The Zatua Hills, which are the subject of this publication, are located between the Bafwasende Degree Square in the north and the Isiro (Paulis) Degree Square in the south, and are situated between 27˚31'48'' and 27˚40'48'' East longitude and between 1˚55'48'' and 2˚01'48'' North latitude. They are located in the Wamba Territory, Haut-Uélé Province in the Democratic Republic of the Congo, approximately 315 km by air from the city of Kisangani (Figure 1).

This area is characterized by a relatively smaller plateau, with altitudes varying between 500 and 1089 m. Intermittent ranges and hills rise up to 300 m above the surrounding plateau.

The target area comprises a series of hills in a NW-SE direction, including the Zatua hills with a maximum altitude of 1300 m. To the south-east, the area is bordered by the Western Rift Valley, including Lakes Albert and Edouard, separated by the Ruwenzori massif (5119 m).

Regional geology is predominantly composed of the Archean-age basement formations covering an area of 500,000 km², including Uganda, South Sudan, and the Central African Republic ( de Wit & Linol, 2015 ; Borg & Shackleton, 1997 ;

Cahen & Snelling, 1966 ). These formations were initially associated with the Tanzanian craton before being separated by the Ruwenzori chain (2 Ga) and the filling of rift basins ( Borg & Shacklet, 1997 ; Lavreau & Navez, 1982 ). These formations include the Bomu amphibolitic and gneissic complex dated 3.4 - 3.7 Ga ( Cahen et al., 1984 ), the West Nile amphibolitic complex dated 3200 - 2728 Ma ( Allibone et al., 2020 ), the Garamba complex, the granitic massifs of the Upper Congo Granites Massifs, which cover a large part of the central area, consisting of magmatic, orthogneissic and various magmatic intrusions, in which the Kibalian greenstone belt formations, including BIFs, are associated, and which form the core of the large Archean units in the region ( Lavreau, 1982 ; Cahen & Snelling, 1966 ; Lavreau, 1982 ; BRGM, 1982 ; Cahen et al., 1984 ; Bird, 2016 ; Allibone et al., 2020 ).

The Zatua Hills are part of the Ngayu belt, which is one of several greenstone belts found in the Nyanza-Kibali greenstone belt, which stretches from Tanzania to the Central African Republic.

The most remarkable formations in the region are the BIFs, which occupy vast stretches compared to other geological formations. Their presence is clearly shown on the map represented in Figure 2. These BIFs are considered as true

markers that define the different structural entities of the region. In some places, these formations can have significant thicknesses of up to 200 m, sometimes in association with certain deformations (folds, faults, laminations) and form relatively high reliefs. BIFs are sometimes found intercalated at the roof by schist rock and at the wall by basalts.

Apart from the BIFs, the region is also characterized by the presence of basaltic formations, intrusions of diorite, quartz monzonite and tonalite in greenstone rocks. All these rocks are covered by formations of Upper Proterozoic age (Lindian), Mesozoic formations and the West African Rift formations of Cenozoic age.

The work consisted of identifying the outcrop zones of BIFs, iron ores (enriched parts) and their host rocks using airborne magnetometric data acquired and interpreted by the company Rio Tinto, which once operated in this area, geological survey work on the relevant sites, collecting samples in the field, and mapping outcrop zones.

These data enabled the confirmation of zones with rich BIFs (% iron ≥ 50), poor BIFs (% iron < 50) and iron ores ((% iron ≥ 60), not forgetting the poor host rock ((% iron < 40).

Several samples were collected in these sites (Figure 3) and some were selected for petrography observations in the laboratory.

The selected samples were used for petrographic observations in the laboratory, namely:

· Macroscopic analysis of the samples collected using the naked eye and a magnifying glass;

· Microscopic analyses using an optical microscope and a scanning electron microscope to examine iron oxides and post-depositional inclusions that could influence the quality of the iron ore encountered.

The results of the petrographic analyses have enabled us to make a metalogenic interpretation of the Zatua Hills BIFs and the associated iron ores, while also taking a look at the associated elements that can influence the metallurgical quality of iron in its various stages through the formation of different alloys.

Of the 24 surface samples collected, we took care to select the petrographic analyses of 6 samples for the purposes of representativeness, details of which are given on the pages below (Figures 3-11).

Several holes were drilled in the Zatua Hills area, including 4 holes selected in order to determine the depth of iron ore enrichment, alteration and the evolution of secondary minerals that could influence the quality of these iron ores.

