Rapid Removal of Cr(VI) from Aqueous Solution Using Polycationic/Di-Metallic Adsorbent Synthesized Using Fe 3+ /Al 3+ Recovered from Real Acid Mine Drainage

: The mining of valuable minerals from wastewater streams is attractive as it promotes a circular economy, wastewater beneﬁciation, and valorisation. To this end, the current study evalu-ated the rapid removal of aqueous Cr(VI) by polycationic/di-metallic Fe/Al (PDFe/Al) adsorbent recovered from real acid mine drainage (AMD). Optimal conditions for Cr(VI) removal were 50 mg/L initial Cr(VI), 3 g PDFe/Al, initial pH = 3, 180 min equilibration time and temperature = 45 ◦ C. Optimal conditions resulted in ≥ 95% removal of Cr(VI), and a maximum adsorption capacity of Q = 6.90 mg/g. Adsorption kinetics followed a two-phase pseudo-ﬁrst-order behaviour, i.e., a fast initial Cr(VI) removal (likely due to fast initial adsorption) followed by a slower secondary Cr(VI) removal (likely from Cr(VI) to Cr(III) reduction on the surface). More than 90% of adsorbed Cr(VI) could be recovered after ﬁve adsorption–desorption cycles. A reaction mechanism involving a rapid adsorption onto at least two distinct surfaces followed by slower in situ Cr(VI) reduction, as well as adsorption-induced internal surface strains and consequent internal surface area magniﬁcation, was proposed. This study demonstrated a rapid, effective, and economical application of PDFe/Al recovered from bona ﬁde AMD to treat Cr(VI)-contaminated wastewater.


Introduction
Despite the socio-economic benefits provided by the gold and coal industries, the mining of these commodities is notorious for inducing significant environmental degradation [1,2]. Mining entails the excavation of large quantities of rock to obtain the targeted minerals or minerals. During the extraction of mineral resources, hard rock (such as gold) or soft rock (such as coal) is exposed to air which typically leads to the oxidation of associated minerals. Generally, coal and gold are associated with sulphide-bearing minerals, e.g., FeS, FeAsS, ZnS, CuS, and NiS amongst others, resulting in the production of metalliferous acidic drainage rich in sulphates [3,4]. Due to the acidic nature of this wastewater stream, minerals in the surrounding geology leach into the surrounding body of water hence increasing the electrical conductivity and total dissolved solids. Acid mine drainage (AMD) or acid rock drainage (ARD) generally contains significant quantities of Al, Fe, Mn, and sulphate as major contaminants. However, the presence of minor, but not insignificant, levels of toxic and hazardous heavy metals, radionuclides, metalloids, oxyanions, and rare K 2 Cr 2 O 7 salt in ultrapure water (18.2 MΩ-cm) in a 1000 mL volumetric flask. The container was then filled to the specified level (mark). Thereafter, serial dilutions were made from the prepared working solution.

Synthesis of PDFe/Al from Authentic AMD
The PDFe/Al was synthesized using the method previous described [1,2]. A known AMD volume was reacted with 30% NaOH at a pH = 4.5 to selectively precipitate trivalent Fe and Al as -OOH compounds and the mixture was stirred for 30 min at room temperature with an overhead stirrer. Subsequently, the mixture was heated to 100 • C while stirring. The precipitates were vacuum filtered (Whatman ® Grade 40 ash-less) and dried. The recovered material was vibratory-ball-milled at 700 rpm and calcined at 800 • C. The milled samples were sieved to a maximum size of 32 µm and stored in a plastic "zip-lock" bag until use.

Optimisation Studies
The synthesised Fe/Al di-metal composite was then used for the removal of Cr(VI) from aqueous solutions. Optimised parameters include initial Cr(VI) concentration (mg/L), adsorbent dose (g), agitation time (g), temperature ( • C), and initial solution pH. In all experiments, 250 mL of Cr(VI) rich solutions was added to 500 mL volumetric flasks. All experiments were performed in triplicate for quality control and quality assurance, and the data were reported as mean ± standard deviations. The impacts of various parameters on the adsorption process were investigated, and the results are summarized in Table 1. The various initial concentration were obtained by diluting a concentrated stock solution of 1000 mg·L −1 Cr(VI) as shown in Table 1. The effect of agitation time was determined by measuring the Cr(VI) concentrations at various time. The influence of initial solution pH was investigated by changing the pH with 0.1 M NaOH/0.1 HNO 3 . Batch adsorption studies were carried out in a thermal shaker/incubator at various temperatures.

