineralogy and structure that XCT may not capture. Thin sections provide valuable information that complements XCT, making them indispensable in certain aspects of comprehensive microstructural characterization.
In this study, the target, matrix phase, and pores are analyzed and visualized to observe their intergrowth and distribution, respectively. In exploring the microstructure prior to any fracturing, the aim is to shed light on a fundamental aspect that influences comminution processes. Beyond the type, intensity, and rate of stress applied during comminution, the fracture characteristics of particles are also linked to their initial microstructure. This study introduces the use of both 2D and 3D methods to achieve a comprehensive and nuanced analysis of a complex microstructure that extends beyond the ideal cases of homogeneously distributed, regularly shaped target grains. Since the properties of the slag, such as grain size, shape, and hardness, can be influenced during slag production, this investigation also aims to identify potential conditions for selective comminution of the slag target phases.

2. Materials and Methods

2.1. Sample and Preparation

For this study, lithium-containing slag was produced at the Institute IME Process Metallurgy and Metal Recycling at RWTH Aachen University, Aachen, Germany. A specific salt composition is selected and brought to a melting temperature. After melt homogenization, the slag is cooled at 50 K/h until reaching ambient temperature. This cooling rate allows the lithium to combine with oxygen and aluminum. Once the slag solidifies, lithium aluminate (γ-LiAlO2) is created as the target phase. X-ray diffraction (XRD) results in Table 1 show four kinds of minerals besides the amorphous residual phase. Gehlenite, glaucochroite, and ß-eucryptite combined with the amorphous phase represent the matrix phase.
Despite the slow cooling rate, the mineral formation process cannot be compared to the natural rock formation process, where the cooling rate is much slower and takes place under higher pressure and temperature. The quality of the slag material therefore varies greatly locally, where a temperature gradient can be observed [38]: the outside of the slag melt cools faster than its center. As a result, the outside has large pores and a very fine dendritic slag structure due to its direct contact with the cooling atmosphere. In general, the dendritic structure is indicative of rapid crystallization since the crystals themselves cannot be fully formed before the melt solidifies. Consequently, branching and ramification occur in the crystal structure.
Since the center of the slag block needs longer cooling times, the fineness of the structure decreases as the solidification of the melt itself is slower, allowing more dissolved ions to combine to form lithium aluminate. Sampling from the center is desired since the properties do not vary as much as the outside of the slag. The structure there is less fine and less porous and is therefore more suitable as test material. Consequently, only the inner slag material is considered in this investigation. The choice of the sampled material offers two significant advantages. Firstly, the cooling strategy is not yet optimized; the cooling rate is still too rapid for this slag system, resulting in a dendritic structure throughout the slag body that represents its microstructure. Secondly, the entire slag body cannot be represented as a whole, but complex structures in slag cannot be avoided due to the challenging conditions for the optimal formation of the target phase. This sample serves as an illustration of an exceptionally complex microstructure that needs to be analyzed.
Figure 1A shows a large broken piece of slag from the center of the slag body. For the microstructural analysis, one sample is prepared in a cubic shape so that the properties are not affected by the different particle sizes. The large piece of slag was sliced using a rock saw (tile cutter). This slice is then cut into a cubic shape using a water jet cutting machine (Figure 1B). It is then chiseled out of the slice and manually shaped to an edge length of 4 mm.

2.2. X-ray Micro Computed Tomography (XCT)

For microstructural analysis, the slag particle was measured using XCT (Zeiss Xradia VERSA 510, Oberkochen, Germany). Since only projection images depending on the rotation angle of the sample can be captured, these images need to be reconstructed. We used the Zeiss XMReconstructor system software (version 11.1.8043). It uses standard filtered back projection coupled with additional beam hardening correction and Gaussian smoothing. More details on the measurement and reconstruction parameters are shown in Table 2.
At 4X optical magnification, a three-dimensional representation of the structure is possible at a voxel size of 5.885 µm. Since the voxel resolution is system-related information that depends on the given mechanics and optics of the XCT and is not precise enough compared to the spatial resolution, objects smaller than 8 voxels are not considered in the analysis.

2.3. Image Processing and Analysis

The open-source software Fiji (ImageJ version 1.53b) was used for image processing [39]. It contains several plugins that are not only suitable for image manipulation but also for microstructure analysis. The raw image stack is first denoised with non-local means denoising. The resulting images are then sharpened at the edges with an unsharp mask, so that the contrast between the phases is improved. Global thresholding is chosen as the segmentation technique to select the target phase [40]. Lines that remain are also removed by removing outliers. With these steps, on the one hand, lone voxels that have arisen due to the natural image noise are avoided. On the other hand, the structures do not stay softened by sharpening edges with the Unsharp mask. Especially with crystalline minerals that can form clear edges and geometrical shapes, sufficient contrast must be ensured during image processing so that the 3D image reflects the real structure. To analyze the target phase, it is segmented from the raw image by using thresholding. Likewise, to examine the matrix phase, the target phase is subtracted from the overall image and turned into a binary stack.
Only the processed images are analyzed microstructurally. For the determination of the thickness of the crystals, the local thickness can be used in 2D and is based on inscribed circles that are generated locally in the target phase. The orientation of the crystals can be determined using the directionality plugin and the output of the histogram and calibration bar. This provides an overview of the local distribution of the target phase alignment and is dependent on the current image orientation or coordinate system.
In 3D, the thickness can be investigated with VGSTUDIO MAX 3.3 (Volume Graphics GmbH in Heidelberg, Germany) by applying the inscribed sphere technique to the structural volume. The latter also offers pore size analysis via the porosity/inclusion analysis module.
Each plugin and module require parameter settings, which are listed in Table 3.

