Accepted Manuscript

Use and misuse of Mg- and Mn- rich ilmenite in diamond

exploration: A petrographic and trace element approach

Montgarri Castillo-Oliver, Joan Carles Melgarejo, Salvador Galí,

Vladimir Pervov, Antonio Olimpio Gonçalves, William L. Griffin,
Norman J. Pearson, Suzanne Y. O'Reilly

PII: S0024-4937(17)30333-X
DOI: doi:10.1016/j.lithos.2017.09.021
Reference: LITHOS 4427
To appear in:
Received date: 21 February 2017
Accepted date: 19 September 2017

Please cite this article as: Montgarri Castillo-Oliver, Joan Carles Melgarejo, Salvador Galí,
Vladimir Pervov, Antonio Olimpio Gonçalves, William L. Griffin, Norman J. Pearson,
Suzanne Y. O'Reilly , Use and misuse of Mg- and Mn- rich ilmenite in diamond
exploration: A petrographic and trace element approach. The address for the
corresponding author was captured as affiliation for all authors. Please check if
appropriate. Lithos(2017), doi:10.1016/j.lithos.2017.09.021

This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.
 ACCEPTED MANUSCRIPT

USE AND MISUSE OF Mg- AND Mn- RICH ILMENITE IN

DIAMOND EXPLORATION: A PETROGRAPHIC AND TRACE

ELEMENT APPROACH

PT
Montgarri Castillo-Oliver* [a,b], Joan Carles Melgarejo [a], Salvador Galí [a],

RI
Vladimir Pervov [c], Antonio Olimpio Gonçalves [d], William L. Griffin [b],

SC
Norman J. Pearson [b] and Suzanne Y. O’Reilly [b]
NU
[a] Departament de Mineralogia, Petrologia i Geologia Aplicada, Universitat

de Barcelona, Martí i Franquès s/n, 08028 Barcelona, Catalonia, Spain

MA

[b] ARC Centre of Excellence for Core to Crust Fluid Systems and GEMOC,

Department of Earth and Planetary Sciences, Macquarie University, NSW 2019,

E D

Australia
PT

[c] Sociedade Mineira de Catoca, Catoca, Lunda Sul, Angola

CE

[d] Departamento de Geologia, Universidade Agostinho Neto de Luanda,

AC

Avenida 4 de fevereiro 71, Luanda, Angola.

Abstract
 ACCEPTED MANUSCRIPT

Magnesian ilmenite is a common kimberlite indicator mineral, although its

use in diamond exploration is still controversial. Complex crystallisation and

replacement processes have been invoked to explain the wide compositional and

textural ranges of ilmenite found in kimberlites. This work aims to shed light on

these processes, as well as their implications for diamond exploration.

PT
Petrographic studies were combined for the first time with both major- and

RI
trace-element analyses to characterise the ilmenite populations found in

SC
xenoliths and xenocrysts in two Angolan kimberlites (Congo-Kasai craton).
NU
A multi-stage model describes the evolution of ilmenite in these pipes

involving: i) Crystallisation of ferric and Mg-rich ilmenite either as metasomatic

MA

phases or as megacrysts, both in crustal and metasomatised mantle domains; ii)

Kimberlite entrainment and xenolith disaggregation producing at least two

D

populations of ilmenite nodules differing in composition; iii) Interaction of both

E
PT

types with the kimberlitic magma during eruption, leading to widespread

replacement by Mg-rich ilmenite along grain boundaries and fractures. This

CE

process produced similar major-element compositions in ilmenites regardless of

AC

their primary (i.e., pre-kimberlitic) origin, although the original enrichment in

HFSE (Zr, Hf, Ta, Nb) observed in Fe3+-rich xenocrysts is preserved. Finally (iv)

formation of secondary Mn-ilmenite by interaction with a fluid of carbonatitic

affinity or by infiltration of a late hydrothermal fluid, followed in some cases by

subsolidus alteration in an oxidising environment.

 ACCEPTED MANUSCRIPT

The complexities of ilmenite genesis may lead to misinterpretation of the

diamond potential of a kimberlite during the exploration stage if textural and

trace-element information is disregarded. Secondary Mg-enrichment of ilmenite

xenocrysts is common and is unrelated to reducing conditions that could favour

diamond formation/preservation in the mantle. Similarly, Mn-rich ilmenite

PT
should be disregarded as a diamond indicator mineral, unless textural studies

RI
can prove its primary origin.

SC
NU
Key words: diamond exploration, ilmenite, Angola, kimberlite, LA-ICP-MS
MA

Corresponding author: montgarri.castillo-oliver@mq.edu.au

E D

1. INTRODUCTION
PT

Ilmenite is a very common phase in kimberlites and related rocks and

CE

consequently it is one of the main kimberlite indicator minerals (KIM) used to

AC

locate new targets (e.g., Mitchell, 1986; Haggerty, 1991; Wyatt et al., 2004).

Because of its abundance in kimberlites, ilmenite has been investigated for use in

diamond exploration (Gurney and Moore, 1993; Griffin and Ryan, 1995). Despite

the successful discovery of some diamond mines (e.g., Ontario kimberlite,

Canada), its use as diamond indicator mineral (DIM) has been debated. The

major-element chemistry of ilmenite as commonly used in mining exploration

 ACCEPTED MANUSCRIPT

(Gurney and Zweistra, 1995; Wyatt et al., 2004) has proven inadequate to

correctly identify the diamond potential of a prospected kimberlite (Schulze et

al., 1995; Robles-Cruz et al., 2009). Recent improvements in in situ techniques

such as LA-ICP-MS have encouraged the use of trace-element chemistry in DIM as

a tool for assessing the diamond grade of the kimberlites, including ilmenite

PT
(Carmody et al., 2014). However, to the authors’ knowledge, the full potential of

RI
this technique has not been applied to kimberlitic ilmenites, and no systematic in

SC
situ study of the trace-element composition of the different textural types of

ilmenite has been published so far.

