Comprehensive characterization and extraction implications of ion adsorption rare earth deposit from a South American source

Abstract

Ion-adsorption rare earth element (REE) deposits are a critical resource for strategic materials, yet their characterization and processing remain complex. This study provides a comprehensive mineralogical, chemical, and geochemical analysis of an ionic clay sample from a South American source, integrating multiple characterization techniques, including XRD, SEM-EDX, XPS, ToF-SIMS, TIMA-X, EPMA, and LA-ICP-MS. The results confirm that kaolinite and micas dominate the matrix, with monazite identified as the primary REE-bearing mineral. Yttrium and heavy REEs are primarily hosted in clays, indicating the necessity of ion-exchange leaching for effective extraction. Liberation studies reveal that monazite is best liberated in finer fractions, suggesting a need for targeted pre-concentration strategies. Surface chemistry analyses demonstrate the presence of REEs as adsorbed species and inner-sphere complexes, supporting the use of selective leaching techniques. The study highlights the economic and environmental considerations of REE extraction from ionic clays and provides insights into optimizing recovery processes while mitigating environmental risks. These findings contribute to the growing body of research aimed at diversifying REE supply sources and improving sustainable extraction methods.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-14891-3.

Introduction

Rare earth elements (REEs)—the 17 lanthanides plus yttrium (Y) and scandium (Sc)—are indispensable to modern clean energy and high-tech applications, most notably as components of permanent magnets in wind turbines and electric vehicle motors¹. A significant portion of the world’s medium and heavy REEs is sourced from ion-adsorption clay (IAC) deposits². These regolith-hosted deposits typically contain 0.05–0.5 wt% total rare earth oxides (REO) with REE ions loosely bound to the surfaces of secondary clay minerals like kaolinite and illite¹. IAC deposits form through intense tropical weathering of REE-enriched granitic rocks, which releases REEs that subsequently migrate and adsorb onto clay mineral surfaces in the soil profile. Unlike hard-rock REE ores (e.g. monazite or bastnäsite) that require aggressive chemical treatment for extraction, IAC ores are relatively easy to mine and process, typically via simple salt leaching at ambient conditions³. This means IAC operations can use low-cost open-pit mining with no complex beneficiation, minimal radioactivity, and little to no tailings, thereby offsetting the lower grade of these clays. Such advantages have made IAC deposits strategically important in the global REE supply chain.

China’s extensive IAC deposits (found mainly in subtropical southern provinces) have historically enabled that country to dominate the REE market⁴. China currently contributes roughly 60% of global REE mine production and controls about 85% of downstream processing capacity, despite holding only a minority of worldwide REE reserves⁴. This supply concentration presents a geopolitical risk and has spurred efforts to diversify REE sources. In recent years, several countries have turned their attention to ion-adsorption clays as alternative REE resources. New discoveries and projects in regions such as Brazil, Chile, Australia, and Uganda highlight a growing international interest in exploiting IAC deposits outside China². Indeed, critical materials strategies worldwide now emphasize the development of these “ionic clay” resources to broaden the REE supply base. For example, leveraging IAC deposits could significantly reduce the capital and environmental costs of REE production while opening new supply opportunities in countries like Australia and Brazil⁵. The relatively light infrastructure and lower waste footprint of IAC mining make these deposits attractive for nations seeking to secure REE access and reduce over-reliance on Chinese supply⁵. In this context, South America, with its extensive granitic weathering crusts, represents an important yet under-characterized frontier for IAC-type rare earth deposits.

The state of occurrence of REEs in weathering crusts provides essential insights into their geochemical behavior during migration, enrichment, and differentiation processes. This understanding is crucial for exploring the genesis of ion-adsorption rare earth deposits and for their effective extraction and use. The continuous chemical extraction method, which analyzes the chemical phase by assessing the solubility and dissolution rates of different minerals or compounds in specific solvents, is commonly used to determine the occurrence state of REEs in weathered crusts⁶. Extensive studies have identified four main forms of ionic rare earth deposits: ion adsorption, colloidal dispersion, independent mineral, and lattice impurity phases. These are categorized into various states such as exchangeable adsorptive, specific adsorptive, colloid adsorbed, gel, supergene mineral, residual mineral, homogeneous, and internal submersible crystal states, with exchangeable adsorbed rare earth typically constituting 80–96% of the total rare earth content^1,3,7.

Over time, weathering processes transform minerals such as feldspar and mica into clay minerals like kaolinite, illite, halloysite, and montmorillonite. The adsorption capacity of these clay minerals varies, influenced by factors such as surface area, structural properties, and the number of active sites. Montmorillonite, halloysite, illite, and kaolinite ranked from high to low in their ability to adsorb rare earth elements. The order of cation adsorption strength for clay minerals is RE³⁺ > Al³⁺ > Ca²⁺ > Mg²⁺ > K⁺ > NH⁴⁺ > Na⁺. For rare earth ions, the adsorption intensity decreases from Sc³⁺ to Lu^3 + 8.

Recent synchrotron radiation studies have advanced the understanding of the occurrence states of REEs, showing that they also exist as outer-sphere or inner-sphere hydrated complexes within clay minerals^9,10. A previous study used X-ray absorption fine structure spectroscopy and they identified yttrium primarily adsorbed on clay mineral surfaces as outer sphere hydrated free ion complexes in Japanese granite weathering crusts⁹. Another study suggested that up to half of the REEs in single clay mineral particles (originated from China) within granite weathering crusts might form inner-sphere hydrated complexes, based on findings from scanning electron microscopy (SEM), transmission electron microscopy (TEM), nanoscale secondary ion mass spectrometry (Nano-SIMS), and desorption tests¹⁰. Further studies using scanning transmission electron microscopy (STEM)-EDS and Nano-SIMS, highlighted illite’s role in enriching and mineralizing REEs in granite weathering crusts¹¹.

