Impurity-bearing ferrihydrite nanoparticle precipitation/deposition on quartz and corundum

Abstract

During ferrihydrite precipitation, metal ions can be sequestered in it to form impurity-bearing ferrihydrite (IBF). Using grazing-incidence small-angle X-ray scattering (GISAXS), heterogeneous precipitation/deposition of pure and IBF nanoparticles on quartz (SiO₂) and corundum (Al₂O₃) was quantified in 0.1 mM Fe³⁺ solutions in the absence and presence of 1 mM Mn²⁺ or Al³⁺ (pH = 3.8 ± 0.1). The impurity ions (Mn and Al) greatly affected ferrihydrite nanoparticle precipitation/deposition on substrates. On SiO₂, ferrihydrite nanoparticle precipitation/deposition was promoted in the presence of Mn but was inhibited in the presence of Al. On Al₂O₃, Mn- and Al-bearing ferrihydrite nanoparticle precipitation/deposition was slower than for pure ferrihydrite. Compared with on SiO₂, pure and IBF nanoparticle precipitation/deposition on Al₂O₃ was significantly inhibited. To understand the mechanisms, interactions among impurity ions, substrates, and precipitates were explored. Surface enrichment of Mn and Al on precipitates was found to increase the zeta potential of ferrihydrite nanoparticles. The changes in surface charges of the precipitates and substrates affected heterogeneous IBF precipitation/deposition significantly. The rates and mechanisms of heterogeneous IBF precipitation/deposition provided here can help predict pollutant transport and design catalyst synthesis.

---

Environmental significance

The importance of impurity-bearing ferrihydrite (IBF) nanoparticle precipitation/deposition cannot be overstated. In many natural soil and aquatic systems, IBF nanoparticle precipitation/deposition controls the fate of many aqueous contaminants. Also, IBF nanoparticles, as an economic and environmentally friendly material with large specific surface area and high reactivity, have been used widely for the removal and degradation of pollutants. To improve their performance, impurity ions are usually doped in ferrihydrite nanoparticles and the nanoparticles are usually coated on supporting substrates. Understanding the fundamentals of interactions among impurity ions, IBF nanoparticles, and substrates, which controls the heterogeneous precipitation/deposition of IBF nanoparticles, is of essential importance.

Introduction

Ferrihydrite is abundant as colloids or coatings on substrates in many natural and engineered aqueous environments, and always contains various impurities (e.g., Al, Mn, Cr, etc.).^1–4 In natural systems, impurity ions can be sequestered in ferrihydrite during its precipitation, and the precipitation of impurity-bearing ferrihydrite (IBF) nanoparticles controls the fate and transport of many aqueous inorganic and organic contaminants.^5–9 Contaminant sequestration during IBF precipitation has been reported at field sites of acid mine drainage (AMD), managed aquifer recharge (MAR), and geologic CO₂ sequestration (GCS).^10,11 For example, at AMD sites, the contaminated water contains high concentrations of Fe³⁺, Al³⁺, and Mn²⁺, which can coprecipitate with ferrihydrite.^10,1213 The IBF nanoparticles can precipitate in solution (homogeneous precipitation) and carry aqueous contaminants (e.g., As) to downstream through colloidal transport, or be precipitated/deposited on rocks (heterogeneous precipitation/deposition) and retard the transport of aqueous contaminants. Impure iron (hydr)oxide nanoparticles have also garnered considerable interest in the fields of material science and chemical engineering. For example, Al, Cr-doped Fe(III) (hydr)oxide nanoparticles have been synthesized to degrade the organic compound hydroquinone and to catalyze water gas shift reactions.^14–19

The homogeneous precipitation of IBF has been studied widely, and impurity ions have been reported to be sequestered in ferrihydrite through surface adsorption, surface precipitation, and structural incorporation.^1–4 Impurity ions have also been reported to affect the morphology, structure, and reactivity of IBF nanoparticles significantly.^{14,16,20–22} For example, the amount of aluminum doping in ferrihydrite can affect the sizes of the nanoparticles significantly, as well as their adsorption capacities for As(III) and As(V) and their redox reactivity for hydroquinone degradation.^14,20,23,24

However, the fundamental knowledge of heterogeneous precipitation/deposition of IBF nanoparticles is much less well understood, and the controlling mechanisms are not clear.^25–28 In our recent studies, different adsorption behaviors of impurity metal ions (e.g., Pb, Cu, Al and Cr) onto quartz (SiO₂) were observed, which affected heterogeneous IBF precipitation/deposition on SiO₂ significantly.^13,29,30 In the work of Legg et al. and Tosco et al., pure ferrihydrite deposition under different ionic strength (IS) on SiO₂ was measured, and the controlling mechanisms were explored.^31,32 IBF precipitation/deposition on different substrates has not been reported.

Also, it is important to quantify the initial IBF precipitation (i.e., nucleation and growth) and deposition from solution. In reactive transport models, the rates of mineral precipitation are predicted as a function of the rate constant (k), solution's saturation ratio (Ω), and mineral surface area (A).³³ The latter is changed by nucleation and growth, and in particular, the initial surface areas of the newly precipitated minerals are calculated based on nucleation rates. Currently, due to a lack of data on initial nucleation rates, great inaccuracy has blighted estimation of the initial surface areas of minerals, leading to inaccuracies in these reactive transport models.

Here, we aimed to fill such important information gaps. The objectives of the present study were to: (1) quantify initial heterogeneous precipitation/deposition kinetics (within 30 min) of Al- and Mn-bearing ferrihydrite on SiO₂ and corundum (Al₂O₃) using grazing incidence small angle X-ray scattering (GISAXS) under acidic conditions (pH = 3.8 ± 0.1); (2) understand fundamental mechanisms by exploring the atomic-level interactions among impurity ions, substrates, and IBF nanoparticles with an integration of interfacial characterization methods and aquatic chemistry measurements.

