1. Introduction
Recent trends in research have sparked a renewed interest in establishing methods which are efficient in processing ores and concentrates using hydrometallurgical and pyrometallurgical techniques [
1]. Mineral processing is often complicated by the presence of impurities and the inability of the method to achieve clean separation. The efficiency of extraction in hydrometallurgical and pyrometallurgical techniques is mainly dependent on the degree of selectivity and the amount of impurities in the sample. In the hydrometallurgical extraction of platinum group elements (PGE) from chromite ore, impurities derived from refractive elements and other elements, such as aluminum, ferric iron, magnesium, and chromium, complicate the mineral processing by producing molten sticky products [
2], which affect the normal operating conditions. Refractive elements occur in chromite ores in different quantities depending on the spinel structure (Mg
2+, Fe
2+) O (Cr
3+, Al
3+, Fe
3+)
2O
3. The removal of base metals is essential to the PGE industry for the removal of interfering base metals, which often limit the normal operating conditions of the blast furnace by forming sticky melts, which cause high operational costs and frequent shutdown of plants.
In different studies, we investigated the isolation of chromium, vanadium, and tungsten from chromite, titanomagnetite (AMIS 0501), and wolframite mineral ore using ammonium phosphate salt as flux. The use of ammonium phosphate as flux was key in isolating these metals as insoluble micro-crystalline c-type metaphosphate M(PO
3)
3 particles (where M = Cr
3+ and V
3+), and soluble phosphates compounds such as WO
2(PO
3)
2 [
3]. From our previous studies, it was shown that the formation of metaphosphate compounds provided an alternative approach for separating chromium and vanadium [
4] from samples with complex matrices, such as mineral ores [
5]. However, elemental analysis of the isolated chromium and vanadium metaphosphate products revealed the presence of Al, Fe, or Ti as impurities. Metaphosphate products of chromium revealed the presence of 5% (Fe), 2% (Al), and <1% (Ti), whilst the metaphosphate product of vanadium revealed the presence of 21% (Fe), 3% (Al), and 45% (Ti). Both metaphosphate products revealed impurities with similar chemical compositions, which prompted a revisit of this technique.
A review of this fusion method was prompted by the desire to improve the selectivity of the ammonium phosphate fusion method and to increase the purity of the metaphosphate’s products. The key drive to this investigation was the need to determine the optimum conditions for the selective separation of pure metaphosphate compounds from samples with complex matrices. To achieve this, a review of the reported metaphosphate compounds of the transition elements was considered, as well as the conditions under which they were formed. This background information was critical to decipher the disparity between the unreacted starting material and the possible by-products. Although, due to the trivalent oxidation state of the transition metals, the formation of metaphosphate compounds is considered unusual in solid-state chemistry. These compounds are stabilized by a dense network of polyphosphate (PO
4)
3− units, which forms an anionic framework that has a high degree of chemical, mechanical, and thermal stability [
6].
Numerous metaphosphate compounds have been reported as belonging to a series of polyphosphate compounds [
7,
8]. These polymorphic compounds can conform into different crystalline forms denoted as A, B, C, D, and E, of which the C-type is the predominant form [
9]. The crystalline structure of chromium [
10] and vanadium metaphosphate compounds obtained in our previous study belong to the C-type compounds. The C-type polyphosphate compounds (M(PO
3)
3 M = Cr
3+ and V
3+) are composed of infinite PO
3− chains and a network of individual MO
6 octahedral structures connected through an infinite number of repeated PO
4 tetrahedral units, as shown in
Figure 1 [
11]. Studies have shown similarities in the chemical and physical properties of these isostructural compounds V(PO
3)
3 [
12], Cr(PO
3)
3 [
6,
13,
14], Al(PO
3)
3 [
15], Ti(PO
3)
3 [
16], and Fe(PO
3)
3 [
6]. The formation of mixtures of metaphosphate products was considered highly inevitable in samples such as chromite (Mg
2+, Fe
2+)O (Cr
3+, Al
3+, Fe
3+)
2O
3 and titanomagnetite Fe
2+(Fe
3+, Ti)
2O
4 mineral ore, which bear different M
3+ ions. Studies have shown similarities in the magnetic properties, thermal stability, high melting points, etc., of these metaphosphate compounds, making separation a daunting task [
17].
It is evident from the preceding analysis that the current fusion method using ammonium phosphate flux is susceptible to the formation of different metaphosphate products. The possibility of co-precipitation is highly probable in mineral ore samples due to the presence of impurities. Therefore, it was decided that alternative operational conditions were necessary to improve the selectivity of this method. The preliminary stage of this investigation involved a review of this fusion process to determine the optimum conditions necessary for the isolation of pure metaphosphate compounds. The reaction mechanism was also examined in an attempt to optimize the selectivity and purity of the products.
3. Results and Discussion
The analysis of the selective isolation of metaphosphate compounds from the inorganic and mineral ore sample (CRM) was first performed by probing the effects of temperature. Temperature was determined in our previous studies to be key in the formation of metaphosphate products and in achieving complete sample dissolution. Although various approaches for the synthesis of metaphosphate compounds have been used at different temperatures [
19], it was yet to be established if temperature changes had an effect on the dissolution as well as the formation of metaphosphate compounds. The optimum temperature for the fusion extraction of metaphosphate compounds using various mineral ore samples was previously determined to be 800 °C. Preliminary investigations at lower temperatures (<800 °C) showed incomplete dissolution, and at higher temperatures (>800 °C), viscous molten melts were obtained, which solidified the melts so that they were difficult to dissolve. The fusion of pure salts (AlCl
3, CrCl
3, FeCl
3, and VCl
3) at 800 °C yielded solid melts which were glassy and crystalline (
Figure 2). It is worth noting that all of the solid melts obtained using this method were hygroscopic, which made them easier to dissolve in water or any other polar solvents.
The different products obtained at temperatures above and below 800 °C were consistent with the research findings conducted by Magda et al. [
20], which shows the decomposition stages of ammonium phosphate salt. The decomposition of ammonium phosphate salt (Equations (1)–(4)) forms different species depending on temperature. The uncertainty of forming a variety of products increased as the temperature increased beyond 800 °C and more phosphate species formed.
3.1. XRD Characterization
Precipitates obtained from the pure salts, CRM mineral ore, and inorganic salts were analysed using the XRD technique to identify their composition as well as the purity of the isolated precipitates. The XRD pattern of the precipitates obtained from pure salts were evaluated and compared with those in the XRD data base. The differences in the XRD pattern confirmed the successful conversion of the starting material to products. It was confirmed that the patterns corresponded to the metaphosphate compounds of Al(PO
3)
3, Cr(PO
3)
3, Fe(PO
3)
3, and V(PO
3)
3 [
12] (
Figure 3a). The XRD patterns for the precipitates obtained from the inorganic salts showed that the following mixtures were present in the metaphosphate compounds: Cr(PO
3)
3 [
11] Fe(PO
3)
3, Al(PO
3)
3, and Ti(PO
3)
3 [
16]. The XRD pattern of the CRM mineral ore positively identified the presence of two major phases in the ore: magnetite (Fe
3O
4) and ilmenite (FeO·TiO
2). These phases in the mineral ore are the major sources of iron (Fe
2+/Fe
3+) and Ti
3+ in the CRM mineral ore. The XRD pattern for the CRM showed dominant peaks in the region of 20–35° (2θ) corresponding to Fe(PO
3)
3, Al(PO
3)
3, and Ti(PO
3)
3, as shown in
Figure 3b.
3.2. IR Characterization
The IR spectra of the precipitates from the inorganic salts and CRM mineral ore exhibit three strong peaks in the region of 1246–1273, 956–964, and 713–767 cm
−1, which were characteristic of the stretching vibration of P–O
ext and the asymmetric stretching frequencies of (P–O
int)PO
2 and P–O–P of the metaphosphates. The spectrum obtained from the CRM precepitate showed an additional sholder peak in the region of 1222 cm
−1, which also corresponds to the P–O
ext vibration of the metaphosphates. The three stretching bands in the region of 964, 767, and 713 cm
−1 were assigned to the stretching mode of the P–O–P bridge bonds. Both spectra of the two products exhibit bands with a poor resolution in the range of 470–615 cm
−1, which is mainly predominated by the antiasymetric stretching modes of the isolated tridimensional network of MO
6 units, and to a lesser extent, by the antisymmetric bending modes of the chain group (O–P–O). A summary of the stretching frequencies of both products is provided in
Table 2. The stretching fequences of both products are compared with those of the isostructural metaphosphate compounds belonging to the C-type. The IR spectra of the precipitates from the inorganic salts showed a number of similar peaks to those of the C-type chromium (III)metaphosphate compound, whilst the precipitate from CRM mineral ore was closely related to the C-type iron (III) metaphosphate. The splitting peaks which can be observed in the region of 665–765 cm
−1 of both spectra is attributable to the high crystallinity of the proucts [
21]. This was confirmed by the sharp distinguished peaks of both precipitates and the complete lack of background noise compared to the pattern of the CRM mineral ore, which was amorphous.
3.3. SEM–EDX and ICP-MS Analysis
3.3.1. Analysis of Products from a Mixture of Inorganic Salts
The morphology and the surface/interface characterization of the precipitates from the inorganic salts and CRM mineral ore sample (AMIS) was conducted using the SEM–EDX technique at a nanospace range of 0–20 keV. The objective of the analysis was to identify the physical and chemical phases in the precipitates, as well as the elemental composition. SEM images (
Figure 4) of the precipitates from the inorganic salts showed micro-crystalline particles that were irregular and clustered. These particles were uniformly distributed throughout the sample and consisted of O (51.3 wt%), P (31.7 wt%), Cr (6.7 wt%), Al (4.6 wt%), Fe (1.8 wt%), and V (2.0 wt%). All the elements added as starting materials (i.e., Al, Cr, Fe, and V) in the form of inorganic salts were retained in the precipitate but in different elemental proportions. A mixture of these elements in the precipitate confirmed the inability of the ammonium phosphate fusion method to selectively precipitate a single product. The order of preference, according to the precipitated inorganic product, was Fe (1.8 wt%) < V (2.0 wt%) < Al (4.6 wt%) < Cr (6.7 wt%), which corresponds to the percentage conversion of 44.37% (Cr), 30.46% (Al), 13.25% (V), and 11.92% (Fe). It is interesting to note that iron was the least preferred amongst the M
3+ metals despite its abundance. The results also suggest chromium has the highest likelihood of being precipitated first as chromium(III) metaphosphate. The aforementioned precipitation order, determined by EDX analysis, was attributed to the differences in the rate of reaction between the phosphate (PO
3−) ions and the M
3+ ions (M = Al, Cr, Fe, and V).
The ICP-MS analysis of the filtrate solutions obtained after the removal of solid precipitates aimed to determine the elemental content of the unreacted M
3+ ions and other ions in the sample. The percentage conversion of the M
3+ ions to metaphosphate compounds was determined by analyzing the corresponding elemental content in the filtrate solutions.
Table 3 shows the quantitative results of the filtrate solutions. The matrix effects caused by the presence of polyatomic ions were circumvented using the KED helium mode. The percentage content of unreacted M
3+ ions in the filtrate solutions were 91.30% (Fe), 88.82% (V), 71.52% (Al), 54.87% (Cr). The high recoveries of Fe in the filtrate solution was assumed to be the result of a slow interaction between the PO
3− and Fe
3+, which led to a reduction in product formation in comparison to the other trivalent ions Cr
3+, Al
3+, and V
3+ ions.
