Reply to Hosseini Monjezi, B.; Lützenkirchen, J. Comment on “Zhou et al. Adsorption Behaviors of Lanthanum (III) and Yttrium (III) Ions on Gibbsite. Minerals 2023, 13, 1530”

The interaction between aluminum (hydr)oxides and inorganic substances has received widespread attention in the fields of ore deposit geochemistry, mineral processing, and environmental protection [1,2,3]. To explore the interaction between aluminum hydroxide minerals and rare-earth elements in bauxite and explain the enrichment and differential mechanisms of rare-earth elements in bauxite, we used batch experiments to investigate the adsorption process of rare-earth elements (La and Y) on gibbsite.

The experimental conditions of our experiment were designed based on a real situation in nature. Since the content of rare-earth elements in bauxite deposits can span from several hundred to thousands of ppm, we chose a high initial rare-earth concentration (3.5 × 10⁻⁴ mol/L). Because of precipitation and seepage, the solid–liquid system in bauxite is mostly a suspension or slurry with a high solids content, so we chose a high solid–liquid ratio (2 g/20 mL). In addition, due to the large number of soluble salts in bauxite, we chose a range of high background electrolyte concentrations (0.005, 0.050, 0.500 mol/L NaNO₃). The adsorption behavior is consistent with the observation that light rare-earth elements are more enriched than heavy rare-earth elements in bauxite. In the comment [4], the authors cite relevant references that question our experimental results, but most of the references cited by the authors are based on studies using low rare-earth concentrations (10⁻⁹ to 10⁻⁵ mol/L), low solid-to-liquid ratios (10⁻³ to 10⁻² g/mL), and relatively low background electrolyte concentrations (0.001 to 0.1 mol/L) [5,6,7]. These references are mainly based on the research paradigm of aqueous chemistry, using dilute solution systems for theoretical modeling. This is because many approximations can be made in the dilute solution system, such as using solution concentration instead of activity, using the currently available electric double-layer model, and so on. It is clearly inappropriate to use the results of these dilute solution systems to evaluate our experimental results, just as a large error would occur when describing real gasses using the ideal gas equation of state. On the contrary, our experimental results are exactly complementary to the previous studies.

In addition, we respond as follows to the main issues mentioned in the comment:

1. In terms of the point of zero charge (PZC) determination’s accuracy, the pH_pzc range reported in the literature for gibbsite is in the range of 8.0–11 [7,8,9], and our result of 10 is clearly within this range. Also, previous researchers have measured the same values as we did [10,11]. Even if there was an error in our experimental results due to the dissolution of gibbsite, the pH_pzc should be lower (e.g., about 8.5 in the reference provided by the authors). This pH was still higher than the pHs used in our adsorption experiments (from 4 to 7), i.e., the surfaces of gibbsite were positively charged during adsorption, which does not affect the interpretation of the results in our experiments.

The authors of the comment suggested that the reason why our titration curves differ from those in previous references was due to the dissolution of gibbsite. In fact, gibbsite dissolves in both acidic and alkaline solutions. Therefore, we deliberately chose a sample with a large particle size (about a few microns) and good crystallinity, which is more consistent with the real situation in bauxite, in the expectation that the effect of dissolution could be reduced. In contrast, the reference mentioned by the authors usually used synthetic samples, which have smaller particle sizes and lower crystallinity, and perhaps this is the main reason for the differences between the titration curves of these samples and those of the samples we used. As stated in the comment, it is necessary to obtain more protons in the surface reaction than in the solution reaction, to ensure that the error of the data is small, so the current studies on the acid-base and/or charge properties of the surface of gibbsite have used synthesized samples with a large specific surface area (tens of m²/g). However, the purpose of this study led us to ultimately select this sample with a small BET specific surface area (2.343 m²/g).

Meanwhile, the authors of the comment believed that in our experimental data of potentiometric titration, the suspension containing 0.05 mol/L NaNO₃ requires more alkali to be added than that of 0.5 mol/L NaNO₃, which contradicts the trend that the proton-related surface charge density of oxide minerals increases with the concentration of monovalent salts. According to our potentiometric titration curve, it can be seen that the volume of alkali required for the suspension containing 0.5 mol/L NaNO₃ titration is significantly larger than that required for the system of 0.05 mol/L NaNO₃, which is not contradictory to the above trend.