Materials and methods

Substrate and solution preparation

Here, SiO₂ and Al₂O₃ were chosen as the substrates for the precipitation/deposition experiments because Si–O and Al–O are the main components of minerals on earth, and because they are used widely as substrates for coating nanomaterials in engineered systems. Also, SiO₂ and Al₂O₃ have quite different surface properties (e.g., dielectric constant, pH_iep), which might affect heterogeneous precipitation/deposition processes.³⁴ Single-crystal SiO₂ (10 × 10 × 0.5 mm, cut along the 10–10 plane, with roughness <5 Å, density of 2.68 g cm⁻², and pH_iep of 3.8) and Al₂O₃ (10 × 10 × 1 mm, cut along the 0001 plane, with roughness <5 Å, density of 3.97 g cm⁻², and pH_iep of 11.58) were purchased from MTI (Salt Lake City, UT, USA). Before the precipitation/deposition experiments, SiO₂ and Al₂O₃ substrates were cleaned, and the detailed cleaning procedures can be found in our previous publications.^{13,29,30,35–37}

All chemicals (Fe(NO₃)₃·9H₂O, Al(NO₃)₃·9H₂O, Mn(NO₃)₃·4H₂O, and NaNO₃) were purchased from Sigma-Aldrich (Saint Louis, MO, USA) and ultrapure water (conductivity <18 mΩ) was used to prepare all solutions (Table 1). Impure ferrihydrite precipitation experiments were conducted with 0.1 mM Fe(III) and 1 mM Al(III) or Mn(II), and the solutions were labeled as “FeAl” or “FeMn”, respectively. Pure ferrihydrite precipitation experiments were also conducted with 0.1 mM Fe(III) and the solution was labelled as “Fe only”. The IS of all solutions was adjusted to be the same, with 5.9 and 2.9 mM NaNO₃ added into Fe only and FeMn solutions, respectively. Using Geochemist's Workbench (GWB; Release 9.0; Aqueous Solutions, Champaign, IL, USA), the pH, IS, and saturation indices (SI) of solutions with respect to Fe(OH)₃ were calculated (Table 1). The FeAl and FeMn solutions were also calculated to be undersaturated with respect to gibbsite (Al(OH)₃) and amorphous Mn(OH)₂.

Table 1 Experimental conditions

Sample name	Mn^2+, mM	Al^3+, mM	Fe^3+, mM	NaNO3, mM	IS,^a mM	pH^b	SI^c
Note:a IS: Ionic strength.b pH: GWB calculated pH values, which were consistent with measured pH (3.8 ± 0.1).c SI: Saturation indices with respect to Fe(OH)3. SI = log(Q/Ksp), where Q is the actual dissolved composition, and Ksp(Fe(OH)3) = 10^–7.22 at 20 °C was used for SI calculations based on GWB database.
Fe only	0	0	0.1	5.9	6.2	3.8	0.36^c
FeMn	1	0	0.1	2.9	6.2	3.8	0.36^c
FeAl	0	1	0.1	0	6.2	3.7	0.23^c

---

In situ GISAXS measurements

Before each GISAXS measurement, a cleaned SiO₂ and Al₂O₃ substrate was placed in our self-designed GISAXS cell, and the substrate surface aligned vertically towards the center of the X-ray beam. To achieve the optimal incident angle for the strongest scattering signal, X-ray reflectivity at the substrate–water interface was considered. With a density value of 2.68 and 3.97 g cm⁻³ for SiO₂ and Al₂O₃, the critical angles for total external reflection of X-rays (with an energy of 14 keV) at SiO₂–water and Al₂O₃–water interfaces were calculated to be 0.105° and 0.136°, respectively.³⁸ Therefore, in our measurements, an incident angle of 0.10° was chosen, and only scattering from particles on substrates was collected.³⁸ More detailed information of calculations of critical angles can be found in Supplementary Information.

After alignment, 0.7 mL of a freshly prepared solution (Fe only, FeMn, or FeAl, Table 1) was injected immediately into the cell and in situ GISAXS measurement started. The deposition and precipitation of IBF nanoparticles on substrates occurred simultaneously immediately after the reaction had started. During the measurements, GISAXS scattering signals caused by particles precipitating and depositing on substrates were collected every 1 min for 30 min. At the end of each 30 min in situ GISAXS experiment, the cell was moved horizontally for 1 mm and a GISAXS image collected at a fresh spot. The GISAXS image collected at the fresh spot was similar to the last image of each in situ GISAXS measurement, indicating no X-ray damage during our measurements. GISAXS experiments were done at Beamline 12 ID-B, Advanced Photon Source at the Argonne National Laboratory (Argonne, IL, USA).

All data analyses were conducted using IRENA, GISAXS SHOP macro, and Igor Pro 6.34 (WaveMetrics, Lake Oswego, OR, USA). Background subtraction was undertaken using the first image of each in situ GISAXS experiment, where no discernible particle scattering was detected. In two-dimensional (2D) GISAXS images, no obvious scattering patterns were shown along vertical directions, indicating that the particles on substrates were not well ordered. Therefore, for simplicity, we assumed the shape of the disordered particles to be a low-resolution and highly symmetric shape, such as a sphere. The subtracted 2D scattering images were then deducted to 1D scattering intensity curves by cutting the long Yoneda wing, where the X-ray scattering signal is the strongest because of the Vineyard effect.^39,40 The deducted 1D GISAXS scattering curves were plotted as scattering intensity (I) vs. scattering vector (q, with units of Å⁻¹, reciprocally related to size) (Fig. 2).

