Low-Rate Bauxite Residue Application Controls Nickel Adsorption, Fractionation, and Mobility in Soils of Different Physicochemical Properties

Abstract

Soils in industrially influenced areas are often exposed to elevated nickel (Ni) levels due to metallurgical and alumina production activities. In this context, this study evaluated bauxite residue (BR) as an amendment to mitigate Ni availability and mobility in five agricultural soils from the Attica region, Greece, selected according to their pH values. Apart from the pH, soil properties were greatly varied. A very small amount of 1% BR (w/w) was incorporated into soils and batch adsorption experiments with eight Ni concentrations ranging between 1 and 90 mg Ni L⁻¹ were performed, followed by the direct application of the Tessier sequential fractionation scheme. BR addition increased the Ni adsorption capacity of soils, particularly those of low and neutral pH. BR increased the pH of acid soils, thus increasing the negatively charged sites on soil colloids. The Langmuir b_L constant provided indications of advanced Ni surface precipitation in the presence of BR. However, the desorption results suggested that, in addition to pH, Fe-Mn free oxides, noticeably those of amorphous form, controlled Ni fractionation in the studied soils. The mobility factor (MF) showed that the availability of Ni was restricted in all soil–BR mixtures. Yet, the distribution of Ni among the chemically active phases was different depending mainly on Fe-Mn free oxide content. Due to its high content in iron oxides, BR assisted the retention of Ni in soils with low Fe-Mn oxide concentration and increased significantly the Ni proportion extracted from the reducible phase. However, in soils richer in Fe-Mn oxides, BR incorporation resulted in enhanced oxidizable and residual fractions, suggesting stronger Ni binding. The results demonstrate that even a low BR application effectively enhances Ni immobilization by increasing adsorption capacity, shifting Ni toward more stable geochemical fractions, and significantly reducing its mobility, highlighting its potential as a sustainable soil amendment for Ni-contaminated soils.

1. Introduction

Soil contamination by heavy metals has emerged as a critical global issue, primarily due to its direct effects on ecosystem processes, agricultural productivity, and food safety. The intensification of industrial activities, mining operations, waste disposal practices, and prolonged agricultural inputs has resulted in the accumulation of potentially toxic elements in soils, frequently surpassing natural geochemical background concentrations [1]. In contrast to various organic contaminants, heavy metals are characterized by their persistence and resistance to degradation, allowing them to remain in soil systems for extended periods. Their accumulation can deteriorate soil quality, impair crop performance, and facilitate entry into the food chain, thereby posing serious risks to human and animal health and undermining long-term food security [2].

Nickel (Ni) is a widely distributed, potentially toxic element that occurs naturally in soils through the weathering of parent materials, particularly those derived from ultra-mafic and mafic rocks, while also being introduced through various anthropogenic sources [3,4]. Nickel concentrations in soils exhibit considerable variability, ranging from trace amounts to several thousand mg kg⁻¹ in soils developed from serpentine materials. From a geochemical standpoint, Ni is predominantly present in the divalent form (Ni²⁺), which represents its most stable and environmentally relevant oxidation state [4]. Although Ni is recognized as an essential micronutrient for plants and microorganisms at low concentrations, excessive levels may result in phytotoxic effects and the disruption of soil biological functions. The behavior and fate of Ni in soils are largely controlled by physicochemical parameters, including soil pH, clay mineral composition, organic matter content, and the presence of iron and manganese (hydr)oxides. Generally, Ni mobility de-creases as soil pH increases due to enhanced sorption, surface complexation, and precipitation processes, whereas acidic conditions tend to promote its solubility and bioavailability [5]. Furthermore, Ni can be partitioned among distinct soil fractions, namely exchangeable, carbonate-bound, oxide-associated, organically bound, and residual, each exhibiting varying degrees of stability and environmental risk [6].

Given the constraints associated with conventional remediation methods, increasing emphasis has been placed on in situ stabilization strategies aimed at limiting metal mobility rather than achieving complete removal. The use of soil amendments represents a cost-effective and environmentally sound approach for immobilizing heavy metals through mechanisms such as sorption, complexation, and precipitation [7,8]. A wide range of organic and inorganic amendments, including biochar, composts, metal oxides, and industrial by-products, have been evaluated, with their effectiveness largely dependent on soil characteristics and contaminant properties [9,10,11]. These amendments may modify soil pH, increase reactive surface area, and introduce additional binding phases, thereby reducing metal transfer to plants and groundwater. A thorough understanding of amendment and soil interactions is therefore essential for predicting the long-term performance and stability of immobilization strategies, particularly for elements such as Ni, that are highly sensitive to soil geochemical conditions [12,13].