Characterisation of the Feedstock and Product Minerals
Inductively Coupled Plasma Mass Spectrometry (ICP-MS) was used to assess the fate of Cr(VI) in aqueous samples (7500ce, Agilent, Alpharetta, GA, USA).

Point of Zero Charge (PZC)
The point of zero charge (PZC) of the PDFe/Al was determined using a method proposed by Smičiklas et al. (2000) [31]. 0.1 g of the PDFe/Al were added to each of nine flasks containing 50 mL 0.1 M KNO 3 solutions adjusted to pH 2-10 (using 0.1 M HNO 3 and/or NaOH) and left to equilibrate for 24 h at ± 25 • C. Subsequently, the suspensions were filtered, and the pH of the filtrates determined.
2.6. PDFe/Al Adsorption Capacity and Removal Efficiency 2.6.1. Adsorption Capacity The adsorption capacities of PDFe/Al were determined using Equation (1) [32]: where Q (mg·g −1 ) is the adsorption capacity; C 0 (mg·L −1 ) is the initial concentration of Cr(VI), C (mg·L −1 ) is the measured concentration of Cr(VI), respectively; V (L) is the volume of the Cr(VI) solution; and m (g) is the dosage of PDFe/Al.

Percentage Removal
The removal efficiency of Cr(VI) by PDFe/Al was determined using Equation (2): where %R e is the removal efficiency of the PDFe/Al; C 0 is the initial Cr(VI) concentration (mg/L); C e is the equilibrium Cr(VI) concentration (mg·L −1 ).

Regeneration Study
Kumari et al. (2006) [32] described a method for studying the regeneration of PDFe/Als. 250 mL of a 150 mg·L −1 Cr(VI) solution was treated with 1 g of PDFe/Al for 90 min in a batch experiment. The adsobent was centrifuged to separate it from the supernatant and the recovered and washed five times with 250 mL ultra-pure water (to remove residual Cr(VI)) and dried. 250 mL of 0.1 M HNO 3 was added to the dried sample at room temperature, the HNO 3 extract was collected and tested for Cr(VI) ions. Equation (3) was used to calculate the regeneration percentage: where C des (mg·L −1 ) is the concentration of Cr(VI) ions in the desorption eluent; C o (mg·L −1 ) is the initial concentration of Cr(VI) ions.