To obtain size information of the particles on substrates, a lognormal model of non-interacting spherical particles was used to fit all GISAXS scattering curves. Good fittings were obtained (the black lines in Fig. 2 are the fitted curves, and the colored dots are the measured data), and the size evolutions of particles (as radius, R, in nm) forming on SiO₂ and Al₂O₃ are plotted in Fig. 3A1 and A2, respectively. Lorentz-corrected intensity curves (i.e., I × q²vs. q) were also plotted (Fig. S1 in ESI†), and the total particle volumes (V, in relative units) on SiO₂ and Al₂O₃ were calculated as the so-called invariant , as shown in Fig. 3B1 and B2, respectively.⁴¹ By assuming a spherical shape for the particles on substrates, using the total particle volume (V) and the mean particle radius (R), the numbers of particles on substrates (N) could be calculated using N = V/R.³³⁶

Mineral phase and chemical composition characterization of the precipitates

Because the amount of precipitates in solution was not sufficient for regular X-ray diffraction (XRD) measurement, synchrotron-based high-resolution X-ray diffraction (HRXRD; Fig. 1) measurements were conducted at Beamline 11-BM, Advanced Photon Source, Argonne National Laboratory. For each experimental condition listed in Table 1, 500 mL solution was freshly prepared and reacted for 30 min. Then, the precipitates were collected using centrifuge filter units (molecular weight, 100 000; Millipore, Burlington, MA, USA), and dried in a desiccator under room temperature. Then, the particles were filled in Kapton capillaries of diameter 0.8 mm and shipped to Advanced Photon Source for measurements. HRXRD results indicated that the precipitates from Fe only, FeMn and FeAl solutions were poorly crystallized 6-line ferrihydrite (Fig. 1).

Fig. 1 HRXRD patterns of the precipitates formed in 0.1 mM Fe³⁺ solutions in the absence (Fe only) and presence of 1 mM Mn²⁺ (FeMn) and 1 mM Al³⁺ (FeAl).

The chemical compositions of precipitates in solutions were investigated by strong acid soak (pH = 0.5 HNO₃) and consecutive dilute acid wash (pH = 3 HNO₃) experiments.^29,30,42,43 Solutions (500 mL) (Table 1) were freshly prepared and let to stand for 30 min. Then, precipitates were collected on the centrifugal filters (Millipore). To obtain the total amounts of impurity ions in the precipitates, particles on centrifugal filters were fully dissolved by soaking in 5 mL of 2% nitric acid (pH = 0.5) overnight. The concentrations of impurity ions (Al and Mn) and Fe were measured by inductively coupled plasma-mass spectrometry (ICP-MS), and the total atomic ratios of Mn/Al over Fe (R_t,M/Fe) in the precipitates calculated.

Studies have reported that impurity ions can be sequestered in ferrihydrite through structural incorporation and surface enrichment.^1–3 To explore the sequestration mechanisms of Al or Mn in the precipitates, consecutive dilute acid wash experiments (pH = 3, HNO₃) were undertaken. Initially, 5 mL of dilute HNO₃ (pH = 3) was used to soak the collected particles on the filter for 10 min, and to dissolve the surface layer of the precipitates. Based on previous studies, dilute acid can desorb the metal ions adsorbed on ferrihydrite surfaces without significant dissolving ferrihydrite particles (<1%) within 30 min, which was also observed here.⁴³ The concentrations of dissolved impurity ions (Mn/Al) and Fe ions were measured by ICP-MS, and the surface atomic ratios of Mn or Al over Fe (R_s,M/Fe) were calculated. Then, consecutively, 5 mL of pH = 3 HNO₃ solutions were added to soak the remaining particles for a longer duration (i.e., 30, 60, and 360 min), until the measured atomic ratios of Mn or Al over Fe did not change significantly, which represented the lattice atomic ratios of Mn or Al over Fe (R_l,M/Fe). For the heterogeneous precipitates formed on SiO₂ and Al₂O₃ surfaces, their amounts were too few for phase identification and chemical composition analyses.

Dynamic light scattering (DLS) measurements

To understand the interactions among metal ions, substrates, and precipitates, DLS using a Zetasizer Nano ZS system (Malvern Instruments, Malvern, UK) was employed to measure the zeta potential. Two sets of experiments were conducted: the first set was to measure the zeta potential of the precipitates, and the second set was to measure the zeta potential of the substrates in the presence of metal ions. In the first set, to measure the zeta potential of pure and IBF colloids, a freshly prepared solution (Table 1) was injected into a zeta cell (DTS1070 folded capillary cell; Malvern Instruments). In the second set, to investigate the potential effects of aqueous Mn²⁺ and Al³⁺ on the surface charges of substrates, the zeta potential of SiO₂ and Al₂O₃ powders was measured at our experimental pH condition (3.8 ± 0.1, adjusted using HNO₃) in 5.9 mM NaNO₃, 1 mM Mn²⁺ (with 2.9 mM NaNO₃), and 1 mM Al³⁺, without the addition of Fe³⁺. All zeta-potential data were collected continuously every 1 min for 1 h at 20 °C, and the mean values and standard deviations were calculated (Table 2). The detailed information on zeta-potential measurements of SiO₂ and Al₂O₃ in the presence of metal ions can be found in our previous publications.³⁶

Table 2 Hydrolysis of metal ions, their sequestration mechanisms in ferrihydrite precipitates, and the surface charge of substrates

Sample name	M(OH)x^n−x,^a, %	R s,^bM/Fe	Rl,^bM/Fe	Rt,^bM/Fe	ζ,^c mV	ζ_SiO2,^d mV	ζ_Al2O3,^d mV
Note:a M(OH)x^n−x, %: percentages of metal ions in hydrolyzed states (e.g. Al(OH)^2+, Al(OH)2^+, Al(OH)3), which were calculated using GWB under our experimental conditions.b Rs,M/Fe, Rl,M/Fe, and Rt,M/Fe: surface, lattice, and total atomic ratios of impurity ions (Mn or Al) over Fe in the precipitates.c ζ: zeta potentials of pure or impurity-bearing ferrhydrite nanoparticles.d ζ_SiO2 and ζ_Al2O3: zeta potentials of quartz and corundum powders in the presence of 1 mM Mn^2+ or Al^3+ (pH adjusted to 3.8 ± 0.1), respectively.
Fe only	N/A	N/A	N/A	N/A	29.1 ± 5.7	−22.4 ± 2.6	20.5 ± 2.8
FeMn	0.5	4.14 ± 1.81	0.0003 ± 0.0001	0.04 ± 0.02	39.9 ± 2.7	−25.1 ± 1.9	27.4 ± 6.8
FeAl	4.4	4.36 ± 0.08	0.03 ± 0.01	0.07 ± 0.01	40.6 ± 2.3	18.1 ± 3.9	34.3 ± 6.4

