Ore Geology Reviews 96 (2018) 181–200

Contents lists available at ScienceDirect

Ore Geology Reviews

journal homepage: www.elsevier.com/locate/oregeorev

Paleoproterozoic manganese and base metals deposits at Kisenge-Kamata T

(Katanga, D.R. Congo)
⁎
Thierry De Puttera, , Jean-Paul Liégeoisa, Stijn Dewaelea,b, Jacques Cailteuxa, Adrian Boycec,
Florias Meesa
a
Royal Museum for Central Africa, Geodynamics and Mineral Resources, Leuvensesteenweg 13, B-3080 Tervuren, Belgium
b
Ghent University, Department of Geology, Mineralogy and Petrology, Krijgslaan 281 S8, B-9000 Gent, Belgium
c
Scottish Universities, Environmental Research Centre, Isotope Geoscience Unit, Rankine Avenue, G75 OQF East Kilbride, United Kingdom

A R T I C LE I N FO A B S T R A C T

Keywords: The Kisenge-Kamata manganese deposit (Katanga, D.R. Congo) belongs to a group of Paleoproterozoic sedi-
Katanga mentary manganese ore occurrences that have been recognized for various parts of Africa. It is a relatively small
Mn carbonate ore deposit with total estimated reserves of ∼12 Mt ore, but it is of major importance for understanding key aspects
Orosirian of the geology of the Kasai Block of the Congo Craton. The available constraints indicate that the deposit formed
Kasai block
in the 2.0–1.9 Ga time interval, in the Orosirian period. The manganese-rich facies of the Kisenge-Kamata de-
Intracontinental graben basin
posits range in composition from manganese carbonate rocks to non-calcareous deposits dominated by man-
ganese-rich garnet. Associated rocks are mainly grey graphitic shales. The Kisenge-Kamata sediments are sig-
nificantly enriched in Cu, Co, Ni, Zn and other accessory metals (V, Mo, Ga). Field constraints and geochemical
data suggest that the Kisenge-Kamata sediments were probably deposited in a relatively small intracratonic basin
(graben) within the Kasai Block, with an inlet allowing periodical access for ocean water influx carrying Mn, REE
and other metals. Primary carbonate ore has relatively low δ13C (−3 to −6‰) and high δ18O (+13 to +20‰)
values, confirming their derivation from marine fluids.

1. Introduction
World-class sedimentary manganese deposits of Paleoproterozoic
age occur in various parts of Africa and in geologically related parts of
South America (cf. Roy, 2006; Kuleshov, 2011; Beukes et al., 2016). The
Kisenge-Kamata manganese deposit in southwestern Katanga, D.R.
Congo (Fig. 1a), is a relatively poorly known member of this group of
deposits. It was first discovered during regional exploration and pro-
spection led by the Belgian geologist Baudour in 1928–30, with iden-
tification of the black ore as manganese oxides towards the end of that
period. More systematic fieldwork and sampling were performed in the
1930s and 1940s by various geologists, such as van der Maesen (in
1935), Van der Stichele and Friedlander (in 1938, 1939), and even-
tually Sekirsky (in 1942) (see overview in Polinard, 1946). The deposit
has later been studied by Polinard (1946), Schuiling and Grosemans
(1956), Marchandise (1958) and Doyen (1973, 1974). In a recent re-
view paper on manganese deposits of Africa, it has been described as a
2.0–2.2 Ga deposit, possibly formed in a back-arc basin setting (Beukes
et al., 2016).
The Kisenge-Kamata deposit has been exploited between 1950 and
1986. A total of 7.4 Mt ore has been extracted during this period, with
production peaks in 1956–65 and 1970–75 (> 300,000 tons of ore
yr−1). The total potential of the deposit was evaluated at about
10 million tons in the 1950’s (Doyen, 1974). The reserves were later re-
evaluated as ∼6.7 Mt, covering both the carbonate ore (1.3 Mt) and
oxide ores (5.4 Mt) (EGMF team, 1998). This figure is considerably
lower than estimates for the Kalahari Manganese Field (hereafter KMF),
the Moanda deposits of Gabon and the Serra do Navio deposit in Brazil
(see e.g. Laznicka, 1992; Varentsov, 1996; Beukes et al., 2016). How-
ever, no systematic detailed prospection has yet been conducted in the
Kisenge-Kamata area, including geophysical prospection that could
help in identifying additional ore lenses at depth (EGMF team, 1998).
The objective of this paper is to reconsider this poorly-known Paleo-
proterozoic Mn deposit at Kisenge-Kamata, based on unpublished data
and new analyses and interpretation. Furthermore, the ore and asso-
ciated rocks are placed in their regional context (the Kasai Block of the
Congo Craton) and compared with contemporaneous Mn ore deposits in
Africa and South America. The presence of base metals in the Kisenge-
Kamata deposit is also assessed.

⁎
Corresponding author.
E-mail address: thierry.de.putter@africamuseum.be (T. De Putter).

https://doi.org/10.1016/j.oregeorev.2018.04.015
Received 7 December 2017; Received in revised form 10 April 2018; Accepted 13 April 2018
Available online 19 April 2018
0169-1368/ © 2018 Elsevier B.V. All rights reserved.
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

Fig. 1. Setting. (a) Location and geological map of the study area (after Lepersonne, 1974); (b) Distribution of ore bodies in the Kisenge area (after Schuiling and
Grosemans, 1956). (c) Cross-section of deposits in the Kamata Main (Principal) open pit mine (after EGMF team, 1998).

2. Geological context and age of the Kisenge-Kamata series
The Kisenge-Kamata area is located in the former Katanga province,
250 km west of Kolwezi (Fig. 1a). Geologically, this area belongs to the
Kasai block, in the southern part of the Congo Craton. The Kisenge-
Kamata deposit hosts sedimentary manganese carbonate and oxide ore
occurrences in outcrops along an East-West trending 6 km-long series of
small hills, including Kisenge, Kamata and Kapolo (Fig. 1b), besides
some localities east of Kapolo. The manganese ore is contained in an
East-West trending lens of shale-dominated deposits (Polinard, 1946;
Schuiling and Grosemans, 1956; Marchandise, 1958; EGMF team,
1998). The basement consists of Archean/Paleoproterozoic granite-
gneiss of the Kasai Block (Cahen and Lepersonne, 1967; Lepersonne,
1974; see Fig. 1a). Shrimp U-Pb zircon dating (Boven et al., 2011)