Results
3.1. Characterisation of PDFe/Al before and after Cr(VI) Adsorption 3.1.1. FTIR Analysis Figure 1 depicts the functional groups of the PDFe/Al before and after Cr(VI) adsorption using a Fourier Transform Infrared Spectrometer (FTIR). Table 2 summarises the peak positions for raw PDFe/Al and Cr-PDFe/Al. For the raw PDFe/Al and Cr-PDFe/Al, significant -OH stretching was measured between 4000 and 3500 cm −1 . At circa 1630 cm −1 and 1100 cm −1 , HOH stretching is observed [33]. After Cr(VI) adsorption, a change in the stretching of HOH group is observed as a shift of wave band 1096.3 to 1103.2 cm −1 [34]. The characteristic peaks for aluminium oxides at~535 cm −1 [35]~606 cm −1 and~705 cm −1 [36], and iron oxides at~606 cm −1 and~795 cm −1 [36,37] provide evidence for the successful synthesis of Fe/Al oxides. In addition, a new absorption band at, corresponding to O-Cr-O, is observed at 610.9 cm −1 [38,39], confirming the successful adsorption of Cr(VI) to the adsorbent surface.
stretching of HOH group is observed as a shift of wave band 1096.3 to 1103.2 cm −1 [3 The characteristic peaks for aluminium oxides at ~535 cm −1 [35] ~606 cm −1 and ~705 cm [36], and iron oxides at ~606 cm −1 and ~795 cm −1 [36,37] provide evidence for the su cessful synthesis of Fe/Al oxides. In addition, a new absorption band at, corresponding O-Cr-O, is observed at 610.9 cm −1 [38,39], confirming the successful adsorption of Cr(V to the adsorbent surface.   Figure 2 depicts the XRD analyses of the PDFe/Al before and after Cr(VI) adsor tion. Iron (Fe) is observed to be the dominant species in the PDFe/Al in the form of go thite/iron(III) oxide hydroxide (FeO(OH)) with clear diffraction peaks showing the cry tallinity of the adsorbent and some amorphousness. In addition, the presence of a minium in the form of aluminium oxide (Al2O3) is observed, thus confirming the co posite nature of the material.
Significant peaks between 2θ ≈ 35° to 50° are observed to have increased after Cr(V adsorption. This is attributed to the diffusion of Cr(VI) ions into the pores of the PDFe/        Table 3 shows the EDS and XRF mole percentage results for raw PDFe/Al and Cr-PDFe/Al. Characterisation before Cr(VI) adsorption shows the presence of Fe and A in the synthesised adsorbent which confirm the dimetallic nature of the material.
The presence of Fe and S shows that AMD was generated from pyrite oxidation. Th presence of Si, Ca, Mg, and other trace elements in the material are impurities that re sulted from co-precipitation from AMD. After Cr(VI) adsorption, it is observed tha chromium is the only element showing a significant difference between the PDFe/Al and Cr-PDFe/Al. Overall, the composition of Fe, Al, O and C was observed to be preserved during adsorption, therefore demonstrating the chemically stability of the material. Figure 4 shows the relationship between 2θ values for PDFe/Al before and afte Cr(VI) adsorption. Significant peaks between 2θ ≈ 35 • to 50 • are observed to have increased after Cr(VI) adsorption. This is attributed to the diffusion of Cr(VI) ions into the pores of the PDFe/Al adsorbent through chemisorption, with the subsequent reduction to Cr(III) (as confirmed by the presence of CrOOH) [42]. Figure 3 shows the diffractogram of the PDFe/Al after Cr(VI) adsorption.
Minerals 2022, 12, x FOR PEER REVIEW 6 of 2 adsorbent through chemisorption, with the subsequent reduction to Cr(III) (as confirmed by the presence of CrOOH) [42].    Table 3 shows the EDS and XRF mole percentage results for raw PDFe/Al and Cr-PDFe/Al. Characterisation before Cr(VI) adsorption shows the presence of Fe and A in the synthesised adsorbent which confirm the dimetallic nature of the material.
The presence of Fe and S shows that AMD was generated from pyrite oxidation. Th presence of Si, Ca, Mg, and other trace elements in the material are impurities that re sulted from co-precipitation from AMD. After Cr(VI) adsorption, it is observed tha chromium is the only element showing a significant difference between the PDFe/Al and Cr-PDFe/Al. Overall, the composition of Fe, Al, O and C was observed to be preserved during adsorption, therefore demonstrating the chemically stability of the material. Figure 4 shows the relationship between 2θ values for PDFe/Al before and afte Cr(VI) adsorption.  Table 3 shows the EDS and XRF mole percentage results for raw PDFe/Al and Cr-PDFe/Al. Characterisation before Cr(VI) adsorption shows the presence of Fe and Al in the synthesised adsorbent which confirm the dimetallic nature of the material.
The presence of Fe and S shows that AMD was generated from pyrite oxidation. The presence of Si, Ca, Mg, and other trace elements in the material are impurities that resulted from co-precipitation from AMD. After Cr(VI) adsorption, it is observed that chromium is the only element showing a significant difference between the PDFe/Al and Cr-PDFe/Al. Overall, the composition of Fe, Al, O and C was observed to be preserved during adsorption, therefore demonstrating the chemically stability of the material. Figure 4 shows the relationship between 2θ values for PDFe/Al before and after Cr(VI) adsorption.
As indicated in Figure Table 4 depicts the porosity of the PDFe/Al before and after Cr(VI) adsorption using Brunauer-Emmet-Teller (BET). It is clear that a significant increase in both the internal surface area and volume was observed The total surface area of the raw PDFe/Al was reported to be around 37.5841 m 2 /g, which then changed to 95.5269 m 2 /g after Cr(VI) adsorption, while an eight-fold increase in internal pore volume (0.003 cm 3 /g to 0.02 cm 3 /g) was observed after Cr(VI) adsorption. These results are consistent with that observed previously for Congo Red adsorption [2].   As indicated in Figure 4, the plot for 2θ values for PDFe/Al before and after Cr( adsorption shows a clear shift in 2θ values of 0.66° thus confirming internal adsorpt induced strains as previously reported in Muedi et al. 2022 [2].  The increased total surface area implies that the adsorbed chemical has the potential to enhance the adsorbent's surface area, indicating the material's potential for subsequent usage after Cr(VI) ion adsorption. Figure 6 depicts the nitrogen adsorption-desorption isotherms of the produced PDFe/Al before and after Cr(VI) adsorption. The adsorbed quantities increased as relative pressure increased, which could indicate that the adsorbent's non-rigid nature and the placement of the distinctive shoulder are compatible with condensate destabilization at the P/P0 ratio, which is limiting the entire process. Furthermore, type IV adsorption isotherm behaviour was observed for both the raw PDFe/Al and the Cr-adsorbed PDFe/Al. tive pressure increased, which could indicate that the adsorbent's non-rigid nature a the placement of the distinctive shoulder are compatible with condensate destabilizati at the P/P0 ratio, which is limiting the entire process. Furthermore, type IV adsorpti isotherm behaviour was observed for both the raw PDFe/Al and the Cr-adsorb PDFe/Al.     at the P/P0 ratio, which is limiting the entire process. Furthermore, type IV adsorpti isotherm behaviour was observed for both the raw PDFe/Al and the Cr-adsorb PDFe/Al.       The morphological features of the synthesized PDFe/Al before and after Cr(VI) adsorption are presented in Figure 7. Figure 7A-C show the morphology of raw PDFe/Als of various sizes, with non-uniform pressed-like structures with irregular agglomerates scattered unevenly. Figure 7D-F show the morphology of PDFe/Als following Cr(VI) adsorption in various sizes, with blood cell-like formations that are irregularly dispersed and lumped together. The shift in structural forms could indicate the presence of chromium heavy metal in the material. Figure 8 shows the mapping of the elemental composition of the PDFe/Al before and after Cr(VI) adsorption using energy dispersive X-ray (EDX). The raw PDFe/Al confirms the co-precipitation of Fe and Al from AMD, (Figure 8A-D). Figure 8E-H illustrates the elemental composition of the material after Cr(VI) adsorption, where chromium is observed to be present.