---

Results and discussion

Hydrolysis of impurity ions affects their structural incorporation in homogenously precipitated ferrihydrite

The chemical compositions of the homogeneously precipitated IBF nanoparticles in solutions were measured by ICP-MS, and the total atomic ratios of the impurity ions (Mn and Al) over Fe (R_t,M/Fe) were calculated. For Mn- and Al-bearing ferrihydrite nanoparticles, the total atomic ratios of Mn (R_t,Mn/Fe, 0.04 ± 0.02 : 1, Table 2) and Al (R_t,Al/Fe, 0.07 ± 0.01 : 1, Table 2) over Fe were low, indicating that small amounts of Mn and Al were sequestered in ferrihydrite precipitates.

Different sequestration mechanisms of impurity ions in ferrihydrite have been reported in experimental and modeling studies.^44–47 Mn and Al ions can substitute Fe(III) ions in the lattice and be sequestered in Fe(III) (hydr)oxides as structural incorporation.^{1–3,44–46} Also, Mn and Al ions can be sequestered through surface adsorption and surface precipitation, resulting in their enrichment on ferrihydrite surfaces.^1–3,46,47 Here, the sequestration mechanisms of Mn and Al in IBF were investigated by comparing the surface (R_s,M/Fe) and lattice (R_l,M/Fe) atomic ratios of the metal ions over Fe. If R_s,M/Fe ≤ R_l,M/Fe, then only structural incorporation occurred. Conversely, if R_s,M/Fe > R_l,M/Fe, then surface enrichment also occurred. As shown in Table 2, R_s,M/Fe (R_s,Mn/Fe = 4.14 ± 1.81 : 1, R_s,Al/Fe = 4.36 ± 0.08 : 1) were much higher than R_l,M/Fe (R_l,Mn/Fe = 0.0003 ± 0.0001 : 1, R_l,Al/Fe = 0.03 ± 0.01 : 1) for FeMn and FeAl, indicating that Mn and Al were enriched on the surfaces of the precipitates. This is because Fe (hydr)oxides have high affinity to various metal ions, including Al and Mn.^47,48

Lattice R_l,M/Fe ratios, which are indicative of structurally incorporated impurity ions in ferrihydrite, showed a ∼100-fold difference for the precipitates formed in FeAl (R_l,Al/Fe = 0.03 ± 0.01 : 1) and FeMn (R_l,Mn/Fe = 0.0003 ± 0.0001 : 1) solutions. To understand the different incorporation behaviors of the impurity ions in ferrihydrite, we first considered the size of these impurity ions. The ionic radii of impurity ions have been reported widely to affect their structural incorporation behaviors into Fe (hydr)oxide nanoparticles at neutral pH conditions. If the ionic radius of the impurity ion is similar to that of Fe³⁺, it is easier for the impurity ion to substitute for the Fe³⁺ and cause less stress in the lattice structure of Fe (hydr)oxide nanoparticles, resulting in easier structural incorporation.⁴⁹ The radii of Al³⁺, Mn²⁺, and Fe³⁺ have been reported to be 0.0535, 0.0645, and 0.0645 nm, respectively.^50–53 If the ionic radius was a major controlling factor, then R_l,Mn/Fe should be much higher than R_l,Al/Fe. However, our measurements showed the opposite result: R_l,Al/Fe was ∼100-fold higher than R_l,Mn/Fe. With regard to the effects of ferrihydrite precipitation rates on incorporation of impurity ions, Cismasu et al. synthesized two series of Al-bearing ferrihydrite particles under different precipitation rates, and the chemical compositions of Al-bearing ferrihydrite were similar.²⁷ For Mn-bearing ferrihydrite, the effects of precipitation rates on Mn incorporation were not reported.

Then, we recalled the pathways of ferrihydrite formation. First, Fe³⁺ were hydrolyzed to form Fe(OH)₃ monomers (Fe³⁺ + 3H₂O → Fe(OH)₃ + 3H⁺). Then, polymerization occurred through continuous olation (2[Fe(H₂O)₅OH²⁺] → [(H₂O)₄Fe-(OH)₂-Fe(H₂O)₄]⁴⁺ + 2H₂O) and oxolation (2[Fe(H₂O)₅OH²⁺] → [(H₂O)₄Fe–O–Fe(H₂O)₅]⁴⁺ + 2H₂O).^54–56 In our previous study, ferrihydrite precipitation experiments were also conducted at pH = 3.7 ± 0.2 with aqueous Al/Fe ratios of 1 and 5, and the total Al/Fe ratios in the precipitates were <1%, much lower than the aqueous Al/Fe ratios.¹³ Under acidic pH conditions (pH = 3.8 ± 0.1) in our previous and current study, ∼95% Al ions were present as unhydrolyzed Al³⁺.¹³ Considering the hydrolysis and polymerization reactions during ferrihydrite formation, we hypothesized that only hydrolyzed metal ions (e.g., Al(OH)_n³⁻ⁿ) could be incorporated in the lattice of ferrihydrite through oxolation and olation reactions with Fe(OH)₃ monomers and polymers. Hydrolysis reactions for As metal were instantaneous, much faster than those for ferrihydrite precipitation. Therefore, the hydrolyzed ion concentrations at equilibrium conditions, calculated by GWB, were used to explain their structural incorporation in ferrihydrite. Under our experimental pH condition (3.8 ± 0.1), 4.4% of Al ions were in their hydrolyzed states (Al(OH)²⁺, Al(OH)₂⁺, Al₂(OH)₂⁴⁺, Al₃(OH)₄⁵⁺, Al(OH)₄⁻, and Al(OH)₃), whereas only 0.5% of Mn ions were hydrolyzed. The lower amount of hydrolyzed Mn ions resulted in the small percentage of lattice Mn/Fe ratio in the precipitates, a result which supported our hypothesis. This hypothesis is also supported by studies of Al incorporation during ferrihydrite precipitation at neutral pH conditions.^57,58 For example, Cismasu et al. synthesized a series of Al-bearing ferrihydrite particles in solutions (pH = 7.5 ± 0.2) with a total metal ion concentration of 2 mM and various aqueous Al/Fe ratios (0.10–0.67).⁵⁸ Under the neutral pH condition, as >99% of Fe and Al ions were hydrolyzed, the atomic ratios of Al/Fe in the precipitates were similar to their aqueous ratios.⁵⁸