3. Results

3.1. Structural Parameters of the Target Phase from 2D-Image Analysis

By means of the XCT measurement of a single slag particle, the inner structure of the target phase can be observed as the gray phase, while air is black, and the matrix is light gray to white (Figure 2). The microstructure of the target phase shows dendritic mineral shapes [46]. A dendrite typically consists of a central trunk from which smaller branches extend and has symmetries in developing branches. Those smaller branches can be arranged in a regular pattern around the central trunk. This can be seen in Figure 2, which shows the image processing steps before the microstructure analysis was conducted. Then, the target phase is extracted as a region of interest for the microstructural analysis by applying skeletonizing. In the binary image, the objects are gradually diluted so that only the central line represents the objects in the image. The line has a width of one pixel while maintaining the connectivity and shape of the object.
By means of a binarized, skeletonized image, in which the target phase is now labeled white, it is possible to determine the branch thickness and the maximum branch length in 2D. In Figure 3A, the maximum branch length is plotted against the number of branches by analyzing its skeleton representation (Figure 3B). The latter represents the number of branches connected to the longest branch in each case. For this slag material, the number of dendrite branches does not increase with the maximum branch length. A lot of dendrites are only connected to one branch. Since the trunk is very thin and shorter than its branches, the maximum branch length informs about the length of the dendrites connected to the trunk.
In Figure 3C, the thickness and the maximum branch length are compared to each other. The thickness is determined by the inscribed circle method (Figure 3D). It involves fitting the largest possible circle within the desired space, with the circle’s diameter providing the measure of the investigated space. As expected, the thickness of dendritic structures is much smaller than the maximum branch length and differs by a factor of up to 110. Therefore, this slag system consists of fine dendrites, which are intergrown by a small number of branches.
The branches of a dendrite have a preferred orientation based on where the trunk lies. In this slag system, the branches are nearly perpendicular to the trunk. The orientation of the branches is shown in Figure 4. Those values are based on the orientation of the images from the XCT measurement. There, most of the branches have an inclination of 44.5° marked in yellow-green. Their trunk is very thin and marked violet, around 90° to its branches. Branches with the same orientation are found in certain dedicated regions. This indicates that those branches originate from the same trunk and stop growing when they are near other branches from a different trunk. Usually, the strongest material flow is found at the corners and edges of a polyhedral crystal. Skeletal growth is only observed when the material flow is faster than the building blocks or ions arrive at their energetically favorable attachment position. The branches of the dendrites grow in the directions where the fastest crystallization is possible, favoring certain planes. The spaces in between are subsequently filled. This circumstance forces other branches to grow in directions where the slag is still liquified and in which the corresponding free ions still exist to form the target phase.
Since the large 4 mm sample led to a limited resolution of 5.8 µm, the target phase appears as intergrown layers. By means of a thin section shown in Figure 5, it is possible to observe more details with higher magnifications. Not all branches are layer-shaped; most branches consist of fine cubic- or cuboid-shaped crystal entities over their length (Figure 5B). Crystals are also connected at the edges and show cross-shaped arrangements that are strung together in a row (Figure 5C). In between those spaces, the matrix phase can be seen. This phase itself is very complex and can vary as a mixture of many very small target phase minerals with non-target phases.

3.2. Structural Parameters of the Matrix Phase from 3D-Image Analysis

Prior to matrix phase analysis, the total volume of slag particles is segmented by thresholding to display only the solid components. The target phase data set (image stack no. 5 of Figure 2) is then subtracted from this total solid volume. This leaves the binary representation of the matrix phase. The wall thickness analysis is generated for both the target phase and the matrix phase. Using the inscribed sphere approach, the largest sphere is inserted into the phase so that its diameter indicates the thickness of the phase. Figure 6 compares the results for the two phases and illustrates the thickness distribution in the corresponding volume.
Due to the crystallization mechanism, the connectivity of the dendrites is low in comparison to their length; there is a layer of non-target material, i.e., matrix, in between the different crystals of the target phase. When compared in 3D, the modal thickness of the crystal, deriving from the inscribed sphere method, of the matrix is the same as the target phase. By visualizing a section of the slag particle, it shows an alternation of the two phases as layers of the same size. Although the matrix phase seems to have more regions of high thickness values, shown in yellow to green, these are mainly concentrated at the ends of differently oriented dendrites. The target phase itself is not needle-shaped but consists of uneven layers interconnected throughout the slag particle, as seen in the 3D representation. Therefore, the matrix is a phase that is not entirely split by the target phase layers, and therefore it is completely connected as a continuous phase.
It cannot be entirely ruled out that parts of the target phase become interconnected within the matrix. As for the small entities in the dendrites, those below the spatial resolution can be assigned to the matrix phase. In general, the structural description of the matrix is complex compared to the target phase since it changes in color, clarity, and crystallinity within the slag particle, as seen in the thin section (Figure 5), but not in the 3D data set.
The thickness of the target phase indicates how small the fragment size must be to obtain a particle consisting only of the target phase. This critical length is compared to the matrix to evaluate how strongly the two phases are intergrown. Due to the similar Mohs hardness and the regular alternation of matrix and target phases, with similar small thickness and strong intergrowth throughout the examined volume, the slag can be considered pseudo homogeneous. This complicates the selective liberation of the target phase.