NU
Such an approach could provide key information about the genesis of
MA

ilmenite in kimberlites, as the complex processes involved can only be

understood by a careful study of both textures and mineral compositions (e.g.,

D

Pasteris, 1980; Robles-Cruz et al., 2009). These processes are often disregarded
E
PT

during the exploration stage, which may result in misleading information about

the diamond grade of the prospected kimberlite.

CE

Therefore, in this study in situ trace-element geochemistry is used for the

AC

first time to identify the processes responsible for the formation of the different

textural populations of ilmenite found in the Cat115 and Tchiuzo pipes (NE

Angola). Additionally, the relationships between ilmenite and diamond are

evaluated, as well as the use of trace elements as tools for diamond exploration.

Although Angola is the fourth largest diamond producer in Africa (Faure, 2010),
 ACCEPTED MANUSCRIPT

very few studies have been published on Angolan kimberlites or their ilmenites

(Boyd and Danchin 1980; Robles-Cruz et al., 2009; Ashchepkov et al., 2012;

Jelsma et al., 2013; Nikitina et al., 2014).

PT
2. GEOLOGICAL SETTING

RI
Angola is located on the SW edge of the Congo-Kasai craton, which was

SC
assembled and stabilised in the Mesoarchean (Pereira et al., 2003). Subsequently,

three main orogenic cycles defined the current geology and structure in the
NU
country: the Eburnean (2-2.2 Ga), the Kibaran (1.4-1 Ga) and the Pan-African
MA

(590 Ma) (Pereira et al., 2003). The subsequent opening of the Atlantic Ocean

during the break-up of Gondwana (~125 Ma; Moore et al., 2008) coincides with a
D

major episode of kimberlite eruption in NE Angola and other African kimberlites

E

(Jelsma et al., 2004; Robles-Cruz et al., 2012; Castillo-oliver et al., 2016). Thick
PT

sedimentary sequences were deposited in the basins created by this extension

CE

(Pereira et al., 2003), and currently cover most of the kimberlites in the country.
AC

Samples of kimberlites, as well as ilmenite xenocrysts, were collected both

from open pits and drill cores of two kimberlite pipes in the Lunda Sul province.

This province, together with Lunda Norte province, is located in the north-

eastern part of Angola and dominates most of the diamond production of the

country (White et al., 1995). As seen in fig. 1, the pipes intruded close to the

Catoca kimberlite: the Cat115 pipe (9°22'40.3"S, 20°17'46.6"E – 7.5 km north of

 ACCEPTED MANUSCRIPT

Catoca) and the Tchiuzo pipe (20 km north of Catoca). The Catoca kimberlite is

currently the sixth diamond producer worldwide. The diamond grade of the

pyroclastic kimberlites of the Cat115 and Tchiuzo pipes are 39% and 46%,

respectively, of the grade of the Catoca pyroclastic kimberlite, which classifies

them as kimberlites of moderate diamond grade. Following the textural

PT
classification established by Cas et al. (2008), most of the samples from the

RI
Catoca pipe were described as pyroclastic (tuffisitic), serpentine-altered, poorly

SC
sorted, crystal-rich, matrix-supported olivine kimberlite. Detailed

characterisation of the mineral phases present in the kimberlite groundmass can

NU
be found in Kotel’nikov et al. (2005) and Robles-Cruz et al. (2009). Samples from
MA

the Cat115 pipe are classified as pyroclastic, fragmental, serpentine-altered,

poorly sorted, moderately crystal-rich, matrix-supported olivine kimberlite.

D

Samples from Tchiuzo are highly serpentinised, but mainly can be described as
E

volcanoclastic, moderately sorted and moderately crystal-rich, matrix-supported

PT

olivine kimberlite. Petrographic description of these two latter kimberlites and

CE

Sr-Nd isotope analyses of perovskite of Cat115 kimberlite enable them to be

AC

classified as group I kimberlites (Castillo-Oliver et al., 2016).

Like most of the Angolan kimberlites and carbonatites, these two intrusions

are found within the Lucapa corridor, a large lineament that crosses the country

from NE to SW (Reis, 1972). The Tchiuzo kimberlite has been dated by U-Pb on

zircon as 121.2±1.8 Ma (Robles-Cruz et al., 2012). The Cat115 kimberlite was

 ACCEPTED MANUSCRIPT

dated by U-Pb on groundmass perovskite (Castillo-Oliver et al., 2016), returning

a lower Cretaceous age (133± 10 Ma). This is consistent with an emplacement

triggered by the breakup of Gondwana, which caused the reactivation of the NE-

SW translithospheric faults that define the Lucapa corridor (Turcotte and

Oxburgh, 1973; Jelsma et al., 2009; Castillo-Oliver et al., 2016). On a local scale,

PT
kimberlite intrusion is thought to be mainly controlled by fault-bounded blocks

RI
with WSW-ENE alignment which are offset by NNW-SSE faults, giving an en-

SC
echelon pattern (Pereira et al., 2003). NU
3. ANALYTICAL METHODS
MA

Electron microscopes (SEM Leica Cambridge S.36 and ESEM Quanta 200 FEI
D

XTE 325/D8395) in the Centres Científics i Tecnològics (CCiT) of the Universitat

E

de Barcelona (UB) were used to define the textures, alteration and zoning
PT

patterns of the ilmenite xenocrysts. A focused beam with an accelerating voltage

CE

of 20 keV was used.