Among inorganic amendments, bauxite residue (BR), commonly referred to as red mud, has gained increasing attention as a promising material for metal immobilization in soils. Bauxite residue is a by-product of alumina production via the Bayer process and is characterized by high alkalinity and elevated contents of iron oxides, aluminum oxides, and aluminosilicate minerals [14,15]. These properties confer a strong capacity to alter soil pH and provide reactive surfaces capable of binding metal cations. Previous research has demonstrated that BR application can significantly enhance Ni retention, particularly in acidic and neutral soils, mainly through pH elevation, increased negative surface charge, and the formation of stable Ni-containing phases [13,16]. Sequential extraction studies have indicated that BR addition promotes the redistribution of Ni from more labile fractions toward oxidizable and residual pools, suggesting stronger binding and reduced environmental mobility. Moreover, mineral phases present in BR, such as cancrinite and iron oxides, have been associated with surface precipitation and the structural incorporation of Ni, further supporting its immobilization potential [16]. However, bauxite residue is highly alkaline and may contain trace elements that pose environmental risks if improperly managed, even though it has potential benefits for metal immobilization. Excessive application rates may alter soil properties, affect microbial activity, and lead to secondary contamination. Therefore, careful assessment of application rates and site conditions is essential before field-scale application, and implementation strategies should always be critically evaluated and designed on a site-specific basis.

The objective of the present study was to assess the effectiveness of low-rate bauxite residue (BR) application as a soil amendment for regulating nickel (Ni) behavior in soils with contrasting physicochemical properties. Specifically, the study aimed to: (i) evaluate the effect of BR 1% (w/w) addition on Ni sorption capacity across different soil types; (ii) assess changes in Ni mobility following amendment incorporation; and (iii) elucidate Ni chemical partitioning and the soil phases responsible for its retention using the Tessier sequential extraction procedure. Through this integrated approach, the study enhances our understanding of amendment–soil–metal interactions while contributing to soil protection strategies and supporting the sustainable reuse of industrial by-products within a circular economy framework.

2. Materials and Methods

2.1. Soils and Bauxite Residue (BR)

Surface soil samples (0–20 cm depth) were collected from five agricultural locations in Attica, Greece, using a stainless-steel auger. The selected soils were collected from non-contaminated agricultural sites and were chosen to represent a range of key physicochemical properties, particularly soil pH, texture, cation exchange capacity, and Fe–Mn oxide contents, that strongly influence nickel adsorption, fractionation, and mobility, while also being highly representative of agricultural soils in xerothermic Mediterranean environments. This approach allowed for the evaluation of bauxite residue performance under controlled conditions and facilitated the comparison of amendment effects across different soil environments. After the collection, the samples were air-dried in the laboratory, gently ground, and passed through a 2 mm sieve before routine physicochemical analyses. Soil pH and electrical conductivity (EC) were measured electrometrically in a soil-to-water suspension (1:1, w/v) using an automated pH meter and a conductivity meter (Selecta 2000, J.P. Selecta S.A., Barcelona, Spain), following ref. [17]. Particle size distribution was determined using the hydrometer method [18]. Cation exchange capacity (CEC) was measured with neutral 1 M sodium acetate solution (Klute & Page, 1982) [17]. Soil organic matter (SOM) content was determined by the Walkley–Black wet oxidation method (Nelson & Sommers, 1982) [19]. Calcium carbonate equivalent (CCE) was measured using a digital calcimeter (FOG L; bd INVENTIONS, Athens, Greece), following Bilias & Barbayiannis (2017) [20]. Free and amorphous Fe and Mn oxides were determined after extraction using the dithionite–citrate–bicarbonate (DCB) method and the acid ammonium oxalate method under dark conditions, respectively [21,22].

Bauxite residue (BR) was obtained from the Mytilineos aluminum plant and was oven-dried at 105 °C overnight to remove moisture. Particle size distribution was measured using a laser diffraction particle size analyzer (Malvern Panalytical Ltd., Malvern, UK). Chemical composition was determined by energy-dispersive X-ray fluorescence using an Xepos instrument (SPECTRO A.I. GmbH, Ottawa, ON K2E 7L2 Canada). Mineralogical analysis of the dried, finely ground BR sample was performed by X-ray powder diffraction (XRPD) using a Bruker D8 Advance diffractometer (Bruker Corporation, Billerica, MA, USA). Measurements were conducted at a scanning rate of 5 min⁻¹ over a 2θ range of 10–75, with graphite-monochromatized Cu Kα radiation (λ = 0.15406 Å). Fourier-transform infrared (FTIR) spectroscopy was performed using a Spectrum 100 FTIR spectrometer (PerkinElmer) equipped with a Zn-Se crystal [13]. The FTIR spectra of the bauxite residue used in this study have been reported in detail in our previous work [13], using the same material and experimental conditions.

2.2. Nickel Sorption by Soils and Soil–BR Mixtures

Soil–BR mixtures were prepared by thoroughly mixing soils with BR at an application rate of 1% (w/w). Both untreated soils and soil–BR mixtures were equilibrated with aqueous Ni solutions at initial concentrations of 1, 5, 10, 20, 30, 50, 70, and 90 mg L⁻¹. All sorption experiments were conducted without adjustment of solution pH or ionic strength.