3.3. Structural Parameters of Pores Recovered from 3D-Image Analysis

Pores are considered defects within the slag particle, which can potentially influence fracture propagation. Porosity cannot be prevented in the slag production process since air can be incorporated during the mixing of the feed in the homogenizing molten state. Due to changes in gas solubility, the crystallization of minerals can also lead to gas generation, which is trapped in the highly viscous liquid slag state. Another reason that pores stay in the slag body is the cooling process. The exterior of the slag body is in contact with the cooling atmosphere and solidifies first. Gas bubbles in these regions tend to be larger, while gas bubbles in the liquid center are fully or partly filled with minerals and appear smaller. In addition, microcracks, or air voids, can occur during cooling due to the shrinkage of the solids. The slag properties and presence of silica alone can promote chemical shrinkage [47].
In Figure 7, the pores are shown within the slag body, especially slag fragments that are larger than 5 mm. In Figure 7A, meso-scale pores can be seen, and in Figure 7B, pores are outlined in white in the depth-of-field image as an example. As minerals form, the shape of pores can change depending on the orientation of the minerals. Since the target phase shows dendritic structures, pores have elongated shapes and are oriented in the same direction as the dendrites. Thus, the geometric parameters of the solid phases influence those of the pores.
To characterize the micro-scale pores within the slag particle, the solids are selected as regions of interest and inverted so that only the gas phases are considered for the pore space analysis. A certain amount of non-spherical-shaped pores, with most of them having a sphericity between 0.53 and 0.62, is shown in Figure 8. Most of the pore radii amount to about 10–30 µm and overlap with the finer thickness range of the dendrites. Figure 8 also shows a 3D visualization of pores within the slag particle, where it can be seen that the pores are distributed randomly. In general, large and fine pores can be found in every region. The size and distribution of the pores are related to crystal growth, the concentration of ions in the melt to grow crystals, and external conditions such as temperature and pressure. Fast-growing crystals may not have enough time to reach their final arrangement, resulting in defects and voids within the crystal structure. These voids act as pores and are randomly distributed in the slag product. Slow-growing crystals or those with a low ion concentration will have sufficient time to arrange properly, which can result in fewer and smaller pores.

4. Discussion

With the help of the slag sample from the center of the whole block, a first overview of their typical internal meso- and micro-structure can be given. Their dendritic nature indicates a too-fast cooling rate to generate defined, compact crystals, which are favorable for mineral processing. With EnAM in particular, it is important to produce slag whose valuable minerals can be broken down and enriched as easily and selectively as possible. Successful liberation is characterized by access to the valuable minerals after, e.g., comminution, which facilitates the subsequent enrichment processes. The aim is to design the production process of the slag in such a way that it can be processed as well as possible so that the valuable minerals, i.e., target phases, can be recovered. No target phases could be found, which are sporadically distributed in the volume and whose shape and size can be described individually. Strongly branched target phases provide a higher surface area than non-dendritic phases. Typically, this may be advantageous if it results in more grain boundaries in the material that potentially act as weak points where cracks can form and grow more easily [15].
The problem for successful liberation is the strong intergrowth of the target phase with the matrix phase. Not only the dendrites themselves are strongly interlocked in the matrix, but also their smaller, cross-shaped entities. The alternation between amorphous and crystalline phases within the matrix complicates the general determination of the degree of intergrowth. For one, crystalline entities must be viewed at multiple scales. The interconnectedness of different minerals within the matrix should be identified through the boundaries of the phases [48]. However, this changes with the transition to the amorphous phase, in which there are no longer any interfaces to be seen.
The small thickness of both phases, in combination with the presence of the matrix phase between the fine branches of the target phase, can cause the overall material to behave homogeneously. This pseudo-homogeneity is supported by the similar hardness of the minerals present, ranging from 6 to 7 on the Mohs scale.
In order to prevent pseudo-homogeneity and influence fracture propagation, it is more favorable if the target phase has a different strength compared to the matrix phase, i.e., either the target or matrix phase has a higher strength [49,50]. To increase the strength, there are two approaches. One is to reduce the defects or voids in the material. This increases the critical stress needed to initiate a crack. The other way is to increase the fracture toughness. In brittle materials, particles or fibers are used to reinforce a specific phase and are commonly found in the manufacture of ceramics [51].
The reduction in defect size can be mainly influenced by the slag production process. In this case, the defect can be represented by pores or microcracks. Pores can never be completely eliminated during slag production, but pore sizes can be reduced by optimizing the cooling strategy. The pores can be filled with further minerals and densify the matrix with a lower cooling rate and, if necessary, holding times. However, this is also related to the composition of the slag feed and its thermodynamic state during cooling, since temperature and pressure determine the type and further growth of specific compounds formed in the slag [51]. Large and very small pores exist in the slag used in this study. Large pores are more likely to be found with respect to the total slag body, while after crushing the body, only fine pores between 10 and 100 µm will be found with particles around 3–4 mm. Due to the high number of fine pores found in this sample, the slag will show locally a lower Young’s modulus and strength [51]. At these pores, the load stress is concentrated and leads to the failure of the material by forming a microcrack between different pores [52]. The higher the pore volume there is in the material, the less fracture energy will be required to obtain a simple or complex mix of target and matrix phase grains [7]. The success of the liberation is then mainly related to the pore size distribution, shape distribution, and their location in the material instead of its mineralogy.
Microcracks are usually already present in the slag due to the production process. With the high number of pores, whose orientation depends on the dendrites, a clear distinction between pore and microcrack is difficult. However, particularly elongated pores can serve as microcracks and can have a positive effect on liberation. In relation to this, microcrack distribution around or within the target phase is desired to initiate boundary-region or preferential breakage [15]. This can be achieved by varying the thermal expansion coefficients of the matrix and target phase. During cooling, strong residual stresses are generated and lead to microcracks in the microstructure. Another way is the formation of microcracks due to shrinkage. These can occur along the grain boundaries, especially if the region around the target phase has fine pores [6,47]. Due to the presence of microcracks or pores around the target phase, the Young’s modulus decreases locally. Incoming cracks then preferentially run in the direction of the target phase [5], which leads to non-random breakage and better access to the target phase in the crushed product.
In the case of a preferential fracture, microcracks are required homogeneously over the entire volume of the target phase. Thus, the Young’s modulus is reduced not only locally but globally [51]. This is not the case for the investigated slag. Due to the uniform distribution of pores throughout the sample volume and the regular arrangement of all dendrites within the free matrix area, there is no correlation as to whether pores are particularly pronounced in one phase. This is further complicated by the multiphase nature of the matrix phase, which consists of various minerals and an amorphous phase. Taking into account the similar hardness, high number of finely distributed pores, strong intergrowth of the matrix and the target phase in the meso- and micro-scales, and the complex composition of the matrix, this investigated slag will exhibit non-selective (random) fracture behavior. Thus, it will be challenging for further processing in terms of enrichment of the target phase due to the small liberation size resulting from random fracture and the finely intergrown nature of the target phase. Optimization of the slag generation (composition and cooling) is needed in order to achieve a coarse-grained product with a simple granular shape. Changing the cooling strategy also changes the resulting microstructure of the slag, which in turn can show a different fragmentation behavior. The interpretation of the fracture behavior must be extended to other microstructural aspects, such as inclusions/impurities, the orientation depending on the stressing conditions, and thermally stressed or already cracked phases due to cooling. These properties influence the local strength distribution of the slag.
With the production of substantial quantities of slag and a tailored cooling strategy to enhance EnAM crystal growth. Considering that the mass-related comminution work for various technical processes generally decreases exponentially with the increasing mean value of the product particle size [7], large EnAM grains should be more energy efficient to expose than small, fine ones and improve the degree of liberation. This approach not only contributes to the optimization of comminution processes but also to further exploration in the field of resource recovery.