Borst et al., employed synchrotron radiation X-ray absorption spectroscopy to compare the states of yttrium (a heavy rare earth) and neodymium (a light rare earth) in clay minerals from China and Madagascar. They observed that these elements could bind as inner-sphere complexes with Al–O and Si–O bonds on or at the edges of clay minerals, indicating the presence of both outer-sphere and inner-sphere complex forms¹². Notably, synchrotron radiation X-ray absorption spectrometry revealed that yttrium is likely present as an 8-coordinated or 9-coordinated outer-sphere hydrated complex on the surface of kaolinite.

In ion-adsorption rare earth ores, it has been estimated that clay minerals, particularly those in the kaolin mineral group (kaolinite and halloysite), adsorb REEs. These clay minerals are common in ion adsorption ores and typically have a low point of zero charge (PZC) (< 4.5 in most cases)¹². The 1:1 layer clay minerals possess a small permanent charge derived from isomorphous substitution (e.g., Al³⁺ for Si⁴⁺), in addition to a surface charge mainly derived from variable pH-dependent charge¹³. The rare earths adsorbed on the surface of these clays, forming outer-sphere complexes, are assumed to be extracted by ion exchange treatment. However, direct analyses of individual minerals have rarely been conducted due to the low concentration of rare earths in ion adsorption ores.

Despite the economic and strategic importance of IAC deposits, high-resolution mineralogical and geochemical data for South American IACs are exceedingly limited. To date, most detailed studies of IAC mineralogy and REE occurrence have focused on Chinese (and a few African or Asian) deposits, while analogous South American occurrences remain poorly documented. The subtle modes of REE occurrence in these clays, whether adsorbed as outer-sphere cations, inner-sphere surface complexes, or contained in discrete REE minerals, can vary with geological setting. As a result, fundamental questions about the host phases, distribution, and chemical state of REEs in South American IAC deposits remain open. Bridging this knowledge gap is crucial for evaluating the potential of these deposits and for designing effective extraction processes tailored to their specific characteristics. The present study addresses this gap by providing a comprehensive characterization of an ion-adsorption clay deposit in South America derived from granitic source rock. This deposit offers a novel case study to compare with well-studied Chinese IACs and to extend the global understanding of REE-bearing regolith clays.

A multi-faceted analytical approach was adopted to fully elucidate the mineralogical, geochemical, and surface chemical properties of the South American IAC sample. Given the complex and fine-grained nature of IAC ores, no single technique is sufficient to capture all aspects of REE occurrence. We therefore integrated complementary methods to obtain a holistic picture of the deposit’s makeup. X-ray diffraction (XRD) was used to identify bulk mineralogy and quantify the major mineral phases present. Scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDX) and electron probe microanalysis (EPMA) provided micron-scale examination of grains, revealing mineral textures and precise chemical compositions of host minerals. An automated mineral analyzer (TESCAN TIMA-X) was employed to map mineral distributions and associations, offering statistical mineralogical data on liberation and grain intergrowth. To investigate trace and distributed elements, laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) was applied, enabling in-situ quantification of REEs within individual mineral phases (e.g. clays, oxides, or phosphates). Critically, the surface chemistry of the clay, where adsorbed REE ions reside, was probed by X-ray photoelectron spectroscopy (XPS) and time-of-flight secondary ion mass spectrometry (ToF-SIMS). These surface-sensitive techniques elucidate the chemical state of REEs and interacting elements on the clay particle surfaces, distinguishing adsorbed species and potential secondary precipitates. By combining these analytical techniques, we capture multiple scales of information: from overall mineral assemblage and bulk chemistry down to micro-textural context and the atomic-level binding environment of REEs. This integrated methodology provides unprecedented detail on how REEs are distributed (adsorbed versus mineral-incorporated), what minerals make up the clay matrix, and how those factors might influence leaching behavior and processing.

In summary, this work delivers the first high-resolution mineralogical and geochemical profile of a South American ion-adsorption clay deposit, filling a significant data gap in the rare earth literature. The objectives of the study are to: (1) characterize the deposit’s mineralogy, clay composition, and bulk chemistry in detail using a suite of analytical techniques; (2) determine the occurrence modes of REEs, specifically, identifying the host phases of REEs and the extent to which they are present as exchangeable surface-adsorbed ions versus as refractory mineral inclusions; and (3) evaluate the implications of these findings for REE extraction and processing. By integrating these insights, the study aims to advance understanding of ion-adsorption clay deposits in South America and contribute to broader efforts in diversifying the global REE supply while optimizing sustainable extraction strategies.

Results and discussion

Alkali fusion followed by ICP-MS/OES results

The elemental composition of the feed is presented in Table 1. The sample contains a total rare earth element (TREE) content of 0.44 wt%, with Ce and Y being the most abundant elements at 0.10 and 0.12 wt%, respectively. The sample also contains 562.6 mg/kg Nd and 162.2 mg/kg Dy. Additionally, notable elements in this sample include Al at 14.5 wt%, Si at 21.4 wt%, and Fe at 6.2 wt%.

Table 1.

Chemical composition of the feed ionic clay.

Element	Concentration (ppm)	Element	Concentration (ppm)
Y	1201.7	Lu	14.9
La	734.1	TREE	4362.9
Ce	1026.2	Th	209.6
Pr	159.7	U	2.9
Nd	562.6	Element	Concentration (wt%)
Sm	95.4	Al	14.5
Eu	4.4	Mg	0.2
Gd	105.5	K	1.4
Tb	19.2	Fe	6.2
Dy	162.2	Mn	0.2
Ho	40.2	Ca	0.1
Er	123.5	Na	0.4
Tm	16.8	P	0.1
Yb	96.7	Si	21.4

XRD results

The mineralogical characteristics of the sample obtained from XRD analysis are presented in Table 2. Bulk and clay XRD analysis indicates that the “<300 µm” sample consists of quartz (22.0%), muscovite (5.2%), biotite (2.7%), anatase (4.0%), albite (1.6%), orthoclase (1.5%), spinel (1.6%). Clay minerals include kaolinite (56.2%), illite (4.2%) and traces of vermiculite.

Table 2.

XRD results for the “<300 µm” sample.

Mineral/compound	< 300 μm
Quartz	22.0
Biotite	2.7
Muscovite	5.2
Anatase	4.0
Albite	1.6
Orthoclase	1.5
Spinel	1.6
Ilmenite	0.3
Clay
Kaolinite	56.2
Illite	4.2
Vermiculite	0.5
Total	100

SEM-EDX results

Elemental mapping of the feed, as illustrated in Fig. 1, revealed a predominance of grains composed of Al, Si, Fe, and O. Backscattered electron (BSE) images and semi-quantitative EDX elemental analysis for the feed sample are displayed in Fig. 2a–d. These images reveal the particle morphology, texture, and size, while the EDX elemental and mapping analyses provide insights into the elemental composition and spatial distribution, respectively. Various particles with REEs inclusions, of varied sizes, were detected. Particles with elemental composition consistent with kaolinite mixed phase (i.e., area 3 in Fig. 2B), monazite (area 1 in Fig. 2D), and muscovite (area 2 in Fig. 2D) were also evaluated.

Fig. 1.

BSE image and elemental maps collected from the feed ionic clay sample. Please note that EDX lines for Ce and La overlap, which can result in similar spatial distribution in the elemental maps.

Fig. 2.

BSE images and EDX analysis of various particles in the feed ionic clay. (A) Particle 1, (B) particle 2, (C) particle 3, (D) particle 4.

Raman spectroscopy results

Raman spectroscopy was used to identify the mineralogical compositions of the ionic clay sample. Quartz, rutile, zircon, ilmenite, and monazite were identified in the feed. Representative Raman spectra of the mentioned mineral phases, along with their respective reference Raman spectra are presented in Fig. 3. Spectra of kaolinite were not produced due to its weak Raman scattering¹⁴.

Fig. 3.

Experimental Raman spectra for the mineral phases identified from the feed sample along with their reference spectra. Please note that the intensities of the spectra have been normalized.

XPS results

Figure 4 shows the representative survey and high-resolution spectra collected from the ionic clay sample. The survey spectrum quantifications indicate that the surface composition primarily consists of O, C, Si, and Al, with small to trace amounts of Fe, N, Mn, P, Mg, and Ca. Trace amounts of K were also detected in the high-resolution C 1s spectrum (not shown); however, could not be quantified.

Fig. 4.

Representative XPS survey and high-resolution spectra collected from the feed ionic clay.

The high-resolution Al 2p and Si 2p spectra showed 2p_3/2 signals at 74.4 eV and 102.7 eV, respectively, aligning with various Al (III) silicate species¹⁵. This is consistent with the XRD and EDX results presented above, which indicate that the feed is primarily composed of kaolinite and muscovite. The high-resolution Fe 2p and Mn 2p spectra displayed complex peak shapes due to the well-known multiplet splitting process¹⁶. Comparison of the experimental data with available reference data suggests that the Fe and Mn signals correspond to FeOOH and Mn₂O₃, respectively.

Ce was not detected in the survey spectrum and only trace levels of P were detected. This is not surprising since the SEM-EDX results highlight that the monazite is sparsely distributed within the kaolinite- and muscovite-rich feed. The same is true for both Nd and Dy.

ToF-SIMS results

Figure 5 presents the Al, Si, Dy, and Nd positive secondary ion images of a selected area from the sample along with selected regions from the generated positive spectrum, highlighting the Dy and Nd peaks. The images along the spectrum demonstrate the detection of the Dy and Nd ions on the surface of the ionic clay sample. These rare earth elements appear to be present as a very thin layer on the surface of the analyzed particles—aligning with the ICP-MS results presented earlier.

Fig. 5.

Selected positive secondary ion images and selected regions of the positive spectrum collected from the feed sample, highlighting the detection of Dy and Nd.

TIMA-X analysis results

Modal mineralogy

A complete and a condensed mineral abundance is given in wt% by sample (Table 3). The condensed mineral list was used to calculate the liberation and association of the minerals. Monazite (0.26% and 0.28% in the “<300 µm” and “<50 µm”, respectively) is the primary REE mineral followed by mixtures of REE phosphates (REE-P) which cannot be properly differentiated because they are fine grained or altered. The matrix is comprised quartz (26.0% and 7.23%), biotite (10.4% and 5.52%), kaolinite (10.8% and 22.7%), smectites (39.1% and 39.0%), muscovite/illite (4.5% and 15.3%), Fe-Oxides (5.31% and 6.34%), ilmenite (2.30% and 1.04%), zircon (0.27% and 0.96%), and trace amounts of other minerals.

Table 3.

Calculated head (wt%) modal abundances of the “<300 µm” and “<50 µm” sample.

Mineral/sample	< 300 μm	< 50 μm	Mineral/sample	< 300 μm	< 50 μm
Monazite	0.26	0.28	Monazite	0.26	0.28
REE-P	0.05	0.11	Other REM/Nb	0.01	0.01
Other REM/Nb	0.01	0.01	Quartz	26.0	7.23
Apatite	0.00	0.00	Biotite	10.4	5.52
Quartz	26.0	7.23	Kaolinite	10.8	22.7
Biotite	10.4	5.52	Smectites	39.1	39.0
Kaolinite	10.8	22.7	Muscovite/Illite	4.50	15.3
Smectites	39.1	39.0	Other silicates	0.04	0.05
Muscovite/Illite	4.50	15.3	Zircon	0.27	0.96
Feldspars	0.23	0.46	Fe-oxides	5.31	6.34
Other silicates	0.04	0.05	Ilmenite/rutile	2.30	1.04
Zircon	0.27	0.96	Mn–Ca–Ox/Carb	0.34	0.60
Sulphides	0.00	0.02	Other	0.00	0.02
Fe-oxides	5.31	6.34	Total	100	100
Ilmenite	2.30	1.04
Rutile	0.45	0.39
Mn–Ca–Ox/Carb	0.34	0.60
Carbonates	0.02	0.05
Other	0.00	0.02
Total	100	100

Mn–Ca–Ox-Carb: refer to Mn–Ca–oxides/carbonates.