182
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

Sm-Nd isochrons; Delhal et al., 1986), and hydrothermal oceanic sea-

floor metamorphism has been dated at c. 1.9 Ga (whole rock Rb-Sr
isochron; André, 1993). Taking all constraints into consideration, it
seems reasonable to consider that deposition of the Kisenge-Kamata
series occurred at c. 1.95 Ga, or at least within the 2.0–1.9 Ga age range,
which is somewhat younger than the 2.2–2.0 Ga range that has pre-
viously been proposed for Kisenge-Kamata (Bekker et al., 2003; Beukes
et al., 2016). Those estimates were, however, only based on similarities
with Francevillian deposits at the north-western margin of the Congo
Craton.
Some fracturation and/or hydrothermal fluid circulation may have
occurred later, for instance during Lufilian (Pan-African) nappe em-
placement in the Katanga Copperbelt region east of Kisenge-Kamata
(John et al., 2004), but this remains uncertain and was probably of
limited intensity.
After the Eburnean orogeny, erosion has most probably removed a
considerable volume of Kisenge-Kamata series deposits. Post-orogenic
Fig. 2. Airborne total magnetic intensity data for the study area, superimposed
uplift and weathering took place in the Katanga area from the Late
to the geological map, with Paleogene Kalahari Sands (KL) and the Archean-
Paleoproterozoic basement (SA) as map units. Note the occurrence of an E-W Mesozoic (c. 80 Ma) onwards (Decrée et al., 2010; De Putter et al.,
alignment of sheets in the study area. 2010, 2015, 2016). This weathering caused considerable metal en-
richment for various types of ores, including heterogenite and mala-
chite deposits in the Copperbelt and Mn oxide deposits at Kisenge-Ka-
indicates that the region is marked by a North-South alternation of East-
mata (De Putter et al., 2015).
West trending sheets of Archean (3 Ga) and Paleoproterozoic rocks
An ongoing reappraisal of Kasai Block metasediments indicates that
(2.1–2.0 Ga), which is also confirmed by airborne geophysical surveys
Mn-rich deposits with similar characteristics occur at other localities,
(Fig. 2). This alternation probably reflects the morphology of Paleo-
such as Mwene Ditu (Fig. 1; Kasai Oriental Province; Morelli and Raucq,
proterozoic sedimentary basins within a metacratonic context (Boven
1961), for which an E-W elongated basin morphology is also re-
et al., 2011; Liégeois et al., 2013). North of the study area (around
cognized. This suggests that Paleoproterozoic Mn-carbonate deposits
latitude 8°S, Northern Katanga; Fig. 1a), the basement comprises c.
and graphitic shales may have been relatively widespread on the
2.83 Ga tonalitic gneisses and c. 2 Ga granites (Delhal and Liégeois,
granite-gneiss Kasai Block.
1982). Further north in the Kasai region, the basement consists of a
Mesoarchean charnockitic complex with a protolith that is older than
3. Materials and methods
3 Ga, in a region for which granulitic metamorphism is known to have
occurred at 2.9–2.8 Ga (720 °C, 6–7 kbar; Bingen et al., 1988), and for
The manganese ore deposits of the Kisenge-Kamata area are docu-
which a more local migmatitic/granitic event has also been recognized
mented by boreholes drilled in the 1950s (Doyen, 1974). Three cores,
(Delhal et al. 1976; Delhal, 1991). More to the south, in Western Ka-
which are today part of the geological collections of the Royal Museum
tanga, a second period of granulitic metamorphism occurred at c.
for Central Africa (RMCA), were used for this study. One core was
2.35 Ga (whole rock and mineral Sm-Nd; Delhal et al., 1986), and more
obtained at Kisenge (Kis-1) (Fig. 3), the other two at Kamata, which is
recent (Luizian) metasediments have been dated at c. 2.2 Ga (Delhal,
located 3 km East of Kisenge (Ka-30, Ka-72) (Fig. 4). Core Kis-1 presents
1991). Late granites intruded at c. 2.02 Ga (Rb-Sr isochrons and U-Pb
an oblique 160 m long section, crosscutting mainly graphitic shales and
zircon ages; Delhal and Liégeois, 1982; Pasteels, 1971), together with
oxidized ore, and reaching micaschist basement rocks in the lower part
pegmatitic dykes at 2.0 ± 0.1 Ga (Delhal and Ledent, 1971). In the
of the core (see also De Putter et al., 2015). Core Ka-72 is an 80 m long
region of Mwene Ditu (Kasai Oriental Province, ∼400 km North of
vertical section through subvertical layers of Mn carbonate ore and thus
Kisenge; Fig. 1a), where Paleoproterozoic manganese ores also occur
shows only little lithological variation (Fig. 4). Core Ka-30, which is
(Morelli and Raucq, 1961), a muscovite-bearing micaschist has yielded
125 m long, crosses the oxidized facies of the carbonate ore in its upper
an age of c. 2.11 Ga (Rb/Sr on muscovite; Ledent et al., 1962), inter-
part (30 m), and then unaltered Mn carbonate and garnetite deposits in
preted to record Eburnean rejuvenation.
lower parts (Fig. 4).
In the Kisenge-Kamata area, Eburnean metamorphism, ranging up-
Thin sections were prepared and examined at the RMCA. XRD mi-
wards from amphibolite facies to the formation of migmatites, that
neral identification was performed at the Department of Geology of the
characterizes the basement (2.1–2.0 Ga; Boven et al., 2011), did not
U Ghent, whereas geochemical analyses for major and trace elements
affect the unconformably overlying Kisenge-Kamata series, which un-
(including REE), measured by using ICP-AES and ICP-MS at the RMCA.
derwent only a low-grade metamorphism (greenschist facies; see
Noble metals were assayed at ActLabs, Canada.
below). The observable geometry of the sedimentary succession is a
Stable isotope analyses (δ13C, δ18O) were performed at SUERC (East
steeply monoclinal disposition without observed major folding (see
Kilbride, Scotland) using an Analytical Precision AP2003 mass spec-
Figs. 3 and 4). This indicates that the Kisenge-Kamata series is late- or
trometer equipped with an acid injector system, after reaction with
post-Eburnean (≤2 Ga). Moreover, the series is crosscut by a pegmatite
105% H3PO4 under He atmosphere at 70 °C. Mean analytical reprodu-
body, dated at 1853 ± 89 Ma (Rb/Sr on muscovite; Ledent et al.,
cibility based on replicates of the SUERC laboratory standard MAB-2
1962). This age provides a poor chronological constraint, but it suggests
(Carrara Marble) was around ± 0.2‰ for both δ13C and δ18O. The
that the Kisenge-Kamata sediments were deposited before 1.8 Ga, re-
results are reported relative to V-PDB for δ13C, and relative to V-SMOW
ducing the time window to 2.0–1.8 Ga (Orosirian). To the north of the
for δ18O.
Kasai granulitic basement, 450 km north of Kisenge, an E-W elongated
graben, whose sediment infilling now consists predominantly of low-
4. Results
grade metamorphic shales (Lulua Group), resembles the Kisenge basin.
Differences are the absence of Mn deposits and the presence of inter-
4.1. Macroscopic features of manganese ore
stratified basalts in the Lulua basin. These basalts (Malafudi tholeiitic
basalts) have been dated at 1.96 ± 0.29 Ga (whole-rock and mineral
Unweathered manganese-rich rocks of Kisenge-Kamata range in

183
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

Fig. 3. Schematic log of the Kis-1 core, and of some other cores crossing the Kisenge deposit (after Doyen, 1974; De Putter et al., 2015); bedding is shown in pale
green shaded areas, based on dip measurements (angle symbols), and is an attempt to link surface outcrops of oxidized ore and mineralized intervals identified in the
cores (green shaded area in Kis-4 to 6 vertical cores and an unnamed horizontal core). The dotted line emphasizes the probable continuity between the bedding and
the position of the unweathered carbonate ore and garnetite level in the Kis-1 core. (For interpretation of the references to colour in this figure legend, the reader is
referred to the web version of this article.)

composition from Mn-carbonate rocks to non-calcareous deposits,
dominated by manganese-rich garnet (spessartine). Associated rocks
are grey graphitic shales with common visible spessartine.
The manganese-rich rocks present two different facies: layered or
massive grey garnetite and dark grey Mn carbonate ore. The latter
appears to overlie the garnet-rich facies, based on the available core
sections. A third manganese-rich facies is cryptomelane-dominated
oxidized ore (see De Putter et al., 2015).
The grey garnetite varies from layered deposits with moderate to
high garnet content to massive deposits that consist almost exclusively
of closely packed garnet crystals. The layered deposits are composed of
mm- to cm-thick layers of dark grey fine-grained garnet-bearing shale
and light grey coarse-grained garnet-rich material (Fig. 5a). The garnet
crystals appear as whitish to light grey round spots, reaching a size of
up to 2 mm. The continuous planar bedding is an expression of varia-
tions in abundance and size of the garnet crystals within the blackish
shale matrix. The layered garnetite has a graphitic aspect in part of the
cores. Massive garnetite is present at the base of the carbonate ore
section in core Ka-30 (Fig. 4), with an upward increase in garnet crystal
size.
The Mn carbonate ore is a dark grey to black fine-grained massive
homogeneous deposit, with sporadic weakly expressed lamination due
to the presence of mm-sized garnet crystals, recognized as whitish to
light-grey spots.

184
T. De Putter et al.
Fig. 4. Schematic log of the Ka-72 and Ka-30 cores, and of some other cores crossing the deposit (after Doyen, 1974); the profile illustrates the typical geometry of
carbonate ore lenses overlying garnet-rich ore (garnetite). Both ore types occur within a sequence of grey graphitic shales. The upper part of the profile has been
altered to supergene oxide deposits (Miocene).
The associated grey graphitic shales are fine-grained grey deposits,
locally with mm-thick wavy lamination emphasized by pale garnet
crystals or garnet crystal moulds (Fig. 5b). A strongly graphitic facies of
these shales is associated with oxidized manganese ore at Mwene Ditu
(Fig. 1), where only this type of ore is known but where observations
have not been extended to subsurface intervals that may include car-
bonate ore intercalations.
The carbonate ore and garnetite deposits of Kisenge-Kamata are
affected by multiple generations of fractures and infillings (Fig. 5c),
whereas the grey shales are significantly less strongly affected. The
fractures are commonly a few mm to a few cm wide, with infillings
composed of various minerals (see further). In the carbonate ore, im-
pregnation commonly occurs along the fractures, occasionally affecting
cm-wide zones and often occurring as coarse-crystalline features (see
further). Strong development of fractures and associated impregnation
occasionally obscure the dark matrix of the carbonate ore (Fig. 5d).
4.2. Mineralogical composition and textural features
The carbonate ore consists predominantly of rhodochrosite
(MnCO3), forming a relatively fine-grained microsparitic groundmass.
This matrix shows some variations in crystal size between samples and
between successive layers of stratified intervals. The carbonate ore
contains a considerable amount of graphite (confirmed by XRD ana-
lysis), which occurs as abundant opaque fine particles, contained as
inclusions in the rhodochrosite mass (see Fig. 6b and e). No braunite
(Mn2+Mn3+ 6 (SiO4)O8) was detected by XRD analysis or thin section
185
Ore Geology Reviews 96 (2018) 181–200
observations, in contrast to earlier reports for the Kisenge-Kamata de-
posits (Doyen, 1974).
The main phenoblast phase in the ore is spessartine
(Mn2+3 Al2(SiO4)3). The size and abundance of the round to polyhedral
spessartine crystals varies between samples and between layers of
stratified sections, typically with narrow size range and locally with
gradual vertical changes in crystal size. Deformation of parallel struc-
tures in the matrix around spessartine crystals is observed where those
structures are accentuated by variations in graphite content (Fig. 6a).
The spessartine crystals commonly contain graphite and/or carbo-
nate inclusions, whereby the carbonates are typically more fine-grained
than the rhodochrosite matrix (see Fig. 6c and e). Both types of inclu-
sions produce patterns determined by sector zoning (Rice and Mitchell,
1991; Rice et al., 2006), which is also expressed in the form of phe-
noblasts with the appearance of radial aggregates of elongated crystals
(Fig. 6b) (cf. Andersen, 1984). Within the growth sectors, parallel
elongated crystals oriented perpendicular to the sides of the phenoblast
are commonly recognized (cf. Burton, 1986). Less common features are
concentric zoning in the outer part of the spessartine crystals and the
concentration of graphite along their sides.
Elongated tephroite (Mn2+2 SiO4) crystals represent a second major
phenoblast phase (Fig. 6c). The same mineral also occurs as short
elongated clusters to longer planar features composed of equant crys-
tals, with orange brown fine material along crystal boundaries and in
part with associated rhodochrosite (Fig. 6d). A peculiar, but common
form of tephroite in the carbonate ore are large crystals with regular
banding, in which parallel layers of graphite-bearing rhodochrosite,
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