AC

Major- and minor-element contents were analysed with a CAMECA SX50

electron microprobe with four wavelength-dispersive spectrometers (WDS), in

the CCiT-UB. The approximate diameter of the electron beam was 1-2 µm.

Analytical conditions were 20 KeV accelerating potential, a beam current of 20

nA and a take-off angle of 40º. Counting times were 10s for the peak and 10s for

the background. Standards used for calibration are the following: Mg ( Kα, TAP,
 ACCEPTED MANUSCRIPT

periclase), Al (Kα, TAP, corundum), Si (Kα, TAP, diopside), Ca (Kα, PET,

wollastonite), Ti (Kα, PET, rutile), V (Kα, LiF, V0), Cr (Kα, PET, Cr2O3), Mn (Kα,

LiF, rhodonite), Fe (Kα, LiF, Fe2O3), Nb (Lα, PET, Nb0), Zr (Lα, PET, ZrO2), Zn (Kα,

LiF, sphalerite), Ni (Kα, LiF, Ni0). The raw data were corrected using the Pouchou

and Pichoir (1984) reduction and the ratio Fe2+/Fe3+ was calculated by

PT
stoichiometry. After the charge-balance calculation, totals between 99 and 101%

RI
were accepted.

SC
Trace-element compositions of the ilmenites were obtained using a New
NU
Wave Research UP-213 laser ablation microsampling system attached to a

Thermo Fisher (XSeries-II) quadrupole ICP-MS at the IBERCRON (Laboratorio de

MA

Geocronología y Geoquímica) of UPV/EHU (Universidad del País Vasco-Euskal

Herriko Unibertsitatea Euskal Herriko Unibertsitatea). It has already been shown

D

that NIST glasses can be used to determine the trace-element composition of

E
PT

ilmenites with better than 10% precision (e.g., Donohue et al., 2012; Carmody et

al., 2014). However, to further confirm the adequacy of the 610 and 612 NIST
CE

glasses as standards, a large and compositionally homogeneous ilmenite

AC

megacryst of kimberlitic origin was cut and a piece was analysed by dissolution

using the Q-ICP-MS. Subsequently, the trace-element composition of the grain

was analysed by ablation of the megacryst calibrating the instrument through the

610 and 612 NIST standards. The results obtained were consistent, and as a
 ACCEPTED MANUSCRIPT

consequence all analyses of the unknown samples were performed in situ using

the NIST glasses as standards.

Fourteen trace elements (Sc, V, Cr, Mn, Co, Ni, Zn, Ga, Zr, Nb, Ta, Hf, W, U)

were analysed in ilmenite grains both on thick polished sections and on polished

PT
epoxy mounts containing individual ilmenite grains. The counting time for each

spot was 90s (30s background, 60s signal) and each analysis was normalised

RI
using the Ti values obtained by the electron microprobe. Samples were analysed

SC
in runs of 14 analyses comprising 10 analyses of unknowns bracketed by two
NU
analyses of the standards at the beginning and the end of each run. Additionally,

two analysis of the ilmenite nodule were included in each run to check the
MA

correct operation of the instrument. Typical detection limits are 6 ppm for Mn

and V, from 1 to 4 ppm for Sc, Cr, Ni and Zn, from 100 ppb to 1 ppm for Co, Ga, Zr,
D

Nb and Ta and from 10 to 100 ppb for Hf, W and U. The typical relative precision
E
PT

and accuracy for a laser microprobe analysis range from 1 to 10%. The nominal

spot size was 55 µm, although in the smaller grains it was reduced up to 10 µm.
CE

The repetition rate and the energy density of the laser beam was 10Hz and 65-
AC

75%, respectively. The data were processed using the Iolite 2.15 software (Paton

et al., 2011). The error associated with each element was also calculated using

this program and is expressed in 2σ (%).

4. ILMENITE PETROGRAPHY
 ACCEPTED MANUSCRIPT

Different ilmenite populations were distinguished, including those found in a

variety of mantle xenoliths sampled by the Angolan kimberlites studied here.

However, the main mode of occurrence of ilmenite in these kimberlites is as

individual xenocrysts within the kimberlitic matrix. Significant textural and

compositional differences were observed among the ilmenite xenocrysts of a

PT
single kimberlite sample. Likewise, several generations of secondary ilmenite,

RI
replacing primary ilmenite grains and/or other titanium oxides, were also

SC
identified. NU
4.1. Ilmenite in mantle xenoliths
MA

Metasomatic ilmenite was found in mantle xenoliths, in the interstitial space

D

between relatively fresh pyroxene and serpentinised olivine grains (fig. 2a).
E

Similar metasomatic ilmenite is commonly found in xenoliths recovered from

PT

kimberlites worldwide, including MARID and PIC rocks (e.g., Dawson and Smith,
CE

1977; Grégoire et al., 2002). This ilmenite can be polycrystalline, which has been

interpreted as the result of stress in the original xenolith. In some of these

AC

xenoliths, ilmenite may show a 5µm zone of Mn enrichment on the boundaries

between ilmenite and the clinopyroxene.

Fe3+-rich ilmenite also occurs as small (0.2-1 mm) rounded or subrounded

grains in a clinopyroxene-olivine peridotite. This ilmenite crystallised together

with other metasomatic minerals such as kaersutite and apatite (fig. 2b), which
 ACCEPTED MANUSCRIPT

replace the primary phases. This xenolith shows significant carbonatization,

represented by the crystallisation of different generations of calcite. Late

hydrothermal processes produced the serpentinisation of olivine and veinlets of

secondary sulphates (barite and celestite). These ilmenite grains can contain

exsolution lamellae of Ti-rich hematite and are partially altered to post-

PT
emplacement minerals such as titanite.