The term “smectites” refers to finely intergrown micas/clays and Fe-oxides/oxyhydroxides that cannot be differentiated at 10 micron pixel resolution with the instrument. It should be noted that the XRD did not detect any Fe-oxides/oxyhydroxides. It is possible that these phases may be poorly crystalline or amorphous.

Other REM/Nb refers to unidentified REE minerals and Nb-bearing phases.

Grain size distribution

The P₈₀ (Table 4) for monazite is 86 μm and 40 μm in “<300 µm” and “<50 µm”, that of kaolinite is 90 μm and 23 μm, respectively, and that of smectites is 140 μm and 26 μm, respectively. The P₈₀ for particle in the overall sample are 211 μm and 30 μm, respectively.

Table 4.

P₈₀ (µm) of selected minerals calculated for the “<300 µm” and “<50 µm” samples.

Sample	< 300 μm	< 50 μm
Monazite	86	40
Other REM/Nb	14	17
Quartz	228	33
Biotite	92	18
Kaolinite	90	23
Smectites	140	26
Muscovite/Illite	57	25
Other silicates	120	25
Zircon	51	41
Fe–oxides	123	25
Ilmenite/Rutile	147	31
Mn–Ca–Ox/Carb	50	19
Other	12	48
Particle	211	30

Liberation, association, and exposure of monazite

The liberation of monazite (pure, free, and liberated) is 32.4% and 65.1% in “<300 µm” and “<50 µm” samples, respectively (Table 5). The remainder occurs as complex particles at 50.3% and 12.4%, middlings with other REM/Nb (0.7% and 5.9%), smectites (1.3% and 7.3%), Fe-oxides (2.7% and 52%), (biotite & kaolinite & smectites: Bt : Knl : Sme (5.1% and 0.2%), and other minerals in minor amounts. The liberation of the monazite by size class for the two samples is shown in Figs. 6 and 7. The graph illustrates the occurrence of monazite by size class. For example, in the “<300 µm” liberated grains account for 22.6% below 100 μm and 9.8% above 100 μm.

Table 5.

Liberation and association of monazite calculated for “<300 µm” and “<50 µm” samples.

Monazite association	< 300 μm	< 50 μm
Pure monazite	0.99	30.0
Free monazite	5.83	6.13
Lib monazite	25.6	29.0
Mnz: Other REM/Nb	0.74	5.95
Mnz: Quartz	2.88	0.20
Mnz: Biotite	0.00	0.00
Mnz: Kaolinite	0.28	0.21
Mnz: Smectites	1.26	7.30
Mnz: Muscovite/Illite	0.00	0.19
Mnz: Other Silicates	0.00	0.37
Mnz: Zircon	0.01	0.00
Mnz: Fe-Oxides	2.65	5.17
Mnz: Ilmenite/Rutile	2.81	0.35
Mnz: Mn-Ca-Ox/Carb	0.08	0.43
Mnz: Other	0.00	0.00
Mnz: Knl : Sme	1.05	1.85
Mnz: Knl : Sme : Ms/Ilt	0.39	0.23
Mnz: Bt : Knl : Sme	5.08	0.22
Complex	50.3	12.4

Fig. 6.

Liberation and association of monazite by size class for “<300 µm” sample.

Fig. 7.

Liberation and association of monazite by size class for “<50 µm” sample.

An image grid showing monazite particles as a function of liberation and association for the samples is illustrated in Fig. 8. The particle maps illustrate the textural features (grain size, complexity) of the monazite in association with other minerals. It should be noted that the liberation/association of the monazite is calculated based on the volume of the mineral and as a function of its mass. The exposure (see below) is calculated based on the free surface of the mineral.

Fig. 8.

Selected particle maps showing monazite particles as a function of liberation and association for “<300 µm” and “<50 µm” samples.

Figure 9 shows that well-exposed monazite (≥ 80%) accounts for 9.3% and 54.0% in “<300 µm” and “<50 µm” samples, respectively. Poorly exposed (< 30%) monazite is 54.6% and 8.2%, while the remainder (36.1% and 37.8%) is variably exposed (≥ 30-<80%). The 30% exposure is used as an empirical threshold for which particles may be recovered (i.e., with flotation).

Fig. 9.

Exposure of monazite for “<300 µm” and “<50 µm” samples.

Liberation of monazite, other REM/Nb, zircon and Fe–Oxides.

The liberation of (pure, free, and liberated) monazite, other REM/Nb, zircon and Fe-oxides is illustrated in Fig. 10. Mineral liberation is higher for all minerals in the “<50 µm” sample.

Fig. 10.

Liberation and association of monazite, other REM/Nb, zircon and Fe-oxides for “<300 µm” and “<50 µm” samples.

Grade recovery calculations

Another, more functional, method of presenting liberation is the mineralogically limiting grade-recovery curves, as shown in Fig. 11. They are based on the calculated mass of minerals and the total mass in each liberation category. Thus, the highest grade (> 80% e.g., monazite) is contained in the > 80% liberated monazite particles. Then the next category (60–80% liberation) is added, and the combined grade is calculated. This is repeated until all the monazite is accounted for. Mineralogically limited grade-recovery analyses provides an indication of the theoretical maximum achievable elemental or mineral grade by recovery, based on individual particle liberation and grade. These results, of course, do not reflect any other recovery factors that could occur in the actual metallurgical process.

Fig. 11.

Grade-recovery of monazite for “<300 µm” and “<50 µm” samples.

Monazite grades of 97% and 81% for grades of 4–45% (excluding the last two points of the curves) are projected for the “<300 µm”; and monazite grades of 99% and 83% for grades of 38–92% are projected for the “<50 µm”.

Mineral chemistry by EPMA and LA-ICP-MS

EPMA was carried out on the thirty monazite grains from the “<300 µm” (Table 6). The results showed that Ce₂O₃ ranges from 28.47 to 29.91% and averages at 29.20%, La₂O₃ ranges from 12.55 to 15.39% and averages at 14.50%, Nd₂O₃ ranges from 11.69 to 13.34% and averages at 12.39%, Pr₂O₃ ranges from 2.95 to 3.40% and averages at 3.18%, Y₂O₃ ranges from 0.34 to 1.48% and averages at 1.00%, Sm₂O₃ ranges from 1.31 to 2.61% and averages at 1.65%, Gd₂O₃ ranges from 0.88 to 1.57% and averages at 1.12%, and ThO₂ ranges from 4.41 to 5.77% and averages at 5.28%. UO₂ is near or below the detection limit of instrument. Monazite is chemically homogeneous. Back scattered electron microscope (BSE) images of the analyzed monazite grains are shown in Fig. 12.

Fig. 12.

BSE Images of the analyzed monazite from “<300 µm” sample.

Table 6.

Range of chemistry for monazite from the EPMA.

One hundred grains of kaolinite, micas and mixtures with Fe-oxyhydroxides were analyzed with both the electron microprobe and the LA-ICP-MS (Supplementary Tables S1 and S2). Back scattered electron microscope (BSE) images of the analyzed monazite grains are shown in Fig. 13. Although a few grains appear to resemble the stoichiometry of kaolinite or muscovite, the majority of the analyses reflect mixtures of clays, micas and Fe-oxy-hydroxides. Thus, Al₂O₃ ranges from 17.79 to 43.45% and averages at 33.72, FeO ranges from 0.55 to 26.30% and averages at 8.42%, and K₂O ranges from 0.01 to 11.30% and averages at 0.61%.

Fig. 13.

BSE images of the analyzed micas/clays from “<300 µm” sample.

The LA-ICP-MS indicates the presence of REE + Y in the same grains. For example, Y ranges from 48 ppm to 4065 ppm and averages at 1620 ppm, La ranges from 14 ppm to 929 ppm and averages at 401 ppm, and Ce ranges from 6 ppm to 899 ppm and averages at 95 ppm among other REE. Thorium averages at 27 ppm. It should be noted that Sc (avg. 21 ppm) and gallium (31 ppm) were also present in the sample.

Cerium and yttrium deportment

The cerium and yttrium distributions shown in Fig. 14 are calculated based on the TIMA-X modal distributions and the average cerium and yttrium concentrations of the minerals from the EMPA and LA-ICP-MS. Monazite and other REE-phosphates account for most of the cerium (94% in both samples), while micas/clays for the remainder 6%. On the other hand, these minerals only host < 2% of the yttrium, while micas/clays host 98% of the yttrium.

Fig. 14.

Ce and Y deportment among the minerals for “<300 µm” and “<50 µm” samples.

Implications and interpretation of characterization results

The comprehensive characterization of the South American ionic clay sample provides significant insights into the mineralogical composition, rare earth element distribution, and extraction potential of the deposit. By integrating the findings from various analytical techniques, a clearer understanding of the ore’s geochemical behavior and its potential for resource recovery can be established.

Mineralogical and chemical composition

The XRD and SEM-EDX analyses confirm that the sample is predominantly composed of kaolinite, micas, and quartz, with minor amounts of biotite, feldspars, and Fe-oxides. The presence of kaolinite and illite is particularly significant, as these clay minerals are known for their cation exchange properties, which facilitate the adsorption of rare earths. The chemical assays (ICP-MS/OES) reveal a TREE content of 0.44 wt%, with Ce and Y as the dominant rare earth elements and considerable amounts of Nd, Pr, and Tb, Dy, highlighting the economic potential of the deposit.

From a processing perspective, the mineralogical composition suggests that ion-exchange leaching would be an essential component of any extraction strategy, particularly to recover Y and heavy REEs hosted in the clay phases. The co-existence of monazite as the primary host for light REEs points to an opportunity for physical pre-concentration methods, such as gravity separation or flotation, to recover monazite before chemical treatment. This could reduce reagent consumption and improve the environmental footprint of downstream processes. The combined presence of monazite and ion-adsorbing clays aligns this deposit with the characteristics of both hard-rock and ionic clay rare earth resources, suggesting that hybrid processing flowsheets may be required.

REE occurrence and deportment

The TIMA-X, EPMA, and LA-ICP-MS analyses provide critical insights into the REE-bearing phases. Monazite is identified as the primary host for Ce and other light rare earths, while Y is predominantly associated with micas and clay minerals. This distinction is crucial for developing an effective extraction strategy, as the separation of monazite may lead to high REE recoveries, whereas the recovery of Y from clays requires different processing techniques, such as ion-exchange leaching.

The decoupling of Y and other heavy REEs from monazite highlights a key challenge in designing an efficient processing flowsheet. Physical separation methods will effectively target monazite but will not recover the majority of heavy REEs, which are largely clay-bound. This highlights the need for a hybrid approach that combines physical pre-concentration of monazite with hydrometallurgical leaching tailored to the ion-adsorption fraction. Additionally, the occurrence patterns observed in this study mirror those reported for Chinese ionic clays, suggesting that established ion-exchange leaching practices may provide a starting point for process development, though site-specific adjustments will be necessary to account for the unique deportment of this deposit.

Particle size and liberation characteristics

The grain size distribution and liberation studies indicate that monazite is better liberated in the finer fraction (< 50 μm), where the liberation increases from 32.4% in the < 300 μm sample to 65.1% in the < 50 μm fraction. This suggests that pre-concentration through size classification or selective grinding could enhance recovery efficiencies by reducing processing costs and improving leach kinetics.