Our hypothesis was further supported by the reported Cr incorporation in ferrihydrite during ferrihydrite precipitation at neutral and acidic pH conditions.^2,59,60 For example, in the work of Tang et al., a series of Cr-bearing ferrihydrite nanoparticles was synthesized with a total metal ion concentration of 0.1 M and different Cr/Fe ratios (0.1–9) at pH ∼ 7.² Under this neutral pH condition, because >99% Fe and Cr ions were hydrolyzed, the atomic Cr/Fe ratios in the Cr-bearing ferrihydrite particles were similar to the initial aqueous Cr/Fe ratios.² Conversely, under an acidic condition (pH = 3.7 ± 0.1), because only ∼36% of aqueous Cr ions were hydrolyzed, the atomic ratios of Cr/Fe (0.06–0.18) in the precipitates were much lower than the aqueous Cr/Fe ratios (0.2–2.5).³⁰

To summarize, all previous studies of Al and Cr incorporation in ferrihydrite under acidic and neutral pH conditions confirmed our hypothesis: i.e., the hydrolysis of impurity ions affected their structural incorporation in ferrihydrite.^{2,13,30,57–60} IBF precipitation experiments have been conducted,^{2,13,30,57–60} and the ionic radii of metal ions were considered to be the controlling factors for their structural incorporation during IBF precipitation. Here, the hydrolysis of impurity ions, which is highly dependent upon pH, was reported, for the first time, to affect their structural incorporation in IBF nanoparticles greatly. The fundamental information provided here can be used to better understand and control incorporation of impurity ions in ferrihydrite during its precipitation under different pH conditions. Also, it would be an interesting future direction to develop a correlation equation of incorporation of impurity ions in ferrihydrite as a function of the radius, oxidation states, and hydrolysis reactions of impurity ions.

Surface enrichment of Mn on ferrihydrite promoted its precipitation/deposition on SiO₂

The GISAXS scattering curves caused by ferrihydrite nanoparticles precipitated/deposited on SiO₂ are shown in Fig. 2A1–A3. Under all of our experimental conditions, the GISAXS scattering intensities increased with reaction time, indicating continuous precipitation/deposition of ferrihydrite nanoparticles on SiO₂. As the reaction proceeded, the peak positions of the GISAXS scattering curves (indicated by red arrows in Fig. 2) shifted from high q range to low q range, indicating an increase in particle size by growth. By fitting the GISAXS scattering curves, the size evolutions of particles on SiO₂ were plotted (Fig. 3A1). At the end of our 30 min precipitation/deposition experiments, the mean radius of pure, Mn- and Al-bearing ferrihydrite nanoparticles was 4.0 ± 0.1, 4.7 ± 0.1, and 3.3 ± 0.1 nm, respectively. The evolution of total particle volumes (indicative of precipitation/deposition rates) is shown in Fig. 3B1. Compared with pure ferrihydrite on SiO₂, the total volume of Mn-bearing ferrihydrite increased to 160%, whereas the total volume of Al-bearing ferrihydrite decreased to 11%.

Fig. 2 GISAXS scattering curves for particles formed on quartz (A1–A3) and corundum (B1–B3) from Fe-only (Fig. A1 and B1), FeMn (Fig. A2 and B2), and FeAl (Fig. A3 and B3) solutions. The black lines are the fitting curves, which match well with the experimental data (colored dots).

Fig. 3 Sizes of particles precipitated/deposited on quartz (Fig. A1) and corundum (Fig. A2), and the total volumes of particles precipitated/deposited on quartz (Fig. B1) and corundum (Fig. B2), from Fe-only, FeMn, and FeAl solutions.

To understand the different effects of impurity ions on heterogeneous ferrihydrite precipitation/deposition, the electrostatic interactions between the precipitates and substrates were considered. The zeta potential of Mn- and Al-bearing ferrihydrite nanoparticles was measured to be 39.9 ± 2.7 and 40.6 ± 2.3 mV, higher than that of pure ferrihydrite (29.1 ± 5.7 mV). Impurity ions, adsorbed or precipitated on ferrihydrite surfaces, have been reported to change the pH_pzc of ferrihydrite nanoparticles.⁵⁸ With a pH_pzc of 8.5 for pure ferrihydrite, the pH_pzc of Si-bearing ferrihydrite nanoparticles decreased due to the low pH_pzc of silica (2–4) on ferrihydrite surfaces. In contrast, the pH_pzc of Mn(OH)₂ and Al(OH)₃ was higher (∼10) than that of ferrihydrite.^61,62 Therefore, the Mn and Al enriched on ferrihydrite surfaces resulted in the increased zeta potential of Mn- and Al-bearing ferrihydrite nanoparticles.