TGA Thermal Stability
The thermal stability of PDFe/Al before and after Cr(VI) adsorption is demonstr in Figure 9. The calcination process was broken down into three stages, with the involving the loss of moisture from the material at temperatures ranging from 100 to °C. At temperatures between 400 and 550 °C, the second step involves the loss of ch cally bonded HOH, while the third stage involves the loss of the hydroxyl group (-OH temperatures over 550 °C.

TGA Thermal Stability
The thermal stability of PDFe/Al before and after Cr(VI) adsorption is demonstrated in Figure 9. The calcination process was broken down into three stages, with the first involving the loss of moisture from the material at temperatures ranging from 100 to 350 • C. At temperatures between 400 and 550 • C, the second step involves the loss of chemically bonded HOH, while the third stage involves the loss of the hydroxyl group (-OH) at temperatures over 550 • C. Minerals 2022, 12, x FOR PEER REVIEW 11 of 21 Figure 9. Thermal stability of the PDFe/Al before and after Cr(VI) adsorption.

Batch Adsorption Experiments
The effects of operational parameters on the removal of Cr(VI), as summarised in Table 1, are illustrated in Figure 10.

Batch Adsorption Experiments
The effects of operational parameters on the removal of Cr(VI), as summarised in Table 1, are illustrated in Figure 10.

Batch Adsorption Experiments
The effects of operational parameters on the removal of Cr(VI), as summarised in Table 1, are illustrated in Figure 10.

Effect of Concentration
As illustrated in Figure 10a, a steep increase in the residual concentration of Cr(VI) is observed after 50 mg/L, thus indicating that the material became oversaturated with Cr(VI) oxyanions. After 50 mg/L, an increase in the residual concentration is observed, thus indicating over-saturation of the adsorbent matrices with Cr(VI) oxyanions. This indicates that the material could not adsorb Cr(IV) oxyanions of concentration >50 mg/L. Therefore, it can be concluded that 50 mg/L is the optimum concentration for Cr(VI) adsorption using PDFe/Al.