These findings point to the potential value of staged or targeted grinding strategies aimed at achieving sufficient monazite liberation while avoiding unnecessary overgrinding of gangue minerals. Such approaches could reduce energy consumption and minimize the production of slimes that might complicate downstream processing. Moreover, given that well-liberated monazite particles are associated with higher grades in mineralogically constrained grade-recovery analyses, focusing on selective liberation could improve the performance of physical separation methods like flotation or gravity separation prior to chemical treatment. This strategy aligns with the dual physical-chemical processing flowsheet indicated by the mineralogical and deportment data.

Surface chemistry and adsorption behavior

The XPS and ToF-SIMS analyses confirm the presence of rare earth-bearing phases on the mineral surfaces, with the detection of Dy and Nd indicating their adsorption onto clay particles. The presence of hydrated outer- and inner-sphere complexes, as indicated by studies in the literature¹²suggests that some REEs may be extracted via ion-exchange techniques, while others are more tightly bound within the mineral lattice.

These results highlight the complexity of REE occurrence at the surface level, with implications for the choice and efficiency of leaching agents. The detection of both outer- and inner-sphere complexes suggests that conventional ion-exchange leaching with ammonium or sodium salts could recover a significant portion of the adsorbed REEs, particularly those present as outer-sphere complexes. However, the presence of inner-sphere complexes indicates that a fraction of the REEs may require more aggressive or modified leaching conditions, such as elevated temperature or tailored reagents, to achieve satisfactory recovery. This mixed occurrence highlights the importance of designing leaching processes that account for both easily exchangeable and more tightly bound REE species.

Extraction and processing considerations

Given the mineralogical and chemical findings, an optimal processing strategy would likely involve a combination of physical and hydrometallurgical techniques. The high monazite content suggests that physical separation methods, such as flotation or gravity separation, may be effective for pre-concentration before chemical processing. Additionally, the presence of rare earths in the clay fraction implies that ion-exchange leaching could be utilized for selective rare earth recovery, particularly for Y and other heavy rare earths.

The dual occurrence of REEs, in both monazite and clay-hosted forms, calls for a hybrid flowsheet that integrates staged grinding, physical separation, and selective leaching. Early recovery of coarse monazite could reduce reagent consumption and limit the generation of leach residues, improving both economic and environmental performance. For the clay-hosted REEs, conventional ion-exchange leaching agents (e.g., ammonium sulfate) could be applied, but the mixed presence of inner- and outer-sphere complexes suggests that process conditions may need to be optimized for this specific material. Importantly, the decoupling of Y from the monazite fraction points to a potential loss of heavy REE value if processing focuses solely on monazite, highlighting the need for integrated recovery approaches.

Economic and environmental implications

The results indicate that while the deposit contains economically viable rare earth concentrations, the separation of monazite from the silicate matrix and the efficient recovery of adsorbed REEs remain key challenges. The findings also emphasize the potential of this deposit as an alternative rare earth source, reducing reliance on traditional suppliers and contributing to global diversification of REE supply chains.

Importantly, the coexistence of monazite and ion-adsorbing clays means that processing will require a balanced approach that integrates both physical and chemical methods. This increases processing complexity but also offers opportunities to design more resource-efficient flowsheets that reduce chemical consumption through early-stage physical separation of monazite. Such strategies could lower operational costs and mitigate environmental impacts related to reagent use and residue management. Additionally, developing effective extraction technologies for this deposit would enhance the economic viability of similar underexplored resources in the region, contributing to regional development and supply security.

By synthesizing data from various characterization techniques, this study provides a comprehensive understanding of the mineralogical and geochemical properties of the South American ionic clay deposit. The results highlight the deposit’s potential for REE extraction and inform future processing strategies aimed at optimizing recovery while minimizing environmental impact. These findings serve as a critical foundation for further metallurgical testing and economic feasibility assessments of this emerging REE resource.

Methodology

Materials and reagents

The ion-adsorption clay sample was obtained from an undisclosed location in South America. Fusion flux (ultrapure, Li₂B₄O₇/LiBO₂/LiBr 49.75 wt%/49.75 wt%/0.50 wt%, SCP Science) was used for alkali fusion. Certified multi-element standard stock solutions from Inorganic Ventures were used to calibrate the analytical instruments.

Sample characterization

Compositional characterization

The samples were fully dried and sieved to less than 300 μm and less than 50 μm. Compositional analysis was conducted using borate (alkali) fusion with a Claisse LeNeo fusion fluxer. The resulting compositions were analyzed for rare earth elements (REEs), Th, U, and bulk elements. ICP-MS with a PerkinElmer NexION 2000 instrument was used to measure the concentrations of REEs, Th, and U. Bulk element concentrations were determined using inductively coupled plasma optical emission spectrometry (ICP-OES) with a PerkinElmer Optima 8000 instrument.

Mineralogical analysis

XRD was conducted with a BRUKER AXS D8 Advance Diffractometer. Test Conditions included Co radiation, 35 kV, 40 mA; Detector: LYNXEYE_XE_T; Regular Scanning: Step: 0.02°, Step time: 0.5s, 2θ range: 6–80°; Clay Section Scanning: Step: 0.01°, Step time: 0.2s, 2θ range: 3–40°. Interpretations were conducted with PDF2/PDF4 powder diffraction databases issued by the International Center for Diffraction Data (ICDD). Diffrac Eva and Topas software. Detection Limit: 0.5-2%. Strongly dependent on crystallinity. The following software was used for analysis: DIFFTAC.EVA, release 2024 Patch 1, version 7.1.0.2, Bruker AXS, TOPAS, version 7.0.1.4 (https://www.bruker.com/en/products-and-solutions/diffractometers-and-x-ray-microscopes/x-ray-diffractometers/diffrac-suite-software/diffrac-eva.html).