In addition, the authors suggested using electrophoresis for pH_iep determination, which was used to verify the accuracy of pH_pzc deduced by titration. Unfortunately, due to the large particle size of the sample we used, the settling rate was too fast to obtain valid data when performing the determination.

Nevertheless, we surely benefited from the authors’ many detailed and highly practical suggestions for characterizing solid surface properties by titration and other methods, such as the further calculations of surface charge.

2. In the part of the experiments in which we detailed the effect of the initial pH and background electrolyte concentration on the adsorption process, our adsorption experiments were carried out under air conditions and did not exclude the effect of carbon dioxide. This is more consistent with the real situation, where carbon dioxide is present under natural conditions. Moreover, carbon dioxide has a greater influence on the adsorption experiments under alkaline conditions, whereas our experiments were mainly conducted under acidic conditions.

Our experimental results indicate that as the initial pH of the solution and background electrolyte concentration increase, the absorption of La³⁺ and Y³⁺ by gibbsite increases. Subsequently, we cited three references to confirm our work. Tochiyama et al. [12] investigated the effect of solution pH and adsorbate concentration on the adsorption of Eu(III) and Co(II) on gibbsite. The results showed that the adsorption of gibbsite on Eu(III) and Co(II) increased with increasing pH in the range of 5–7; Ahmed et al. [13] investigated the effects of sulfate, pH value, and salt concentration on the adsorption of Ca²⁺ on gibbsite and kaolinite through batch experiments. The research results showed that the adsorption of sulfate and the enhancement of pH value (4–7) at a low Ca²⁺ concentration (0.5 mmol/L) reduced the positive charge of calcium and the accompanying electrostatic repulsion. Both sulfate and surface net charge can affect the adsorption of calcium sulfate on gibbsite. When the calcium concentration in the solution is high (10 mmol/L), and the sulfate concentration increases to the threshold of calcium precipitation, the adsorption of calcium on minerals will be inhibited, and the adsorption will decrease with the increase in pH value. Due to the high concentration of Ca²⁺, which is much higher than our rare-earth ion concentration, and the possibility of precipitation with the added background ions, this will not be discussed and compared; Saeki et al. [14] investigated the adsorption of Fe²⁺ and Mn²⁺ on gibbsite, and the experimental results showed that under constant background electrolyte concentration conditions (0.1 mol/L NaCl), the adsorption of Fe(II) and Mn(II) on gibbsite increased with increasing pH. Therefore, our experimental results are not contradictory to the experimental phenomena in the cited references. Although, we have also found that some literature shows opposite adsorption patterns, and some literature suggests that the background electrolyte concentration does not affect the adsorption of metal ions on gibbsite [15]. However, other references show the same trend, which is consistent with our experiment. For example, the research results of Girvin et al. [16] show that in the range of background electrolyte concentration of 0.01–1 mol/L NaClO₄, the adsorption of Co(II) on gibbsite increases with the increase in ion strength.

The differences in the adsorption rates of metal ions on gibbsite under different pH conditions in different references may be influenced by various factors, such as the concentrations of adsorbate and adsorbent, the physical and chemical properties of adsorbent, etc. The authors of the comment, based on the collected literature data, concluded that at a constant solid concentration, an increase in adsorbent concentration pushes the adsorption edge to a higher pH. In our experiments, the effect of different initial adsorbent concentrations on adsorption was not carried out, so we did not discuss the comparison. However, this could be a good research direction.

We use 450 nm polyethersulfone (PES) membrane filters for solid–liquid separation rather than centrifugal settling, because centrifugation is more likely to introduce particles into the solution to be measured when decanting or aspirating the supernatant after centrifugation. Also, high-speed centrifugation tends to result in the fragmentation of the particles, and the adsorption of small particles will again adsorb rare-earth ions from the solution. During the adsorption experiments, we have 2–3 blank parallel experiments for each batch of experiments (without the addition of gibbsite), and the absorbance measured from the blank sample is basically the same as that of the initial rare-earth solution, which indicates that the adsorption of rare-earth ions by the PES membrane is negligible.