Effect of Initial Solution pH
As illustrated in Figure 10b, a significant amount of Cr(VI) oxyanions were removed from the aqueous system at a pH of 3. This corroborates the point of zero charge (PZC) of the adsorbent which was found to be pH PZC = 3.02. Moreover, [43] also reported that the adsorption of Cr(VI) on magnetite (Fe 3 O 4 ) decreases with an increase in pH after level 3. An optimal pH = 3 was also obtained by [44]. In addition, it was observed that the final pH in all experimental runs were between 2.07 and 2.79 indicating that in all cases, except for the run in which the initial pH was 2, a decrease in pH was observed.

Effect of Adsorbent Dosage
As illustrated in Figure 10c, a steep increase is observed from 0.1 to 3 g during Cr(VI) removal, after which the trend takes a gentle slope. This could indicate the depletion of available sites in the adsorbent; hence, the material could not adsorb more Cr(VI) oxyanions. This behaviour is also observed in Figure 10c, where the maximum adsorption capacity of 1 g PDFe/Al in 250 mL of 10 mg/L Cr(VI) solution was observed to be approximately 1 mg/g, hence 3g of PDFe/Al nanocomposite is optimum dosage for Cr(VI) removal. Gürü et al. 2008 [45] also reported that the highest removal of chromium was observed at 3 g using diatomite.

Effect of Agitation Time
As illustrated in Figure 10d, it was observed that the removal of Cr(VI) from an aqueous system increases with time. After 180 min of agitation, the graph takes a gentle slope, hence no significant removal. Most studies reported 4 h as the optimal time for removal of Cr(VI) from water [46,47]. From the results, 180 min is the optimal agitation time for Cr(VI) removal.

Effect of Temperature
As illustrated in Figure 10e, it was observed that the removal of Cr(VI) from an aqueous system increased with an increase in temperature, particularly between 25-35 • C. Temperatures higher than 35 • C make Cr(VI) to go back into solution. Additionally, [48] also reported that maximum removal of Cr(VI) was achieved at 40 • C. Therefore, 35 • C is an ideal temperature for maximum removal of Cr(VI) from an aqueous system