RI
SC
Anhedral ilmenite grains also can occur in garnet pyroxenite xenoliths,
NU
typically surrounded by garnet (fig. 2c). However, in contrast to the

intergranular ilmenite veinlets, this type of ilmenite may contain exsolved

MA

lamellae of hematite. The clinopyroxene in these garnet pyroxenites also contains

very thin ilmenite lamellae (<5 µm) oriented parallel to its (010) plane.
E D

However, no further studies were performed on these ilmenite lamellae.

PT
CE

4.2. Ilmenite xenocrysts

AC

i. Nodular ilmenite

This group is composed of rounded ilmenite macrocrysts, with diameters

between 0.1 and 2 cm. Nodular ilmenite is often polycrystalline and the grains

typically show curved boundaries and triple junctions (fig. 2d) generated by

recrystallisation.
 ACCEPTED MANUSCRIPT

Some ilmenite nodules from the studied kimberlites have a typical Mg-

enrichment along the fractures and near the grain boundaries of the smaller

grains. This replacement by more magnesian ilmenite was also observed in

nodular ilmenite from the Catoca kimberlite (Robles-Cruz et al., 2009). Locally

this Mg enrichment is followed by a late increase in Mn, which was also found in

PT
Catoca (Robles-Cruz et al., 2009). Nodular ilmenite in the Cat115 and Tchiuzo

RI
kimberlites is usually not altered or replaced by any other secondary phase, such

SC
as the titanite, ulvöspinel or magnetite identified in nearby intrusions such as

Lucapa (Castillo-Oliver et al., 2016).

NU
MA

ii. Ferrian ilmenite xenocrysts with hematite exsolutions

D

This type of ilmenite is common in the kimberlites studied here. The grain
E

size of the xenocrysts ranges from 50 to 800 microns. The ilmenite hosts lamellae
PT

of hematite that have been partially or totally altered (fig. 2e-f). With some
CE

exceptions, these lamellae are very thin (<5 microns wide) and occupy most of

the grain. However, in some cases the boundaries of the grains are lamellae-free.
AC

As observed in other ilmenite worldwide (Haggerty, 1991b), the hematite

lamellae can be partially dissolved. Most of the grains are replaced by

symplectitic intergrowths of Mg-rich ilmenite, perovskite, titanite and ulvöspinel

along the grain boundaries (see next section).

 ACCEPTED MANUSCRIPT

4.3. Secondary ilmenite

i. Symplectitic ilmenite

This ilmenite typically replaces ferric ilmenite with hematite exsolution,

resulting in a symplectitic texture similar to that defined in the Catoca kimberlite

PT
(Robles-Cruz et al., 2009) (fig.2e). In this work, however, symplectitic ilmenite is

restricted to the grain boundaries of primary ilmenite xenocrysts, with only a few

RI
exceptions. This replacement is mainly driven by crystallographic discontinuities,

SC
exsolution planes, and subgrain or grain boundaries. Symplectitic ilmenite shows
NU
significant compositional variations and is enriched in Mg, which results in

darker shades in the BSE images than for the other ilmenite groups. This
MA

replacement process is often accompanied by the crystallisation of some typical

kimberlitic phases, such as perovskite, as well as ulvöspinel and titanium oxides.

E D

Among them, rutile is the predominant Ti oxide replacing ilmenite, as previously

PT

recorded in many other kimberlite samples worldwide (Pasteris, 1980; Mitchell,

CE

1986; Haggerty, 1991b; Tappe et al., 2014). Alteration of this assemblage

resulted in the formation of late titanite and magnetite.

AC

ii. Tabular ilmenite

This type of ilmenite has been recognised in both the Catoca (Robles-Cruz et

al., 2009) and Tchiuzo pipes. It forms small (10-50 µm) euhedral plates in the
 ACCEPTED MANUSCRIPT

kimberlite matrix (fig. 2g). This ilmenite is significantly enriched in Mn. It is

compositionally homogeneous and shows no alteration to any other phase.

iii. Secondary Mn-rich ilmenite

PT
Mn-rich ilmenite replaces earlier Ti-Fe phases, mainly ilmenite, rutile and

RI
ulvöspinel (fig.2g-h). This type of ilmenite occurs only in the Tchiuzo kimberlite,

SC
although similar compositions were also found at the boundaries of some

ilmenite xenocrysts in the Catoca kimberlite (Robles-Cruz et al., 2009). However,

NU
in the Tchiuzo kimberlite, the replacement is much more important than in
MA

Catoca, and in some grains it almost completely replaces the original xenocryst.
D

5. MAJOR-ELEMENT COMPOSITION
E
PT

Representative major-element compositions of each ilmenite type are

presented in table 1. Although more than 312 analyses were performed, for the
CE

clarity of the graphs and the following interpretation, only a selection (180
AC

points) is included in the figures. In order to ensure a comprehensive study, only

the duplicate analyses of compositionally homogeneous grains were excluded

from the graphs.

In the MgO-TiO2 diagram (fig. 3a) most of the grains plot within the

kimberlitic field defined by Wyatt et al. (2004), which is consistent with their
 ACCEPTED MANUSCRIPT

kimberlitic origin. The exceptions include the Mn-rich ilmenites and those found

in garnet pyroxenites and xenoliths with the apatite-kaersutite-ilmenite

assemblage. In the ilmenite-geikielite-pyrophanite-hematite end-member

ternary diagrams (fig. 3c), these ilmenites show compositions like those of the

xenocrysts described from the Catoca diamondiferous kimberlite (Robles-Cruz et

PT
al., 2009). Negative correlations of both Ti and Mg with Fe3+ are observed (fig.4a-

RI
b). Symplectitic ilmenite is usually slightly more enriched in Mg. The major-

SC
element composition of polycrystalline ilmenite is identical to the rest of the

grain or nodule (not shown), which indicates that the process that produced
NU
recrystallization was mechanical and did not involve any additional fluid.
MA

However, as already observed in the SEM images, a late enrichment in Mg occurs

in some of the smaller grains along grain boundaries and fractures.

