- Title
- A reappraisal of the origin of the Hotazel Fe-Mn Formation in an evolving early Earth system through the application of mineral-specific geochemistry, speciation techniques and stable isotope systematics
- Creator
- Mhlanga, Xolane Reginald
- ThesisAdvisor
- Tsikos, Harilaos
- ThesisAdvisor
- Lyons, Timothy W
- Subject
- Manganese ores -- South Africa -- Hotazel
- Subject
- Manganese ores -- Geology
- Subject
- Iron ores -- South Africa -- Hotazel
- Subject
- Iron ores -- Geology
- Subject
- Geochemistry -- South Africa -- Hotazel
- Subject
- Isotope geology -- South Africa -- Hotazel
- Subject
- Geology, Stratigraphic -- Archaean
- Subject
- Geology, Stratigraphic -- Proterozoic
- Subject
- Transvaal Supergroup (South Africa)
- Subject
- Great Oxidation Event
- Date
- 2020
- Type
- Thesis
- Type
- Doctoral
- Type
- PhD
- Identifier
- http://hdl.handle.net/10962/146123
- Identifier
- vital:38497
- Description
- Marine chemical sediments such as Banded Iron Formations deposited during the Archean-Palaeoproterozoic are studied extensively because they represent a period in the development of the Earth’s early history where the atmospheric O₂ content was below the present levels (PAL) of 21%. Prior to the Great Oxidation Event (GOE) at ca. 2.4 Ga, highly ferruginous and anoxic marine environments were dominated by extensive BIF deposition such as that of the Griqualand West Basin of the Transvaal Supergroup in South Africa. This basin is also thought to record the transition into the first rise of atmospheric O₂ in our planet, from the Koegas Subgroup to the Hotazel Formation dated at ca. 2.43 Ga (Gumsley et al., 2017). Two drill cores from the north eastern part of the Kalahari Manganese Field characterized by a well-preserved and complete intersection of the cyclic Mn-Fe Hotazel Formation were studied at a high resolution (sampled at approximately one-meter interval). Such high-resolution approach is being employed for the first time in this project, capturing in detail the three manganese rich layers intercalated with BIF and the transitions between these lithofacies. The micro-banded BIF is made up of three major phases, namely Fe-Ca-Mg carbonates (ankerite, siderite and calcite), magnetite, and silicates (chert and minor Fe-silicates); laminated transitional lutite consist of mainly hematite, chert and Mn-carbonates, whereas the manganese ore layers are made up of mostly calcic carbonates (Mn-calcite and Ca-kutnahorite) in the form of laminations and ovoids, while Mn-silicates include dominant braunite and lesser friedelite. All three lithofacies are very fine grained (sub-mm scale) and so petrographic and mineralogical observations were obtained mostly through scanning electron microscope analysis for detailed textural relationships with focus on the carbonate fraction. Bulk geochemical studies of the entire stratigraphy of the Hotazel Formation have previously provided great insights into the cyclic nature of the deposit but have not adequately considered the potential of the carbonate fraction of the rocks as a valuable proxy for understanding the chemistry of the primary depositional environment and insights into the redox processes that were at play. This is because these carbonates have always been attributed to diagenetic processes below the sediment-water interface such as microbially-mediated dissimilatory iron/manganese reduction (DIR/DMR) where the precursor/primary Fe-Mn oxyhydroxides have been reduced to result in the minerals observed today. The carbonate fraction of the BIF is made up of ankerite and siderite which co-exist in a chert matrix as anhedral to subhedral grains with no apparent replacement textures. This suggests co-precipitation of the two species which is at apparent odds with classic diagenetic models. Similarly, Mn-carbonates in the hematite lutite and manganese ore (Mn-calcite, kutnahorite, and minor rhodocrosite) co-exist in laminae and ovoids with no textures observed that would suggest an obvious sequential mode of formation during diagenesis. In this light, a carbonate-specific geochemical analysis based on the sequential Fe extraction technique of Poulton and Canfield (2005) was employed to decipher further the cyclic nature of the Hotazel Formation and its primary versus diagenetic controls. Results from the carbonate fraction analysis of the three lithofacies show a clear fractionation of iron and manganese during primary – rather than diagenetic - carbonate precipitation, suggesting a decoupling between DIR and DMR which is ultimately interpreted to have taken place in the water column. Bulk-rock concentration results for minor and trace elements such as Zr, Ti, Sc and Al have been used for the determination of either siliciclastic or volcanic detrital inputs as they are generally immobile in most natural aqueous solutions. These elements are in very low concentrations in all three lithofacies suggesting that the depositional environment had vanishingly small contributions from terrigenous or volcanic detritus. In terms of redox-sensitive transition metals, only Mo and Co appear to show an affinity for high Mn facies in the Hotazel sequence. Cobalt in particular attains a very low abundance in the Hotazel BIF layers at an average of ~ 4 ppm. This is similar to average pre-GOE BIF in South Africa and worldwide. Maxima in Co abundance are associated with transitional hematite lutite and Mn ore layers, but maxima over 100ppm are seen in within the hematite lutite and not within the Mn ore proper where maxima in Mn are recorded. This suggests a clear and direct association with the hematite fraction in the rocks, which is modally much higher in the lutites but drops substantially in the Mn layers themselves. The similarities of bulk-rock BIF and modern-day seawater REE patterns has been used as a key argument for primary controls in REE behaviour and minimal diagenetic modification. Likewise, the three lithofacies of the Hotazel Formation analysed in this study all share similar characteristics with a clear seawater signal through gentle positive slopes in the normalised abundance of LREE versus HREE. Negative Ce anomalies prevail in the entire sample set analysed, which has been interpreted before as a proxy for oxic seawater conditions. However, positive Ce anomalies that are traditionally linked to scavenging and deposition of primary tetravalent Mn oxyhydroxides (e.g., as observed in modern day ferromanganese nodules) are completely absent from the current dataset. The lack of a positive Ce anomaly in the manganese ore and peak Co association with ferric oxides and not with peak Mn, suggests that primary deposition must have occurred within an environment that was not fully oxidizing with respect to manganese. The use of stable isotopes (i.e., C and Fe) was employed to gain insights into redox processes, whether these are thought to have happened below the sediment-water interface or in contemporaneous seawater. At a small scale, all lithofacies of the Hotazel Formation record bulk-rock δ¹³C values that are low and essentially invariant about the average value of -9.5 per mil. This is independent of sharp variations in overall modal mineralogy, relative carbonate abundance and carbonate chemistry, which is clearly difficult to reconcile with in-situ diagenetic processes that predict highly variable δ¹³C signals in response to complex combinations of precursor sediment mineralogy, pore-fluid chemistry, organic carbon supply and open vs closed system diagenesis. At a stratigraphic scale, the carbonate δ¹³C (-5 to -13‰) variations between the different lithologies could instead represent temporal changes in water-column chemistry against well-developed physico-chemical gradients, depth of deposition and biological processes. The low iron isotope values recorded in the hematite lutite and manganese ore samples can be attributed to fractionation effects of initial oxidation of ferrous iron to form Fe-oxyhydroxides in the shallow parts of the basin, from an already isotopically highly depleted aqueous Fe-pool as proposed previously. The slightly higher but still negative bulk-rock δ⁵⁶Fe values of the host BIF can be attributed to water-column Fe isotopic effects at deeper levels between primary Fe oxyhydroxides and an isotopically heavier Fe(II) pool, which was subsequently preserved during diagenetic recrystallization. All above findings were combined into a conceptual model of deposition for the three different lithologies of the Hotazel Formation. The model predicts that free molecular oxygen must have been present within the shallow oceanic environment and implicates both Mn and Fe as active redox “players” compared to classic models that apply to the origin of worldwide BIF prior to the GOE. The deposition of the Hotazel strata is interpreted to have occurred through the following three stages: (1) BIF deposition occurred in a relatively deep oceanic environment above the Ongeluk lavas during marine transgression, where a redoxcline and seawater stratification separated hydrothermally sourced iron and manganese, in response to an active Mn-shuttle mechanism linked to Mn redox cycling. Abundant ferrous iron must have been oxidized by available oxygen but also by oxidised Mn species (MnOOH) and possibly even some soluble Mn(III) complexes. Through this process, Mn(III) was being effectively reduced back into solution along with cobalt(III), as Mn(II) and Co(II) respectively, thus creating maxima in their concentrations. A drawdown of Fe(OH)₃ particles was therefore the only net precipitation mechanism at this stage. Carbonate species of Fe and the abundant magnetite would possibly have formed by reaction between the ferric hydroxides and the deeper Fe(II) pool, while organic matter would also have reacted in the water-column via DIR, accounting for the low δ¹³C signature of Fe carbonate minerals. (2) Hematite lutite formation would have occurred at a relatively shallower environment during marine regression. At this stage, reductive cycling of Fe was minimal in the absence of a deeper Fe(II) reservoir reacting with the ferric primary precipitates. Therefore, DIR progressively gave way to manganese reduction and organic carbon oxidation (DMR), which reduced MnOOH to form Mn(II)-rich carbonates in the form of kutnahorite and Mn-calcite. Co-bearing Fe(OH)₃ would have precipitated and was ultimately preserved as Co-bearing hematite during diagenesis. (3) Deposition of manganese-rich sediment occurred at even shallower oceanic depths (maximum regression) where aerobic organic carbon oxidation replaced DMR, resulting in Ca-rich carbonates such as Mn-bearing calcite and Ca-kutnahorite, yet with a low carbon isotope signature recording aerobic conditions of organic carbon cycling. Mn(III) reduction at this stage was curtailed, leading to massive precipitation of MnOOH which was diagenetically transformed into braunite and friedelite. Simultaneous precipitation of Co-bearing Fe(OH)₃ would have continued but at much more subdued rates. Repeated transgressive-regressive cycles resulted in the cyclic BIF-hematite lutite- manganese ore nature of the Hotazel Formation in an oxidized oceanic environment at the onset of the Great Oxidation Event, which was nonetheless never oxic enough to drive Mn(II) oxidation fully to its tetravalent state. The mineralogy and species-specific geochemistry of the Hotazel strata, and more specifically the carbonate fraction thereof, appear to faithfully capture the chemistry of the primary depositional environment in a progressively evolving Earth System. This project opens the door for more studies focusing on better constraining primary versus diagenetic depositional 2020 Hotazel Fe and Mn deposition mechanisms of iron and manganese during the period leading up to the GOE, and possibly re-defining the significance of Fe and Mn as invaluable redox proxies in a rapidly changing planet.
- Format
- 180 pages, pdf
- Publisher
- Rhodes University, Faculty of Science, Geology
- Language
- English
- Rights
- Mhlanga, Xolane Reginald
- Hits: 5214
- Visitors: 5358
- Downloads: 442
Thumbnail | File | Description | Size | Format | |||
---|---|---|---|---|---|---|---|
View Details Download | SOURCE1 | MHLANGA-PHD-TR20-288.pdf | 12 MB | Adobe Acrobat PDF | View Details Download |