The reason for using the absorbance method to test the concentration of rare earth instead of ICP-MS is that the concentration of rare-earth ions in the solution after adsorption is high, and the use of ICP-MS requires several dilutions and may result in a large error.

3. There are many discussions on adsorption kinetics models [17,18]. Some references [19,20] suggest that nonlinear fitting should only consider partial kinetic data before the system approaches equilibrium for quasi-second-order kinetic models, which is more accurate than linear fitting when using all the data. Therefore, we selected experimental data that from 0 to 2 hours for fitting the quasi-second-order kinetic model.

For the consistency of the data in Figures 5 and 6 in the article, firstly, Figure 5 is the experimental data measured from the adsorption experiment, while Figure 6 is the kinetic model fitting based on the experimental data. The Q_e obtained in Table 1 is the fitted data, and it is normal for there to be a certain difference between the data and the actual experimental data. On the other hand, Figures 4 and 5 are both actual experimental data, and the adsorption rates (about 70% and 65% for La³⁺ and Y³⁺, respectively) in Figure 4 can be compared with the adsorption rates (between 70% and 75% and 60% and 65% for La³⁺and Y³⁺, respectively) in Figure 5 under the condition of solution pH 7 without adding background electrolyte. The experimental results are close to consistency.

Based on pre-experiments, we determined that the initial pH range and background electrolyte concentration of the solution, as well as the equilibrium time used in thermodynamic experiments, were all 72 h. Among them, the unit of K_F in the Freundlich model is mg^1−1/n g⁻¹ L^−1/n.

In terms of the unit of experimental data, perhaps as the authors of the comment pointed out, the number of moles per square meter is a commonly used unit in adsorption research in the fields of environment, materials, or chemistry, which facilitates comparability between research. However, the research purpose of our experiment was to explore the interaction between aluminum hydroxide minerals and rare-earth elements enriched in bauxite. Under geological conditions, it is very common to use ppm for element content. Therefore, in order to compare the content of rare-earth elements in bauxite under geological conditions, we used ppm as the unit.

4. The surface complexation model is a commonly used computational model in adsorption research. However, as we mentioned above, our experiments use high rare-earth ion concentrations, high solid-to-liquid ratios, and high background electrolyte concentrations, which will encounter many problems when performing surface complexation modeling, such as how the activity coefficients should be determined, how the electric double-layer model should be chosen, and so on. Therefore, we only perform a preliminary analysis and the fitting of the data to explain the rare-earth enrichment process in natural systems. In addition, as the commentator pointed out, it requires a lot of funds and effort to use the surface complexation model well. This is, of course, where we need to make efforts for future research.

5. Section 2.2 of the article (Adsorption experiments) provides the sources of the samples and chemical reagents used, as follows: the gibbsite sample was purchased from Shanghai Macklin Biochemical Technology Co., Ltd. (Shanghai, China). La(NO₃)₃, Y(NO₃)₃, and arsenazo III were purchased from Aladdin Scientific Corp. (Cambridge, UK).

The specific surface area of gibbsite calculated by the Brunauer–Emmett–Teller (BET) method based on the nitrogen adsorption–desorption isotherms was 2.343 m²/g. We did not provide these data in the article because the subsequent experimental analysis did not involve the specific surface area.

As previously mentioned, the adsorption experiment was also conducted in the atmosphere and did not exclude the influence of carbon dioxide, since carbon dioxide is widely present under natural geological conditions and may participate in the enrichment process of rare-earth elements in bauxite. Theoretically, rare-earth ions will undergo a large amount of precipitation, achieving a 100% adsorption rate. However, according to our experimental data, 100% adsorption has not occurred in various situations. According to the blank experiment (without gibbsite), no significant precipitation phenomenon occurred within the pH and background electrolyte concentration range of the experiment, and the REE concentration measured by the spectrophotometer was almost the same as the initial concentration. Therefore, we believe that the model obtained based on the assumption of an ideal dilute solution is not suitable for calculations in the systems with high solid–liquid ratios which we have adopted.

Finally, we would like to thank the authors of the comment for their critical evaluation and helpful discussion of our work.