5. Conclusions

A new workflow for the 2D and 3D characterization of slag was presented. In this case, a lithium-containing slag sample is examined for its microstructure. The XCT analysis results were supported and presented with thin sections and depth of field images. Based on the single slag particle, the following conclusions can be drawn with a view to facilitating the recovery of the lithium-containing target phase via comminution processes:
  • The dendritic appearance of the target phase within the slag matrix indicates a too-fast cooling rate. With regard to the overall aim of separating defined crystals from the slag matrix, the cooling process must be further optimized to form isolated, compactly shaped target phase grains desirable for comminution.
  • The matrix is composed of multiple minerals and changes crystallinity. This makes a future correlation of its microstructure to its fracture behavior much more difficult since a clear and sufficient phase- and interface-based description is not achievable in a simple and time-efficient manner without additional resources.
  • The matrix phase and the target phase cannot be considered independently. The strong intergrowth on several scales, coupled with the similar hardness of the minerals, can create a pseudo-homogeneity of the slag material and favor the random fracture behavior.
  • Pores and the target phase cannot be considered independently. Since they can be found in the whole sample, independent of the local existence of target and matrix phases, this can amplify the tendency to random fracture behavior. In the future, therefore, attention should also be paid to tailored pore formation through optimized slag generation and cooling processes to make the comminution of this slag type more selective and energy efficient.

Author Contributions

T.T.V.: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Visualization, Writing: original draft. T.L.: Supervision, Writing—review, and editing. U.A.P.: Supervision, Writing—review and editing, Funding acquisition, Project administration, Resources. All authors have read and agreed to the published version of the manuscript.

Funding

The research presented herein was supported by grants from the German Research Foundation (DFG, Deutsche Forschungsgemeinschaft, grant number 470551727) and is part of the priority program 2315, “Engineered Artificial Minerals (EnAM): A Geo-Metallurgical Tool to recycle critical elements from waste streams”. The responsibility for the contents of this publication lies with the authors.

Data Availability Statement

The original and processed data presented in this study, as well as the full XCT settings, are openly available in the OpARA online repository of TU Dresden and TU Bergakademie Freiberg at http://dx.doi.org/10.25532/OPARA-24 (accessed on 11 September 2023).

Acknowledgments

During the preparation of this work, the authors used ChatGPT 3.5 and DeepL in order to enhance the readability of the text. After using these tools, the authors reviewed and edited the content as needed and took full responsibility for the content of this publication. The authors would like to thank all colleagues in the department of Mechanical Process Engineering and Mineral Processing, TU Bergakademie Freiberg for all their support, especially R. Ditscherlein for proofreading. Also, for sample preparation, the authors would like to thank T. Hutsch from Fraunhofer Institut für Fertigungstechnik und angewandte Materialforschung IFAM in Dresden.