Adsorption Kinetics
Adsorption kinetics for the adsorption of Cr(VI) by PDFe/Al was studied to demonstrate the mechanisms and rates of adsorption, as shown in Table 5 and Figure 11. Figure 11 shows different kinetic models fitted to the kinetic data for the adsorption of hexavalent chromium (Cr(VI)) onto the adsorbent. As shown in Figure 11a, a good trend was obtained from the pseudo-first order (PFO) model (R 2 = 0.945). However, there seems to be a lot of uncertainty depicted by the model the future observations in respect of Cr(VI) adsorption application, thereby making it risky to apply PFO kinetic model.
In Figure 11b, a good trend was obtained from the pseudo-second order (PSO) model, where results were obtained in the same manner as PFO. However, the PSO depends on the adsorbed amount of the adsorbate. The results obtained show a better trend than PFO (R 2 = 0.962) with great prediction of the adsorption rate of Cr(VI).  Table 5.
In Figure 11b, a good trend was obtained from the pseudo-second order (PSO) model, where results were obtained in the same manner as PFO. However, the PSO depends on the adsorbed amount of the adsorbate. The results obtained show a better trend than PFO (R 2 = 0.962) with great prediction of the adsorption rate of Cr(VI).
In Figure 11c, the Langmuir kinetics show a trend with good results (R 2 = 0.929) for the adsorption of Cr(VI); however, there is uncertainty in the diffusion and extrapolation of the model.
In Figure 11d, the two-phase pseudo-first order adsorption (TPA) model, which is based on two parallel adsorption processes: rapid and slow adsorption mechanisms, outperforms the PFO and PSO kinetic models. In comparison to previous models, the findings obtained demonstrate the best trend for Cr(VI) adsorption (R 2 = 0.993). As a result, it demonstrates how Cr(VI) is rapidly adsorbed and gradually slows down as saturation approaches.
In Figure 11e, the Crank diffusion model was investigated to determine pore diffusion since the diffusion coefficient (De) remains constant under the condition that the diffusion is uniform within the sphere. The results obtained show that an effective diffusion coefficient De = 2.61 × 10 −13 m 2 ·s −1 provides a good prediction of the Cr(VI) adsorption process. This indicates that the system is significantly limited by mass transport; the molecular diffusion coefficient of chromate is 1.4494 × 10 −9 m 2 ·s −1 [49] which is four orders of magnitude greater that the effective diffusivity.  Table 5.
In Figure 11c, the Langmuir kinetics show a trend with good results (R 2 = 0.929) for the adsorption of Cr(VI); however, there is uncertainty in the diffusion and extrapolation of the model.
In Figure 11d, the two-phase pseudo-first order adsorption (TPA) model, which is based on two parallel adsorption processes: rapid and slow adsorption mechanisms, outperforms the PFO and PSO kinetic models. In comparison to previous models, the findings obtained demonstrate the best trend for Cr(VI) adsorption (R 2 = 0.993). As a result, it demonstrates how Cr(VI) is rapidly adsorbed and gradually slows down as saturation approaches.
In Figure 11e, the Crank diffusion model was investigated to determine pore diffusion since the diffusion coefficient (D e ) remains constant under the condition that the diffusion is uniform within the sphere. The results obtained show that an effective diffusion coefficient D e = 2.61 × 10 −13 m 2 ·s −1 provides a good prediction of the Cr(VI) adsorption process. This indicates that the system is significantly limited by mass transport; the molecular diffusion coefficient of chromate is 1.4494 × 10 −9 m 2 ·s −1 [49] which is four orders of magnitude greater that the effective diffusivity.
Weber Morris' intra-particle diffusion model was used to study the effects of interparticle and intra-particle diffusion on Cr(VI) adsorption in Figure 11f. The many phases of adsorption visible in the multilinear fit of the data are depicted in this model. The first phase depicts in-particle diffusion, the second phase depicts in-particle diffusion, and the last phase depicts adsorption to the adsorbent. A good fit of the data was obtained (R 2 = 0.991) and the results showed a D e value between D e1 = 7.02 × 10 −12 m 2 ·s −1 and D e2 = 5.11 × 10 −13 m 2 ·s −1 . These results correlate well with the predicted effective diffusivity from the Crank diffusion model.  Two phase adsorption [52][53][54] dQ slow dt

Adsorption Isotherms
The adsorption isotherms of Cr(VI) adsorption by PDFe/Al dimetal composite were studied on various mechanisms that determine the adsorption process. The isotherm models are summarized in Table 6 and illustrated in Figure 12. Figure 12a depicts the Langmuir adsorption isotherm, with a maximum adsorption capacity (Q max ) of 6.67 mg·g −1 , which is acceptable for a mineral based adsorbent, especially considering that the adsorbent was recovered from authentic industrial waste AMD. To compare, Alemu et al. [56] and Panda et al. [57] tested the removal of Cr(VI) using industrially obtained basalt rock (Q max = 0.079 mg·g -1 ) and dolochar (Q max = 0.904 mg·g -1 ), respectively. In addition, the maximum adsorption capacities for comparable Fe/Al materials were in the range of 2.3-59.9 mg·g −1 , indicating that the current study compares well with results from the literature [26][27][28][29][30].
The Freundlich adsorption isotherm is shown in Figure 12b, where the intensity parameter indicates the adsorption favorability, with K F values of K F (298 K) = 0.8477; K F (318 K) = 1.15; K F (328 K) = 1.13; K F (338 K) = 1.098. The adsorption of Cr(VI) by PDPDFe/Al from the aqueous system is highly favorable in this investigation, with n F = 2.48.
The two-surface Langmuir adsorption isotherm is depicted in Figure 12c, which posits that the adsorbent's surface contains various surface types with varying adsorption capabilities. The results show that Q max,1 = 1.80 mg·g −1 for one surface and Q max,2 = 5.096 mg·g −1 for the other, for a total maximum adsorption of Q = 6.896 mg·g −1 . Table 6 shows a summary of tested adsorption isotherms, fitted parameters, coefficient of determination (R 2 ), and root-mean-square error (RMSE) as a measure of goodness of fit for various isotherm models.

