The Dependence of Ilmenite’s Dissolution Rate in Hydrochloric Acid on the Fe(III)/Fe(II) Ratio, with Fe K-Edge XANES Pre-Edge Peak Analysis

Abstract

The complete dissolution of the Ti and Fe content of ilmenite is a necessary first step for the production of TiO₂ directly from ilmenite. Hydrochloric acid is one of the possible solubilizing agents. However, the ability to dissolve ilmenite in hydrochloric acid depends on the nature of the source material. Here, we consider the effect that the oxidation state of Fe in the ilmenite has on the dissolution rate. Three placer ilmenite fractions from New Zealand and India were dissolved in concentrated hydrochloric acid in a stirred vessel. The dissolution rate constants for Fe and Ti for each fraction based on a shrinking sphere model were compared with the Fe(III)/Fe(II) ratio. Pre-edge Fe K-edge XANES as a measure of Fe(III)/Fe(II) has been shown to extend to ilmenite, which has a complex pre-edge region due to the involvement of Ti electronic levels. It was found that there is a relationship between the oxidation state of Fe and the dissolution rate, with a higher Fe(II) content resulting in more rapid dissolution. A higher Fe(II) content reflects a younger, less weathered material, closer to the “standard” stoichiometry of ilmenite. These data and the presented correlation may support the design of industrial processes to digest ilmenite in hydrochloric acid from varying feedstocks.

1. Introduction

Titanium has been declared a critical material by the EU [1], the US [2], and other jurisdictions. The main source of titanium is now ilmenite [3]. The critical titanium materials are metal and the pigment TiO₂. To produce these from ilmenite, first, either the ilmenite is upgraded to a Ti-rich material or TiO₂ is produced directly from the dissolution of ilmenite and direct hydrolysis. Ilmenite occurs in many deposits around the world. It is distributed between sedimentary (37%), igneous (43%), lateritic (19%), and metamorphic deposits (1%). Placer deposits of ilmenite are commonly exploited, because they are easier to extract than hard rock sources and sometimes have lower levels of impurities due to both provenance and a higher level of weathering, which can leach some undesirable components [4]. Placer deposits used in this study were taken from Barrytown in New Zealand [5], where the source of the ilmenite was the metamorphic rocks of the Southern Alps [6], and from the Kanyakumari and Tuticorin regions of Tamil Nadu, India [7].

Hydrochloric acid is one option as a solubilizing agent for ilmenite [8,9,10,11,12]. It can be used either for complete dissolution of the Ti and Fe components of ilmenite, maintaining both these elements in solution, or for the selective leaching of Fe. Selective leaching of Fe is normally achieved by the complete dissolution of ilmenite with the concurrently uncontrolled hydrolysis of Ti so that only the Fe component remains in solution. It is not presently used in any direct pigment production processes, but is used in upgrading the Ti content of ores to form so called “synthetic rutile”, a feedstock richer in Ti than the original ilmenite. Hydrochloric acid digestion has two advantages over sulfuric acid digestion. It is possible to precipitate rutile hydrate directly from hydrochloric acid [9] (the preferred form for pigment material), whereas sulfuric acid always forms anatase directly from solution. Hydrochloric acid can be easier to recycle than sulfuric acid [13].

However, the solubility of ilmenite in hydrochloric acid can be different to that in sulfuric acid. Many ilmenites are relatively insoluble in hydrochloric acid and must first undergo oxidation followed by a reduction step to make them sufficiently reactive [10,14,15,16], while a few are relatively soluble directly in hydrochloric acid [8,12]. Factors that influence these differences in solubility are not well established.

Factors affecting the rate of ilmenite digestion in hydrochloric acid have been studied, including the usual considerations of temperature, acid concentration, time, and agitation, as described in a comprehensive review [13]. Factors specific to ilmenite have received less attention, such as the development of pores and other structures assisting dissolution [10,17] or avoiding the hydrolysis of titanium [18,19].

Here, the goal was to try to understand factors specific to different ilmenites that affect their solubility in hydrochloric acid. The hypothesis tested was that the dissolution rate depends directly on the Fe(II)/Fe(III) ratio. Dissolution rates were measured for a selection of ilmenites with different levels of iron oxidation. As an adjunct, to investigate a spectroscopic method to measure the Fe(II)/Fe(III) ratio in ilmenite, pre-edge XANES is one option that has proved useful for iron containing minerals [20,21]. We provide an extension of the Fe(II)/(III) method to ilmenite (and probably similar minerals).