Conflicts of Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  1. Motz, H.; Geiseler, J. Products of steel slags an opportunity to save natural resources. In Waste Management Series: Waste Materials in Construction Wascon 2000; Woolley, G.R., Goumans, J., Wainwright, P.J., Eds.; Elsevier Science Ltd.: Oxford, UK, 2000; pp. 207–220. ISBN 1478-7482. [Google Scholar]
  2. Joulazadeh, M.H.; Joulazadeh, F. Slag; Value Added Steel Industry Byproducts. Arch. Metall. Mater. 2010, 55, 1137–1145. [Google Scholar] [CrossRef]
  3. Ren, W.; Yang, Z.; Sharma, R.; Zhang, C.; Withers, P.J. Two-dimensional X-ray CT image based meso-scale fracture modelling of concrete. Eng. Fract. Mech. 2015, 133, 24–39. [Google Scholar] [CrossRef]
  4. Huang, Y.; Yang, Z.; Ren, W.; Liu, G.; Zhang, C. 3D meso-scale fracture modelling and validation of concrete based on in-situ X-ray Computed Tomography images using damage plasticity model. Int. J. Solids Struct. 2015, 67–68, 340–352. [Google Scholar] [CrossRef]
  5. Zabler, S.; Rack, A.; Manke, I.; Thermann, K.; Tiedemann, J.; Harthill, N.; Riesemeier, H. High-resolution tomography of cracks, voids and micro-structure in greywacke and limestone. J. Struct. Geol. 2008, 30, 876–887. [Google Scholar] [CrossRef]
  6. Renard, F.; Bernard, D.; Desrues, J.; Ougier-Simonin, A. 3D imaging of fracture propagation using synchrotron X-ray microtomography. Earth Planet. Sci. Lett. 2009, 286, 285–291. [Google Scholar] [CrossRef]
  7. Schubert, H. Aufbereitung Fester Mineralischer Rohstoffe, 1975th ed.; VEB Deutscher Verlag für Grundstoffindustrie: Leipzig, Germany, 1977; ISBN 3-342-00289-1. [Google Scholar]
  8. Yashima, S.; Kanda, Y.; Sano, S. Relationships between particle size and fracture energy or impact velocity required to fracture as estimated from single particle crushing. Powder Technol. 1987, 51, 277–282. [Google Scholar] [CrossRef]
  9. Rumpf, H. Grundlegende physikalische Probleme bei der Zerkleinerung. Chem. Ing. Tech. 1962, 34, 731–741. [Google Scholar] [CrossRef]
  10. Tavares, L.M. Chapter 1 Breakage of Single Particles: Quasi-Static. In Handbook of Powder Technology: Particle Breakage; Salman, A.D., Ghadiri, M., Hounslow, M.J., Eds.; Elsevier Science B.V: Amsterdam, The Netherlands, 2007; pp. 3–68. ISBN 0167-3785. [Google Scholar]
  11. Kaya, M. Recovery of metals and nonmetals from electronic waste by physical and chemical recycling processes. Waste Manag. 2016, 57, 64–90. [Google Scholar] [CrossRef]
  12. Spooren, J.; Binnemans, K.; Björkmalm, J.; Breemersch, K.; Dams, Y.; Folens, K.; González-Moya, M.; Horckmans, L.; Komnitsas, K.; Kurylak, W.; et al. Near-zero-waste processing of low-grade, complex primary ores and secondary raw materials in Europe: Technology development trends. Resour. Conserv. Recycl. 2020, 160, 1–18. [Google Scholar] [CrossRef]
  13. Gaudin, A.M. Principles of Mineral Dressing. J. Phys. Chem. 1940, 44, 532–533. [Google Scholar] [CrossRef]
  14. Mariano, R.A.; Evans, C.L.; Manlapig, E. Definition of random and non-random breakage in mineral liberation—A review. Miner. Eng. 2016, 94, 51–60. [Google Scholar] [CrossRef]
  15. King, R.P.; Schneider, C.L. Mineral liberation and the batch communition equation. Miner. Eng. 1998, 11, 1143–1160. [Google Scholar] [CrossRef]
  16. Macherauch, E.; Zoch, H.-W. Quantitative Gefügeanalyse. In Praktikum in Werkstoffkunde: 91 Ausführliche Versuche aus Wichtigen Gebieten der Werkstofftechnik; Macherauch, E., Zoch, H.-W., Eds.; Vieweg+Teubner: Wiesbaden, Germany, 2011; pp. 125–134. ISBN 978-3-8348-9884-5. [Google Scholar]
  17. Dobó, Z.; Dinh, T.; Kulcsár, T. A review on recycling of spent lithium-ion batteries. Energy Rep. 2023, 9, 6362–6395. [Google Scholar] [CrossRef]
  18. European Commission; Directorate-General for Internal Market, Industry, Entrepreneurship and SMEs; Grohol, M.; Veeh, C. Study on the Critical Raw Materials for the EU 2023: Final Report; Publications Office of the European Union: Luxembourg, 2023. [Google Scholar]
  19. Werner, D.; Peuker, U.A.; Mütze, T. Recycling Chain for Spent Lithium-Ion Batteries. Metals 2020, 10, 316. [Google Scholar] [CrossRef]
  20. Wuschke, L.; Jäckel, H.-G.; Leißner, T.; Peuker, U.A. Crushing of large Li-ion battery cells. Waste Manag. 2019, 85, 317–326. [Google Scholar] [CrossRef] [PubMed]
  21. Kaas, A.; Wilke, C.; Vanderbruggen, A.; Peuker, U.A. Influence of different discharge levels on the mechanical recycling efficiency of lithium-ion batteries. Waste Manag. 2023, 172, 1–10. [Google Scholar] [CrossRef] [PubMed]