The various analysis carried out on surface and borehole samples in the Zatua Hills (Figure 3 and Figure 12) have revealed that, both at surface and at depth, the same genetic processes that led to the enrichment of the BIFs in exploitable rich iron ore. These analyses clearly show that for BIFs samples whose quartz has been largely leached and enriched in iron, the other deleterious elements or iron impurities (Al, P and Si) are less represented in the samples (Figure 9, Figure 13(b) and Figure 13(d)). This leaching is possible along zones of weakness linked to rock joints, fractures and folding leading to the departure of primary quartz and the concentration of iron ores (magnetite, martite and hematite). Where the voids left by the departure of the primary silica are not filled, secondary hematite, goethite, gibbsite, kaolinite, cadwaladerite, strengite, variscite and other forms of ferruginous clay fill the pores as a matrix, consolidating the rock fragments (in the case of mineralised breccias). These secondary minerals are also the real sources of deleterious elements such as Si (residual silica and kaolinite), P (strengite-variscite solid solution), Al (gibbsite, cadwaladerite and kaolinite), Cl (cadwaladerite); and the hydration level of the rock.

All of these replacements took place through a succession of genetic processes that were sometimes cancelled out by other events. The supergene process is the last to be perceptible in the samples observed and through which meteoric waters are largely responsible for the transformation or leaching of the rock leading to its enrichment in iron ( Morris, 1980 , 1985 , 2002 ; Harmsworth et al., 1990 ).

This observation remains as far as possible in favour of the modified supergene-hypogen process associated with metamorphism ( Beukes et al., 2002 ; Gutzmer et al., 2006 ; Lascelles 2002 , 2007 ), whereby hydrothermal solutions, waters from the deconvolution of rocks hydrated by retrograde metamorphism and meteoric waters played a major role in the leaching of silica and the enrichment of BIFs in exploitable rich iron ores. The presence of Ti in borehole samples (Figure 13(d)) clearly shows that it is associated with both phyllites and BIFs. Phyllites (chlorites, sericite schists) are considered to be low-temperature metamorphic rocks derived from the argillites found in this sedimentary basin. The affinity of titanium to this type of rock shows that the form of titanium oxide associated with it would be anatase, which corresponds to this type of low-temperature, low-pressure metamorphism.

The samples selected for petrographic analysis helped us to put forward arguments concerning the magnetite-martite-hematite paragenesis, silica leaching, the development of microplaty hematite, the development of goethite and other secondary minerals considered to be deleterious to iron ores.

1) Magnetite-Martite-Hematite paragenesis

Referring to the macroscopic and microscopic observations of the surface and borehole samples in the preceding pages, we can see that the initial iron oxide in the BIFs of the Zatua Hills was magnetite, ideally (Fe²⁺Fe³⁺₂O₄) which was deeply oxidised to martite (Fe₂O₃) as expressed in the following equation (Figure 4(b), Figure 7(a) and Figure 7(b), Figure 10(b) and Figure 13(a), Figure 13(c) and Figure 13(e)):

2Fe 3 O 4 + 1 / 2 O 2 ⇒ 3Fe 2 O 3

The oxidation process began around the grains and along the fractures from which it progresses parallel to various directions of the crystallographic axes of the magnetite. This oxidation process leads to the development of martite aggregates in the form of lattice networks reflecting the positions of the magnetite crystallographic planes in relation to the oxygen cation planes observed in the magnetite cubic and hematite rhombohedral networks (Figure 7(a) and Figure 7(b)).

As the oxidation process begins, the lamellar martite crystals gradually enlarge and become more abundant until the remaining magnetite crystals are completely oxidised and replaced by martite aggregates. These martite lamellae and lattice structures are commonly preserved and serve as witnesses to confirm the presence of magnetite as a protore mineral (Figure 7(a) and Figure 7(b)).

Locally, small magnetite crystals survive within the cores of martite aggregates. The fine martite crystals produced during the oxidation process commonly form pseudomorphic substitutions of magnetite, preserving its detail in the primary microstructures (Figure 4(b)).