2. Materials and Methods

2.1. Ilmenite

Ilmenite concentrate from Barrytown (BT), Westland, New Zealand was supplied by Rio Tinto Pty Ltd. Ilmenite concentrates in two grades from the Kanyakumari (TIS) and Tuticorin (TVP) regions of Tamil Nadu, India, were supplied by VV Minerals Ltd. All three ilmenite fractions were from placer deposits. The ilmenite was ground in small batches (35 g for 2 min) in a 150 mm disc mill (Tema, Siebtechnik Tema Ltd., Daventry, UK) for the digestion experiments. The particle size of the ground ilmenite was measured with a Malvern Mastersizer 3000 (Malvern Instruments, Malvern Panalytical, Malvern, UK).

2.2. SEM/EDS

Samples of unground ilmenite for scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS) were mounted in Epo-tek 301-1b resin (Epoxy Technology Inc., Billerica, MA, USA), polished at a final polishing powder size of 1 µm, then coated with carbon. Images were recorded using a FEI QUANTA 200 SEM (ThermoFisher, Hillsboro, OR, USA) and EDS spectra were collected employing a spot size of 4.0 (instrument setting—no units) at a count rate of 2000–3000 counts per second and a dead time of 30%.

2.3. Dissolution Procedure

The ground ilmenite was digested in 38% w/w HCl (10.7 mol kg⁻¹ HCl or 12.7 M HCl at 20 °C) at a temperature of approximately 84 °C. An acid/ilmenite ratio of 1.44 to 1 was used, where this was calculated as kg HCl to kg ilmenite in the ore. The equivalent ilmenite in the ore was obtained by treating Fe₂O₃ as FeO, then calculating it as (mol Ti + mole Fe)/2. In this work, 6.12 mol HCl to 1 mol “FeTiO₃” was used in digestion, noting that the ratio of Ti to Fe in the ilmenite was not exactly 1:1.

The reaction of ilmenite with HCl can be represented by Equation (1), where the concentration of titanium species in solution depends on the concentration of hydrochloric acid.

FeTiO_3(s) + 4HCl_(aq) → Fe²⁺_(aq) + TiOCl_2(aq) + 2Cl⁻_(aq) + 2H₂O

(1)

The digestion of ground ilmenite in HCl was performed in a 1.5 L round-bottom glass flask placed on a heating mantle and fitted with a reflux condenser. This slurry was agitated with a glass paddle stirrer on a glass shaft. The HCl was preheated to 40 °C before the addition of ilmenite.

The content of Ti and Fe in the digestion liquor with time was obtained by the reduction of a diluted aliquot of filtered liquor with Al metal followed by quantitative oxidation by titration with cerium sulfate solution. Titration endpoints for Ti were indicated by methyl blue and for Fe by phenanthranilic acid.

2.4. Dissolution Kinetics

Dissolution kinetics were fitted to a shrinking sphere model using Sigmaplot (Systat Software Inc., San Jose, CA, USA, Version 16) to obtain standard errors.

2.5. X-Ray Absorption Spectroscopy (XAS)

XAS spectra were recorded at the Fe K-edge using the bending magnet MEX1 beamline at the Australian Synchrotron. Samples of ground ilmenite were diluted (to 5 wt%) with cellulose and pressed into pellets for analysis in transmission mode at ambient temperature. The beam energy was controlled using a Si(111) double crystal monochromator providing (ΔE/E = 1.4 × 10⁻⁴). The beam size on the sample was 0.5 × 3 mm, with a photon flux around 3 × 10¹⁰ ph sec⁻¹. An Fe metal foil placed before the last ion chamber was used as an energy reference with each sample.

Data processing for XANES used Larch software (2024.12.0) running under Larix [22]. The energy was calibrated to Fe foil run in series with each sample, with E₀ for the metal defined as 7110.7 eV [23,24]. A baseline was subtracted for the pre-edge region using a Victoreen function. The pre-edge region was then fitted with Gaussian peaks.

3. Results and Discussion

3.1. Particle Size Distribution

Each of the ilmenite fractions had a different particle size distribution with a median size varying between 150 and 250 μm (Figure 1a). After grinding, the particle size distribution was more similar between the three processed fractions, with a median size of ca. 50 μm (Figure 1b). This served two purposes: 1. to reduce differences in the dissolution kinetics resulting from different particle sizes, and 2. to achieve a faster dissolution rate to simplify the experimental work.

