Pedogenesis of Lateritic Soils and the Enrichment of Critical Metals: A Study from Southeast Sulawesi, Indonesia

Lateritic soil is a prospective source of metals termed critical or strategic metals due to their use in high-technology in - dustries. Critical metals include rare earth elements (REEs). In this study, two profiles of lateritic outcrops from the Ni-producing area of Southeast Sulawesi in Indonesia were sampled for magnetic susceptibility as well µ-XRF analyses to identify how the concentration of critical metals changed during pedogenesis. The results show that there are three dif - ferent patterns from the bottom layer up. The first pattern is enrichment experienced by Sc, La, Cr, and Ti. The second pattern is depletion experienced by Ni, Nd, and Ho. The third pattern is the localization of high concentration at a certain depth, as experienced by Co, Ce, and V. The concentration of certain critical metals correlates either positively or nega - tively with magnetic susceptibility, inferring that magnetic susceptibility might be used as a proxy indicator for critical metal concentration in lateritic soil.


Introduction
Certain metals, including rare earth elements (REEs), platinum group metals (PGM), indium (In), tungsten (W), tin (Sn), niobium (Nb), thallium (Tl), gallium (Ga), and magnesium (Mg), are termed critical metals due to their high demands and market values (Olivetti et  The differentiation of REE contents in lateritic soils and its relationship with pedogenesis, however, are not known with certainty, as research on these matters is rather limited. In this study, we examine the REEs content of lateritic soil outcrops in nickel mines in the Indonesian province of Southeast Sulawesi (see Figure 1a). This province is one of the top nickel producers in Indonesia. The objective is to identify differentiation processes such as enrichment, depletion, and the localization of REE in lateritic outcrops. Apart from geochemical analyses, susceptibility measurements were also carried out to test whether the magnetic susceptibility values correlated with the contents of several critical metals, as reported in a review by Santoro et al. (2022).

Geological Setting
Tectonic conditions on Sulawesi Island are complex and resulted from a highly complicated tectonic system during the Late Cretaceous and Early Tertiary in which the Philippine, Australian, and Eurasian plates were in interaction (Hall and Wilson, 2000;Hall, 2012). As a result of this tectonic process, Sulawesi Island possesses extremely complex geological features that are represented in its distinctive K-shaped form (see Figure 1a). The island's uniqueness leads to metallogonic richness (nickel laterite deposit), particularly the abundance of ultramafic rocks as the source rock for the formation processes (Kadarusman et al., 2004;Surono, 2013;Choi et al., 2021). Peridotite (lherzolite, harzburgite, and dunite), gabbro, microgabbro, pyroxenite, diabase, serpentinite, and pelagic make up the group of oceanic crustal rocks known as ophiolite (see Figure 1b). This study area in South Konawe Regency of Southeast Sulawesi Province is located in a mafic-ultramafic ophiolite strip that is under the control of a localised fault system (Surono, 2013;Zhang et al., 2020). Based on the onsite geological observations in the study area (see Figure 1c), the outcrops are located under the layers of (d) peridotites that were found in the study area; and (e) serpentinites that were found in the surrounding study area

Materials
Sixteen samples were collected for examination from two profiles on the same outcrop. The two profiles (termed A and B, respectively) are located about 3 metres from each other. The outcrop is about 240 cm thick in depth, and samples were taken at an interval of 30 cm in each profile to obtain detailed changes in both magnetic susceptibility as well as pedogenic processes. The profiles cover the following layers (from top to bottom): red limonite, yellow limonite, and saprolite (see Figure  2). The profiles were carefully scraped to expose a fresh surface layer. Sampling was carried out by taking about 3 kg of material from each position. The samples were then air dried to remove the remaining water content. The dried samples were ground with a mortar and pestle.  (2022), ground samples were then sieved with a 200-mesh-size sieve (74 µm) to obtain grain sizes that are suitable for XRF analyses. Some of the sieved samples were placed in an 8 cm 3 standard cylindrical plastic holder for magnetic susceptibility analyses. Other samples were prepared for geochemical analyses by pressing them into pellets that are 4 cm in diameter and 4 mm in thickness.

Methods
Magnetic susceptibility analyses were carried out using a Bartington MS-3 magnetic susceptibility system (Bartington Ltd., Oxford, UK) that uses an MS-3 metre and MS-2B dual frequency sensors. The parameters measured are mass-specific magnetic susceptibility at low (470 Hz) and high (4700 Hz) frequencies, referred to as χ LF and χ HF , respectively. The third parameter, χ FD% is calculated using the equation (Dearing, 1999;Tamuntuan et al., 2015). Parameters χ LF and χ HF indicate the quantity of magnetic minerals, while χ FD% indicates the contribution of superparamagnetic (SP) grains. For magnetite (Fe 3 O 4 ), which is the most common magnetic mineral in nature, SP grains are grains that are < 30 nm. SP grains are similar paramagnetic grains with much higher magnetic susceptibility (Dearing, 1999). Low χ FD% values indicate the absence or small contribution of SP grains, while high χ FD% values indicate a higher contribution of SP grains (Dearing, 1999). Magnetic analyses were performed at the Institut Teknologi Bandung's Laboratory of Characterization and Modeling of Physical Rock Properties. The geochemical analyses were carried out using EDAX Orbis Micro X-ray Fluorescence, or EDXRF, equipment (EDAX Ltd., Mahwah, NJ, USA) to generate chemical maps of the samples. These analyses were conducted at the Research Center for Nanoscience and Nanotechnology (RCNN), Institut Teknologi Bandung. The results of these analyses are the concentrations of major oxides, transition metals (Ni, Co, Sc, Cr, and V), and REEs (Sc, La, Ce, Nd, Gd, and Ho). This type of XRF, termed µ-XRF, has a spatial resolution diameter that is many orders of magnitude smaller than conventional XRF (Bran-Anleu et al., 2018; Sharps et al., 2021). Table 1 shows the values of the χ LF and χ FD% measurements of the two profiles. The χ LF and χ FD% values The bottom parts of the profiles only show yellow limonite and its transition zone towards saprolite.  changed as they increased from the bottom layer to the top. In each profile, it is clear that there is a change in χ LF and χ FD% values at a depth of 120-150 cm. For more details, the depth profile for the ratio can be seen in Figure  3a. Table 1 shows that the values of χ LF /χ FD% ratio could be grouped into 3 clusters, which can be seen more clearly in Figure 3b, namely: cluster 1 has a high χ LF value with a high χ FD% (χ FD% ≥5%); cluster 2 has a high χ LF value with a low χ FD% percentage χ FD% <5%; and cluster 3 has low χ LF and χ FD% values. Table 2 shows the results of the geochemical analysis. Certain oxides have increased concentrations from the bottom up, namely Fe, Cr, and Al; there are oxides that experience concentration depletion from the bottom up, namely Si and Mg; and there are also other oxides that do not change significantly with depth, such as Mn.

Results
In addition, Table 2 also shows the concentration values of critical metals and REEs. Certain metals have increased in concentration from the bottom up, namely Sc, La, Cr, and Ti; there is also a group of metals that experience concentration depletion from the bottom up, namely Ni, Ho, and Nd; and there are also other groups that do not experience significant changes with depth, such as Ce, Co, and V. For more details, see

Discussion
The results shown in Table 1 and Figure 3 regarding the variations in χ LF and χ FD% values in laterite profiles are influenced by variations in magnetic minerals, especially Fe minerals, which are formed during the pedogenesis process. From the results of cluster division (see Figure 3b), the characteristics of cluster 1 are indicated due to the presence of ultrafine superparamagnetic minerals produced during the pedogenesis process, which are formed in the uppermost layer of laterite (the red limonite zone). Cluster 2 occurs because magnetite SP grains are unstable, so they experience oxidation processes and change into hematite and goethite. In both profiles (A and B), cluster 2 is found in the yellow limonite layer. Cluster 3 has low χ LF and χ FD% in the transition zone between yellow limonite and saprolite. This cluster forms because the layer above the transition zone is stable, so mobilisation of Fe minerals in the lower layer is minimal but labile minerals are added. Generally, these labile minerals are diamagnetic, including Si and Mg. As discussed by Szuszkiewicz et al. (2021), iron oxide plays an important role in pedogenic processes, including the transfer and transformation of iron (Fe) minerals in the soil.
During the pedogenesis process, not only the mobilisation of Fe minerals occurs but also changes in critical metal concentrations, which result in the emergence of patterns of enrichment, depletion, and localization. The enrichment pattern shows an increase in the concentration of major oxides and minerals from the bottom layer up. This pattern has similarities with the laterite pedogenesis process, where there are several metal elements and REE that experience enrichment near the surface, or the same with the abundance of Fe minerals. One of the REE metals that has experienced enrichment with a clear trend is Sc (Figure 4a). The relationship between the presences of Sc and Fe has been reported in earlier studies (Chassé et (Choi et al., 2021).
During the pedogenesis process in laterite cases in this study, several factors might have occurred that affected the enrichment of critical metals at the top, espe- (2) Sc has increased in the upper and lower limonite zones; and (3) Sc and several other REEs have the same pattern as the pedogenesis process during the transformation of magnetite SP grains to goethite and hematite. When there is a lot of weathering, the limonite zone forms because water keeps moving downward, carrying unstable (highly mobile) elements and leaving stable (not mobile) elements in the uppermost horizon. As long as the water supply stays the same, this pedogenesis process will lead to in situ enrichment because Fe, Al, Cr, and Ti are more stable or immobile and will stay in the top layer (Choi et al., 2021). The depletion pattern is the opposite of enrichment during the pedogenesis process. This pattern is associ-ated with the presence of labile minerals (minerals that readily mobilize). One of the minerals found in profiles A and B experiencing depletion is Ni. The maximum concentration of Ni is found in the saprolite zone (see      reported that the maximum V concentration was in weathered rock fragments termed the iron duricrust horizon. Several elements detected in profiles A and B have a correlation with the presence of other elements, so in studying certain REE characteristics, it is necessary to look at the positive and negative correlations between several elements. Table 3 shows the Pearson correlation coefficients (r) between χ LF , the concentration of major elements, and critical metals. Since each profile has only 8 samples, i.e., n = 8, the critical value for r at the 0.05 level of significance is 0.622 (Nino-Zarazua, 2012). In Table 3, r values that are higher than 0.622 (or lower than -0.622) are marked in bold typeface, indicating significant positive (or negative) correlations between the parameters. Significant positive correlations were found between the values of χ LF and Fe 2 O 3 , Al 2 O 3 , Ti, Cr, La, and Sc. These correlations are more clearly seen in On the other hand, Sc is in conflict with the abundance of MgO and SiO 2 concentrations. SiO 2 and MgO levels decrease from the bedrock to near the surface. Moreover, the concentration of χ LF was negatively correlated with the presence of Ni, MgO, SiO 2 , Ho, and Nd (see Figure 6). The abundance of magnetic minerals during the pedogenesis process has a significant influence on the presence of REE in lateritic soils. From the results discussed, it was found that magnetic parameters can be used as a proxy to observe the pedogenesis process in laterite; besides that, they can be used as an indicator of the presence of some metals, such as high Ni concentration associated with low χ LF , and low Sc concentration associated with high χ LF .

Conclusion
The pedogenesis process of lateritic soils is controlled by the abundance of iron oxide minerals, which are characterised by an increase in the value of magnetic susceptibility from the source rock to the upper profile of lateritic nickel deposits. During the pedogenesis process, the magnetite mineral changes grain size from coarse to fine, as shown by changes in χ FD% values that are low at the bottom and high at the top. The magnetite SP grains in the uppermost zone have changed into iron oxide minerals (hematite and goethite). Based on the studies conducted, there are three patterns of changes in critical metal concentrations during the pedogenesis process: 1) The enrichment pattern has the same pattern as the pedogenesis process (increasing concentration from the bottom layer up), as seen in the metal concentration patterns of Sc, La, Cr, and Ti; 2) the depletion pattern shows decreased concentration from the bottom layer up due to remobilization with labile minerals (Mg, Si) and deposits in the bottom layer, as seen in the concentration pattern of Ni, Nd, and Ho; and 3) the localization pattern, in which high concentrations of certain metals (Co, Ce, and V) occurred at a certain depth. From the results of geochemical and magnetic susceptibility tests, it is known that the concentration of critical metals in lateritic soils depends on the amount of magnetic minerals. The increase in susceptibility value indicates an increase in magnetic mineral concentration. This can be seen in the increase in the concentration of the critical metals Sc, Co, and La in the limonite zone, which has a positive correlation with the χ LF value. However, it is inversely proportional to Ni, Nd, and Ho, which are negatively correlated with χ LF . Thus, the rock magnetism method can be considered a proxy to see the pattern of the presence of critical metals in the laterite profile.