  22. Werner, D.M.; Mütze, T.; Kaas, A.; Peuker, U.A. Chapter 22—Mechanical and physical processes of battery recycling. In Nano Technology for Battery Recycling, Remanufacturing, and Reusing: Micro and Nano Technologies; Farhad, S., Gupta, R.K., Yasin, G., Nguyen, T.A., Eds.; Elsevier: Amsterdam, The Netherlands, 2022; pp. 455–486. ISBN 978-0-323-91134-4. [Google Scholar]
  23. Abdollahifar, M.; Doose, S.; Cavers, H.; Kwade, A. Graphite Recycling from End-of-Life Lithium-Ion Batteries: Processes and Applications. Adv. Mater. Technol. 2023, 8, 2200368. [Google Scholar] [CrossRef]
  24. Diekmann, J.; Hanisch, C.; Froböse, L.; Schälicke, G.; Loellhoeffel, T.; Fölster, A.-S.; Kwade, A. Ecological Recycling of Lithium-Ion Batteries from Electric Vehicles with Focus on Mechanical Processes. J. Electrochem. Soc. 2017, 164, A6184. [Google Scholar] [CrossRef]
  25. Willard, R.J.; McWilliams, J.R. Microstructural techniques in the study of physical properties of rock. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 1969, 6, 1–12. [Google Scholar] [CrossRef]
  26. Richard, J.; Gratier, J.-P.; Doan, M.-L.; Boullier, A.-M.; Renard, F. Time and space evolution of an active creeping zone: Competition between brittle and ductile deformations, new insights from microstructure studies of SAFOD (San Andreas Fault Observatory at Depth) samples. J. Geophys. Res. Solid Earth 2013, 119, 8132–8153. [Google Scholar] [CrossRef]
  27. Talovina, I.; Aleksandrova, T.; Popov, O.; Lieberwirth, H. Comparative analysis of rocks structural-textural characteristics studies by computer X-ray microtomography and quantitative microstructural analysis methods. Obogashchenie Rud 2017, 56–62. [Google Scholar] [CrossRef]
  28. Popov, O.; Talovina, I.; Lieberwirth, H.; Duriagina, A. Quantitative Microstructural Analysis and X-ray Computed Tomography of Ores and Rocks—Comparison of Results. Minerals 2020, 10, 129. [Google Scholar] [CrossRef]
  29. Lin, C.L.; Miller, J.D. Cone beam X-ray microtomography—A new facility for three-dimensional analysis of multiphase materials. Min. Metall. Explor. 2002, 19, 65–71. [Google Scholar] [CrossRef]
  30. Cnudde, V.; Boone, M.N. High-resolution X-ray computed tomography in geosciences: A review of the current technology and applications. Earth-Sci. Rev. 2013, 123, 1–17. [Google Scholar] [CrossRef]
  31. Ueda, T.; Oki, T.; Koyanaka, S. Experimental analysis of mineral liberation and stereological bias based on X-ray computed tomography and artificial binary particles. Adv. Powder Technol. 2018, 29, 462–470. [Google Scholar] [CrossRef]
  32. Evans, C.L.; Wightman, E.M.; Yuan, X. Quantifying mineral grain size distributions for process modelling using X-ray micro-tomography. Miner. Eng. 2015, 82, 78–83. [Google Scholar] [CrossRef]
  33. Garcia, D.; Lin, C.L.; Miller, J.D. Quantitative analysis of grain boundary fracture in the breakage of single multiphase particles using X-ray microtomography procedures. Miner. Eng. 2009, 22, 236–243. [Google Scholar] [CrossRef]
  34. Withers, P.J.; Bennett, J.; Hung, Y.-C.; Preuss, M. Crack opening displacements during fatigue crack growth in Ti–SiC fibre metal matrix composites by X-ray tomography. Mater. Sci. Technol. 2006, 22, 1052–1058. [Google Scholar] [CrossRef]
  35. Löwer, E.; Makowlew, C.; Leißner, T.; Peuker, U.A. Wetting behavior of porous structures: Three-dimensional determination of the contact angle after filter cake dewatering using X-ray microscopy. Adv. Water Resour. 2021, 151, 103894. [Google Scholar] [CrossRef]
  36. Ramandi, H.L.; Mostaghimi, P.; Armstrong, R.T.; Saadatfar, M.; Pinczewski, W.V. Porosity and permeability characterization of coal: A micro-computed tomography study. Int. J. Coal Geol. 2016, 154–155, 57–68. [Google Scholar] [CrossRef]
  37. Babout, L.; Maire, E.; Fougères, R. Damage initiation in model metallic materials: X-ray tomography and modelling. Acta Mater. 2004, 52, 2475–2487. [Google Scholar] [CrossRef]
  38. Buchmann, M.; Borowski, N.; Leißner, T.; Heinig, T.; Reuter, M.A.; Friedrich, B.; Peuker, U.A. Evaluation of Recyclability of a WEEE Slag by Means of Integrative X-Ray Computer Tomography and SEM-Based Image Analysis. Minerals 2020, 10, 309. [Google Scholar] [CrossRef]
  39. Schneider, C.A.; Rasband, W.S.; Eliceiri, K.W. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 2012, 9, 671–675. [Google Scholar] [CrossRef] [PubMed]
  40. Sezgin, M.; Sankur, B. Survey over image thresholding techniques and quantitative performance evaluation. J. Electron. Imaging 2004, 13, 146–168. [Google Scholar] [CrossRef]
  41. Buades, A.; Coll, B.; Morel, J.-M. Non-Local Means Denoising. Image Process. On Line 2011, 1, 208–212. [Google Scholar] [CrossRef]