Fig. 5. Different facies of the ore and associated shales of the Kisenge-Kamata deposits: (a) alternation of dark grey fine-grained garnet-bearing shale and light grey
coarse-grained garnet-rich deposits; (b) grey graphitic shale with sporadic wavy lamination emphasized by pale garnet crystals or garnet crystal moulds; (c) typical
aspect of the carbonate ore, with multiple generations of fractures and veins; (d) iron-rich interval, crossed by caryopilite veins (greenish).

similar to matrix material, separate parallel tephroite sections that are
in optical continuity (Fig. 6e). These crystals can be isolated or clus-
tered, with parallel to oblique relative orientations. Banding is never
deformed around enclosed spessartine crystals (see Fig. 6e).
The carbonate ore is crossed by various types of veins. Tephroite
veins, marked by a greenish grey colour in hand specimens, are com-
posed of large anhedral crystals, without banded patterns, and typically
contain associated dolomite. Pyroxmangite (Mn2+SiO3), with bright
reddish colours in hand specimens, is less common as vein material,
typically occurring as elongated crystals with fine-grained inclusions
along their central axis where it occurs in neighbouring parts of the
matrix.
In several samples, veins of yellowish brown caryopilite
(Mn2+3 Si2O5(OH)4) occur, with associated transformation of spessartine
and other silicates in neighbouring parts of the matrix. Other common
vein types are rhodochrosite and quartz veins, both with variable tex-
tures, whereby quartz veins generally crosscut all other types of veins.
In the vicinity of quartz and rhodochrosite veins, tephroite phenoblasts
are commonly altered.
Garnetite intervals are marked by high contents of opaque material
in the matrix (Fig. 6f), with variable concentrations of a recognizable
carbonate fraction. The spessartine crystals commonly show concentric
patterns caused by variations in abundance of opaque inclusions (see
Fig. 6f). Cementation by xenotopic quartz is pervasive at several levels.
The associated shales, which were not systematically studied using
thin sections, consist largely of muscovite and kaolinite, with dispersed
fine non-sulphidic opaque material, whereby kaolinite appears to be an
alteration product of chlorite.
The study of polished sections confirmed the presence of various
sulphides in carbonate ores of core Ka30, which had previously been
reported by Doyen (1974). Although it was not possible to establish a
paragenetic relationship between different sulphide minerals, a large
variety of different fine-grained (< mm) Cu-, Co-, Ni, Pb and Zn-sul-
phides has been identified. Minerals like cobaltite, nickeline, pyrite,
chalcopyrite, chalcocite, galena and sphalerite occur dispersed in the
recrystallized matrix of the carbonate rocks, whereas mm-size galena
has also been identified in cross-cutting rhodochrosite veins. The Cu-
sulphides are partly altered to digenite, and the Fe-sulphides to iron
oxides (goethite, hematite) (Fig. 7).
The Kisenge-Kamata deposits are most comparable with various
occurrences of Paleoproterozoic manganese ores in Brazil, for which
petrographic data are relatively scarce (e.g. Candia and Girardi, 1979;
Costa et al., 2005; Chisonga et al., 2012). Manganese carbonate ores
have also been documented for Nsuta deposits of Ghana (e.g.
Kleinschrot et al., 1994), but petrographic information for that occur-
rence mainly concerns specific carbonate facies (e.g. Nyame et al.,
2003; Nyame and Beukes, 2006; Nyame, 2008). Another similar African
deposit (Kampumba, Zambia) is hardly documented (Fockema and

186
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

Fig. 6. Petrographic features of the Kisenge-Kamata deposits. (a) Graphite-rich lenses following the contours of spessartine phenoblasts, in a graphite-bearing
microsparitic rhodochrosite matrix (Kamata, core Ka72, RG 79644, 37.20–38.50 m; frame width 2.6 mm); (b) Spessartine crystal with strongly pronounced sector
zoning (Kamata, core Ka72, RG 79664, 78.30–79.30 m; fw 1.3 mm); (c) Tephroite and spessartine phenoblasts, in rhodochrosite matrix with coarser crystal size than
carbonate inclusions in spessartine (Kamata, core Ka72, RG 79658, 70.90–71.70 m; fw 2.6 mm); (d) Undulating intercalation composed of tephroite and rhodo-
chrosite (Kamata, core Ka72, RG 79665, 79.30–80.70 m; fw 2.6 mm); (e) Tephroite phenoblast composed of isolated parallel sections that are in optical continuity,
separated by matrix-related graphite-bearing rhodochrosite, with enclosed spessartine crystals (Kamata, core Ka72, RG 79644, 37.20–38.50 m; fw 1.3 mm); (f)
Garnetite, composed of zoned spessartine in matrix of opaque fine material (Kisenge, core Kis1, RGM 8246, 136.30 m; fw 2.6 mm).

Austen, 1956).
4.3. Chemical composition
Chemical analyses were performed for 70 samples, including seven
samples of local basement rocks (Table S1). The latter include micas-
chists (57.7–59.4 wt% SiO2) (n = 3) characterized by high Al2O3 con-
tents (18.9–21.4 wt%), basic rocks (49.1–52.6 wt% SiO2) (n = 3) and
pegmatites (68.1 wt% SiO2) (n = 1). For the Kisenge-Kamata deposits,
some samples yielding high LOI values have a the sum of oxides that is
considerably < 100% (95–97%), which we attribute to LOI under-
estimation. Therefore, the oxide percentages reported in this study were
not recalculated.
In terms of major element composition, the often graphitic shales
that are part of the Kisenge-Kamata series mostly have high SiO2 con-
tents (60.5–69.0 wt%) and variable Al2O3 contents (11.0–21.8 wt%)
(n = 8), whereas some shales display lower SiO2 contents (42.1–55.1 wt
%) and more variable Al2O3 contents (9.9–33.2 wt%) (n = 4). The
carbonate ore is mainly characterized by high MnO contents
(42.3–58.7 wt%), high LOI values (24.5–33.2 wt%), and low SiO2 con-
tents (2.7–11.5 wt%) (n = 22). Several carbonate ore intervals are
characterized by higher SiO2 contents (22.1–67.5 wt%), partly due to
higher spessartine content, and they also yield more variable MnO
contents (8.5–50.8 wt%) and lower LOI values (7.1–18.4 wt%)
(n = 12). These deposits are referred to as high-silica carbonate ore in
this paper, distinct from the more common low-silica carbonate ore.
The oxidized ore has overall a variable composition (0.8–67.0 wt%
SiO2, 1.2–13.4 wt% Al2O3, 3.1–70.6 wt% MnO; n = 10).

187
T. De Putter et al.
Fig. 7. Petrographic features of opaque minerals the Kisenge-Kamata deposits (Kamata, core Ka30). (a) Cobaltite (Cob) and nickeline (Nic) (RGM 8392; PPL, frame
width 650 µm). (b) Cobaltite (Cob) and pyrite (Py) (RGM 8392; PPL, fw 325 µm). (c) Sphalerite (Sph), galena (Gn) and pyrite (Py) (RGM 8392; PPL, fw 650 µm). (d)
Chalcopyrite (Cp), bornite (Bn) and digenite (Dig) (RGM 8392; PPL, fw 325 µm).
Concerning transition elements, the various analysed materials
show different types of enrichment relative to the upper crust average
(Rudnick and Gao, 2003; Table S2). The basement rocks are variably
enriched in certain elements, without exceeding two to three times the
upper crust value: 29–239 ppm Cr (vs. 92 ppm), 27–52 ppm Co (vs.
17 ppm), 20–147 ppm Ni (vs. 47 ppm), 1–92 ppm Cu (vs. 28 ppm),
71–152 ppm Zn (vs. 67 ppm), 100–324 ppm V (vs. 17 ppm). The shales
of the Kisenge-Kamata series are similarly enriched (53–343 ppm Cr:
2–160 ppm Co: 7–114 ppm Ni: 2–218 ppm Cu: 11–187 ppm Zn:
76–312 ppm V), except for a greater degree of Pb enrichment
(19–105 ppm). The low-silica carbonate ore is enriched in Co
(142–748 ppm), Ni (420–1290 ppm), and Zn (169–860 ppm), but it is
poor in Cr (< 20 ppm) and mostly also in Cu and Pb. The high-silica
carbonate ore is similarly enriched in Ni (202–710 ppm) and Zn
(181–980 ppm), depleted in Cr (< 20 ppm, with one exception), and
depleted to weakly enriched in Pb, whereas Cu and Co show varying
degrees of (mutually independent) enrichment. The oxidized ore is
variably enriched in Ni (116–3321 ppm), Zn (105–2622 ppm), Co
(32–987 ppm), Cu (127–1119 ppm, with two exceptions) and V
(145–453 ppm), generally depleted in Cr (< 85 ppm) and variably de-
pleted to enriched in Pb.
Au, Pt and Pd concentrations, both in low-silica carbonate ores
(n = 17) and high-silica carbonate ores (n = 3), are high compared to
mean upper crust values, with 11–92 ppb Au (vs. 1.5 ppb), 7–20 ppb Pt
(vs. 0.5 ppb) and 1–10 ppb Pd (vs 0.52 ppb) for the low-silica carbonate
ore, and much higher Au concentrations in the high-silica carbonate ore
(44–791 ppb).
188
Ore Geology Reviews 96 (2018) 181–200
4.4. Stable isotope composition (δ13C, δ18O)
Four of the five analysed low-silica carbonate ore samples of the
Kisenge-Kamata deposits, composed of rhodochrosite with minor or no
dolomite as carbonate minerals (see Table 1), have weakly negative
δ13C values (-3.8 to -6.3‰ δ13CPDB) and low18O values (17–21.1‰
δ18OSMOW). The fifth sample (Ka72/45p) has a more strongly negative
δ13C value (−10.6‰) and a higher δ18O value (24.7‰). The high-silica
carbonate ores, with only rhodochrosite as carbonate phase, show a
more variable composition, with δ13C values of –5.9 to –18.9‰ and
δ18O values of 20.4–33.3‰, but they are overall characterized by ra-
ther strongly negative δ13C and high δ18O. Rhodochrosite veins have
rather strongly negative δ13C values (−12.3 to −14.6‰) and high
δ18O values (22.7–26.9‰).
5. Discussion
5.1. General significance of textural features
The carbonate ore is the result of recrystallization of fine-grained
carbonate deposits during burial and later deformation, without pro-
ducing a coarse grain size or pressure twinning, suggesting a relatively
low degree of metamorphism. Carbon species were transformed to
dispersed graphite particles, masking the intrinsically pinkish colour of
the rhodochrosite matrix. The presence of a silicate fraction resulted in
spessartine and tephroite formation, whereby variations in silicate
content of the protolith appear to be recorded by variations in pheno-
blast abundance and size. Spessartine phenoblasts formed by largely
displacive growth, as registered by inclusion patterns and occasional
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

graphite-rich linings (Rice and Mitchell, 1991). The carbonate inclu-

pyroxmangite, mica, graphite

pyroxmangite, mica, graphite
sions that they contain, with finer grain size than the rhodochrosite
matrix, appear to record early spessartine development and protection
graphite, kaolinite of enclosed carbonates from recrystallization.
The banded patterns shown by tephroite phenoblasts were inter-
preted by Doyen (1974) as stromatolitic structures, whose occurrence
kaolinite

graphite
has been documented for Paleoproterozoic deposits of the region (e.g.

Quartz, Tephroite
Litherland and Malan, 1973), but this interpretation does not account
for several aspects of the phenoblasts. Possible alternative explanations
XRD (minor)

Spessartine,
Spessartine,
Spessartine,
Spessartine,

Spessartine,

include alteration of lamellar intergrowths of tephroite and allegha-

Quartz

Quartz

Quartz
nyite (Mn2+5 (SiO4)2(OH)2) (Campbell Smith et al., 1944; see also
Rogers, 1919), and alteration of intergrowths formed by exsolution,
which has been reported for tephroite (e.g. Hurlbut, 1961) and which
produces two intersecting sets of lamellae in olivine group minerals (see
Rhodochrosite (no XRD conf.)

Rhodochrosite (no XRD conf.)

Rhodochrosite (weathered)
also Mikouchi et al., 1995). However, derivation from this type of
Rhodochrosite, spessartine

Rhodochrosite, spessartine
Rhodochrosite, graphite

precursor does not explain the matrix-related nature of the bands se-
parating the tephroite sections. Enclosed spessartine phenoblasts,
Rhodonite, quartz

without associated deformation of tephroite banding, confirm early

Rhodochrosite

Rhodochrosite
Rhodochrosite

Rhodochrosite

Rhodochrosite
Rhodochrosite

Rhodochrosite
XRD (major)

garnet development.
Dolomite
Dolomite

The Mn-silicate assemblage at Kisenge-Kamata (pyroxmangite, te-

phroite, scarce rhodonite) is compatible with metamorphism at high
pressure (∼2 kbar) and low temperature (∼250 °C), constrained by the
stability of tephroite and pyroxmangite (Dasgupta et al., 1993;
Mohapatra and Nayak, 2005). Tephroite is stable at low pCO2, whereas
δ18O VSMOW

pyroxmangite is a high-pCO2 phase (XCO2 > 0.2), so pyroxmangite

formation could be due to local conditions of higher pCO2, related to
17.0
19.7
24.7
17.4
18.3
25.5
16.6
19.2
22.4
33.3
20.4
21.1
23.9
26.9
22.7
23.0

rhodochrosite breakdown. Among common Mn-silicate minerals, the

high-temperature phase, at any pressure, is rhodonite (Maresch and
Mottana, 1976), which is scarce in the Kisenge-Kamata deposits.
δ18O VPDB
Stable isotope (δ13C, δ18O) composition of bulk carbonate ore and rhodochrosite veins of the Kisenge-Kamata deposits.

The Kisenge-Kamata deposits were affected at various stages by

−13.5
−10.8

−13.1
−12.2

−13.9
−11.3

−10.2

fracturing and vein development with rhodochrosite, various Mn sili-

−6.0

−5.3

−8.2

−9.5
−6.8
−3.9
−7.9
−7.6
2.3

cates and late-stage quartz as mineral phases. However, the deposits

have been only affected by low-grade metamorphism and display little
or no evidence for subsequent intense hydrothermal alteration. The
δ13C VPDB

veins are, therefore, probably the result of local dissolution-re-

−10.6

−10.5

−14.8
−10.7
−18.9

−14.2
−12.3
−12.8
−14.6
−6.0
−4.6

−5.1
−4.0

−6.3
−3.8

−6.0

precipitation, producing a pyroxmangite-tephroite-(rhodonite) asso-

ciation (Maresch and Mottana, 1976).

5.2. Geochemical characteristics and enrichment processes

Sample type

Host-rock
Pink vein
Pink vein
Pink vein
Pink vein

The Kisenge-Kamata ores are rich in MnO, ranging from 9 to 59%

Ore
Ore
Ore
Ore
Ore
Ore
Ore
Ore
Ore
Ore
Ore

with an average content of 42% MnO for the primary ore and from 3 to
68%, with a mean of 37% MnO for the oxidized ore. The Kis-1 core
samples (Fig. 3) illustrate how the elements are enriched within the
Graphite and garnet-bearing rhodochrosite marble

No separate analysis of vein, weathered material

different zones (Fig. 8a and b). In this core, the supergene Mn ore in-
terval is much thicker (c. 40 m) than the remaining Mn carbonate ore
(c. 5 m), but the Mn enrichments are similar (Fig. 8a). In both cases, Mn
Garnet and graphite-bearing dolomite
Olivine-bearing rhodochrosite marble

Garnet-bearing rhodochrosite marble

Garnet-bearing rhodochrosite marble

Garnet-bearing rhodochrosite marble

enrichment is accompanied by enrichment in Ni, Zn and Co, whereby

Graphite and garnet-bearing shale

Ni and Zn are most strongly enriched in the supergene ore. In contrast,

Fe and Cr are depleted in both Mn ore types. These elements are en-
No separate analysis of vein

riched in the oxidized horizons bracketing the Mn carbonate ore

Rhodochrosite Marble
Rhodochrosite marble

Rhodochrosite marble

Rhodochrosite marble
Rhodochrosite marble

(Fig. 8b) and in the supergene Mn ore, especially in its upper part
Rhodochrosite vein

Rhodochrosite vein
Sample description

(Fig. 8a), as well as in the micaschist basement. Cu is not enriched

except at the base of the supergene Fe-Cu horizon. The Kisenge deposit
is overall poor in iron. In the carbonate ore, the iron oxide content
varies between 0.6 and 3.8%, with a mean of 1.8% Fe2O3t (Table S1). In
the oxidized ore, Fe2O3t is generally higher, with values between 0.5
and 9% and with a mean of 5.2%, if two Fe-rich samples (13 and 21%)
are not considered (Table S1). The SiO2 content of the Mn carbonate ore
Sample id. (Ka-72 core)

is lower than that of the basement rocks, but still relatively high
(22–68%, Table S1), in a core in which only the silica-rich carbonate
ore type is present (Fig. 3).
Ka72/72_PKV
Ka72/45_PKV
Ka72/46_PKV
Ka72/63_PKV

In the Kis-1 core, Fe and Cr are positively correlated for samples of

Ka 30/107
Ka72/103
Ka72/107
Ka72/45p

Ka72/75b

Kis1/262
Kis1/268
Kis1/271
Ka72/21

Ka72/69

Ka30/55
Ka72/2

the oxidized upper part (Fig. 9a), showing a trend that is different from
Table 1

that for shale and basement rock samples. The latter represents the
normal crustal trend with strong Cr enrichment relative to Fe. On the

189
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

Fig. 8. Variations of selected metal concentrations with depth in the Kis-1 core. (a) Variations over the full core length; the carbonate ore and associated shales have
high Cu (shale), Co, Ni and Zn (carbonate) concentrations, and metals are also concentrated in the Cenozoic supergene ore horizon (Co-Ni-Zn in Mn-oxide ore, Cu in
mixed Mn-Fe oxide ore). (b) Zoom on the Mn carbonate ore interval at ∼135 m along core (see also Fig. 3): the carbonate ore, with high Co, Ni and Zn, is framed by
two thin oxidized intervals, composed of Mn oxide ore (below) and Fe-oxide-rich veined carbonate deposits (above).

other hand, a single trend exists for Ni-Zn data representing all rock
types (Fig. 9b).
In the Kis-1 core from Kisenge, only high-silica carbonate ores are
represented; therefore the data need to be combined with those for
Kamata (Ka30, Ka72) to obtain a complete view of carbonate facies
composition. At Kamata, the cores mainly crosscut low-silica carbonate
ore (0.8–11.5% SiO2, with 1–3.3% Al2O3), but high-silica carbonate ore
is also present (21.5–39.9% SiO2, with higher Al2O3 content) (Table
S2). The strong negative correlation between MnO and SiO2 (Fig. 10a)
suggests dilution of manganese sedimentation by input of detrital sili-
cates, as no indications for hydrothermal silica enrichment exist. MnO
is not clearly correlated with Fe2O3t (Fig. 10b, see also Fig. 8a), re-
cording a difference in enrichment processes for both elements.
Low-silica carbonate ores, high-silica carbonate ores and oxidized