The zeta potential of SiO₂ powder in the absence and presence of NaNO₃ was similar, indicating no adsorption of Na⁺ onto SiO₂ powder under our experimental conditions (pH = 3.8 ± 0.2). The SiO₂ surfaces in the absence and presence of 1 mM Mn(II) ions under our experimental pH conditions (3.8 ± 0.1) were both negatively charged, with a zeta potential of −22.4 ± 2.6 and −25.1 ± 1.9 mV, respectively. Conversely, in the presence of 1 mM Al(III) ions, the SiO₂ surfaces became positively charged (zeta potential = 18.1 ± 3.9 mV). The different effects of impurity ions on the surface charges of SiO₂ may be related to the different behaviors of metal ion adsorption onto substrates, which is an important topic that we are investigating now, but it is not the focus of the present study. The surface zeta potential of the precipitates and substrates, which determined their electrostatic interactions, explained well the observed trend of heterogeneous ferrihydrite precipitation/deposition. In the presence of Mn, stronger attractive forces existed between the more positively charged Mn-bearing ferrihydrite (39.9 ± 2.7 mV) and negatively charged SiO₂ surfaces (−25.1 ± 1.9 mV), resulting in the promoted Mn-bearing ferrihydrite precipitation/deposition on SiO₂. Conversely, in the presence of Al, repulsive forces existed between the positively charged Al-bearing ferrihydrite precipitates (40.6 ± 2.3 mV) and SiO₂ surfaces (18.1 ± 3.9 mV), thereby hindering Al-bearing ferrihydrite precipitation/deposition on SiO₂. The changes in zeta potential of the substrates in FeMn and FeAl solutions could be caused by adsorption of Mn²⁺ and Al³⁺ or heterogeneous precipitation of Mn(OH)₂ and Al(OH)₃ on substrates. Further discussion about possible adsorption of Mn²⁺ and Al³⁺vs. Mn(OH)₂ and Al(OH)₃ precipitation on substrates can be found in ESI.†

Inhibited heterogeneous IBF precipitation/deposition on Al₂O₃

The GISAXS scattering curves due to pure and IBF nanoparticles precipitated/deposited on Al₂O₃ surfaces are shown in Fig. 2B. Compared with those for SiO₂, the precipitation/deposition of pure, Mn- and Al-bearing ferrihydrite nanoparticles were all inhibited on Al₂O₃. Compared with the precipitates on SiO₂ from Fe only, FeMn, and FeAl solutions, at the end of the 30-min experiments, the particle sizes on Al₂O₃ were smaller (2.5 ± 0.1, 3.4 ± 0.2, and 3.0 ± 0.2 nm, respectively, Fig. 3A2), and the total particle volumes were lower (67%, 31%, and 50% of the particle volumes for pure, Mn- and Al-bearing ferrihydrite precipitates on SiO₂, respectively). Using the total particle volume and mean particle radius, the numbers were calculated. Under Fe only, FeMn, and FeAl solution conditions, the numbers of nanoparticles on Al₂O₃ were calculated to be ∼3 ± 1-, ∼0.8 ± 0.4- and ∼0.7 ± 0.4-fold those on SiO₂. Hence, compared with on SiO₂, the mean sizes of particles on Al₂O₃ were smaller under all three solution conditions; whereas the total particle number was larger on Al₂O₃ under a Fe-only solution. Heterogeneous precipitation (including heterogeneous nucleation and growth) and deposition on substrates can occur simultaneously, and growth controls particle size. The growth of ferrihydrite on substrates was controlled by the attachment of Fe hydroxide polymers onto the substrates, which is affected by the electrostatic forces between the polymers and substrates.^13,36 To understand the reason for the smaller sizes of ferrihydrite nanoparticles on Al₂O₃ than on SiO₂, electrostatic interactions between the polymers and Al₂O₃/SiO₂ surfaces were considered. In Fe-only and FeMn solutions, the Al₂O₃ surfaces were positively charged (zeta potential = 20.5 ± 2.8 and 27.4 ± 6.8 mV), whereas SiO₂ surfaces were negatively charged (zeta potential = −22.4 ± 2.6 and −25.1 ± 1.9 mV). In FeAl solution, the surface of Al₂O₃ was more positively charged (zeta potential = 34.3 ± 6.4 mV) than that of SiO₂ (zeta potential = 18.1 ± 3.9 mV). All precipitates were positively charged so, assuming that the Fe hydroxide polymers carried similar charges to those of the precipitates, the stronger electrostatic repulsive forces between the polymers and Al₂O₃ surfaces inhibited the heterogeneous growth of IBF on Al₂O₃, resulting in smaller particles on Al₂O₃.

Heterogeneous nucleation and deposition can increase the total particle number on substrates. The deposition behavior can be predicted by the Derjaguin, Landau, Vervey, and Overbeek (DLVO) theory. For a Fe-only solution, stronger electrostatic repulsive forces were present between more positively charged Al₂O₃ surfaces and ferrihydrite surfaces, which should have inhibited ferrihydrite deposition onto Al₂O₃ to a greater extent than on SiO₂. Therefore, if deposition was the dominant process, the number of ferrihydrite particles (N) on Al₂O₃ should be less than that on SiO. However, the number of pure ferrihydrite particles on Al₂O₃ was calculated to be ∼2-fold higher than that on SiO₂. Therefore, heterogeneous nucleation, instead of deposition, should be the dominant pathway for the pure ferrihydrite nanoparticle formation on Al₂O₃.

The promoted nucleation of pure ferrihydrite on Al₂O₃ rather than on SiO₂ has been reported in our previous work of pure ferrihydrite precipitation on SiO₂ and Al₂O₃.³⁶ The latter has higher hydrophobicity, which indicates higher substrate–water interfacial energy than for SiO₂,³⁶ so the interfacial energy barrier for heterogeneous ferrihydrite nucleation on Al₂O₃ was lower than that on SiO₂ according to classic nucleation theory.³⁶ Therefore, heterogeneous ferrihydrite nucleation was promoted on Al₂O₃.³⁶

The total volume of nanoparticles (representing precipitation/deposition rates) on Al₂O₃ surfaces was also compared (Fig. 3B2). Compared with pure ferrihydrite precipitation on Al₂O₃, Al- and Mn-bearing ferrihydrite precipitation/deposition was inhibited. Such inhibition can be explained well from the zeta-potential differences of the substrates and nanoparticles in the absence and presence of impurity ions. In the presence of Al and Mn, the zeta potential of Al₂O₃ and precipitate surfaces was more positively charged, and the stronger electrostatic repulsive forces between the substrate and IBF nanoparticles resulted in slower IBF precipitation/deposition on Al₂O₃. In the presence of Al and Mn, the increase in the zeta potential of IBF nanoparticles was caused by Al/Mn enrichment on the particle surfaces, as explained above. The increase in the zeta potential of Al₂O₃ might have been caused by adsorption of Mn and Al ions onto Al₂O₃. Further understanding of the mechanisms controlling ion adsorption onto substrates is not the focus of the present paper, but is a very interesting research topic which we are currently exploring.