3.2. Ilmenite Composition

The elemental analysis of the ilmenite fractions was provided by the supplier for the TIS and TVP fractions and was available in the literature for the BT fraction [8] (

Table 1. Ilmenite composition (supplied by VV Minerals for TIS and TVP and from (Haverkamp et al., 2016 [8]) for BT).

Element as Oxide	Barrytown BT (wt%)	Kanyakumari TIS (wt%)	Tuticorin TVP (wt%)
TiO2	47.1	51.5	46.1
Fe2O3	3.0	12.7	15.8
FeO	37.5	32.9	34.4
Al2O5	2.3	0.5	0.4
SiO2	5.2	1.1	1.1
P2O5	0.2	<0.1	<0.1
V2O5	0.1	0.2	0.3
MnO	1.7	n.a.	n.a.
CaO	1.1	n.a.	n.a.
K2O	0.3	n.a.	n.a.
MgO	0.2	n.a.	n.a.
Sum	98.7	98.9	98.0

n.a. not analysed.

). These fractions were chosen to have a range of Fe₂O₃ content relative to FeO; the ratios for the three samples were 0.08, 0.385, and 0.459 (using the oxide masses), or using Fe(III)/Fe(II) molar ratios, the ratios were 0.072, 0.347, and 0.413. This range was large enough to cover most of the natural range of Fe(III)/Fe(II) in ilmenite.

The structure of ilmenite grains differs between ores. The Barrytown, NZ, ore contained a high proportion of small inclusions, mostly aluminosilicates, hence the high elemental composition of these elements (Figure 2a). It also contained some apatite inclusions, and so had a relatively high phosphorus content [5]. The Kanyakumari and Tuticorin ores contained significantly fewer inclusions (Figure 2b,c), and those that existed appeared to be more higher atomic mass compositions, probably iron oxide, although elsewhere a more thorough investigation of the composition of ilmenite from these deposits has shown the presence of ilmenite–hematite and magnetite–ilmenite intergrowths with some titanomagnetite [7].

3.3. Dissolution Kinetics

The dissolution of the ground ilmenite was carried out under conditions that enabled the solubilisation of both the iron and titanium components, without the hydrolysis of the Ti component, which would enable the digestion liquors to subsequently be used for the controlled hydrolysis of TiO₂ [9]. The dissolution curves show no drop off in Ti concentration (Figure 3), which indicates that no hydrolysis of Ti took place.

These dissolution kinetics have been shown previously to follow a shrinking sphere model where the rate is determined by the rate of chemical reaction at the surface of the ilmenite particles [8], which is represented by Equation (2)

$$kt=1-{(1-a)}^{1/3}$$

(2)

where k is the rate constant, t is dissolution time, and a is the extent of reaction. Dissolution data measured here were fitted to these kinetics (Figure 3b) to give rate constants (with standard deviations in parentheses) for BT, TIS, and TVP, respectively, of 1.985 (0.038), 1.569 (0.037), and 1.200 (0.044) × 10⁻³ min⁻¹. These rate constants incorporated the effect of the changing surface area as the particles dissolved.

3.4. Fe(III)/Fe(II) Measurement by XANES

The pre-edge region of the Fe K-edge XAS spectra for the ilmenites showed a feature that could be resolved into four peaks. In a simple iron oxide there are two peaks present (in some cases resolvable to three peaks), corresponding to the 1s → 3d transition in Fe. The centroid of these peaks can be used to quantitatively determine the Fe(III)/Fe(II) ratio [20,21,25].

In ilmenite, the pre-edge region is composed of four peaks [26,27]. The two additional peaks, at higher energy, result from electric dipole transitions to excited Fe* 4p states mixed with the neighboring Ti 3d states [26,28]. The two lower energy peaks due to the Fe 1s → 3d transition can be resolved into three peaks with 1s2p RIXS (resonant inelastic X-ray scattering).