190
T. De Putter et al.
Fig. 8. (continued)
ores display different REE diagrams (normalized to PAAS; Fig. 11a–c),
but they are similar in having rather flat patterns with positive Ce and
Eu anomalies. The Ce anomalies are true anomalies, unrelated to La
anomalies, as shown by the negative correlation between Ce/Ce∗ and
Pr/Pr∗ (Fig. 11d; Bau and Dulski, 1996). The Ce and Eu anomalies are
not correlated (Fig. 11e), which means that they are caused by different
processes, as expected for elements whose non-trivalent form is dif-
ferent (Ce4+ vs. Eu2+), and hence requires different redox conditions
(oxidizing for Ce4+, reducing for Eu2+).
The Ce anomalies are positively correlated with MnO (Fig. 11f),
which can be due to adsorption of oxidized Ce4+ on Mn-oxide/
191
Ore Geology Reviews 96 (2018) 181–200
hydroxide colloids and on organic particles (Sholkovitz et al., 1994;
Holser, 1997). Mn oxide colloids precipitate when redox conditions
allow oxidation of Mn2+ to Mn4+, the latter forming oxides/hydroxides
with much lower solubility than Mn2+. Elements adsorbed by these
colloids, such as cerium in the form of Ce4+, which forms in similar Eh-
pH conditions as Mn4+ (Sholkovitz et al., 1994), can be incorporated
during crystallisation of minerals from the colloidal deposits
(Koschinsky and Halbach, 1995). The Ce anomaly can, therefore, have
developed during initial Mn precipitation in the form of Mn oxides, as
precursor of Mn carbonates that formed later by mineral transforma-
tion.
T. De Putter et al.
Fig. 9. Transition metal plots. (a) Cr vs. Fe2O3t plot, showing a distinct oxidation trend and a crustal trend; (b) Zn vs. Ni plot, illustrating that the data represent a
single set showing positive correlation between both elements.
The positive Eu anomaly, formed during mineral crystallization in
reducing conditions, developed at a different stage than the Ce
anomaly. Excluding the possibility of hydrothermal processes, it can be
related to diagenesis by interaction with anoxic alkaline pore solutions
(see Xiao et al., 2017). At Kisenge-Kamata, this can have occurred
during transformation of Mn oxides to carbonates, in a CO2-rich en-
vironment. This diagenetic change took place without causing re-
mobilization of Ce as Ce3+. Further studies are needed to give an
adequate explanation, but strong attraction of LREE by Mn oxides (Usui
et al., 2017) and variable kinetics are possible factors. In the context of
a graben, as proposed for the Kisenge-Kamata basin, a minor con-
tribution of hot (> 250 °C) hydrothermal fluids (Bau and Dulski, 1996;
Beukes et al., 2016) can, however, also have had a role in generating
positive Eu anomalies.
192
Ore Geology Reviews 96 (2018) 181–200
In spider-diagrams (Fig. 12, normalized to pelagic clay; Li and
Schoonmaker (2014), the low-silica carbonate ores, high-silica carbo-
nate ores and oxidized ores are all marked by positive Mn peaks; the
carbonate ores also show a positive Ca peak, which is compatible with
their mineralogical composition. All are depleted in Ba and most are
depleted in Sr, implying interaction with modified marine brines rather
than ocean water preserving its original composition in terms of Ba-Sr
content. The spider-diagrams show that the low-silica carbonate ores
are the most homogeneous, the others being much more variable (ex-
cept LILE for the high-silica carbonate ores).
The stable isotope composition of bulk carbonate ore and rhodo-
chrosite veins of the Kisenge-Kamata deposits (Table S1) are compared
in this study with published data for other Paleoproterozoic Mn ore
deposits from Africa and South America (Table S2): Serra do Navio
T. De Putter et al.
Fig. 10. MnO vs. SiO2 and vs. Fe2O3 plots. (a) MnO vs. SiO2, showing negative
correlation for the Kisenge-Kamata ores (b) MnO vs. Fe2O3, showing much
higher enrichment in Mn than in Fe in the low-Si ores; (c) Fe2O3 vs. SiO2.
(Chisonga et al., 2012), Moanda (Hein and Bolton, 1993), Nsuta (Yeh
et al., 1995; Nyame and Beukes, 2006), and three KMF deposits, i.e.
Voëlwater (Schneiderhan et al., 2006), Hotazel (Tsikos et al., 2003),
and Mamatwan (Gutzmer et al., 1997). In comparison with other Pa-
leoproterozoic manganese deposits in Africa and Brazil (Fig. 13a), δ13C
values for Kisenge-Kamata ores show a similar spread, whereas δ18O
values are higher than most reference deposits but similar to those re-
ported for KMF ores (Gutzmer et al., 1997). The low-silica carbonate
ores have generally less negative δ13C values than the high-silica ores
and the rhodochrosite veins. One low-silica carbonate ore sample, al-
beit affected by recent weathering (Ka72/45p), has a value similar to
those obtained for rhodochrosite veins, suggesting that the matrix was
locally transformed during vein development. This may also apply for
those high-silica ore samples whose δ13C values are close to those of the
vein occurrences, whereby part of the silicate fraction of those macro-
scopically unaltered deposits could in fact be vein-related, rather than
exclusively controlled by protolith-determined spessartine content. The
latter is suggested by relatively high residual SiO2 concentrations when
part of the total SiO2 content is attributed to spessartine by assigning all
Al2O3 to that mineral, with assumed stoichiometric composition.
Whether the abundance of pure Mn-silicates, such as tephroite, is un-
related to protolith composition remains to be confirmed. The results
for Kisenge-Kamata are comparable with those for Nsuta (Nyame and
Beukes, 2006) in presenting two clusters, one with weakly negative
δ13C, low δ18O and limited spread, and the other with more negative
δ13C, higher δ18O and greater spread, but the attribution of the second
193
Ore Geology Reviews 96 (2018) 181–200
group to interaction with later fluids (Nyame and Beukes, 2006) is not
clearly applicable to Kisenge-Kamata.
Sample Ka72/45p has a more negative δ13C than the other low-
silica carbonate ores. When considering the MnO content, this sample is
close to the rhodochrosite veins (Fig. 13b). As a whole, the low-silica
carbonate ores plot on a straight line with the rhodochrosite veins. This
indicates that the late events, either fracture filling or metamorphism,
have lowered the δ13C values. Knowing that marine carbonates that
formed at 1.95 Ga have a δ13C values around +1% and that marine
organic carbon of the same period has δ13C values around −28% (Kah
et al., 2004; Nyame and Beukes, 2006; Krissansen-Totton et al., 2015;
Bindeman et al., 2016), the weakly negative δ13C values of the low-
silica carbonate ore points to a predominant role of marine fluids with
little influence of organic matter. The trend towards lower values si-
milar to the rhodochrosite veins can be attributed to the later event(s)
that generated these veins.
The high-Si primary ores plot below the line marked by vein rho-
dochrosite and low-silica carbonate ores due to dilution of Mn by sili-
cates. They have also strongly negative δ13C values, more negative than
vein rhodochrosite, suggesting that silica enrichment occurred at a
different (later) stage. This second group overall displays more positive
δ18O and more negative δ13C values, both with relatively variable
ranges (Fig. 9). As mentioned above, these samples comprise secondary
minerals pointing to decreasing metamorphism and/or alteration.
For base metal origin, the correlation of metal concentration with
SiO2 content can be considered, whereby the strong negative correla-
tion between MnO and SiO2, interpreted as indicating a marine Mn
source and detrital Si carrier, is used as reference (Fig. 14a). Cobalt and
nickel show a pattern that is similar to that of MnO (Fig. 14b and c),
indicating a marine origin for these two elements. Zn displays a similar
pattern, but with some additional enrichment in the high-silica carbo-
nate ores, pointing to a marine origin with superimposed later enrich-
ment whose nature still needs to be determined. In contrast to the
previous elements, chromium concentrations are low in the low-silica
carbonate ores and in oxidized ores that are poor in silica, and they can
be high in the associated shales and in local basement rocks, indicating
a continental origin for this element. The high-Si carbonate ores and
high-Si oxidized ores show enrichments, reflecting a varying admixture
of bedrock-derived material. Copper shows no variation with SiO2
content or rock type, suggesting a mixed origin. Current copper con-
centrations are also partly determined by supergene processes, with
strong enrichment in the oxidized ores.
5.3. Manganese ore deposition and paleogeographic context
The manganese ore at Kisenge-Kamata is a chemical sediment un-
conformably deposited on the Archean/Paleoproterozoic Kasai Block of
the Congo Craton during the Orosirian period (2.0–1.9 Ga). The pri-
mary ores are low-grade metamorphic carbonate rocks dominated by
rhodochrosite and spessartine, alternating with graphitic shales, which
are partly spessartine-bearing. Their deposition occurred at or just after
the end of the Eburnean orogeny that shaped the Kasai Block. No
younger orogenic major events affected the Kisenge-Kamata area or any
other part of the Kasai Block. Later tectonic movements only tilted the
series to a subvertical position, with no evidence of folding. This sug-
gests that the current geometry of the Kisenge-Kamata basin is close to
its original disposition, which is a narrow E-W elongated in-
tracontinental graben basin, based on field and geophysical data
(Figs. 1 and 2). A comparable setting is known to have existed in the
northern part of the Kasai Block, where sediments were deposited in a
graben at c. 1.96 Ga (Lulua graben; Delhal et al., 1989). This area was
affected by subsidence, controlled by active faults along the basin
margins, after the Eburnean metamorphic event (André, 1993). This
suggests that the Kisenge-Kamata basin developed in an overall exten-
sional setting, in an opening graben, affected by tectonically-driven
subsidence, in an intracontinental setting. This specific intracratonic
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

Fig. 11. PAAS-normalized (Taylor and McLennan, 1985) REE patterns for (a) low-silica carbonate primary ore; (b) high-silica carbonate primary ore, and (c) oxidized
ore; (d) Ce/Ce* vs. Pr/Pr* plot, confirming the existence of true Ce anomalies for the various ore types; (e) Ce/Ce* vs. Eu/Eu* plot, showing a lack of correlation
between Ce and Eu anomalies, and hence suggesting that they relate to distinct processes; (f) Ce/Ce* vs. MnO plot, showing a positive correlation that results from the
adsorption of Ce4+ on Mn ore.

194
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

Fig. 12. Pelagic-clay-normalized spider diagrams (Li and Schoonmaker, 2014) spider-diagrams for (a) low-silica carbonate ore; (b) high-silica carbonate ore and (c)
oxidized ore.

195
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

Fig. 13. Stable isotope data plots: (a) δ13C vs δ18O, with results for the Kisenge-Kamata deposits: low-silica carbonate ore, high-silica carbonate ore and rhodochrosite
veins and published data for other Paleoproterozoic African deposits (Kalahari Manganese Fields, Nsuta, Moanda) and a related Brazilian deposit (the Serra do Navio)
deposit (Brazil); (b) MnO vs. δ13C plot: low-silica carbonate ore and rhodochrosite veins plot on a straight line.

location does not support the recently proposed back-arc basinal de-
positional setting for the Kisenge-Kamata ores (Beukes et al., 2016).
Geochemical characteristics, such as positive Ce anomalies and
weakly negative δ13C values of the carbonate ore, indicate contact with
open-marine water. Ocean waters had a δ13C values of c. + 1‰ at
1.95 Ga (Krissansen-Totton et al., 2015). At this time, ocean waters
were still only weakly oxidized. Atmospheric pO2 had risen sub-
stantially from c. 2.4 Ga to c. 2.0 Ga (e.g. Holland and Beukes, 1990),
causing massive deposition of banded iron formations until 1.8 Ga, but
the degree of oxidation of ocean waters probably remained low, even
throughout the Mesoproterozoic (Sperling et al., 2015). This implies
that during the Orosirian, ocean waters were globally reduced and in
consequence rich in manganese. However, progressive oxidation of the
oceans resulted in stratification, with an oxidized upper water body and
an anoxic (sulfidic) lower interval (Anbar and Knoll, 2002). In shallow
basins, or shallow marginal parts of ocean basins, manganese that

196
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

Fig. 14. Plots of metal contents (MnO, Co, Zn, Cr, Cu, Ni), vs. SiO2: (a) MnO vs. SiO2; (b) Co vs. SiO2; (c) Ni vs. SiO2; (d) Zn vs. SiO2; (e) Cr vs. SiO2; (f) Cu vs. SiO2.
Mn, Co, Ni and Zn (a–d), indicating the relative importance of the marine vs. terrigenous origin of the considered elements.

precipitated at this time would be preserved, in contrast to deep-ocean
environments, where Mn4+ species would be dissolved upon settling in
reducing conditions. In intracontinental basins, such as the one inferred
for Kisenge-Kamata, shallow water conditions can be predominant and
they can therefore be preferential sites for manganese accumulation,
under favourable conditions. Such conditions can exist along con-
tinental margins, where pure Mn ores would accumulate in distal parts,
while deposits with an admixture of detrital material would form in
more proximal parts. Subsidence coupled with successive transgression-
regression cycles could produce a sedimentary succession as observed
for the Kisenge-Kamata deposits (Fig. 4). However, this model does not
take into account the specific setting of the Kisenge-Kamata basin,
which is an intracontinental elongated graben, and the nature of the
deposits, with important differences in composition between carbonate
ores and associated shales. The latter have low concentrations of Mn
(< 0.2%) and base metals (Co, Ni, Zn) (Table S1). Only a few thin
spessartine-bearing shale intervals adjacent to the carbonate ore have
higher MnO contents (Kis1/157 and Kis1/111 with c. 5.5% MnO, and
Kis1/121 with 1.25% MnO; Fig. 8). This indicates that these metals did
not precipitate during deposition of fine-grained detrital material,
which is in better agreement with overall Mn-poor water compositions
during periods dominated by detrital sedimentation. In conditions with
continuously Mn-rich waters, deposits that accumulated at stages with
abundant clastic sediment input are typically rich in manganese, de-
pending on sedimentation rates (e.g. Herbosch et al., 2016).
We propose that the Kisenge-Kamata graben was most of the time a
land-locked basin with supply of Mn-poor water and sediments by
rivers. When the barrier isolating the basin from the ocean was lowered,
for instance due to vertical tectonic movements linked to graben for-
mation, Mn-rich ocean water invaded the basin, resulting in sudden

197
T. De Putter et al. Ore Geology Reviews 96 (2018) 181–200

Fig. 15. Depositional context: (a) schematic block diagram (modified from Prosser, 1993) of the graben basin at Kisenge-Kamata. Stage (b) corresponds to the closed
basin in which Mn-free detrital clay is deposited, together with metals having a terrigenous origin; stage (c) corresponds to the basin connected with the ocean,
bringing in Mn and associated metals; stage (d) again represents the closed basin, with detrital material deposited over the Mn ore that formed during the previous
stage.

198
T. De Putter et al.
deposition of Mn and associated metals during mixing with oxic basin
waters, whereby Paleoproterozoic ocean waters are assumed to have
been only weakly oxidized (Planavsky et al., 2009, 2010). If further
exchange with the ocean is limited, due to a narrow ocean connection
or to basin closure following rise of the barrier, the basin water will
rapidly attain low Mn concentrations, in equilibrium with poorly so-
luble Mn4+ compounds, ending the episode of Mn precipitation. The
model that we propose (Fig. 15) accommodates the development of
high-silica carbonate ores and garnetite deposits, in parts of the basin
with significant settling of detrital material.
In the schematic presentation of the favoured model (Fig. 15), the
barrier is positioned along the longitudinal edge of the graben (current
SW boundary), but it could be also located at the end of the graben
(current SE edge), which would not change the functioning of the
model. Within the basin, proximal deposits were dominated by fine-
grained detrital material, originating from the Kasai Block (Fig. 15a).
These deposits may not have contained any authigenic Mn compounds.
The clayey sediments were deposited in conditions with high carbon
species concentration (Catling and Claire, 2005), ultimately resulting in
graphitic shale as weakly metamorphic facies. This stage is accom-
panied by deposition of Cr and probably some Cu (Fig. 15b). When the
barrier between ocean and graben is submerged, marine water invades
the basin bringing in Mn, Co, Ni, Zn and some Cu in reduced state
(Fig. 15c). All ores display positive Ce anomalies, controlled by ocean
water composition, whereby the anomaly is more pronounced with
increasing Mn content (Fig. 11F). When the intracontinental graben
basin was again isolated, the terrigenous detrital influx once more
predominated, corresponding with low-Mn shale intervals in the Ki-
senge-Kamata series (Fig. 15d).
The rhodochrosite deposits trap Mn and also concentrate Ce relative
to other LREE (La, Pr) (Fig. 11; see also Beukes et al., 2016). The weak
positive Eu anomaly that is shown by the Kisenge-Kamata carbonate ore
(Fig. 11) most likely resulted from diagenetic interaction with pore
water in reducing conditions (Xiao et al., 2017). The Kisenge-Kamata
ore also confirms the overall trend for Mn deposits to display significant
enrichment in manganese (relative to potential source rocks), but not in
iron (Fig. 9a and b), with Mn-Fe partitioning reaching values exceeding
those recorded for other Paleoproterozoic deposits (MnO/Fe2O3
ratio > 40; Maynard, 2010; see also Green and Daly, 1982).
6. Conclusion
The Kisenge-Kamata Mn deposit has been first reported in the 1930s
and the earliest descriptions were published in the 1940s (Polinard,
1946). Following the work of Doyen (1973), the Kisenge-Kamata de-
posit has not been re-investigated, and it is only briefly mentioned in
review papers dealing with manganese deposits (e.g. Roy, 2006,
Kuleshov, 2011; Beukes et al., 2016). The present study provides ad-
ditional information about this poorly known Paleoproterozoic
(2.0–1.9 Ga) manganese deposit, and about the Kasai Block, which
hosts the Kisenge-Kamata deposit and possibly several comparable oc-
currences, as suggested by showings of oxidized Mn ores at Mwene-Ditu
(Kasai).
Overall, the issue of the mineral potential of the Kasai Block base-
ment and overlying metasediments has only rarely been addressed. This
study shows that late- or post-Eburnean metasedimentary Mn carbonate
ores contain significant quantities of other metals (Co, Ni, Zn, Cu),
whereas associated shales are enriched in Cr and Cu. Manganese and a
significant part of these base metals are derived from Paleoproterozoic
ocean waters, through periodic invasion of intracratonic graben basins,
trapping ocean-ridge-related metals rather than recycling metals from
the underlying basement. The latter might however be the case for Cr
and possibly also for Cu.
The metal content of the Kisenge-Kamata graben and other similar
basins within the Kasai Block has possibly played a significant role in
feeding younger Proterozoic ore deposits located around the Congo
199
Ore Geology Reviews 96 (2018) 181–200
Craton. A likely candidate for trapping of metals derived from Archean
and Paleoproterozoic formations is the Neoproterozoic Central African
Copperbelt (CAC), located close to the east of the Kisenge-Kamata de-
posit area.
Acknowledgements
The authors are grateful to Ariel Boven for providing new un-
published geochronological constraints, to Mines d’Or de Kisenge
(MDDK) for permitting use of their airborne geomagnetic data, and to
Jacques Navez and Laurence Monin (RMCA) for performing chemical
analyses. Finally, we thank the two anonymous reviewers for their
detailed reviews that allowed us improve the overall quality of the
manuscript.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at http://dx.doi.org/10.1016/j.oregeorev.2018.04.015.
References
Anbar, A.D., Knoll, A.H., 2002. Proterozoic ocean chemistry and evolution: a bioinorganic
bridge? Science 297, 1137–1142.
Andersen, T.B., 1984. Inclusion patterns in zoned garnets from Mageroy, north Norway.
Mineral. Mag. 48, 21–26.
André, L., 1993. Age Rb-Sr protérozoïque inférieur du magmatisme continental du groupe
de la Lulua (Kasai, Zaïre): ses implications géodynamiques. Ann. Soc. Géol. Belgique
116, 1–12.
Bau, M., Dulski, P., 1996. Distribution of yttrium and rare-earth elements in the penge
and kuruman iron-formations, transvaal supergroup, South Africa. Precambr. Res. 79,
37–55.
Bekker, A., Karhu, J.A., Eriksson, K.A., Kaufman, A.J., 2003. Chemostratigraphy of pa-
leoproterozoic carbonate successions of the wyoming craton: tectonic forcing of
biogeochemical change? Precambr. Res. 120, 279–325.
Beukes, N.J., Swindell, E.P.W., Wabo, H., 2016. Manganese deposits of Africa. Episodes
39, 285–317.
Bindeman, L.N., Bekker, A., Zakharov, D.O., 2016. Oxygen isotope perspective on crustal
evolution on early Earth: a record of Precambrian shales with emphasis on
Paleoproterozoic glaciations and Great Oxidation Event. Earth Planet. Sci. Lett. 437,
101–113.
Bingen, B., Demaiffe, D., Delhal, J., 1988. Aluminous granulites of the Archean craton of
kasai (Zaire): petrology and P-T conditions. J. Petrol. 29, 899–919.
Boven, A., Liégeois, J.P., He, H., Jang, J., Jelsma, H.A., Armstrong, R.A., 2011. The
Southern Kasai shield: A Metacratonic Boundary of the Congo craton? International
Conference on Craton Formation and Destruction, Beijing, China.
Burton, K.W., 1986. Garnet-quartz intergrowths in graphitic pelites: the role of the fluid
phase. Mineral. Mag. 50, 611–620.
Cahen, L., Lepersonne, J., 1967. The Precambrian of the Congo, Rwanda and Burundi. In:
In: Rankama, K. (Ed.), The Precambrian Volume 3. Wiley Interscience, New York, pp.
143–290.
Campbell Smith, W., Bannister, F.A., Hey, M.H., Groves, A.W., 1944. Banalsite, a new
barium-feldspar from Wales. Mineral. Mag. 27, 33–47.
Candia, M.A.F., Girardi, V.A.V., 1979. Aspectos metamórficos da Formação Lafaiete em
Morro da Mina, Distrito de Lafaiete, MG. Boletim do Instituto de Geociências da
Universidade São Paulo 10, pp. 19–30.
Catling, D.C., Claire, M.W., 2005. How Earth’s atmosphere evolved to an oxic state: a
status report. Earth Planet. Sci. Lett. 237, 1–20.
Chisonga, B.C., Gutzmer, J., Beukes, N.K., Huizenga, J.M., 2012. Nature and origin of the
protolith succession to the Paleoproterozoic Serra do Navio manganese deposit,
Amapa Province, Brazil. Ore Geol. Rev. 47, 59–76.
Costa, L.M., Fernandez, O.J.C., Requelme, M.E.R., 2005. O depósito de manganês do Azul,
Carajás: estratigrafia, mineralogia, geoquímica e evolução geológica. In: Marini, O.J.,
Queiroz, E.T., Ramos, B.W. (Eds.), Caracterização de Depósitos Minerais em Distritos
Mineiros da Amazônia. ADIMB, Brasília, pp. 227–334.
Dasgupta, S., Sengupta, P., Fukuoka, M., Roy, S., 1993. Contrasting paragenesis in the
manganese silicate – carbonate rocks from Parseoni, Sausar Group, India and their
interpretation. Contrib. Miner. Petrol. 114, 533–538.
Decrée, S., Deloule, E., Ruffet, G., Dewaele, S., Mees, F., Marignac, C., Yans, J., De Putter,
Th., 2010. Geodynamic and climate controls in the formation of Mio-Pliocene world-
class oxidized cobalt and manganese ores in the Katanga Province, D.R Congo. Miner.
Depos. 45, 621–629.
Delhal, J., 1991. Situation géochronologique 1990 du Précambrien du Sud-Kasai et de
l’Ouest-Shaba. Rapport Annuel du Département de Géologie et Minéralogie du Musée
Royal de l’Afrique Centrale. Années 1989–1990, 119–125.
Delhal, J., Ledent, D., 1971. Ages U/Pb et Rb/Sr et rapports initiaux du Sr du Complexe
gabbro-noritique et charnockitique du bouclier du Kasai (République démocratique
du Congo et Angola). Annal. Soc. Géol. Belgique 94, 211–221.
Delhal, J., Liégeois, J.P., 1982. Le socle granito-gneissique du Shaba occidental (Zaïre).
T. De Putter et al.
Pétrographie et géochronologie. Annal. Soc. Géol. Bel. 106, 296–301.
Delhal, J., Ledent, D., Torquato, J.R., 1976. Nouvelles données géochronologiques re-
latives au complexe gabbro-noritique et charnockitique du bouclier du Kasaï et à son
prolongement en Angola. Annal. Soc. Géol. Belgique 99, 211–226.
Delhal, J., Deutsch, S., Denoiseux, B., 1986. A Sm/Nd isotopic study of heterogeneous
granulites from the Archean Kasai-Lomami gabbro-norite and charnockite complex
(Zaire, Africa). Chem. Geol. 57, 235–245.
Delhal, J., Deutsch, S., Snelling, N.J., 1989. Datation du complexe sédimentaire et vol-
canique de la Lulua (Protérozoïque inférieur, Kasaï, Zaïre). Musée Royal de l’Afrique
Centrale, Tervuren (Belgique). Département de Géologie et Minéralogie, Rapport
annuel 1987–88, pp. 93–99.
De Putter, Th., Bayon, G., Mees, F., Ruffet, G. & Smith, Th., 2016. Cenozoic sedimentation
history of the Congo Basin revisited. In Programme and Abstracts of the SGF
Workshop Source to Sink: A Long Term Perspective of Sediment Budgets and Sources
Characterization. Rennes, France, pp. 51–52.