For each experiment, 1 g of adsorbent material (soil or soil–BR mixture) was placed in a 50 mL polyethylene centrifuge tube and mixed with 25 mL of Ni solution at the selected concentration. The suspensions were shaken on a reciprocal shaker at 120 rpm for 1 h at room temperature (20 ± 2 °C). The equilibration time was selected based on preliminary kinetic experiments. After shaking, samples were centrifuged at 4000 rpm for 10 min using a benchtop centrifuge (K241, Centurion Scientific, Chichester, UK). The supernatants were then filtered through Whatman No. 42 filter paper.

Nickel concentrations in the filtrates were measured using an atomic absorption spectrometer (AA240FS, Varian, Middelburg, The Netherlands). All adsorption experiments were performed in duplicate, and mean values were reported. The amount of Ni adsorbed at equilibrium (q_e) was calculated from the difference between the initial and equilibrium Ni concentrations, according to Equation (1) [23,24].

$${\mathrm{q}}_{\mathrm{e}}={\displaystyle \frac{ \left(Co-Ce\right)\times V}{m}}$$

(1)

2.3. Sequential Extraction of Ni and Mobility Factor (MF)

After completion of the sorption experiments, nickel fractionation in all Ni-treated soils and soil–BR mixtures was determined using a modified five-step sequential extraction scheme originally proposed by ref. [25]. The applied extraction procedure and operational conditions are summarized in

Table 1. Tessier sequential extraction scheme.

Step	Fraction	Reagents	Procedure
1	F1 Exchangeable	8 mL 1.0 M MgCl2 (pH = 7)	1 h shaking 120 rpm room temperature
2	F2 Carbonate-bound	8 mL 1 Μ sodium acetate (NaOAc) (pH = 5 with HOAc)	5 h shaking 120 rpm room temperature
3	F3 Reducible—bound to Fe/Mn oxides	20 mL 0.04 M NH2OH.HCl in 25% (v/v) HOAc (pH = 2.0)	6 h shaking 120 rpm T = 96 ± 0.5 °C
4	F4 Oxidizable—bound to organic matter and sulfides	3 mL 0.02 M HNO3 + 5 mL 30% H2O2 (pH = 2 with HNO3)	1 h digestion room temperature occasionally shaking
		3 mL 30% H2O2 (pH = 2)	1 h digestion T = 85 ± 5 °C occasionally shaking
		5 mL 3.2 M NH4OAc in 20% (v/v) HNO3 and the mixture diluted to 20 mL	3 h digestion T = 85 ± 5 °C occasionally shaking 30 min shaking 120 rpm room temperature
5	F5 Residual	9 mL conc. HCl + 3 mL conc. HNO3	Digested in a microwave 15 min until T = 200 °C 15 min at 200 °C

. According to the modified Tessier protocol, Ni was operationally divided into five fractions: exchangeable (F1), carbonate-bound (F2), reducible (F3), oxidizable (F4), and residual (F5). After each extraction step, identical procedures were followed for sample shaking, centrifugation, and filtration of the extractants. All extraction steps were carried out in duplicate, and Ni concentrations in the resulting solutions were determined using atomic absorption spectrometry (AA240FS, Varian, Middelburg, The Netherlands).

Nickel mobility was evaluated using the mobility factor (MF), which represents an index of metal binding strength and is widely used to assess the potential mobility of heavy metals in soils and sediments [26]. The MF was calculated from the sequential extraction results according to ref. [27], as shown in Equation (2):

$$\mathrm{MF} = {\displaystyle \frac{\mathrm{F}1+\mathrm{F}2}{\mathrm{F}1+\mathrm{F}2+\mathrm{F}3+\mathrm{F}4+\mathrm{F}5}} \%$$

(2)

where F1, F2, F3, F4, and F5 represent the Ni concentrations in the respective fractions obtained from the sequential extraction procedure.

2.4. Adsorption Isotherms

Sorption equilibrium isotherms provide essential physicochemical information for evaluating the performance and applicability of an adsorption system. Isotherm models are used to describe the surface properties and adsorption affinity of the adsorbent, as well as the relationship between the amount of adsorbate retained by the adsorbent and its equilibrium concentration in the aqueous phase. Although several isotherm models can be applied to describe adsorption processes in solution, the Langmuir [28] and Freundlich [29] isotherms are the most widely used.

2.4.1. Langmuir Adsorption Isotherm

The Langmuir adsorption isotherm assumes monolayer coverage of the adsorbent surface and negligible interactions between adsorbed species. The linearized Langmuir Equation (3) is expressed as follows:

$${\displaystyle \frac{C_e}{q_e}}= {\displaystyle \frac{1}{b_L{q}_m}}+ {\displaystyle \frac{Ce}{q_m}}$$

(3)

where $q_m$ (mg g⁻¹) is the maximum adsorption capacity of the adsorbent for Ni, and $b_L$ (L g⁻¹) is the Langmuir equilibrium constant related to the affinity of binding sites. The parameters $q_m$ and $b_L$ were calculated from the slope and intercept of the $C_e$/$q_e$ versus $C_e$ linear graph.

2.4.2. Freundlich Adsorption Isotherm

The Freundlich adsorption isotherm is an empirical model used to describe adsorption on heterogeneous surfaces and assumes the possibility of multi-layer adsorption [29]. The linearized form of the Freundlich equation (Equation (3)) is given by the following:

$$log{q}_e=log{K}_f+\left({\displaystyle \frac{1}{n}}\right)log{C}_e$$

(4)

where $K_f$(mg g⁻¹) is the Freundlich constant related to adsorption capacity, and $n$is an empirical parameter representing adsorption intensity. The values of $K_f$ and $n$ were determined from the slope and intercept of the linear plot of $\log q_e$versus $\log C_e$. Adsorption is considered favorable when the value of $n$lies within the range $1<n<10$.

2.5. Statistical Analysis

For the purpose of this study, the software STATISTICA for Windows (StatSoft, Inc., Tulsa, OK, USA, 1995, Version 10) was used and comparisons of mean values were performed using a T test.

3. Results and Discussion

3.1. Nickel Adsorption

3.1.1. Soils

Nickel adsorption was lower on acid soils (S1 and S3) than on neutral and alkaline soils (S2, S4 and S5) (Figure 1). The main physicochemical properties of the soils are presented in

Table 2. Selected physicochemical soil properties of studied soils. O.M.: organic matter, CEC: cation exchange capacity, Fed: free iron oxides, Feo: amorphous iron oxides, Mn_d: free manganese oxides, Mn_o: amorphous manganese oxides.

Soil Properties			Soil
	S1	S2	S3	S4	S5
Clay %	19.0	21.0	27.0	37.0	23.1
Silt %	28.0	37.50	20.0	20.0	38.4
Sand %	53.0	41.50	53.0	43.0	38.5
O.M. %	1.11	0.99	1.53	1.86	1.56
CEC (cmolc kg−1)	13.39	18.78	22.17	30.26	18.35
pH (1:1)	5.41	7.58	6.03	6.89	7.56
CaCO3 eq.	-	1.6	-	-	2.4
Fed %	2.19	1.13	2.91	2.84	0.95
Feo%	0.14	0.09	0.33	0.37	0.12
Mnd %	0.19	0.04	0.17	0.15	0.04
Mno %	0.31	0.04	0.21	0.18	0.03
Soil–BR pH (1:1)	6.82	7.61	6.70	7.20	7.55

. It is well-accepted that metal adsorption on soil increases as soil pH increases. Ref. [30] reports that metal adsorption on polluted soils was higher for the alkaline than for the acid soils, and ref. [31] concluded that alkaline tropical soils adsorbed higher amounts of Ni compared to acid soils. A possible mechanism is that, at higher alkaline pH levels, Ni sorption is enhanced by the precipitation of Ni²⁺ and NiOH⁺ with hydroxides [32]. Soil characteristics other than pH can also produce this effect.

Clay, organic carbon and iron oxide content (particularly those of low crystallinity) and redox conditions all affect metal adsorption on soils. Ref. [33] found that higher organic matter, oxide, and clay content in soil enhanced the quantity of adsorbed metals. However, no clear relation between these soil properties and adsorption capacity of the studied soils was observed.

3.1.2. Soils and Bauxite Residue

Bauxite residue incorporation in the studied soils increased Ni adsorption. Compared to soils, adsorbed Ni concentration was constantly higher for all soil–BR mixtures and over the whole range of Ni loadings (Figure 1). In

Table 3. Range of $q_e$ values in mg kg⁻¹ for soils and soil–BR mixtures (N = 8). Mean $q_e$ concentration values in parenthesis.

	Soil	Soil–BR
S1	21.20–688.50 (425.01)	27.38–1785.00 (719.78)
S2	23.08–1752.50 (730.48)	27.53–2101.75 (799.47)
S3	20.83–1227.50 (593.51)	26.78–1736.25 (686.48)
S4	20.43–1863.00 (768.46)	27.60–2271.00 (836.19)
S5	22.58–1859.50 (762.32)	28.23–2049.00 (801.96)

, the range and the average values of adsorbed Ni at equilibrium ($q_e$) on soils and soil–BR mixtures are presented. Maximum $q_e$ values were obtained for the highest Ni loading. The maximum adsorbed amount of Ni after the addition of BR increased impressively, by almost three times in soil S1, by 30% in soil S3 and between 9 and 17% in soils S2, S4 and S5. Furthermore, for all soil–BR mixtures, higher qe values were also obtained for the lowest Ni initial solution concentration of 1 mg L⁻¹, pointing to the positive impact of bauxite residue on the Ni retention capability of the soils at very low Ni loadings (Table 3). Though BR addition increased Ni adsorption on all soils, the BR effect was more pronounced for acid (S1 and S3) than for neutral and alkaline soils (S2, S4 and S5). Due to its highly alkaline nature (pH = 10.68), BR incorporation into soils increased the pH of acid soils (Table 2). In particular, pH increased by 1.4 and 0.7 units in S1-BR and S3-BR mixtures, respectively, compared to S1 and S3 pH values. A small pH rise of 0.3 units was also observed for the neutral S4 soil with the addition of BR, whereas the pH of alkaline soils was not affected. Though the amount of bauxite residue added to soils was very small, it was able to increase the pH of acid soils, leading to the occurrence of new negatively charged sites on soil colloids that attracted the freshly added Ni. Higher pH also led to lower competition between Ni ions and H⁺ for adsorption sites and altered the hydrolysis state of Ni in the solution [34]. Many authors report on the effect of BR addition on soil pH. Ref. [35] recorded an increase in soil pH from 0.1 to 4.8 pH units and [36] Lee et al. (2011) observed that the addition of 5% w/w BR raised soil pH from 5.2 to 9.4. Brown et al., (2005), Gray et al., (2006) [37,38] and Lombi et al. (2002) [39] concluded that the addition of 2% bauxite residue increased soil pH by between 0.7 and 2.3 units. The application of 5% BR in five arable soils increased their pH by 0.4 to 1.6 units in the short term [40]. In a recent study conducted by Zafeiriou et al. (2025) [13], the incorporation of 20% BR in six soils increased their pH by 1.6 to 2.8 units; five of these soils were used in the present work.

Apart from the increased Ni adsorption due to the positive pH effect on the acid S1 and S3 soils, bauxite residue, even at the low rate of 1%, offered active sites for Ni sorption, further increasing the adsorptive capacity of all soils (Figure 1, Table 3). The FTIR spectra of the bauxite residue used in this study were altered by the addition of 5, 40 and 90 mg Ni L⁻¹ [13]. The observed shifts were attributed to hydrogen bond formation and the precipitation of Ni(OH)₂ on the BR surface, and Ni sorption on the carbonate phase, on Si bearing minerals and on cancrinite. Ref. [41] suggested that, when sources of accessible Al are restricted and soluble Ni compounds are added rapidly to soils, Ni precipitation occurs in forms with characteristics like Ni(OH)₂. Furthermore, refs. [32,42] report that mixed Ni–Al hydroxide surface precipitates are formed in the presence of soluble Al sources and a-Ni(OH)₂ is produced when Ni is sorbed on non-Al-bearing minerals (talc, silica).

3.1.3. Adsorption Isotherms

As indicated by the significant regression coefficients, experimental data for both soil and soil–BR mixtures were sufficiently explained by the Langmuir and Freundlich equations (

Table 4. The parameters of Langmuir and Freundlich isotherm models. ** p < 0.01, *** p < 0.001.

Soil		Langmuir				Freundlich
	$\max q_e$ mg g−1	$q_m$ mg g−1	$b_L$ L g−1	$R^2$	$R_L$	n	$K_f$	$R^2$
S1	0.69	0.74	0.37	0.99 ***		2.08	8.16	0.81 **
S1-BR	1.79	2.10	0.12	0.97 ***		1.42	9.59	0.96 ***
S2	1.75	2.02	0.25	0.95 ***		1.49	12.19	0.89 **
S2-BR	2.21	2.65	0.13	0.94 ***		1.32	11.05	0.96 ***
S3	1.23	1.42	0.17	0.97 ***		1.59	9.04	0.87 **
S3-BR	1.74	1.83	0.13	0.93 ***		1.54	9.33	0.93 ***
S4	1.86	2.65	0.17	0.95 ***		1.20	12.18	0.96 ***
S4-BR	2.27	4.21	0.07	0.89 **		1.17	11.06	0.99 ***
S5	1.86	2.28	0.26	0.94 ***		1.37	13.00	0.90 ***
S5-BR	2.05	2.06	0.26	0.97 ***		1.50	12.52	0.93 ***

). For either the soils or the soil–BR mixtures, the maximum experimental Ni adsorption concentration ($q_e$) and that which was predicted by the Langmuir equation maximum Ni capacity were very close, pointing to the applicability of the Langmuir model on the experimental adsorption data. The only exception noticed is for the S4-BR mixture, that showed a much higher $q_m$ value compared to the experimental maximum adsorption value $q_e$ and the weaker fit of the Langmuir model according to the $R^2$ value (Table 4).

BR addition to soils almost tripled and doubled the calculated maximum adsorptive capacity ($q_m$) for S1 and S4 soils, respectively, whereas it considerably increased the $q_m$ values for soils S2 and S3. The higher $q_m$ values can be attributed to the considerable increase in pH of the S1 and S3 soils due to the incorporation of BR, which created additional negative charge on soil colloids, reinforcing Ni adsorption. Soil S4 is a neutral soil and the small pH increase caused by the addition of BR appears to not be able to completely explain the impressive $q_m$ rise. Nevertheless, compared to the other soils, S4 has the highest clay, organic carbon and amorphous oxide content, colloids with pH-dependent charges (i.e., increasing negative surface charges as the pH increases), features that might explain the observed qm rise (Table 2).

The $b_L$ Langmuir constant decreased in almost all soil–BR mixtures, suggesting that the addition of BR affected the affinity of binding sites and the rate of adsorption [43]. This probably points to alterations in the adsorption stages that, when related to the higher adsorption capacity of soil–BR mixtures, provide indications of increased precipitation of Ni, most likely in the form of Ni hydroxides. As supported by refs. [44,45], higher values of the Langmuir affinity constant $b_L$ indicate stronger binding of the adsorbate on the surface of the adsorbent that prevents a metal from being susceptible to precipitation or, inversely, lower $b_L$ values imply higher metal availability to participate in other processes as precipitation.

According to the calculated K_f and n Freundlich constants, the addition of BR into soils did not significantly influence the adsorption of Ni. The K_f values, as a measure of sorption capacity, did not actually differ between soils and soil–BR mixtures, whereas n values were slightly lower for most soil–BR mixtures, indicating a limited effect of BR addition on the adsorptive capacity of the soils. However, for both soils and soil–BR mixtures, n values > 1 were observed, showing a favorable adsorption process. The Freundlich equation describes Ni sorption on the S4-BR mixture better than the Langmuir equation, suggesting that a heterogeneous, multi-layer process involving chemical and physical adsorption occurred.

3.2. Sequential Extraction

Ni percentage desorption from the five fractions of the Tessier sequential extraction protocol for all soils and soil–BR mixtures are presented in Figure 2. Mean percentage desorption from the five fractions for all soils and soil–BR mixtures are included in

Table 5. Mean fraction percentage values for soils and soil–BR mixtures. Different lowercase letters indicate significant differences within each soil and fraction (N = 8).

	F1%	F2%	F3%	F4%	F5%
S1	50.79 a (p < 0.01)	6.64 a	23.21 a	3.38 a	15.97 b (p < 0.05)
S1-BR	25.88 b	7.38 a	28.00 a	7.38 a	30.75 a
S2	25.57 a (p < 0.01)	6.77 a (p < 0.001)	14.19 b (p < 0.001)	1.32 b (p < 0.001)	52.14 a
S2-BR	4.88 b	1.50 b	37.50 a	19.00 a	36.63 a
S3	30.62 a	10.53 a (p < 0.05)	42.80 a	3.60 b	11.92 b
S3-BR	26.38 a	6.25 b	36.25 a	6.88 a (p < 0.01)	24.25 a (p < 0.05)
S4	11.71 a	9.50 a	48.85 a	8.03 b	21.89 a
S4-BR	12.38 a	5.00 a	48.25 a	10.75 a (p < 0.01)	23.62 a
S5	26.46 a (p < 0.01)	6.83 a (p < 0.01)	2.49 b	1.78 b	62.44 a (p < 0.01)
S5-BR	6.88 b	2.12 b	39.63 a (p < 0.001)	13.50 a (p < 0.001)	38.00 b

.

3.2.1. Soils

In all soils, the exchangeable and the acid-soluble fractions increased by increasing Ni solution concentration, whereas the sum of the reducible, the oxidizable and the residual fractions decreased (Figure 2). At low Ni loadings, most of the added Ni sorbed on high-affinity sites of soil colloids that limit Ni mobility in the soil ecosystem. The progressive saturation of these sites by higher Ni loadings led to greater amounts of added Ni remaining in relatively soluble forms that were extracted in the F1 and F2 fractions. Considering that the sequential extraction procedure was applied immediately after the batch adsorption experiments, there was no time for aging processes to occur, which would otherwise promote stronger binding of Ni on soil colloids, diffusion into mineral structures, and gradual transformation into more stable forms over time [45]. As a result, the lack of aging is more evident at higher Ni additions. Nevertheless, different patterns of Ni distribution among the more stable fractions were still observed between the studied soils. For soils S2 and S5, characterized by a low content of free and amorphous Fe oxides, Ni is predominately associated with the residual fraction, indicating its incorporation within primary minerals or strongly bound lattice sites, rather than association with reactive secondary phases (Table 2, Figure 2a,e). In contrast, soils S3 and S4, which exhibit elevated Fed and Feo content, show a clear dominance of the reducible fraction, suggesting that Ni is mainly retained through sorption and/or co-precipitation with Fe (hydr)oxides (Table 2, Figure 2d,c). This pattern supports the role of Fe oxides as key geochemical carriers controlling Ni partitioning, in agreement with Liu et al. (2025) [46], who demonstrated preferential incorporation of Ni isotopes into Fe (hydr)oxides in basalt-derived paddy-soils. Mn oxides may have also played an important role on Ni retention through strong surface complexation mechanisms, explaining the elevated Ni concentrations extracted in the reducible fraction (Table 2), (Figure 2a). The combined influence of Fe and Mn oxides on Ni binding is consistent with the findings of Ettler et al. (2025) [47], who demonstrated that Ni Fe oxyhydroxides and Mn oxides are important carriers for Ni in soils developed on Ni-laterites. The results of Fernández-Martínez et al. (2024) [48] showed that the presence of Fe and Mn oxides regulate the extractability of heavy metals in soils. Although Laha et al. (2024) [49], applying the Tessier sequential extraction procedure on contaminated playground soils in India, reported a dominant residual Ni fraction (53%), the considerable proportions of Ni recovered from the reducible (27.6%) and oxidizable (14.5%) fractions further highlight the role of Fe-Mn oxides as dynamic sorbents controlling Ni partitioning. The presence of Ni found in the oxidizable fraction could have originated from the incomplete dissolution of Fe-Mn oxide-bound Ni during the reducible extraction step, as hydroxylamine hydrochloride, used as an extractant in the third step of the Tessier scheme, does not fully recover metals associated with these phases [50,51]. It should be noted that the sequential extraction procedure applied in this study provides operationally defined fractions that depend on the selectivity of the reagents used rather than strictly representing discrete mineralogical phases. Partial overlap between fractions, re-adsorption during extraction, and redistribution of metals may occur, which can influence the apparent metal speciation. The Tessier scheme was selected because it includes a specific carbonate-bound fraction, which was particularly relevant given that soil pH was a key criterion in soil selection and that carbonate-related processes were expected to influence Ni behavior. In this context, the Tessier method was considered more appropriate than alternative schemes (e.g., BCR) for addressing the objectives of the present work. Moreover, within the scope of this study, the method serves as a reliable comparative tool, since all soils and treatments were analyzed under identical experimental conditions, allowing consistent evaluation of relative changes in Ni partitioning. Therefore, the present fractionation results should be interpreted in conjunction with adsorption behavior, soil properties, and complementary evidence.

3.2.2. Soils and Bauxite Residue

Bauxite residue incorporation in the studied soils affected Ni extraction from the five chemically active phases, indicating a shift in the dominant retention mechanisms. Across all soils, BR addition promoted the redistribution of Ni from the more labile fractions (F1 and F2) toward less mobile forms (F3, F4, and F5), reflecting enhanced immobilization processes (Figure 2, Table 5).

As indicated from the mean percentage Ni desorption values (Table 5), in the acidic soils S1 and S3, BR addition significantly enhanced Ni bound with the residual fraction, indicating stronger fixation through incorporation into stable mineral phases. In S1, the pronounced decrease in exchangeable Ni suggests rapid surface complexation and subsequent stabilization of Ni in the presence of BR. In S3, the decrease in acid-soluble Ni coupled with an increase in oxidizable Ni implies a progressive transfer of Ni from carbonate- or weakly sorbed forms to associations with organic matter and oxide-related phases. BR addition in the alkaline S2 and S5 soils significantly enhanced Ni retention in the reducible and oxidizable fractions while simultaneously lowering the Ni amount extracted from the exchangeable and the acid-soluble fractions. This redistribution indicates that newly introduced reactive surfaces from BR, particularly Fe–Mn oxides, acted as effective sorbents for Ni, favoring adsorption and co-precipitation mechanisms. Similarly, in the neutral S4 soil, BR addition significantly increased Ni recovered from the oxidizable fraction, suggesting stabilization through association with organic matter and/or oxide-bound phases.

The observed fractionation patterns are mechanistically consistent with the chemical composition of the bauxite residue used in this study, which is dominated by iron oxides (46.7%), mainly goethite, hematite and rutile [13]. Nickel, like all metals, shows a strong tendency to sorb onto iron oxides [52]. Therefore, enriching soils with iron oxides, higher desorption of Ni is expected from the reducible F3 fraction, which is associated with Fe-Mn oxides.

Similar immobilization effects following BR application have been reported by Pavel et al. (2015) [53] for Zn, Cd, and Pb in heavily contaminated soils. The obtained higher oxidizable and residual Ni concentrations can be attributed to the limited effectiveness of hydroxylamine to fully dissolve Fe oxides, especially the well-crystallized phases [50,51]. Probably, the quantity of Ni that was not extracted from the Fe oxides was recovered in the oxidizable and residual fractions of the Tessier scheme.