D

Based on these observations, four main compositional groups can be

E
PT

distinguished depending on their Fe3+, Mg, Cr and Mn contents.

CE

5.1. Ilmenite sensu strictu

AC

This group is represented exclusively by the ilmenite found in the

amphibole-ilmenite-apatite metasomatic assemblage in peridotites. It has very

low abundances of Mg, Mn and Fe3+ (0.4-2.4 wt% MgO; < 1 wt% MnO; < 20 wt%

Fe2O3) and is the group with composition closest to ilmenite sensu strictu

(Ilm=0.73-0.82).
 ACCEPTED MANUSCRIPT

5.2. Fe3+-rich ilmenite

Ferric ilmenite (“hemoilmenite”) compositions are observed in the grains

having hematite exsolution, and to a significantly lesser extent, in exsolution-free

PT
ilmenite nodules. Their Mg contents are variable (2.4-9.3 wt% MgO), but usually

low (average 3.1 wt%). Cr concentrations are commonly below 1.5 wt% Cr 2O3

RI
(fig. 3b). Mn contents rarely exceed 0.25 wt% MnO. In contrast, the Fe 3+ content

SC
of these ilmenites is significantly higher (6.9-45.7 wt% Fe2O3, average 39 wt%)
NU
than in the other populations.
MA

5.3. Mg-rich ilmenite

D

Intergranular ilmenite found in xenoliths and most of the nodules of ilmenite

E
PT

used in this study can be defined as Mg-rich ilmenites (7.6-17.0 wt% MgO,

average 12.5 wt%). Ilmenites belonging to this compositional group are also
CE

often characterised by higher Cr contents (1-6.15 wt% Cr2O3) and lower Fe2O3
AC

abundances (2.0-20.7 wt% Fe2O3, average 11.4 wt%). The late Mg enrichment

along grain boundaries and fractures found in some of the nodules, as well as the

symplectitic ilmenites, typically produces the highest Mg values. Manganese

contents in these textural groups are also slightly higher.

 ACCEPTED MANUSCRIPT

5.4. Mn-rich ilmenite

This group is composed of the tabular ilmenite and the late Mn-rich ilmenites

that replace earlier ilmenite xenocrysts and other Ti phases. Their manganese

contents are very high compared to the other ilmenites (up to 40.8 wt% MnO),

PT
especially in the secondary ilmenite, which is almost pure pyrophanite (Pph≈88)

(fig 3c and fig.4c). They are also typically poor in Mg, Cr and Nb (<0.4 wt% MgO,

RI
<0.1 wt% Cr2O3 and <0.3 wt% Nb2O5). In the MgO-TiO2 diagram they plot in the

SC
“non-kimberlitic” field (fig. 3a), despite being of a kimberlitic origin. This
NU
discrepancy was partially addressed by Wyatt et al. (2004), who circumvented

this issue by stating that usually most Mn-rich groundmass ilmenite would be
MA

filtered out from the heavy concentrates due to its small size, leaving the MgO-

TiO2 criterion still valid for diamond prospecting purposes. Even if a similar
D

approach could be confidently taken for tabular ilmenite in this study (<50 µm),
E
PT

some reasonable doubts can be set for the effect of secondary Mn-rich ilmenite

that pervasively replaces large, previous Ti-rich phases. The role of ilmenite with
CE

such compositions in diamond exploration is further discussed in section 7.4.

AC

Representative compositional profiles were done accross several ilmenite

nodules. Mg-rich ilmenite nodules lack zoning in terms of major-element

composition, only becoming enriched in Mg very close to grain boundaries or

fractures. In contrast, symplectitic ilmenite replacing nodular ferric ilmenite is

 ACCEPTED MANUSCRIPT

typically zoned. As shown in fig. 5, the element profiles and maps for symplectitic

ilmenite reveal that the main variations are in their Mg and Cr contents.

6. TRACE-ELEMENT COMPOSITION

PT
The trace-element compositions of the analysed ilmenites are compared with

RI
data from other kimberlitic ilmenite worldwide in figure 6. Representative

SC
analyses of the trace-element composition of each ilmenite group have been

included in table 2. Because of the fineness of the ilmenite-hematite exsolutions,

NU
it was not possible to ablate only ilmenite without analysing hematite lamellae at
MA

the same time. Therefore, in order to avoid data representing a mixture of both

phases, trace-element analysis of this textural group was only performed in the
D

few grains with exsolution-free areas.

E

Data from the Angolan kimberlites studied here are consistent with the few
PT

available trace-element analyses of ilmenite nodules from kimberlites worldwide

CE

(Moore et al., 1992; Griffin et al., 1997; Kostrovitsky et al., 2004; Kaminsky and
AC

Belousova, 2009). Ilmenite grains from the Angolan kimberlites show relatively

high Nb contents (310-3220 ppm) similar to those found in other kimberlitic

ilmenites worldwide, as well as a significant enrichment in the other HFSE (45-

410 Ta ppm, 10-2430 Zr and up to 100 Hf ppm). This composition clearly differs

from those of ilmenites derived from other geological settings, which usually

have very low contents of these elements (Jang and Naslund, 2003; Dare et al.,
 ACCEPTED MANUSCRIPT

2012; Kryvdik, 2014). However, it should be noted that ilmenite found in

ultramafic lamprophyres is also commonly enriched in HFSE (Tappe et al., 2004,

2006). Similarly, the Nb and Ta compositional range of carbonatitic ilmenite

extends towards even higher values (Kryvdik, 2014).

PT
Most of the ilmenite populations show a positive correlation between Nb and

the other HFSE (fig. 6). This trend coincides with that found in ilmenites from

RI
carbonatites and ultramafic lamprophyres (Tappe et al. 2004, 2006; Kryvdik,

SC
2014). A slight covariance is also seen between Nb and V. In contrast, there is an
NU
inverse relationship between Nb and Mg contents in the ilmenites.

The ilmenite populations with the highest Mg contents (i.e. symplectitic,

MA

secondary Mg enrichment in the ilmenite nodules) typically show significantly

higher contents of HFSE, Cr and V (table 2) at a given Nb content, and thus

E D

diverge from the general trends described previously. In contrast, ilmenite in

PT

garnet pyroxenites has very low contents of HFSE, Cr and Sc, but
CE

characteristically high Zn and V contents (>700 ppm Zn, > 3000 ppm V), as well

as high Nb/Ta (>17). These values are closer to those from non-kimberlitic
AC

ilmenites, where Nb/Ta is typically >10 (Jang and Naslund, 2003; Dare et al.,

2012; Kryvdik, 2014). In contrast, Nb/Ta ratios in ilmenites from kimberlites are

usually lower and fall within a narrow range (5-10).

7. DISCUSSION
 ACCEPTED MANUSCRIPT

7.1. Ilmenite evolution

The occurrence and composition of ilmenite in the Angolan kimberlites can

only be understood by integrating textural characterisation and chemical

analysis of the ilmenite populations. Textural evidence indicates a pre-kimberlite

PT
origin for most of the ilmenite nodules. Additionally, the large compositional

range shown by the ilmenite xenocrysts, as well as the textural relationships

RI
between the different populations, are also indicators of disequilibrium between

SC
the xenocrysts and the host kimberlite. Therefore, none of the ilmenite
NU
populations are considered to be kimberlitic liquidus phases.

A multi-stage model is proposed in fig. 7 to explain ilmenite evolution in the

MA

Angolan kimberlites. The textures and compositions observed in the ilmenite

xenocrysts have resulted from a combination of complex processes involving

E D

crystallisation of ilmenite at both mantle and crustal depths, entrainment, and

PT

interaction with the kimberlitic magma(s) and late subsolidus replacement by

CE

Mn-rich ilmenite.

i. Primary crystallisation of Fe3+ and Mg-rich ilmenite

AC

Although most ilmenite grains found in the Angolan kimberlites are small

(<1 cm), some of them are large enough to be considered part of the megacryst

suite (Harte and Gurney, 1980). Cr-rich and Cr-poor (0.05-6.15 wt% Cr2O3)

ilmenite coexist in both kimberlites, with the higher values reported for the

Tchiuzo kimberlite. Even if ilmenite usually belongs to the low-Cr megacryst

 ACCEPTED MANUSCRIPT

suite, coexistence of Cr-rich and Cr-poor ilmenite within the same kimberlite has

also been reported in few other localities worldwide(Eggler et al., 1979; Shee and

Gurney, 1979; Kopylova et al., 2009; Pivin et al., 2009).

The origin of megacrysts in kimberlites is still a matter of debate (i.e.,

PT
cognate vs not-cognate), but it is commonly agreed that they crystallised in the

mantle shortly prior to –or during- kimberlite eruption (e.g., Nixon and Boyd,

RI
1973; Schulze, 1984; Moore and Belousova, 2005). However, the nature of their

SC
parental magma/fluid is more controversial and different origins have been
NU
proposed, including: i. a proto-kimberlitic magma (e.g., Griffin et al., 1989;

Golubkova et al., 2013), ii. mantle metasomatic fluids with a genetic link with the
MA

proto-kimberlite magma (e.g., Kopylova et al., 2009); iii. the kimberlitic magma

itself (Harte and Gurney, 1981; Schulze, 1987; Shee and Gurney, 1979; Moore
D

and Lock, 2001; Nowell et al., 2004; Moore and Belousova, 2005; Tappe et al.,
E
PT

2011), and iv. other melts such as basanites, picrites, meimechites and MORB-

derived melts (Harte, 1983; Jones, 1987; Moore et al., 1992; Griffin et al., 1997).
CE

Recent Hf and Lu-Hf studies reported similar isotopic signatures and ages for
AC

the magnesian ilmenite megacrysts and the host kimberlite, which supports a

genetic link between them (Nowell et al., 2004; Tappe et al., 2011). However, the

occurrence of abundant and compositionally diverse ilmenite in the mantle

xenoliths sampled by the kimberlites may also encourage seeking for alternative

origins for the ilmenite nodules in Angola. Here we propose that disaggregation
 ACCEPTED MANUSCRIPT

of mantle rocks that had been previously metasomatised by Ti- and Fe-rich fluids

may have also contributed to the current ilmenite populations found in the

kimberlites. A link between the magmas forming megacryst suite and those that

introduced clinopyroxene, garnet and ilmenite in cratonic xenoliths has already

been proposed (e.g., Kopylova et al., 2009). In order to test this hypothesis, the

PT
composition of the ilmenites found in the xenoliths sampled by the kimberlites is

RI
compared next with that of the magnesian and ferric ilmenite nodules.

SC
Ilmenite in garnet pyroxenite xenoliths shows compositions that plot outside
NU
of the kimberlite field defined by Wyatt and coauthors (2004) and commonly

displays evidence of hematite exsolution. It is extremely poor in elements typical

MA

of kimberlites as Mg, Cr, but it is enriched in V and Zn. Other metasomatic

products in mantle xenoliths comprise the following mineral associations: a) Mg-

D

poor ilmenite with kaersutite and apatite; and b) ilmenite (intermediate Mg,
E
PT

slightly enriched in HFSE) with Cr-rich diopside. Similar ilmenite associations

have been found in many metasomatised xenoliths in Angolan kimberlites (i.e.