In summary, the hydrolysis of impurity ions affected their structural incorporation into ferrihydrite lattices. Also, enrichment of impurity ions on IBF surfaces can affect the zeta potential of IBF nanoparticles. Furthermore, the surface charges of substrates can be altered greatly in the presence of impurity ions. Through all these three mechanisms, impurity ions can affect the heterogeneous precipitation/deposition of IBF on mineral surfaces significantly.

Conclusions

The present study provided new insights on the rates and mechanisms of heterogeneous IBF precipitation/deposition on different substrates. The hydrolysis of impurity ions, which is affected by pH, was found to impact their structural incorporation in IBF nanoparticles. The information provided here can help predict and control the incorporation of metal ions in IBF nanoparticles under different pH conditions. Also, the interactions of the impurity ions and IBF nanoparticles with Al₂O₃ were observed to be totally different to those of SiO₂, which altered the zeta potential of the precipitates and substrates, thus affecting heterogeneous IBF precipitation/deposition considerably. The fundamental information provided here can be used to better predict the fate and transport of metal ions in natural and engineered aqueous systems. It can also be used to better design metal-doped ferrihydrite nanoparticles as adsorbents or precursors of Fe oxide catalysts for industrial applications.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Chong Dai and Juanjuan Liu were supported by a Faculty Start-up Fund from the University of Houston to conduct the GISAXS, DLS, and ICP-MS experiments. Chong Dai was also supported by the US Department of Energy, Office of Science, Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division, to analyze the experimental data and prepare the manuscript. We thank Dr. Xiaobing Zuo for valuable discussion on GISAXS experiments and data analyses at beamline 12-ID-B, Dr. Saul H. Lapidus for HRXRD analyses at beamline 11-BM, Advanced Photon Source, Argonne National Laboratory, and Dr. James D. Kubichi and Dr. Nadine Kabengi for valuable discussion of ion adsorption onto substrates. Use of the facilities at beamlines sectors 12-ID-B and 11-BM at Advanced Photon Source was supported by the US Department of Energy, Office of Science, Office of Basic Energy Science, under contract number DE-AC02-06CH11357.