Regeneration Study
As shown in Figure 13, a regeneration study was carried out to determine the possibility of recovering and reusing PDFe/Al following Cr(VI) adsorption. tion capabilities. The results show that Qmax,1 = 1.80 mg·g −1 for one surface and Qmax,2 = 5.096 mg·g −1 for the other, for a total maximum adsorption of Q = 6.896 mg·g −1 . Table 6 shows a summary of tested adsorption isotherms, fitted parameters, coefficient of determination (R 2 ), and root-mean-square error (RMSE) as a measure of goodness of fit for various isotherm models.  A desorption investigation was carried out to regenerate the material after Cr(VI) adsorption, as shown in Figure 13. The material had the ability to be utilized for Cr(VI) adsorption more than four times. During the first four cycles, it was observed that the material matrices could still adsorb the oxyanions from an aqueous system. However, after four reuse cycles, it was observed that the material started losing efficacy, probably due to the loosening of the material matrices, which probably lost layers in the process of adsorption-desorption. Table 6. Summary of tested adsorption isotherms, fitted parameters, coefficient of determination (R 2 ), and root-mean-square error (RMSE) as a measure of goodness of fit for various isotherm models.

Regeneration Study
As shown in Figure 13, a regeneration study was carried out to determine the possibility of recovering and reusing PDFe/Al following Cr(VI) adsorption. A desorption investigation was carried out to regenerate the material after Cr(VI) adsorption, as shown in Figure 13. The material had the ability to be utilized for Cr(VI) adsorption more than four times. During the first four cycles, it was observed that the material matrices could still adsorb the oxyanions from an aqueous system. However, after four reuse cycles, it was observed that the material started losing efficacy, probably due to the loosening of the material matrices, which probably lost layers in the process of adsorption-desorption.

Summary of Results
The results obtained in the study provide clear evidence for the potential of PDFe/Al synthesized from authentic AMD for the adsorption of Cr(VI) from aqueous solution. The surface characterization results demonstrated that the adsorption of Cr(VI) to the

Summary of Results
The results obtained in the study provide clear evidence for the potential of PDFe/Al synthesized from authentic AMD for the adsorption of Cr(VI) from aqueous solution. The surface characterization results demonstrated that the adsorption of Cr(VI) to the adsorbent involved several surface interactions with the presence of Cr(VI) and Cr(III) observed after adsorption. In addition, the adsorption of Cr(VI) resulting in adsorption induced strains within the adsorbent matrix which resulted in an increase in both the internal surface area and surface volumes. The surface analyses of the adsorbent showed the presence of Fe, O, Al, S, Cr (only after adsorption), and other minor constituents, confirming the heterogenous nature of the AMD used for the synthesis.
The batch adsorption results demonstrated that the adsorption followed a two-phase adsorption process with an initially fast process followed by a significantly slower adsorption.
The isotherm results were best described by a two-surface Langmuir kinetic model with distinct thermodynamic and saturation properties. The material further demonstrated good reusability with Cr(VI) in excess of 90% possible after five adsorption/desorption cycles.

Comparison of Fe/Al Dimetal Nanocomposite for Removal of Cr(VI) from an Aqueous System
Comparisons of rate constants, times to reach 99% of equilibrium [36], and maximum Cr(VI) adsorption capacities of different studies employing Fe-and Al-based adsorbents from the literature to the current material are made in Table 7. From Table 7, it can be concluded that the synthesized Fe/Al dimetal nanocomposite demonstrated a maximum adsorption capacity comparable to other Fe and Al based adsorbents (within the same order of magnitude). In addition, it was observed that the PDFe/Al achieved equilibrium nearly an order of magnitude faster (279 min) than most of the studies (1978-5952 min), except for the acetic acid-modified kaolinite that achieved equilibrium within 513 min. These results are significant, as they demonstrate the rapidity of the adsorbent in removing the Cr(VI), a requirement for the application of an adsorption system industrially. To illustrate the point, the mesoporous iron-zirconium bimetal oxide [27] demonstrated a maximum adsorption capacity of circa 60 mg·g −1 ; however, the time to 99% equilibrium took 2619 min (~44 h). This slower adsorption rate would invariably affect the residence times required for successful water treatment which negates the higher adsorption capacity displayed by the adsorbent.
In addition, the fact that the PDFe/Al dimetal nanocomposite was recovered and synthesised from authentic acid mine drainage, as opposed to synthetic chemicals, further supports the potential impact of the adsorbent as it provides an avenue for AMD valorisation through PDFe/Al synthesis and subsequent application for water treatment.