CE

Robles-Cruz et al., 2009) and in other kimberlites worldwide (e.g., Dawson and
AC

Smith, 1977; Dawson, 2002). They show a wide range of compositions and

mineral assemblages, suggesting a diversity of metasomatic processes reflecting

the action of different types of magmas and/or conditions of crystallisation.

Major and trace elements of magnesian ilmenite nodules found in the

kimberlite matrix are comparable to those of the intergranular ilmenite of the

 ACCEPTED MANUSCRIPT

xenoliths. Moreover, some of these nodules show a characteristic enrichment in

Mg close to fractures and cracks, which already suggests disequilibrium with the

host kimberlite. Therefore, it seems reasonable to suggest that at least part of

these nodules could have derived from the disaggregation of these xenoliths

during the ascent of the kimberlite magma. This observation is relevant, since it

PT
may also suggest a genetic link between the kimberlite melt and the fluids that

RI
metasomatised the mantle. This possibility was already envisaged by Dawson

SC
and Smith (1977), who proposed that MARID rocks crystallised from a

kimberlite magma before being entrapped by the host kimberlite. More recently,
NU
MARID rocks have also been considered potential source rocks for both group I
MA

and group II kimberlites (Konzett et al., 1998; Giuliani et al., 2015).

In contrast, the available data are not enough to constrain the source of the
D

Fe3+-rich ilmenite xenocrysts. After comparing their trace-element composition

E
PT

with the data on all the mantle xenoliths recovered from the kimberlite, as well

as with data from the literature, no clear source could be identified. However, the
CE

abundant hematite exsolution observed in most of these xenocrysts suggests a

AC

crustal origin. The formation of oriented hematite lamellae within an originally

homogeneous ilmenite nodule has been reported in many kimberlites worldwide

and is commonly believed to be caused by a decrease of temperature or increase

in oxygen fugacity under subsolidus conditions (e.g., Mitchell, 1986; Haggerty,

1991a).
 ACCEPTED MANUSCRIPT

ii. Formation of Mg-rich ilmenite: interaction with the kimberlitic magma

Although most of the ilmenite nodules originally had relatively high Mg

contents (~8.8 wt% MgO), they were often further enriched in this element (>11

wt% MgO) along grain boundaries and fractures. This increase in Mg is very

PT
common in kimberlitic ilmenites (Mitchell, 1986), and it has been attributed to

an interaction with the kimberlitic magma (e.g., Pasteris, 1980), with Mg-rich

RI
metasomatic fluids (Boctor and Boyd, 1981) or with a late Mg-rich fluid not

SC
necessarily related to the kimberlitic magma as discussed in Robles-Cruz et al.
NU
(2009). In situ analysis shows that the Mg enrichment is commonly accompanied

by a slight increase in the HFSE (Ta, Zr and Hf) contents of the original nodule
MA

(fig. 6). Consequently this enrichment is interpreted here as the result of

interaction with a carbonate-bearing kimberlite magma, which would have high

D

contents of these elements (Kamenetsky et al., 2014; Castillo-Oliver et al., 2016;

E
PT

Tappe et al., 2017).

CE

This interaction with the kimberlitic magma becomes more evident in the

ferric ilmenite xenocrysts found in the Tchiuzo and Cat115 kimberlites. In this
AC

case, symplectitic ilmenite replaces the xenocryst along grain boundaries and/or

the exsolution planes. This second generation of ilmenite is typically enriched in

Mg (>10 wt% MgO) and it has very high HFSE contents, which seem to be mostly

inherited from the original ferric ilmenite. Additionally, symplectitic ilmenite

crystallises together with characteristic kimberlite phases such as perovskite and

 ACCEPTED MANUSCRIPT

rutile (Mitchell, 1986; Tappe et al., 2014), further supporting a kimberlitic origin

for this second generation of ilmenite. The differences in the Mg content of the

secondary Mg-rich ilmenites indicate variable degrees of interaction with the

kimberlite magma. More advanced stages of this replacement process have been

observed in the nearby Catoca kimberlite (fig. 4a-c in Robles-Cruz et al., 2009).

PT
Therefore, even if the efficiency of this replacement process could not be

RI
quantified, complete replacement of ferric ilmenite by Mg-rich (symplectitic)

SC
ilmenite in the studied kimberlites cannot be dismissed. This interaction

probably took place during the ascent of the kimberlitic magma and not prior to
NU
entrainment, in which case more homogeneous compositions might be expected.
MA

iii. Crystallisation of Mn-rich ilmenite

The next stage is represented by the crystallisation of Mn-rich ilmenite,

E D

either replacing previous generations of ilmenite or as small plates of manganoan

PT

ilmenite in the kimberlitic matrix. This type of ilmenite is characteristic of the

CE

Tchiuzo kimberlite and it is significantly more enriched in manganese than the

Mn-rich ilmenites from the nearby Catoca kimberlite (Robles-Cruz et al., 2009)
AC

and other pipes of the same kimberlitic cluster (Ashchepkov et al., 2012).

Manganoan ilmenites have also been reported in other kimberlites of the same

cluster (Robles-Cruz et al., 2009) and worldwide (e.g., Wyatt, 1979; Pasteris,

1980; Tompkins and Haggerty, 1985; Chakhmouradian and Mitchell, 1999),

although they are more common in carbonatite and lamprophyre intrusions (e.g.,
 ACCEPTED MANUSCRIPT

Mitchell, 1978; Gaspar and Wyllie, 1983; Kryvdik, 2014; Tappe et al., 2017).