The oxidation process responsible for the martitisation of the magnetite in the Zatua Hills samples studied is deep and extensive; only traces of the unoxidised magnetite crystals are still visible in the samples (Figure 4(b), Figure 7(a) and Figure 7(b), Figure 10(b) and Figure 13(a), Figure 13(c) and Figure 13(e)). Under low-temperature (near-surface) oxidation conditions, the oxidation products of the martite have lattice structures. They are, however, often present locally in the outcrops studied and, where they are present, they commonly show large crystals and interspersed lamellae (Figure 7(a) and Figure 7(b)). In many cases, the pseudomorphic martite consists of finely grained aggregates showing the early stages of recrystallization and grain development. In these cases, the martite particles are recrystallized and grain development processes are initiated in the conversion of martite lamellae to equidimensional crystals. These aggregates typically show a strong irregularity in the contours of the crystals formed, indicating a lack of crystallographic equilibrium.

The low degrees of recrystallization observed in most of the samples suggest that there was a slight increase in temperature at some stage after the oxidation process. These temperatures were, however, neither too high nor for a long period to allow the martite recrystallization process to continue. One exception to this observation is the development of very elongated hematite crystals in this zone. This reflects the recrystallization of the finely grained martite along the small shear planes of the iron ore (Figure 6(b) and Figure 13(f)-(h)).

2) Silica leaching process and development of microplaty hematite

The textural evidence described in the petrographic section suggests that most of the BIFs samples in the suite studied consist of quartz bands and magnetite. These BIFs are considered to be prototypical iron ores and contain around 50% or less Fe₂O₃ with silica predominating. Leaching of quartz led to the development of pores in the rock, which still persist in some of the samples analysed. The presence of abundant pores increases permeability and provides a favourable space for the precipitation of secondary minerals.

The first stage in filling these pores with matrix constituents is the development of microplaty hematite. Typically, crystals of microplaty hematite form a core on the surface of the pseudomorphic martite and begin to grow.

These crystals are favourably oriented in the direction of leaching from the cavities, and the pores are then able to develop elongated or scale-like crystals of microplaty hematite, also known as specularites, on their outer surface (Figure 6(b) and Figure 13(f)-(h)).

The abundance and extent of these crystals in the rock varies from sample to others. In some samples, abundant, well-developed hematite crystals are found, while in others they are virtually absent. The development of these crystals contributes to the cementation of the pores in the martite-rich fraction.

The primary nature of the iron fragments and other forms of breccia is marked by the development of a low rate of microplaty hematite crystals developed in the matrix between the well mineralized fragments, thus preserving their morphologies.

The development of microplaty hematite crystals reflects the introduction into the system of a variable rate of additional iron leading to varying degrees of iron enrichment. Microplaty hematite is an important constituent of iron-rich ores, but virtually absent in iron-poor ores, and it probably develops with a slight increase in temperature, which is responsible for the early recrystallization of martite ( Lascelles, 2007 ).

These types of hematite crystals are cemented by well-formed crystals of secondary hematite associated with goethite and other secondary minerals deposited later, and form a botryoidal or concretion texture inside the pores, but sometimes an acicular texture (Figure 8, Figure 9 and Figure 13(a)).

Microplaty hematite is the result of a thorough leaching process of BIFs. This process is on the one hand supergene, marked by the oxidation of magnetite, by meteoric fluids and by the recrystallization of pre-existing martite; and on the other hand hypogene, marked by the precipitation of hematite from the oxidation of hydrothermal fluids ( Powell et al., 1999 ; Lascelles, 2007 ) and the dehydration of goethite ( Morris, 1985 , 2002 ).

The supergene process and the dehydration of the goethite require oxidation, due to alteration, followed by burial and Low Temperature Metamorphism (LTM).

3) Development process of goethite and other secondary minerals

It is true that secondary minerals are produced by precipitation in pores, fractures and other permeable zones under low-temperature, near-surface conditions. These minerals largely consist of goethite associated with secondary hematite, kaolinite, gibbsite, strengite-variscite solid solutions and cadwaladerite. The petrographic analyses presented in the previous sections have attempted to provide more details, but here we will attempt to give summaries.

The porous and permeable nature of many samples allows free circulation of aquifers and provides suitable sites for the precipitation of one or more secondary minerals. These minerals commonly show the development of fine bands, with botryoidal or concretionary textures, which indicate precipitation along the pores and lead to the formation of a hard crust. Unfortunately, these secondary minerals, notably goethite, kaolinite, gibbsite and the strengite-variscite solid solution serie, are the source of a significant proportion of the deleterious elements or impurities in the majority of samples.