De Putter, Th., Mees, F., Decrée, S., Dewaele, S., 2010. Malachite, an indicator of major
Pliocene Cu remobilization in a karstic environment (Katanga, Democratic Republic
of Congo). Ore Geol. Rev. 38, 90–100.
De Putter, Th., Ruffet, G., Yans, J., Mees, F., 2015. The age of supergene manganese
deposits in Katanga and its implication for the Neogene evolution of the African Great
Lakes Region. Ore Geol. Rev. 71, 350–362.
Doyen, L., 1973. The manganese ore deposit of Kisenge-Kamata (Western Katanga).
Mineralogical and sedimentological aspects of the primary ore. In: Amstutz, G.C.,
Bernard, A.J. (Eds.), Ores in Sediments. Springer-Verlag, Berlin, pp. 93–100.
Doyen, L., 1974. Étude métallogénique des gisements de manganèse de Kisenge-Kamata-
Kapolo, Shaba, Zaïre. Unpublished PhD thesis. Université Libre de Bruxelles, pp.
493.
EGMF team, 1998. Kisenge Feasibility Study. Directed, compiled and written by J.
Cailteux, chief of the project. Entreprise Générale Malta Forrest (EGMF) unpublished
Report DRD/KGE03-98 for Cluff Mining, 197p.
Fockema, R.A.P., Austen, A.L.S., 1956. Manganese deposits of Northern Rhodesia. In
González Reyna, J. (ed.), Symposium sobre Yacimientos de Manganeso. Tomo II.
Africa, pp. 273–291.
Green, D., Daly, M.C., 1982. Manganese mineralization in Zambia. Trans. Inst. Min.
Metall. B 91, 33–41.
Gutzmer, J., Beukes, N.J., Yeh, H.-W., 1997. Fault-controlled metasomatic alteration of
Early Proterozoic sedimentary manganese ore at Mamatwan mine, Kalahari
Manganese Field, South Africa. S. Afr. J. Geol. 100 (1), 53–71.
Hein, J.R., Bolton, B.R., 1993. Composition and origin of the Moanda manganese deposit,
Gabon. Extended abstracts book, 16th International Colloquium on African Geology
(Ezulwini), pp. 150–152.
Herbosch, A., Liégeois, J.P., Pin, C., 2016. Coticules of the Belgian type area (Stavelot-
Venn Massif): limy turbidites within the nascent Rheic oceanic basin. Earth-Sci. Rev.
159, 186–214.
Holland, H.D., Beukes, N.J., 1990. A paleoweathering profile from Griqualand West,
South Africa: evidence for a dramatic rise in atmospheric oxygen between 2.2 and 1.9
bybp. Am. J. Sci. 290A, 1–34.
Holser, W.T., 1997. Evaluation of the application of rare-earth elements to paleoceano-
graphy. Palaeogeogr. Palaeoclimatol. Palaeoecol. 132 (1–4), 309–323.
Hurlbut, C.S., 1961. Tephroite from Franklin, New Jersey. Am. Mineral. 46, 549–559.
John, T., Schenk, V., Mezger, K., Tembo, F., 2004. Timing and P-T evolution of whites-
chist metamorphism in the Lufilian arc-Zambezi belt orogen (Zambia): implications
for the assembly of Gondwana. J. Geol. 112, 71–90.
Kah, L.C., Lyons, T.W., Frank, T.D., 2004. Low marine sulphate and protracted oxyge-
nation of the Proterozoic biosphere. Nature 431, 834–838.
Kleinschrot, D., Klemd, R., Bröcker, M., Okrusch, M., Franz, L., Schmidt, K., 1994.
Protores and country rocks of the Nsuta manganese deposit (Ghana), Neues Jahrbuch
für Mineralogie. Abhandlungen 168, 67–108.
Koschinsky, A., Halbach, P., 1995. Sequential leaching of marine ferromanganese pre-
cipitates: genetic implications. Geochim. Cosmochim. Acta 59, 5113–5132.
Krissansen-Totton, J., Buick, R., Catling, D.C., 2015. A statistical analysis of the carbon
isotope record from the Archean to Phanerozoic and implications for the rise of
oxygen. Am. J. Sci. 315, 275–316.
Kuleshov, V.N., 2011. Manganese deposits: communication 2. Major epochs and phases of
manganese accumulation in the Earth’s history. Lithol. Min. Resour. 46, 546–565.
Laznicka, P., 1992. Manganese deposits in the global lithogenetic system: quantitative
approach. Ore Geol. Rev. 7, 279–356.
Ledent, D., Lay, C., Delhal, J., 1962. Premières données sur l’âge absolu des formations
anciennes du ‘socle’ du Kasai (Congo méridional). Bull. Soc. Belge Géol. 71, 223–237.
Lepersonne, J., 1974. Carte Géologique du Zaïre au 1:2,000,000 et Notice Explicative.
Département des Mines, République du Zaïre, Direction de la Géologie, pp. 66.
Li, Y.H., Schoonmaker, J.E., 2014. 9.1 Chemical Composition and Mineralogy of Marine
Sediments. second ed. Treatise on Geochemistry, Elsevier, pp. 1–32.
Liégeois, J.P., Abdelsalam, M.G., Ennih, N., Ouabadi, A., 2013. Metacraton: nature,
genesis and behavior. Gondwana Res. 23, 220–237.
Litherland, M., Malan, S.P., 1973. Manganiferous stromatolites from the Precambrian of
Botswana. J. Geol. Soc. London 29, 543–544.
Marchandise, H., 1958. Le gisement et les minerais de manganèse de Kisenge (Congo
belge). Bull. Soc. Belge Géol. 67, 187–211.
Maresch, W.V., Mottana, A., 1976. The pyroxmangite-rhodonite transformation for the
200
Ore Geology Reviews 96 (2018) 181–200
MnSiO3 composition. Contrib. Miner. Petrol. 55, 69–79.
Maynard, J.B., 2010. The chemistry of manganese ores through time: a signal of diversity
of earth-surface environments. Econ. Geol. 105, 535–552.
Mikouchi, T., Takeda, H., Miyamoto, M., Ohsumi, K., McKay, G.A., 1995. Exsolution la-
mellae of kirschsteinite in magnesium-iron olivine from an angrite meteorite. Am.
Mineral. 80, 585–592.
Mohapatra, B.K., Nayak, B., 2005. Petrology of Mn carbonate – silicate rocks from the
Gangpur Group, India. J. Asian Earth Sci. 25, 773–780.
Morelli, B., Raucq, P., 1961. Lambeaux d’une série métamorphique manganésifère entre
Mwene-Ditu et Luputa (Kasai). Bull. Séances de l’Acad. R. Sci. d’Outre-Mer 7,
908–923.
Nyame, F.K., 2008. Petrography and geochemistry of intraclastic manganese-carbonates
from the ∼2.2 Ga Nsuta deposit of Ghana: significance for manganese sedimentation
in the Paleoproterozoic of West Africa. J. Afr. Earth Sci. 50, 133–147.
Nyame, F.K., Beukes, N.J., 2006. The genetic significance of carbon and oxygen isotopic
variations in Mn-bearing carbonates from the Palaeo-Proterozoic (∼2.2 Ga) Nsuta
deposit in the Birimian of Ghana. Carbonates Evaporites 21, 21–32.
Nyame, F.K., Beukes, N.J., Kase, K., Yamamoto, M., 2003. Compositional variations in
manganese carbonate micronodules from the Lower Proterozoic Nsuta deposit,
Ghana: product of authigenic precipitation or post-formational diagenesis ? Sed.
Geol. 154, 159–175.
Pasteels, 1971. Age du granite de la Lunge (près de Kamina, Katanga). Rapport Annuel du
Département de Géologie et Minéralogie du Musée Royal de l’Afrique Centrale,
Année 1970, 41.
Planavsky, N., Bekker, A., Rouxel, O.J., Kamber, B., Hofmann, A., Knudsen, A., Lyons,
T.W., 2010. Rare Earth Element and yttrium compositions of Archean and
Paleoproterozoic Fe formations revisited: new perspectives on the significance and
mechanisms of deposition. Geochim. Cosmochim. Acta 74, 6387–6405.
Planavsky, N., Rouxel, O.J., Bekker, A., Shapiro, R., Fralick, P., Knudsen, A., 2009. Iron-
oxidizing microbial ecosystems thrived in late Paleoproterozoic redox-stratified
oceans. Earth Planet. Sci. Lett. 286, 230–242.
Polinard, E., 1946. Le minerai de manganèse à polianite et hollandite de la Haute Lulua.
Institut Royal Colonial Belge, Section des Sciences Naturelles et Médicales, Mémoires,
In-8°, 16 (1), 41 pp.
Prosser, S., 1993. Rift-related linked depositional systems and their seismic expression. In:
Williams, G.D., Dobb, A. (Eds.), Tectonics and seismic sequence stratigraphy.
Geological Society Special Publication, pp. 35–66.
Rice, A.H.N., Mitchell, J.I., 1991. Porphyroblast textural sector-zoning and matrix dis-
placement. Mineral. Mag. 55, 379–396.
Rice, A.H.N., Habler, G., Carrupt, E., Cotza, G., Wiesmayr, G., Schuster, R., Sölva, H.,
Thöni, M., Koller, F., 2006. Textural sector-zoning in garnet: theoretical patterns and
natural examples from Alpine metamorphic rocks. Austrian J. Earth Sci. 99, 70–89.
Rogers, A.F., 1919. An interesting occurrence of manganese minerals near San Jose,
California. Am. J. Sci. 48, 443–449.
Roy, S., 2006. Sedimentary manganese metallogenesis in response to the evolution of the
Earth system. Earth Sci. Rev. 77, 273–305.
Rudnick, R.L., Gao, S., 2003. Composition of the continental crust. In: In: Rudnick, R.L.
(Ed.), Treatise on Geochemistry Volume 3. Elsevier, Amsterdam, pp. 1–64.
Schneiderhan, E.A., Gutzmer, J., Strauss, H., Mezger, K., Beukes, N.J., 2006. The che-
mostratigraphy of a Paleoproterozoic MnF-BIF succession – the Voëlwater Subgroup
of the Transvaal Supergroup in Griqualand West, South Africa. S. Afr. J. Geol. 109,
63–80.
Schuiling, H. & Grosemans, P., 1956. Les gisements de manganèse du Congo belge. In
González Reyna, J. (ed.), Symposium sobre Yacimientos de Manganeso. Tomo II.
Africa, pp. 131–142.
Sholkovitz, E.R., Landing, W.M., Lewis, B.L., 1994. Ocean particle chemistry: the frac-
tionation of rare earth elements between suspended particles and seawater. Geochim.
Cosmochim. Acta 58, 1567–1579.
Sperling, E.A., Wolock, C.J., Morgan, A.S., Gill, B.C., Kunzmann, M., Halverson, G.P.,
Macdonald, F.A., Knoll, A.H., Johnston, D.T., 2015. Statistical analysis of iron geo-
chemical data suggests limited Late Proterozoic oxygenation. Nature 523, 451–454.
Taylor, S.R., McLennan, S.M., 1985. The Continental Crust: Its Composition and
Evolution. Blackwell, Oxford, pp. 312.
Tsikos, H., Beukes, N.J., Moore, J.M., Harris, C., 2003. Deposition, diagenesis, and sec-
ondary enrichment of metals in the Paleoproterozoic Hotazel Iron Formation,
Kalahari Manganese Field, South Africa. Econ. Geol. 98, 1449–1462.
Usui, A., Nishi, K., Sato, H., Nakasato, Y., Thornton, B., Kashiwabara, T., Tokumaru, A.,
Sakaguchi, A., Yamaoka, K., Kato, S., Nitahara, S., Suzuki, K., Iijima, K., Urabe, T.,
2017. Continuous growth of hydrogenetic ferromanganese crusts since 17 Myr ago on
Takuyo-Daigo Seamount, NWPacific, at water depths of 800–5500 m. Ore Geol. Rev.
87, 71–87.
Varentsov, I.M., 1996. Manganese Ores of Supergene Zone: Geochemistry of Formation.
Kluwer Academic Publishers, Dordrecht, pp. 342.
Yeh, H.-W., Hein, J.R., Bolton, B.R., 1995. Origin of the Nsuta manganese carbonate
proto-ore, Ghana: carbon- and oxygen-isotope evidence. J. Geol. Soc. China 38 (4),
397–409.
Xiao, J.F., He, J.Y., Yang, H.Y., Wu, C., 2017. Comparison between Datangpo-type
manganese ores and modern marine ferromanganese oxyhydroxide precipitates
based on rare earth elements. Ore Geol. Rev. 89, 290–308.