SEM-EDX analyses

To prepare for SEM-EDX and Raman analyses, the powder samples were homogenized by stirring and encapsulated in epoxy to create polished cross-section mounts. These mounts were polished using diamond suspension then coated with a thin gold layer to reduce charging artifacts during analysis. A Hitachi SU3500 Variable Pressure SEM, coupled with an Oxford AZtec X-Max50 SDD X-ray analyzer, was used for the analyses. An accelerating voltage of 15 kV was chosen for the SEM-EDX analyses. EDX spectroscopy, capable of detecting elements from C to U within a few microns depth and with a detection limit of around 0.1 wt%, was utilized for semi-quantitative analysis, which was focused on identifying the REEs present in the feed and adsorbed samples. SEM imaging was employed using a back-scattered electron (BSE) detector to capture images that highlight elemental distribution based on atomic numbers, with heavier elements appearing brighter than lighter elements. Note that certain elements have overlapping EDX lines, such as Fe with Dy and Ce with La (along other REEs) which may interfere with the semi-quantitative data however, including certain elements regardless of the overlap was required for adequate spectral fitting.

Raman spectroscopy

For Raman spectroscopy, over 50 Raman spectra were gathered for the Feed sample using a Renishaw InVia Raman Spectrometer, equipped with a 633 nm wavelength laser. The analyses were conducted at 10% or 50% laser power, with each exposure lasting 20 s. A 50× objective lens focused the laser onto the sample surface, targeting analysis spot size of approximately 1 μm in diameter. Raman spectroscopy, which elucidates material structures through laser-induced molecular vibrations, was used to confirm the mineralogical composition of the particles. The generated Raman spectra were compared with multiple spectral databases to identify the mineral phases in the feed sample. The presented reference spectra were obtained from the Rruff database. Additional information is as follows:

Laser power: 50,000 mW.

Spectral range: 130–4000 cm^− 1 and 101–1328 cm^− 1 (which is 750 cm^− 1 centre).

All were centred at 750 cm^− 1, except for quartz, where the stated larger range was used.

Spectral resolution: ~ 1 cm^− 1.

Accumulations: 1.

Processing: Baseline subtraction was performed.

XPS surface characterization

To prepare for XPS analysis, the powder samples were pressed into adhesive tape, with the surrounding area masked using copper tape. The analyses were performed with a Kratos AXIS Supra X-ray photoelectron spectrometer, utilizing a monochromatic Al Kα source (15 mA, 15 kV). The instrument work function was calibrated to give a binding energy of 83.96 eV for the Au 4f_7/2 line for metallic gold, and the spectrometer dispersion was adjusted to give a binding energy of 932.62 eV for the Cu 2p_3/2 line of metallic copper. The Kratos charge neutralizer system was used on all specimens.

Survey scan analyses were conducted with an analysis area of 300 μm × 700 μm, a pass energy of 160 eV, and a step size of 1 eV. High-resolution analyses were performed with the same analysis area but with a pass energy of 20 eV and a step size of 0.1 eV. Spectra were charge corrected to the main line of the C 1s spectrum (adventitious carbon) set to 284.8 eV. Data were analyzed using CasaXPS software (version 2.3.26). Fitting parameter used to fit the Fe 2p and Mn 2p spectra have been reported elsewhere¹⁶.

ToF-SIMS surface characterization

The ionic clay sample received was separated into various size fractions, ranging from 250 to 500 μm to < 63 μm. Various swatches of the fine fraction were mounted on an indium substrate, whereas for the larger size fractions, random particles were selected under an optical stereomicroscope and mounted on the indium substrate prior to the time-of-flight secondary ion mass spectrometry (ToF-SIMS) analysis. The data of the most representative/relevant particles are presented in the paper.

The sample was analyzed by ToF-SIMS using an ION-TOF (GmbH) TOF-SIMS IV equipped with a BiMn cluster liquid metal ion source. The secondary ions of positive polarity were extracted from the sample surface at 128 × 128 pixels over a rastered area, mass separated and detected via a reflectron-type of time-of-flight analyzer. Current was set at ~ 0.3 pA. The data was acquired from 50 scans per area, 2 shots per pixel. A positive ion mass spectrum was initially calibrated by H⁺, C⁺, CH₃⁺, and C₃H₅⁺.

TESCAN TIMA-X characterization

TIMA-X is an acronym for TESCAN Integrated Mineral Analyzer. It is based on four Energy Dispersive X-Ray (EDX) silicon drift detectors (SDD) attached to a TESCAN MIRA (field-emission gun—FEG) platform which also includes a backscattered electron (BSE) and secondary electron (SE) detectors. The TIMA system utilizes both the EDX and BSE signals to identify minerals at each measurement point, and it is optimized to deal with rapidly acquired low-count spectra. These EDX (and BSE) spectra (and BSE data) are compared with entries in a mineral library on a first match principle to identify the mineral phase, where this mineral library is based on theoretical mineral/phase composition or created by the user based from BSE, X-ray spectral window counts, and/or ratios. TIMA-X has four X-ray analysis scanning modes to identify mineral/compounds including the High-Resolution Mapping (THRM) mode. The THRM collects a BSE signal and an X-ray spectrum at a set resolution to map the particles, and collect modal and textural information (i.e., liberation, exposure).

The mode of TIMA-X analysis used for this project was Dot Mapping (TDM). The TDM analysis mode uses a BSE grid at a predetermined pixel spacing to segment areas of homogenous BSE intensities and identifies the center of the greatest inscribed circle (similar to the point spectroscopy), it then creates a grid for the X-ray acquisition with the specified resolution spacing the same as the BSE. The X-ray data from zones of similar BSE and EDS signals are summed to produce a single higher quality spectrum for each final segment, which is used for the mineral identification. This analysis mode is good for modal mineralogy, grain size, and liberation analysis.