References

1. C. E. Martinez and M. B. McBride, Environ. Sci. Technol., 1998, 32, 743–748.
2. Y. Tang, F. M. Michel, L. Zhang, R. Harrington, J. B. Parise and R. J. Reeder, Chem. Mater., 2010, 22, 3589–3598.
3. A. C. Vajpel, A. Rousset, K. Uma, I. Chandra, P. Saraswat and V. K. Mathur, Solid State Ionics, 1989, 32–33(Part 2), 741–748.
4. C. Gu, Z. Wang, J. D. Kubicki, X. Wang and M. Zhu, Environ. Sci. Technol., 2016, 50, 8067–8076.
5. H. Zeng, B. Fisher and D. E. Giammar, Environ. Sci. Technol., 2007, 42, 147–152.
6. H. Zeng, A. Singh, S. Basak, K.-U. Ulrich, M. Sahu, P. Biswas, J. G. Catalano and D. E. Giammar, Environ. Sci. Technol., 2009, 43, 1373–1378.
7. Z. Wang, K.-U. Ulrich, C. Pan and D. E. Giammar, Environ. Sci. Technol. Lett., 2015, 2, 227–232.
8. M. Tong, S. Yuan, Z. Wang, M. Luo and Y. Wang, J. Hazard. Mater., 2016, 305, 41–50.
9. M. A. Hinkle, Z. Wang, D. E. Giammar and J. G. Catalano, Geochim. Cosmochim. Acta, 2015, 158, 130–146.
10. G. Lee, J. M. Bigham and G. Faure, Appl. Geochem., 2002, 17, 569–581.
11. C. W. Neil, Y. J. Yang and Y.-S. Jun, J. Environ. Monit., 2012, 14, 1772–1788.
12. P. Wu, C. Tang, C. Liu, L. Zhu, T. Pei and L. Feng, Environ. Geol., 2009, 57, 1457–1467.
13. Y. Hu, Q. Li, B. Lee and Y.-S. Jun, Environ. Sci. Technol., 2013, 48, 299–306.
14. T. L. Jentzsch and R. L. Penn, J. Phys. Chem. B, 2006, 110, 11746–11750.
15. G. Doppler, A. X. Trautwein, H. M. Ziethen, E. Ambach, R. Lehnert, M. J. Sprague and U. Gonser, Appl. Catal., 1988, 40, 119–130.
16. J. Moutinho, T. D. Souza and M. D. C. Rangel, React. Kinet. Catal. Lett., 2004, 83, 93–98.
17. S. Natesakhawat, X. Wang, L. Zhang and U. S. Ozkan, J. Mol. Catal. A: Chem., 2006, 260, 82–94.
18. A. L. C. Pereira, G. J. P. Berrocal, S. G. Marchetti, A. Albornoz, A. O. D. Souza and M. D. C. Rangel, J. Mol. Catal. A: Chem., 2008, 281, 66–72.
19. C. Ratnasamy and J. P. Wagner, Catal. Rev.: Sci. Eng., 2009, 325–440.
20. T. L. Jentzsch, C. L. Chun, R. S. Gabor and R. L. Penn, J. Phys. Chem. C, 2007, 111, 10247–10253.
21. E. B. Quadro, M. D. Dias, A. Maria, M. Amorium and M. D. Rangel, J. Braz. Chem. Soc., 1999, 10, 51–59.
22. J. Liu, C. Liang, H. Zhang, Z. Tian and S. Zhang, J. Phys. Chem. C, 2012, 116, 4986–4992.
23. Y. Masue-Slowey, R. H. Loeppert and S. Fendorf, Geochim. Cosmochim. Acta, 2011, 75, 870–886.
24. G. Waychunas, B. Rea, C. Fuller and J. Davis, Geochim. Cosmochim. Acta, 1993, 57, 2251–2269.
25. S. Krehula and S. Musić, J. Alloys Compd., 2006, 426, 327–334.
26. K. Rout, A. Dash, M. Mohapatra and S. Anand, J. Environ. Chem. Eng., 2014, 2, 434–443.
27. A. C. Cismasu, F. M. Michel, J. F. Stebbins, C. Levard and G. E. Brown, Geochim. Cosmochim. Acta, 2012, 92, 275–291.
28. B. M. Sass and D. Rai, Inorg. Chem., 1987, 26, 2228–2232.
29. C. Dai and Y. Hu, Environ. Sci. Technol., 2015, 49, 292–300.
30. C. Dai, X. Zuo, B. Cao and Y. Hu, Environ. Sci. Technol., 2016, 50, 1741–1749.
31. B. A. Legg, M. Zhu, L. R. Comolli, B. Gilbert and J. F. Banfield, Environ. Sci. Technol., 2014, 48, 13703–13710.
32. T. Tosco, J. Bosch, R. U. Meckenstock and R. Sethi, Environ. Sci. Technol., 2012, 46, 4008–4015.
33. C. I. Steefel and P. Van Cappellen, Geochim. Cosmochim. Acta, 1990, 54, 2657–2677.
34. J. Liu, X. Wu, Y. Hu, C. Dai, Q. Peng and D. Liang, J. Chem., 2016, 2016, 11.
35. Y. Hu, B. Lee, C. Bell and Y.-S. Jun, Langmuir, 2012, 28, 7737–7746.
36. Y. Hu, C. Neil, B. Lee and Y.-S. Jun, Environ. Sci. Technol., 2013, 47, 9198–9206.
37. C. Dai, A. G. Stack, A. Koishi, A. Fernandez-Martinez, S. S. Lee and Y. Hu, Langmuir, 2016, 32, 5277–5284.
38. B. L. Henke, E. M. Gullikson and J. C. Davis, At. Data Nucl. Data Tables, 1993, 54, 181–342.
39. B. Lee, S. Seifert, S. J. Riley, G. Tikhonov, N. A. Tomczyk, S. Vajda and R. E. Winans, J. Chem. Phys., 2005, 123, 074701.
40. G. Renaud, R. Lazzari, C. Revenant, A. Barbier, M. Noblet, O. Ulrich, F. Leroy, J. Jupille, Y. Borensztein and C. R. Henry, Science, 2003, 300, 1416–1419.
41. Y. Li, T. Gao and B. Chu, Macromolecules, 1992, 25, 1737–1742.
42. R. G. Ford, K. Kemner and P. M. Bertsch, Geochim. Cosmochim. Acta, 1999, 63, 39–48.
43. M. Edwards and M. M. Benjamin, J. - Water Pollut. Control Fed., 1989, 61, 481–490.
44. A. Manceau and W. Gates, Clay Miner., 2013, 48, 481–489.
45. M. Alvarez, E. E. Sileo and E. H. Rueda, Chem. Geol., 2005, 216, 89–97.
46. R. Cornell and R. Giovanoli, Clays Clay Miner., 1987, 35, 11–20.
47. C. Hansel, D. Learman, C. Lentini and E. Ekstrom, Geochim. Cosmochim. Acta, 2011, 75, 4653–4666.
48. X. Wang, M. Zhu, S. Lan, M. Ginder-Vogel, F. Liu and X. Feng, Chem. Geol., 2015, 415, 37–46.
49. R. M. Cornell and U. Schwertmann, The iron oxides: structure, properties, reactions, occurrences and uses, John Wiley & Sons, 2003.
50. B. Singh, D. Sherman, R. Gilkes, M. Wells and J. Mosselmans, Clay Miner., 2002, 37, 639–649.
51. D. Schulze, Clays Clay Miner., 1984, 32, 27–39.
52. W. Stiers and U. Schwertmann, Geochim. Cosmochim. Acta, 1985, 49, 1909–1911.
53. A. Y. Ramos, H. C. N. Tolentino, J. Enzweiler, S. M. Netto and M. D. C. M. Alves, Am. Mineral., 2003, 88, 876–882.
54. C. M. Flynn, Chem. Rev., 1984, 84, 31–41.
55. A. L. Rose and T. D. Waite, Environ. Sci. Technol., 2003, 37, 3897–3903.
56. M. Zhu, C. Frandsen, A. F. Wallace, B. Legg, S. Khalid, H. Zhang, S. Mørup, J. F. Banfield and G. A. Waychunas, Geochim. Cosmochim. Acta, 2016, 172, 247–264.
57. M. S. Massey, J. S. Lezama-Pacheco, F. M. Michel and S. Fendorf, Environ. Sci.: Processes Impacts, 2014, 16, 2137–2144.
58. A. C. Cismasu, C. Levard, F. M. Michel and G. E. Brown, Geochim. Cosmochim. Acta, 2013, 119, 46–60.
59. J. E. Amonette and D. Rai, Clays Clay Miner., 1990, 38, 129–136.
60. N. Papassiopi, Z. Gaitanarous and A. Xenidis, Fresenius Environ. Bull., 2012, 21, 2399–2405.
61. J. Rosenqvist, P. Persson and S. Sjöberg, Langmuir, 2002, 18, 4598–4604.
62. H. Jung and Y.-S. Jun, Environ. Sci. Technol., 2015, 50, 105–113.

---

Footnote

† Electronic supplementary information (ESI) available: Details of GISAXS calculations of critical angles, XPS measurements, and discussion on adsorption of Mn²⁺ and Al³⁺ ion vs. Mn(OH)₂ and Al(OH)₃ precipitation on substrates, as well as GISAXS Lorentz-corrected curves (Fig. S1) and XPS measurements of Mn oxidation states (Fig. S2). See DOI: 10.1039/c7en00835j

---