Moreover, the concentration of Fe-Mn oxides and particularly that of amorphous Fe oxides in soils substantially controlled the impact of bauxite residue addition on Ni fractionation. Among the different iron and manganese oxides present in soils, the non-crystallized amorphous iron oxides (Feo) are considered the most chemically active due to their large surfaces that provide high-affinity sites for metal immobilization [54]. Soils S3 and S4, characterized by high Fed and Feo contents, showed no significant change in the reducible fraction following BR addition, indicating that their oxide sorption capacity was already near saturation. In contrast, S2 and S5 soils, with low Fe-Mn oxide contents, exhibited a pronounced increase in the reducible Ni fraction upon BR amendment, reflecting the introduction of new reactive Fe oxide surfaces (Figure 2, Table 5).

Overall, the results of the present study indicate that, in soils already enriched in Fe-Mn oxides, 1% BR addition on freshly added Ni fractionation is limited and cannot be detected in the reducible fraction but is expressed mainly through redistribution into oxidizable and residual factions, whereas in oxide-poor soils, BR addition preferentially enhances Ni association with the reducible fraction.

Beyond Fe-Mn oxide content, the Al- and Ca bearing phases present in the BR, such as diaspore, boehmite, and calcite [13], may have further contributed to Ni immobilization through additional sorption and precipitation mechanisms, consistent with previous findings highlighting the role of BR mineralogy and Al, Fe, and Ca contents in metal stabilization [55,56].

3.3. Mobility Factor

In Figure 3, the mobility factors for the soils and for the soil–BR mixtures over the whole range of Ni initial solution concentrations are presented. Mobility factors are related to the mobility and bioavailability of metals in soils; high MF values indicate high mobility and bioavailability [57,58]. The mobility factors of all soils increased as the Ni loading increased. Higher MFs were obtained for the acid soils S1 and S3 and were lower for the alkaline S2 and S5 soils. According to refs. [59,60], MF values < 1% pose no risk to the environment, MF values between 1 and 10% suggest a low risk, MF ranges from 10 to 30% and from 30 to 50% point to medium and high risk, respectively, and an MF > 50% indicates very high risk. Following this classification, for Ni loadings > 5 mg L⁻¹, soluble forms of Ni in soil S1 represent >45%, showing high to very high risk for the environment. For S3 soil, high to very high risk due to Ni mobility was observed for initial Ni solution concentration > 20 mg L⁻¹. For the alkaline soils S2 and S5, high risk can be expected for Ni loadings > 30 mg L⁻¹. The lowest MFs, < 37% even for the highest amount of added Ni, were calculated for the neutral soil S4. This may be due to the substantial iron oxide content in this soil, especially that of the amorphous forms, that promoted Ni immobilization [54] (Table 2).

The addition of BR in the studied soils reduced MFs, yet markedly only for the alkaline S2 and S5 soils. This remarkable reduction in Ni soluble forms in S2-BR and S5-BR mixtures, portrayed by MF values < 15% at the most, indicates that, in soils with very low Fe and Mn oxide content, the presence of BR can actively reduce Ni mobility in the soil ecosystem, mitigating potential hazards posed by Ni effluents to the environment and human health. However, the incorporation of 1% BR in the acid S1 and S3 soils though decreased MFs did not seem adequate to effectively restrict Ni mobility in the soil environment, particularly for Ni loadings > 10 mg L⁻¹.

4. Conclusions

A very low rate of 1% BR incorporation in the studied soils that received various Ni loadings increased Ni adsorption and decreased Ni mobility. The emerged mechanisms involved because of the BR addition are the higher pH values of acid and neutral soils that raised the pH-dependent negative charge promoting Ni adsorption, and the impact of the initial Fe-Mn free oxide content in the soils that affected the distribution of adsorbed Ni among the soil phases. Compared to soils, lower mobility of Ni was recorded for all soil–BR mixtures but, while BR addition impressively increased Ni binding on the reducible phase of soils with low Fe-Mn oxide concentration, in soils with higher Fe-Mn oxide content, the reducible fraction did not alter and the oxidizable and/or the residual fractions were increased, pointing to stronger Ni retention. The results of this study strongly support that BR can be considered as an effective soil amendment to prevent adverse Ni phenomena in soil ecosystems, by reducing Ni mobility and toxicity in soils, therefore enhancing soil health and preventing Ni transfer to the food chain. Future research should expand the present work by investigating Ni adsorption and fractionation in a wider range of soils and soil–BR mixtures, using bauxite residues of different chemical compositions and application rates. Although this study demonstrates the effectiveness of low-rate BR application under controlled laboratory conditions, further research is required to evaluate its practical feasibility in real field environments. Such studies should include pot- and field-scale experiments, long-term monitoring of soil and crop responses, the assessment of potential microelement leaching risks, and comprehensive economic analyses incorporating industrial management costs and potential benefits of by-product reuse. These integrated investigations will provide a more complete understanding of the agronomic, environmental, and economic sustainability of bauxite residue application in soil remediation.