Proposed Mechanism
The results from the study indicate that the adsorption of Cr(VI) by PDFe/Al is significantly diffusion controlled with a predicted effective diffusivity of between 7.02 × 10 −12 m 2 ·s −1 and 4.79 × 10 −13 m 2 ·s −1 which is between 3 and 4 orders of magnitude less than the molecular diffusion coefficient of chromate (1.4494 × 10 −9 m 2 ·s −1 [49]). The adsorption isotherms indicated that the surface likely consisted of at least two surfaces with distinct adsorption properties. The one surface had exothermic adsorption properties with a corresponding decrease in entropy while the other supported endothermic adsorption with an increase in entropy predicted. The decrease in entropy on the first surface likely demonstrates the increased organisation resulting from adsorption of chromium from solution, while the increased entropy on the second surface result from the displacement of protons from the surface resulting in the observed decrease in pH [1]. It is interesting to note that the standard enthalpy and entropy values for the single surface Langmuir model were between the corresponding values from the two surface Langmuir model, indicating the single surface Langmuir model represented a net result of the two-surface model.
The results further showed that the adsorption of Cr(VI) induced significant internal strain and corresponding deformation within the support matrix which result in the formation of cracks therefore increased surface area and volume. Finally, it was observed that the adsorbed Cr(VI) was reduced to Cr(III) on the surface of the adsorbent. This likely corresponds to the oxidation of the Iron(II/III) oxide observed in the XRD profile after Cr(VI) adsorption [60]. The reduction of Cr(VI) to Cr(III) is known to have a standard Gibbs free energy of −390 kJ/mol [61] while the oxidation of Iron(II/III) oxide has a Gibbs free energy of −42 kJ/mol [61]. The resulting oxidation-reduction reaction has a standard Gibbs free energy of −450 kJ/mol and is therefore highly favourable and spontaneous.

Conclusions
The synthesis of polycationic di-metal iron/aluminium nanocomposite was executed through selective precipitation followed by calcination and vibratory ball milling. The material was successfully recovered and synthesised from industrial acid mine drainage and applied in the removal of hexavalent chromium (Cr(VI)) from an aqueous system. Adsorption parameters were optimized using the one-factor-at-a-time (OFAAT) approach on a batch experimental set-up. The fate of Cr(VI) in the aqueous solution, as well as the surface of the material, was determined; the results demonstrated not only the successful adsorption of Cr(VI) from solution, but also confirmed the presence and in situ reduction of Cr(VI) to Cr(III) on the surface.
Recorded optimised conditions for Cr(VI) adsorption are: 50 mg/L initial Cr(VI) concentration; 3 g of PDFe/Al; pH 3; 180 min agitation time; and temperature of 35°C. The synthesized PDFe/Al was observed to have a Langmuir adsorption capacity of Q max = 6.67 mg·g −1 for Cr(VI) with >95% removal.
Adsorption kinetics followed a two-phase pseudo-first-order as opposed to pseudofirst or pseudo-second-order behaviour. The adsorption process followed two-surface Langmuir adsorption model. The PDFe/Al achieved more than 90% Cr(VI) recovery after five adsorption/desorption cycles, thereby demonstrating the potential reusability.
It was proposed that a diffusion limited adsorption mechanism involving two surfaces with distinct adsorption characteristics were responsible for the adsorption. The mechanism further involved the reduction of Cr(VI) to Cr(III) while simultaneously significant internal adsorption induced strains resulted in the formation of internal cracks within the adsorbent matrix resulting in increased pore surface area and pore volume.
Comparison of the maximum adsorption capacities and time required to reach equilibrium of the PDFe/Al to literature studies demonstrated that the PDFe/Al exhibited a comparable adsorption capacity while a superior adsorption rate was measured.
This study demonstrates the industrial potential of the synthesized PDFe/Al for the simultaneous valorisation of AMD and the treatment of Cr(VI), thereby directly contributing towards sustainable mining.