Approximately 450,000 particles and greater than 800,000 particles were analyzed from the two samples. A pixel spacing of 3 μm and dot spacing of 9 μm were used for the analysis. Beam energy was 25 kV, probe current 11.83 nA and beam current 11.44 nA.

The liberation and association characteristics of monazite, other REM/Nb, zircon and Fe-Oxides were examined. For the purposes of this analysis, particle liberation is defined based on 2D particle area percent. Particles are classified in the following groups (in descending order) based on mineral-of-interest area percent: pure (100% of the total particle volume), free (≥ 95%), and liberated (≥ 80%). The non-liberated grains have been classified according to association characteristics, where binary association groups refer to particle area percent greater than or equal to 95% of the two minerals or mineral groups. The complex groups refer to particles with ternary, quaternary, and greater mineral associations including the mineral of interest.

EMPA characterization

Wavelength dispersive analyses were performed using the JEOL JXA-8230 electron microprobe housed at Queen’s University, Kingston ON, Canada. For the analyses of clay minerals, accelerating potential was 15 kV, beam current was 10 nA, and the beam was defocused to 10 microns. Standards used in the analyses were albite (Si, Na), anorthite (Al), rutile (Ti), synthetic fayalite (Fe), rhodonite (Mn), olivine (Mg), wollastonite (Ca), sanbornite (Ba), adularia (K), synthetic fluorophlogopite (F), and tugtupite (Cl). For the analyses of monazite, the accelerating potential was 15 kV, beam current was 30 nA, and the beam was defocused to 3 microns. Standards used were synthetic rare earth and yttrium orthophosphates (REE, Y, P), wollastonite (Si, Ca), synthetic thorianite (Th), and synthetic uraninite (U). Where necessary to avoid interferences (Pr, Nd, Sm, Gd, Dy), Lβ lines of the rare earths were used instead of Lα lines. In addition, UMβ was used instead of UMα in order to minimize interference from Th; an overlap correction was applied. The dead time-corrected X-ray data were processed using the PAP atomic number and absorption corrections of Pouchou and Pichoir (1991)¹⁷, the characteristic fluorescence correction of Reed (1990)¹⁸, the continuum fluorescence correction of Springer (1971)¹⁹, and the mass absorption coefficients of Heinrich (1987)²⁰. Detection limits were determined according to the method of Williams (1987)²¹, but with matrix corrections applied.

LA-ICP-MS characterization

Samples were analyzed at the Queen’s Facility for Isotope Research (QFIR) at Queen’s University in Kingston ON, Canada, using a Thermo Scientific iCAP TQs inductively coupled plasma-mass spectrometer (ICP-MS) coupled with an Elemental Scientific/NWR-193 excimer ArF laser. Laser analyses used a 20 μm spot size, 10 Hz repetition rate, 70% energy (~ 5 J/cm²), 30 s ablation time, and 5 s laser warm up. Between ablations, a washout time of 20 s was used to ensure no material was carried over and to determine background counts. The ICP-MS was tuned before each analysis using the standard reference material NIST612 to achieve maximum sensitivity while minimizing oxide and doubly-charged ion production (²³⁸U¹⁶O < 1%, ¹³⁷Ba⁺⁺ < 6%). Certified reference materials NIST610, NIST612, NIST614, MASS-1, PTC-1b, CCu-1e, FeS4 and FeS5 were ablated bracketing each set of sample analyses for calibration and to account for instrumentation drift. Total analyte measurement time was 1.0 s with a dwell time of 0.013 s per mass. The Iolite software package (v4.8.2) was used for data normalization and drift correction.

Conclusions

In this study, a sample of ionic clay from the granite granitoid deposits in South America was thoroughly characterized. The sample matrix is predominantly composed of kaolinite, micas, and quartz, with other minerals present in minor quantities. Notably, the micas and clays are strongly associated with Fe-oxides and oxyhydroxides, as evidenced by the high Fe₂O₂ content.

The primary REE-bearing mineral is monazite, which is locally altered and exhibits chemical and textural heterogeneity, forming REE-phosphates. Liberation of monazite was measured at 32% for particles under 300 μm and 65% for particles under 50 μm. Monazite shows a coarser grain size of 86 μm in the < 300 μm fraction compared with 40 μm in the < 50 μm fraction. The data indicate the potential for recovering coarse monazite before re-grinding, allowing for improved liberation from middling particles.

Electron probe microanalysis (EPMA) confirms that monazite analyzed is chemically homogeneous. Additionally, EPMA and LA-ICP-MS analyses reveal that micas and clays host significant quantities of REE and Y, along with trace elements such as Sc and Ga. Elemental deportment analysis suggests that Y is decoupled from the other REE. Cerium is primarily hosted by monazite and its altered phases, while Y is mainly associated with micas and clays.

Although monazite recovery is expected to yield good REE grades and recoveries, the decoupling of Y indicates that it may be lost to the silicate fraction. This assumption should be verified through metallurgical test work, as the mineralogical data alone cannot substitute for metallurgical testing outcomes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

S.C. performed investigation; methodology; X.R.D.; Formal analysis; and writing-review and editing. T.G. performed TIMA-X analysis; formal analysis; and writing-review and editing. B.A. performed SEM-EDX; Raman spectroscopy; ToF-SIMS; formal analysis; and writing-review and editing. J.D.H. performed X.P.S.; Formal analysis; and writing-review and editing. G.A. performed conceptualization; funding acquisition; investigation; methodology; project administration; resources; supervision; validation; writing-original draft; and writing-review and editing.

Data availability

The datasets generated and/or analyzed during the current study are available in the Zenodo repository, https://zenodo.org/badge/DOI/10.5281/zenodo.15838933.svg. Please contact the corresponding author, Dr. Gisele Azimi (g.azimi@utoronto.ca) if you require additional information regarding the data.

Declarations

Competing interests

The authors declare no competing interests.