  42. Darbon, J.; Cunha, A.; Chan, T.; Osher, S.; Jensen, G. Fast nonlocal filtering applied to electron cryomicroscopy. In Proceedings of the 2008 5th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, Paris, France, 14–17 May 2008; pp. 1331–1334. [Google Scholar] [CrossRef]
  43. Lee, T.-C.; Kashyap, R.L.; Chu, C.-N. Building skeleton models via 3-D medial surface/axis thinning algorithms. CVGIP Graph. Models Image Process. 1994, 56, 462–478. [Google Scholar] [CrossRef]
  44. Hildebrand, T.; Rüegsegger, P. A New Method for the Model-Independent Assessment of Thickness in Three-Dimensional Images. J. Microsc. 1996, 185, 67–75. [Google Scholar] [CrossRef]
  45. Woolley, A.J.; Desai, H.A.; Steckbeck, M.A.; Patel, N.K.; Otto, K.J. In situ characterization of the brain-microdevice interface using device-capture histology. J. Neurosci. Methods 2011, 201, 67–77. [Google Scholar] [CrossRef]
  46. Chopard, B.; Herrmann, H.J.; Vicsek, T. Structure and growth mechanism of mineral dendrites. Nature 1991, 353, 409–412. [Google Scholar] [CrossRef]
  47. Mac, M.J.; Yio, M.; Wong, H.S.; Buenfeld, N.R. Analysis of autogenous shrinkage-induced microcracks in concrete from 3D images. Cem. Concr. Res. 2021, 144, 106416. [Google Scholar] [CrossRef]
  48. Exner, H.E.; Hougardy, H.P. Einführung in die Quantitative Gefügeanalyse; DGM Informationsges: Frankfurt, Germany, 1986; ISBN 9783883551081. [Google Scholar]
  49. Shojaei, A.K.; Voyiadjis, G.Z. Dynamic fracture mechanics in rocks with application to drilling and perforation. In Porous Rock Fracture Mechanics; Shojaei, A.K., Shao, J., Eds.; Woodhead Publishing: Cambridge, UK, 2017; pp. 233–256. ISBN 978-0-08-100781-5. [Google Scholar]
  50. Bradt, R.C.; Lin, C.L.; Miller, J.D.; Chi, G. Interfacial fracture of multiphase particles and its influence on liberation phenomena. Miner. Eng. 1995, 8, 359–366. [Google Scholar] [CrossRef]
  51. Rösler, J.; Harders, H.; Baeker, M. Mechanisches Verhalten der Werkstoffe, 2nd ed.; Teubner Verlag/ GWV Fachverlage GmbH: Wiesbaden, Germany, 2006; ISBN 978-3-8348-1818-8. [Google Scholar]
  52. Liu, D.; Šavija, B.; Smith, G.E.; Flewitt, P.E.J.; Lowe, T.; Schlangen, E. Towards understanding the influence of porosity on mechanical and fracture behaviour of quasi-brittle materials: Experiments and modelling. Int. J. Fract. 2017, 205, 57–72. [Google Scholar] [CrossRef] [PubMed]
Minerals 14 00042 g001
Figure 1. Lithium aluminate slag: (A) large slag piece is crushed out of the center of the slag body. The lithium-aluminate phase is finely distributed over the whole slag piece. (B) Waterjet cutting of the sliced slag into a cubic-shaped sample.
Figure 1. Lithium aluminate slag: (A) large slag piece is crushed out of the center of the slag body. The lithium-aluminate phase is finely distributed over the whole slag piece. (B) Waterjet cutting of the sliced slag into a cubic-shaped sample.
Minerals 14 00042 g001
Minerals 14 00042 g002
Figure 2. Image processing steps for a lithium aluminate slag particle are exemplary in one image of the XCT image stack: (A) original, raw image; (B) non-local means denoising; (C) edge sharpening; (D) thresholding; (E) removing outliers; (F) skeletonizing.
Figure 2. Image processing steps for a lithium aluminate slag particle are exemplary in one image of the XCT image stack: (A) original, raw image; (B) non-local means denoising; (C) edge sharpening; (D) thresholding; (E) removing outliers; (F) skeletonizing.
Minerals 14 00042 g002
Minerals 14 00042 g003
Figure 3. Dendritic structure analysis. (A) maximum branch length in relation to the number of branches connected to the longest branch; (B) analyzing the skeletonizing result; (C) thickness determined by (D) inscribed circles and max. branch length as a number distribution.
Figure 3. Dendritic structure analysis. (A) maximum branch length in relation to the number of branches connected to the longest branch; (B) analyzing the skeletonizing result; (C) thickness determined by (D) inscribed circles and max. branch length as a number distribution.
Minerals 14 00042 g003
Minerals 14 00042 g004
Figure 4. The orientation of the dendrites is determined by means of a processed and binarized XCT image.
Figure 4. The orientation of the dendrites is determined by means of a processed and binarized XCT image.
Minerals 14 00042 g004
Minerals 14 00042 g005
Figure 5. (A) thin section of a lithium aluminate slag particle with zoomed-in sections showing (B) small cuboid-shaped or fine entities; and (C) dendrites consisting of cubical crystals that are connected in a cross arrangement.