The secondary minerals in these samples are marked by the composition of silicon (quartz or chert, goethite and kaolinite), aluminium (gibbsite, variscite, cadwaladerite, goethite and kaolinite) and phosphorus (variscite) in the analyses ( Makuku et al., 2018 ).

a) Goethite

It has a ubiquitous phase, and is the most abundant of the secondary minerals in the samples processed. It is developed in the form of fine bands, with a botryoidal or concretional texture on the external surfaces of leached cavities, along fractures, in other permeable spaces in the rock and serves as a matrix of the rock cementing the rock fragments (case of breccias mentioned above).

In some cases, goethite appears to flood the rock, filling most of the pores. The very fine nature of the goethite bands indicates variations in the degree of crystallinity, levels of hydration of hematite or martite and impurities ( Morris, 1980 , 1985 , 2002 ; Harmsworth et al., 1990 ).

The goethite inter-grows with the martite crystals in the samples and becomes problematic as its content increases. In addition to its water content, petrographic and geochemical analyses show that it contains a certain amount of Si and Al, as well as P ( Makuku et al., 2023 ). A small percentage of Ti is also present. In some samples rich in Al and gibbsite, goethite can also contain Al and be classified as Al-goethite (Figure 9).

b) Secondary hematite

Secondary hematite is far less abundant than goethite among the minerals precipitated in the pores at the last phase (botryoidal or concretion textures). Where it is present, it is often associated in layers with goethite and, in a way, represents the dehydration product of goethite (Figure 9, Figure 13(c), Figure 13(g), Figure 13(h)).

Unlike the associated goethite, scanning electron microscope (SEM) analyses of this type of hematite show that it is close to the theoretical Fe₂O₃ composition, and is essentially free of the impurities associated with goethite.

c) Kaolinite

Kaolinite is present in very small quantities, but can vary in some samples where it occupies, alone or in association with other impurities, pores and fractures. It also shows an unusual mode of occurrence in which it is present as small rounded bodies that may be induced by goethite or other secondary minerals (Figure 9 and Figure 13(h)).

Kaolinite commonly shows varying degrees of secondary ferruginization, marked by a reddish colour and thousands of hematite inclusions. Scanning Electron Microscope (SEM) analysis of kaolinite shows the presence of Si, Al and often a small amount of Fe. Kaolinite does not contain all the deleterious elements such as Si and Al, although its degree of crystallinity and/or water content is variable.

d) Gibbsite

Gibbsite shows the same pattern of occurrence as kaolinite, but remains the most common mineral. It occurs alone or in association with goethite and/or kaolinite (Figure 9 and Figure 13(b)). SEM analysis shows that gibbsite consists mainly of Al, but its water content and degree of crystallinity are variable. Gibbsite is often a very late phase of precipitation and is located at the core of the goethite in the pores or at the center of the transgressive goethite in the fractures.

e) Anatase (Titane Oxide)

The presence of Ti in borehole samples clearly shows that it is associated with both phyllites and BIFs. Phyllites (chlorites, sericite schists) are considered to be the low-temperature metamorphic rocks derived from the argillites found in this sedimentary basin. The affinity of titanium to this type of rock shows that the form of titanium oxide associated with, it would be anatase, which corresponds to this type of low-temperature, low-pressure metamorphism (Figure 13(d)) ( Makuku et al., 2023 ).

f) Strengite-Variscite solid solutions

Some samples analyzed in this zone show high phosphorus content, which is the highest possible found in associations of elements hosted by goethite. In most cases, the source of the high phosphorus content can be traced to the presence of strengite-variscite solid solution [ideally formulated: FePO₄∙2H₂O - AlPO₄∙2H₂O]. The composition of these minerals is variable for Al-rich solid solutions associated with gibbsite. On the other hand, the iron-rich elements of the solid solution (strengite) are commonly present in zones devoid of Al-rich minerals. These minerals can contain between 38% and 55% P₂O₅ and their low level in the sample has enormous consequences for the metallurgical quality of the iron ore ( Makuku et al., 2023 ).

g) Cadwaladerite and associated phases

Cadwaladerite [AlCl(OH)₂∙4H₂O] is the chlorine-rich mineral identified in various iron deposits, most notably in the West Zatua samples (Figure 9). This phase occurs as fine grains inter-growing with gibbsite and/or strengite-variscite solid solutions, but its composition is variable and reflects an amorphous mineral. It is possible that traces of cadwaladerite are also present in the goethite and probably in the clay minerals.

It has been observed that the presence of abnormal levels of P₂O₃ in the borehole samples is linked to fracture zones where the circulation of clay minerals is regular in association with phosphorous minerals, which circulation reflects a supergene process par excellence ( Makuku et al., 2023 ).