3.3.2. EDX Analysis of Precipitates from CRM (AMIS 06038) Mineral Ore Sample
In this part of the study, we investigate whether the elemental content in the ore has any bearing influence in the selective precipitation of Al, Cr, Fe, Ti, and V as metaphosphates. According to the certificate, the percentage content of these elements in the CRM were reported as Fe
2O
3 (75.51%), TiO
2 (13.91%), Al
2O
3 (3.64%), V
2O
3 (1.52%), Cr
2O
3 (0.24%). Magnetite (Fe
2+(Fe
3+)
2O
4) and rutile (TiO
2) were identified by the XRD pattern as the dominant phases in the mineral ore. The fusion of this CRM sample using ammonium phosphate flux (sample: flux, 1:25) yielded a grey powdered precipitate. SEM–EDX analysis of this product revealed micro-particles that were crystalline and irregular in shape and size. The elemental compositions of these particles, as revealed by EDX analysis (
Figure 5), were in the order of Cr (<0.5 wt%) < V (0.5 wt%) < Al (2.6 wt%) < Ti (3.5 wt%) < Fe (17.8 wt%). Interestingly, the elemental contents in the precipitate were ordered according to the quantities in the CRM. Ions in high concentration, e.g., Fe and Ti, were precipitated in higher proportions compared to those with lower concentrations. The proportions of the metal concentrations precipitated from the CRM mineral ore were Fe (68.45%), Ti (9.46%), Al (2.37%), V (0.98%), Cr (<0.1%). These results confirmed the influence of (M
3+) ion content (either formed or naturally occurring +3 ions) in the original sample towards the precipitation of metaphosphate compounds. The results obtained in this part of the study also demonstrated the influence of M
3+ content on selectivity, i.e., the higher the M
3+ ion content in a sample, the higher the metaphosphate product.
Analysis of the CRM filtrate solutions using ICP-MS revealed an average percentage recovery of Fe (7.06%), Ti (4.45%), Al (1.27%), V (0.54%), Cr (0.24%) from the original source. The percentage content of the M
3+ ions retained in the filtrate solution differed from those obtained from the inorganic solutions. The higher percentage proportion of iron in the filtrate solution was presumed to be the result of Fe
2+ ions from the starting material (Fe
2+(Fe
3+,Ti)
2O
4). Since iron (Fe
3+/Fe
2+) occupied over ~76% of the CRM, it was inevitable that most of the Fe
3+ ions in solution reacted with the PO
3− ions, thereby limiting the chance (~24%) of the other elements (Ti
3+, Al
3+, V
3+, and Cr
3+ ions) interacting with PO
3- ions. Other elements identified in the CRM filtrate solution included Na, Mg, K, Ca, Si, Mn, Co, Ni, Cu, and Ba, and their respective percentage contents are listed in
Table 3. The presence of only Fe, Ti, Al, V, and Cr in the precipitate further confirmed the preference (selectivity) of this method towards the ions with a trivalent oxidation state.
4. Conclusions
The selective precipitation of base metals (M = Al, Cr, Fe, Ti, and V) was achieved using the ammonium phosphate fusion method. This procedure was determined to be ideal for the extraction of base elements as metaphosphate compounds. The chemical composition of the isolated metaphosphate compounds contained metal ions in a trivalent oxidation state. Therefore, it can be concluded that this extraction technique was ideal for metals that can form a stable trivalent oxidation state, as no impurities with different oxidation states were found. The isolated metaphosphate products were insoluble in both polar and organic solvents. The fusion method was characterized by poor selectivity as mixtures of metaphosphate from the M3+ compounds were precipitated. Selectivity was shown to be dependent on the percentage content of M3+ ions in the original sample or formed during the fusion process. It was also observed that selectivity improved in samples, such as the CRM, which contained high iron content (75.51%, presumably Fe3+/Fe2+ ions).