For the ilmenite samples analysed here, which were mixed mineral samples dominated by ilmenite-like phases, it was possible to fit four peaks to the pre-edge region (Figure 4). We fitted Gaussian peaks. No improvement in the fit quality was obtained using a pseudo-Voigt peak shape. A physically realistic fit was obtained for the BT and TVP samples with an unconstrained four-peak fit. For the TIS ilmenite, an unconstrained fit resulted in the Fe(III) peak becoming large and very broad, clearly a non-physical fit. Therefore, constraints were applied to the TIS fit with the result that the TIS centroid, although with a good statistical fit, was not considered reliably determined here because it was dependent on the choice of realistic constraints. The centroid for the first two lower energy fitted pre-edge peaks were BT 7111.8 (<0.1) eV, TIS 7112.2 (error ill defined) eV, and TVP 7112.3 (0.1) eV. The centroid value presented for TIS is represented by the constrained fit shown in Figure 4 with error bars assigned very conservatively based on the range of peak positions possible when different constraints were chosen. The fit for the four Gaussian peaks to data produced a reduced χ² of 4.2 × 10⁸, 5.1 × 10⁸, and 6.8 × 10⁸ for BT, TIS, and TVP, respectively (Figure 5a).

This was compared with data for iron-containing basalt presented by Berry et al. [25] where the additional pre-edge peaks due to the interaction with Ti were not present (Figure 5b). Considering the differences in the nature of the pre-edge, the match with the Berry data is surprisingly close (and a similar match can be made to data from Cotterell) [21]. This suggests that this method of determining Fe(III)/Fe(II) in minerals is able to be extended more broadly than has been applied to date, to cases where electric dipole transitions to excited Fe* 4p states mixed with the neighboring Ti 3d states are also present.

3.5. Relationship Between Oxidation State and Dissolution Kinetics

The ilmenite containing lower Fe(III) was found to have a higher rate constant for dissolution in hydrochloric acid than ilmenite with more Fe(III) present (Figure 6).

The reason for this difference could be in part due to the degree of alteration of the ilmenite since its formation from magma or metamorphosis. The alteration of ilmenite is likely to involve oxidation of the Fe(II) to Fe(III) [29,30,31], and therefore the ratio of Fe(III)/Fe(II) can be a proxy, in some cases, for weathering and alteration. It is not clear why alteration should result in slower dissolution kinetics—the loss of structural integrity and lattice stability might be thought to encourage dissolution through a greater density of dislocation sites and fracture channels.

Ilmenite from a wider range of sources would be useful to expand the test of this hypothesis. Not included here were any hard rock sources. These might be expected to have less weathering and environmental alteration. However, Tellnes ilmenite from Norway, as an example, has 6.5% Fe₂O₃ and 40.87% FeO [10], giving a low oxide ratio of 0.16 (or 0.14 Fe(III)/Fe(II)), but this is not as reduced as the Barrytown ilmenite included in this study. It has however, been shown to be quite soluble in HCl [10], unlike many ilmenite deposits, and this is therefore likely to support the relationship discovered here between the Fe(III)/Fe(II) ratio and HCl solubility.

An alternative method to consider mineral alteration and its effect on solubility might be to use high resolution X-ray powder diffraction to measure the lattice strain within ilmenite, which could be expected to increase with weathering, including the result of Fe(II) being oxidised to Fe(III), and this lattice strain might correlate with solubility. However, it is generally accepted that increased lattice strain and increased disorder leads to greater solubility in many systems [32], not lower solubility, as might be inferred from data presented here. This apparent anomaly deserves further investigation.

These data and the presented correlation may support the design of industrial processes to digest ilmenite in hydrochloric acid from varying feedstocks. Where a feedstock contains a higher Fe(III)/Fe(II) ratio, slower digestion kinetics might be expected, and this results in slower throughput and greater risk of hydrolysis of the Ti component during digestion. Where a lower Fe(III)/Fe(II) ratio is present, more rapid dissolution could be expected, and therefore, a more rapid rise in temperature due to the exothermic reaction; this needs to be included in the reactor design.

4. Conclusions

It has been shown with a selection of ilmenites that the dissolution rate of ilmenite in hydrochloric acid is correlated with the Fe(III)/Fe(II) ratio, with a smaller fraction of Fe(III) resulting in more rapid dissolution in hydrochloric acid. This is contrary to what might be expected, where more Fe(III) should lead to greater structural disorder, which would therefore be expected to lead to more rapid dissolution. Determining the mechanism behind this change would be an interesting subsequent study that might be informative more broadly for mineral dissolution. It has also been demonstrated that the complex Fe K-edge XANES pre-edge features of ilmenite can be used to measure this Fe(III)/Fe(II) ratio, while previously this analysis method was restricted to a simple pair of pre-edge peaks. This understanding of the solubility of ilmenite can support recent interest in titanium extraction from ilmenite using hydrochloric acid, with pre-commercialization work currently being undertaken.