Figure 5. (A) thin section of a lithium aluminate slag particle with zoomed-in sections showing (B) small cuboid-shaped or fine entities; and (C) dendrites consisting of cubical crystals that are connected in a cross arrangement.
Minerals 14 00042 g005
Minerals 14 00042 g006
Figure 6. Comparison of the thickness according to the matrix and target phase of lithium aluminate slag. The thickness values for three-dimensional structures are determined by the inscribed sphere method provided by VGSTUDIO MAX.
Figure 6. Comparison of the thickness according to the matrix and target phase of lithium aluminate slag. The thickness values for three-dimensional structures are determined by the inscribed sphere method provided by VGSTUDIO MAX.
Minerals 14 00042 g006
Minerals 14 00042 g007
Figure 7. Meso-scale pores in the Lithium aluminate slag body. (A) thin section showing large pores (white space); (B) depth-in-field within a pore (white outlined).
Figure 7. Meso-scale pores in the Lithium aluminate slag body. (A) thin section showing large pores (white space); (B) depth-in-field within a pore (white outlined).
Minerals 14 00042 g007
Minerals 14 00042 g008
Figure 8. Pore radii of the nearly cubic-shaped slag particle (edge length 4 mm) were determined based on XCT measurements, a 3D representation of the pores (blue) within the used slag sample (grey), and the pore radii and shape distribution of the investigated sample.
Figure 8. Pore radii of the nearly cubic-shaped slag particle (edge length 4 mm) were determined based on XCT measurements, a 3D representation of the pores (blue) within the used slag sample (grey), and the pore radii and shape distribution of the investigated sample.
Minerals 14 00042 g008
Table 1. XRD results of lithium-containing slag and its corresponding Mohs hardness. Lithium aluminate (γ-LiAlO2) is the target phase, and the remaining components are assigned to the matrix.
Table 1. XRD results of lithium-containing slag and its corresponding Mohs hardness. Lithium aluminate (γ-LiAlO2) is the target phase, and the remaining components are assigned to the matrix.
GehleniteGlaucochroiteβ-Eucryptiteγ-LiAlO2Residual (Amorphous)
amount in wt.%25.81817.522.516.2
Mohs hardness5–666.57.55–6.5
Table 2. XCT measurement and reconstruction settings of the cubical lithium aluminate slag particle.
Table 2. XCT measurement and reconstruction settings of the cubical lithium aluminate slag particle.
Measurement Settings
parametervalue
source distance in mm−90
detector distance in mm13
optical magnification4X
acceleration voltage in kV80
electrical power in W7
source filter (Zeiss standard)LE4
voxel size in µm5.885
camera binning2
number of projections1601
exposure time in s10
angle range (°)360
scan time (h:mm)06:30
Reconstruction Settings
functionparameter
reconstruction algorithmfiltered back projection
smoothinggaussian, 0.5
defect correctionnone
byte scalingmanual (−0.0538, 0.147)
beam hardening constant0.27
Table 3. Image processing workflow for pre-processing and microstructural analysis in 2D in Fiji (ImageJ 1.53b) and 3D in VGSTUDIO MAX 3.3.
Table 3. Image processing workflow for pre-processing and microstructural analysis in 2D in Fiji (ImageJ 1.53b) and 3D in VGSTUDIO MAX 3.3.
Image Processing StepParameterRef.
2D preprocessing (Fiji [39])original image (16-bit)-
non-local means denoisings|50, sf|50[41,42]
unsharp maskr|1.5, mw|0.8
threshold27,756–49,087
8-bit conversion-
remove Outliersr|1.5, th|50
2D microstructure
processing and analysis (Fiji)		skeletonizing-[43]
local thickness-[44]
directionality-[45]
3D analysis
(VGSTUDIO MAX)		porosity/inclusion analysisprobability th|0
wall thickness analysis (sphere method)probability th|0
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Võ, T.T.; Leißner, T.; Peuker, U.A. Utilizing X-ray Computed Tomography for Lithium Slag: A Guide to Analyzing Microstructure and Its Potential Influence on Liberation. Minerals 2024, 14, 42. https://doi.org/10.3390/min14010042

AMA Style

Võ TT, Leißner T, Peuker UA. Utilizing X-ray Computed Tomography for Lithium Slag: A Guide to Analyzing Microstructure and Its Potential Influence on Liberation. Minerals. 2024; 14(1):42. https://doi.org/10.3390/min14010042

Chicago/Turabian Style

Võ, Thu Trang, Thomas Leißner, and Urs A. Peuker. 2024. "Utilizing X-ray Computed Tomography for Lithium Slag: A Guide to Analyzing Microstructure and Its Potential Influence on Liberation" Minerals 14, no. 1: 42. https://doi.org/10.3390/min14010042

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop