1 Introduction

For an iron ore deposit to be considered economically recoverable, it must contain at least 25% iron [1]. However, ores with iron content lower than 25% can still be economically exploited if the ore deposit is large (economy of scale), can be concentrated (beneficiation), and can be transported cheaply [2]. Out of more than 300 iron-bearing minerals, five (haematite, magnetite, maghemite, goethite, and siderite) are the primary sources of iron ore minerals for steelmaking [1]. Phosphorous is reported to be usually present in secondary iron oxide minerals, such as limonite, ochre, goethite, secondary haematite, and alumina-rich minerals such as clay and gibbsite, and in apatite/hydroxyapatite in magnetite ores [3]. The acceptable levels of P in hot metal range from 0.08 to 0.14 [4]; hence, iron ores for steelmaking with very low phosphorus contents are desirable. Iron ores can be classified with respect to their phosphorus content as low-phosphorus ore (< 0.07% P), medium-phosphorus ore (0.07–0.10% P), or high-phosphorus ore (> 0.10% P) [5].

High-phosphorus iron ores abound worldwide and constitute the bulk of “low grade” iron ores. For instance, China is reported to have a proven reserve of approximately 4.0 billion tons of low-grade (Fetotal 35–50%), high-phosphorus (0.4–1.8% P) oolitic haematite ore [6, 7], while Nigeria is reported to have a proven reserve of 1.25 billion tons of low-grade (Fetotal 45.6–54.2%), high-phosphorus (0.76–2.69% P) oolitic goethite ore [8,9,10]. Australia’s reserve of low-phosphorus ore (< 0.07% P) in Western Australia is estimated [11] to be about 20.4 billion tons and projected [12] to be depleted by 2030, while there are 8.0 billion tons of high-phosphorus ore (> 0.10% P) [5]. High phosphorus content is reported to account for the non-exploitation of Moncorvo iron ores located in the Trás-os-Montes region, northeast Portugal, and estimated to exceed 1000 million tons of iron ore reserves, arguably among the largest iron ore reserves in Europe [13,14,15]. The Moncorvo iron ores of Portugal are reported to be low grade ranging in iron content from 35 to 43 wt% Fe with an average iron content of 37 wt% Fe, coupled with a high phosphorus content that ranges from 0.3 to 0.7 wt% P, though phosphorus content in some samples in excess of 1.2 wt% P has also been reported [14]. The Moncorvo iron ores of Portugal appear to be the subject of some ongoing valorization research efforts aimed at bio-desphosphorization [16]. Table 1 summarizes information on some of the high-phosphorus ores worldwide.

Table 1 High-phosphorus iron ores worldwide

Iron ores with phosphorus content lower than 0.07% P are commercially desired as they are suitable for use in iron production. As a consequence, a lot of research efforts have been focused on phosphorus removal from high-phosphorus iron ore [20, 24, 32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48] in a bid to reduce the P content to acceptable levels. Processes that had been employed for phosphorus removal from iron ores include thermal treatment [7, 21, 24, 38,39,40,41, 49,50,51,52,53], microwave treatment [38, 54,55,56], ultrasonic treatment [54], acid leaching [12, 17, 20, 49, 56,57,58,59,60], alkaline leaching [5, 59], bioleaching [30, 61,62,63], agglomeration methods [64,65,66,67], and froth flotation [31, 52, 68, 69]. For a more detailed appreciation of the methods employed to process and valorize high-phosphorus-content ores, the recent reviews by Quast [70] and by Pereira and Papini [71] are recommended. With an ever-growing demand for natural resources and rapid depletion of high-quality sources for minerals, there arises a need for research on lower-grade minerals to ensure their economic exploitation. In this millennium, the secret of continued availability of some mineral resources might depend on the development of new methods and systems, for refining and using hitherto unusable and hence unexploited mineral grades. In order to achieve this goal, a good understanding of the constitution of each of these mineral sources is vital. Taylor et al. [72] had opined that exploitation of lower-grade iron deposits will probably supplement the dwindling reserves of high-grade ores in the future.

Agbaja iron ore deposit estimated at 1250–2000 million metric tons of ore reserve constitutes the largest known Nigerian iron ore deposit, but is currently under-exploited due to its high phosphorus content. It has been reported to consist of oolitic and pisolitic structures rich in iron oxides, in a predominantly clay matrix [73, 74]. The principal constituent mineral has been reported to be goethite, with minor quantities of haematite, maghemite, siderite, quartz, kaolinite, pyrite, and an average of 0.09% S [8, 9, 73]. The Agbaja iron ore is also reported to have a high phosphorus content in the range of 0.76–2.13% [25, 26, 75] which makes its use in iron making in the blast furnace without reduction of the phosphorus content impractical. To overcome this challenge and exploit this huge reserve, a thorough understanding of the ore composition and the partitioning of the elements/compounds present is necessary. In this work therefore, characterization of Agbaja iron ore has been done using a combination of material characterization techniques ranging from optical microscopy, scanning electron microscopy with energy-dispersive X-ray spectroscopy, powder X-ray diffraction, and thermal gravimetric analysis and differential scanning calorimetry.

The phosphorus in high-phosphorus haematite ore is reported to occur predominantly in the form of apatite (Ca3(PO4)2) [6]. According to Fisher-White et al. [5], although microscopic inclusions of apatite have been reported in some Australian ores, phosphorus is reported to occur predominantly within the goethite (α-FeOOH) fraction [76,77,78] in concentrations that are thought to be indicative of the original amounts of the mineral apatite (Ca5(PO4)3OH) in the ore [79]. Phosphorus in the iron ore is reportedly capable of being attached to the goethite, and FeO(OH) in the form of a solid solution [76]. Wells and Ramanaidou [80] studied the occurrence and mineralogical association of phosphorus in Australian bedded iron ore deposits and concluded that P is associated mainly with goethite most likely as adsorbed PO4, in the absence of direct experimental evidence of Fe replacement by P within the structure of synthetic goethite that is detectable as changes in lattice parameters. Earlier works based on Australian iron ores had reported [81] that phosphorus can occur in iron ores in wide variety of ways: as microscopic inclusions of hydroxyapatite Ca5[PO4]3(OH) [82], as secondary phosphates such as vivianite, Fe32+[PO4]2.8H2O, and wavellite, Al3[PO4]2(OH,F)3 [83], and as rare-earth-bearing phosphates [76, 84]. Phosphorus can also occur in iron ores not as discrete mineral phases but incorporated within goethite [78, 85], a scenario that is more likely when the average levels of P are of the order of 0.1 per cent or more [79, 81, 84, 86].

An earlier study [8] on Agbaja iron ore had classified it as a low-grade (54.17 wt% Fe), high-phosphorus, acidic ore and based on their XRD results reported to be comprised of the minerals: magnetite, goethite, corundum, and quartz, with traces of iron and aluminium phosphates. However, this report is silent regarding the distribution of the phosphorus, a question that is resolved in the present study. In spite of the various reports of phosphorus removal from high-phosphorus iron ores using various techniques mentioned earlier herein, not much is known about the local chemical environment around the phosphorus atom(s) in high-phosphorus iron ores. This knowledge is prone to be valuable for effective and targeted phosphorus removal from the different varieties of high-phosphorus iron ores. The local environment around the phosphorus atom in high-phosphorus iron ores is bound to be dependent on its bonding with other elements (such as Fe, Si, O, and Al present in the ore) and hence its chemical presentation/phase (compound). Due to the low percentage of phosphorus (usually less than 3% P) in these ores classified as high-phosphorus iron ores, powder XRD often employed to determine and quantify crystalline phases in ores, is seldom useful in providing the vital insight into the bonding and local chemical environment around the phosphorus atom(s) in high-phosphorus iron ores.

2 Experimental

A bulk sample of the ore weighing about 1200 g was obtained, and 1000 g of it was mechanically pulverized to powder. From the remainder, a good and representative cross section was obtained, mounted on epoxy resin, and polished progressively with abrasive paper of grit sizes 400, 600, 800, and 1000 prior to optical and scanning electron microscopy, and elemental mapping using energy-dispersive spectroscopy. XRF analysis was carried out using the XRF fusion method (lithium borate) on ore sample that had been pulverized by agate milling until more than 85% of sample passes through 75-micron sieve. Powder X-ray diffraction was carried out at room temperature using Cu-Kα radiation on a Rigaku Geigerflex X-ray powder diffractometer at a speed of 1o/min from 10o to 120o. High-temperature X-ray diffraction was also carried out on powder samples at temperatures between 25 and 1220 °C using Cu-Kα radiation on same diffractometer with same parameters. In addition, thermal analysis (TGA and DSC) was carried out simultaneously with a Netzsch STA 409 PC/PG Simultaneous TGA–DSC machine after calibration with 176.500 mg of the ore powder sample using an air flow rate of 100 ml/min and nitrogen flow rate of 18 ml/min, and heating from 20 to 1250 °C with a heating rate of 1 °C/min. The results obtained from the thermal analysis were analysed using Netzsch Proteus Thermal Analysis Software. Scanning electron microscopy, elemental mapping, and energy-dispersive X-ray analysis of the ore were carried out using a Hitachi SU-70 scanning electron microscope with EDS capability provided by a Bruker Quantax 400 EDS system.

3 Results and discussion

3.1 Optical microscopy results

Optical microscopy results from a polished cross section of the ore (Fig. 1) show the oolitic structure of the ore, composed of lighter spherical grains in concentric layers embedded in a darker matrix. The presentation of the ore is consistent with the structure of oolitic ores.

Fig. 1
figure 1

Optical microscopy image of a bulk of Agbaja iron ore sample mounted on epoxy and polished

3.2 Scanning electron microscopy results

Figure 2a shows the morphology of the grains of the pulverized iron ore, the sharp edges being consistent with mechanical pulverization. The SEM image of a polished cross section of Agbaja iron ore shown in Figure 2b reveals with more clarity the oolitic nature of the ore. In this image, the darker areas are predominantly the gangue minerals, while the lighter areas comprise mostly of the iron ore.

Fig. 2
figure 2

SEM images of a powdered Agbaja iron ore sample and b piece of Agbaja iron ore mounted on epoxy and polished (the darker areas are predominantly the gangue minerals, while the lighter areas comprise the iron ore)

3.3 Scanning electron microscopy with energy-dispersive X-ray spectroscopy with elemental mapping results

The EDS results (Fig. 3 and supplementary material S-1) confirm the presence of iron, aluminium, silicon, oxygen, and most importantly phosphorus, whereas no minerals containing phosphorus were detected in the powder XRD results, most probably due to the fact that its concentration in the ore (measured to be 1.395% P; Table 2) is below 3% w/w, the detection limit with XRD. The non-detection of hydrogen is obvious as EDS is not suitable for detecting the first five (and light) elements.

Fig. 3
figure 3

a SEM image of Agbaja iron ore (2 is predominantly gangue minerals and 1, 3, 4 are iron-rich minerals), b combined energy-dispersive X-ray elemental mapping of selected elements; separate energy-dispersive X-ray elemental mapping of c iron, d phosphorus, e calcium, and f aluminium, showing the distribution of elements in a sample of Agbaja iron ore

Table 2 Chemical composition of Agbaja iron ore

Energy-dispersive X-ray elemental mapping of selected elements (Fe, P, Ca, and Al) in Fig. 3 shows that aluminium, a distinct component of the gangue (kaolinite—Al4(OH)8Si4O10), is partitioned into the lighter areas of the SEM image, whereas phosphorus is observed in both the lighter areas (gangue minerals) and the darker areas (iron ore). This observed lack of partitioning of phosphorus is indicative of the futility of attempting to reduce phosphorus content in this ore by physical separation processing methods. This result is consistent with the results from analysis of different fractions of Agbaja iron ore by Mücke et al. [87] which showed the presence of phosphorus in all the different ore fractions.

3.4 X-ray fluorescence spectroscopy results

Results of chemical analysis by X-ray fluorescence (Table 2) indicate that the Agbaja iron ore sample used in this work contains 1.395% P (equivalent to 3.196% P2O5) which is consistent with earlier reports [8, 9, 25,26,27] and 53.1% Fe.

3.5 Powder X-ray diffraction results

The powder X-ray diffraction results of the as-received ore (Fig. 4) show the Agbaja iron ore sample to comprise mainly of goethite (α-FeOOH) and kaolinite (Al4(OH)8Si4O10). The XRD results obtained after thermal analysis up to 1220 °C (Figs. 4 and 5) indicate the presence of haematite (Fe2O3), together with silica (SiO2) and corundum (Al2O3), both products of kaolinite (Al4(OH)8Si4O10) thermal decomposition. Inability to detect phosphorus in the XRD results in spite of its detection in the EDX results (Fig. 3 and supplementary material S1) is attributed to the probable existence of phosphorus in the iron ore in an amorphous form and/or to the fact that its concentration in the ore, as can be observed from XRF data (Table 2) being quite below 3 wt% (around the detection limit of XRD), makes the detection of the crystalline phase containing phosphorus difficult.

Fig. 4
figure 4

Powder X-ray diffraction pattern of as-received Agbaja iron ore identifying it as mainly iron oxide hydroxide (FeO(OH)) with some kaolinite (Al4(OH)8Si4O10) and same ore after thermal treatment up to 1220 °C via the TGA/DSC tests identifying its transformation/decomposition to haematite, silica and corundum (G goethite, K kaolinite, Q quartz/silica, H haematite, and C corundum)

Fig. 5
figure 5

Powder X-ray diffraction patterns for Agbaja iron ore at different temperatures as sample is heated up from room temperature and maintained at the respective measurement temperatures from 25 to 1220 °C (G goethite, K kaolinite, Q quartz/silica, H haematite and C corundum)

In the light of this, it is plausible that phosphorus might exist in Agbaja iron ore as an amorphous phase, and/or as interstitials in the iron oxide/hydroxide structure and probably also in the structure of the aluminosilicates of the gangue. The later possibility has been posited to account for phosphorus in some Australian iron ores [79, 84, 86]. Furthermore, Uwadiale [88] had mentioned that in certain phosphoriferous ore deposits, in which the phosphate mineral(s) are extremely fine grain sized, these phosphate mineral(s) may not be readily identified, even with the use of qualitative SEM with EDAX. Mücke and Farshad [89] in their work on comparison and differentiation of types and subtypes of phanerozoic ooidal ironstones had reported their inability to associate the phosphorus content of Agbaja iron ore with apatite, as they could not detect apatite in spite of the 2.6 wt% P2O5 content (equivalent to 1.135 wt% P content) in the Agbaja ore sample studied. It is worthy to note that Adedeji and Sale [8] had uniquely reported the presence of traces of iron and aluminium phosphates in Agbaja iron ore which has yet not been confirmed by the reports of later researchers. These reports and the results from the present study favour a postulation that the phosphorus in Agbaja ore is most probably not presenting as a discrete crystalline phase.

By acquiring XRD data as the ore sample is heated up from room temperature, the evolution of crystalline phases as a function of sample temperature was monitored. The results of the high-temperature XRD measurements (Fig. 5) are in agreement with the results of XRD measurements on iron ore samples at ambient temperature, and after heating up to 1220 °C during TGA/DSC test (Fig. 6), showing that the crystalline phases in Agbaja iron ore at elevated temperatures are haematite, silica, and corundum with no indication of any phosphorus containing phases. Significant changes in the composition of the crystalline phases present in the ore were observed from temperatures higher than 400 °C. These results are in agreement with the report of Uwadiale and Whewell [9] who in their work on the effect of temperature on magnetizing reduction of Agbaja iron ore heated Agbaja ore together with coal between 100 and 1000 °C, collected XRD data, and reported the presence of haematite, magnetite, kaolinite, and carbon between 400 and 500 °C; haematite, magnetite, and carbon between 600 and 700 °C; haematite, magnetite, wustite, and carbon between 800 and 900 °C (with some metallic iron at 900 °C); and magnetite, metallic iron, and carbon at 1000 °C.

Fig. 6
figure 6

Combined TGA/DSC plots for Agbaja iron ore with peak analysis results

3.6 TGA/DSC results

The thermal analysis results (TGA/DSC) are presented in Fig. 6 and Table 3 (and as supplementary materials S-2 and S-3). The most prominent endothermic peak observed in the DSC plot (Table 3; Fig. 6, S-2 and S-3) with onset around 257 °C and ending at 328 °C with a peak around 285 °C is attributed to the hydroxylation of the goethite with the loss of the OH group in its structure resulting to the formation of haematite (Fe2O3). The dehydroxylation of goethite to form haematite has been reported [90] to occur through intermediate hydroxylated phases, mainly between 250/260 and 400 °C [91, 92], the actual temperature range and kinetics being strongly dependent on such factors as particle size or crystallinity [93], composition (aluminium substitution [94], and water content [95] included), and testing variables like heating rate and duration.

Table 3 Results from the analysis of observed peaks in DSC measurements

In the obtained thermogravimetric results, the presence of other less intense peaks, particularly at 502 °C (endothermic) and 938 °C (exothermic), both below 1000 °C up to which haematite is thermally stable, is attributed to the major second phase in the ore—kaolinite (Al4(OH)8Si4O10)—detected in the XRD results. The peak at 502 °C is hence attributed to the dehydroxylation of kaolinite and possible formation of metakaolinite, which has been reported to occur in the temperature range 400–540 °C [96]. The relatively low observed temperature for this reaction (502 °C) is indicative of poor ordering in the crystalline structure of the kaolinite [96, 97]. The exothermic peaks observed at 938, 1038, and 1202 °C are attributed to recrystallization and transformation of dehydration products to metakaolinite, corundum, and silica [97].

4 Conclusions

Agbaja iron ore, a high-phosphorus iron ore, has been characterized using a variety of analysis techniques in an attempt to unravel the presentation and distribution of phosphorus in the iron ore (between the iron-rich phase and the gangue mineral(s) which will be very useful in planning effective beneficiation and phosphorus removal strategies). From the results of this study, the presence of phosphorus is confirmed, its concentration in the sample studied is determined to be 1.395 wt% P, and its spatial distribution in the ore is demonstrated to be in both the iron-rich phase and the gangue minerals. However, the actual presentation of phosphorus in the ore could not be determined as the XRD results did not show the presence of any phosphorus containing crystalline phases (minerals). This might be indicative of the plausible presence of phosphorus in Agbaja iron ore as an amorphous phase, and/or as interstitials in the iron oxide/hydroxide structure, and probably also in the structure of the aluminosilicates of the gangue mineral(s).