However, there is still a lack of consensus about their origin. The processes

suggested to have led to their formation include primary crystallisation under

superdeep conditions (Kaminsky and Belousova, 2009), crystallisation from a

late-stage fraction of the kimberlitic melt (e.g., Tompkins and Haggerty, 1985),

PT
interaction with a CO2–rich fluid or a carbonatite-like magma (Gaspar and Wyllie,

RI
1984; Robles-Cruz et al., 2009; Tappe et al., 2017) and even late hydrothermal

SC
activity related to serpentinisation (Robles-Cruz et al., 2009) or groundwater

circulation (Wyatt, 1979). In the kimberlites studied here, manganoan ilmenite is

NU
produced at a very late stage in the mineral sequence of the kimberlite, and may
MA

replace all the other ilmenite generations and previous Ti-oxides. A recent study

has shown that carbonatitic magmas could result from by fractionation of a

D

silicocarbonatitic magma will ultimately be devoid of HFSE (Tappe et al., 2017).

E

If this model applies, the near absence of HFSE in the Mn-ilmenites studied here
PT

could be explained by interaction with a late carbonatitic magma which triggered

CE

their crystallisation. This would be consistent with previous works (Gaspar and
AC

Wyllie, 1984; Robles-Cruz et al., 2012) and multi-stage processes of magma

mingling/mixing which have been previously proposed for kimberlites

(Kamenetsky et al., 2014; Castillo-Oliver et al., 2016) and lamprophyres (Tappe

et al., 2017). Alternatively, the composition and occurrence of the Mn-rich

ilmenite in Angola could also be explained by circulation of groundwater

 ACCEPTED MANUSCRIPT

circulation (Wyatt, 1979); or hydrothermal fluids related to serpentinisation as

also discussed by Robles-Cruz et al. (2009).

In any case, the development of both magnesian and manganoan ilmenites in

the Angolan kimberlites cannot be ascribed to only one process; instead, it

PT
suggests the interaction of the xenocrysts with different types of kimberlitic-

carbonatitic magmas and hydrothermal fluids. Mixing of compositionally distinct

RI
magmas within a single kimberlite has already been proposed in other Angolan

SC
kimberlites (Castillo-Oliver et al., 2016), as well as in kimberlites from other
NU
cratons (Chakhmouradian and Mitchell, 2001; Ogilvie-Harris et al., 2009).
MA

7.2. Trace-element composition as a tracer of the source of the ilmenite

D

xenocrysts
E

As discussed above, ilmenite xenocrysts of different origin may eventually

PT

acquire similar major-element compositions (i.e. symplectitic ilmenite and Mg-

CE

rich ilmenite nodules). This compositional homogenisation is suggested to result

AC

from the interaction of the xenocrysts with the kimberlitic magma during

entrainment, and this is supported by the co-crystallisation of ilmenite and

perovskite. Although in this study symplectitic ilmenite only partially replaces

existing ilmenite xenocrysts, totally replaced nodules have been found in the

Catoca kimberlite (Robles-Cruz et al., 2009). As a consequence, we suggest that

replacement of an originally Fe3+-rich ilmenite could lead to compositionally

 ACCEPTED MANUSCRIPT

homogeneous Mg-rich ilmenite xenocrysts by interaction with the kimberlitic

magma. Therefore, caution should be taken when analysing ilmenite

concentrates without consideration of their texture and/or their trace-element

composition. A high Mg content in ilmenite would not necessarily indicate the

reducing conditions within the diamond stability field commonly regarded as

PT
favourable for diamond preservation. Instead, this enrichment could be a good

RI
indicator of the degree of interaction of the ilmenite with the kimberlitic magma.

SC
Trace-element chemistry is useful for constraining the source of the ilmenite

xenocrysts by identifying whether the high Mg contents result from the

NU
interaction with the kimberlitic magma or earlier metasomatism. In this study,
MA

the characteristically high contents in HFSE in the Fe3+-rich ilmenites clearly

differ from the trace-element signatures of the other ilmenite xenocrysts found in
D

the kimberlites. This has allowed us to follow them through the stages of their
E

subsequent replacement during interaction with the kimberlite.

PT
CE

7.3. Links between diamond and ilmenite

AC

Diamond formation is commonly agreed to occur as a consequence of

complex metasomatic processes in the mantle, mainly near the base of the SCLM

(Stachel et al., 2004; Malkovets et al., 2007; Smart et al., 2017). Infiltration of

carbon-bearing fluids, enriched in Sr, Zr, Nb and LREE, is thought to be

responsible for diamond growth (Sobolev et al., 1997; Araújo et al., 2009).
 ACCEPTED MANUSCRIPT

Similarly, it is widely accepted that ilmenite is a typical metasomatic phase in the

mantle (O’Reilly and Griffin, 2013), commonly found in MARID (Dawson and

Smith, 1977) and related rocks (Grégoire et al., 2002).

Nevertheless, the link between ilmenite and diamond in kimberlites remains

PT
unclear. In contrast with other DIMs, ilmenite is a very rare inclusion in diamond

(Harris et al., 1984). However, some evidence pointing to a link between the two

RI
phases has been found, either as ilmenite inclusions in diamonds or as

SC
intergrowths with silicate phases (Meyer and Svisero, 1975; Mvuemba Ntanda et
NU
al., 1982; Jacques et al., 1989; Sobolev et al., 1997; Kaminsky et al., 2000, 2001).

The composition of the ilmenite inclusions in diamond is heterogeneous, ranging

MA

from magnesian ilmenites (9-16 wt% MgO; Sobolev et al., 1997 and references

therein) to Mg-poor (0.04-0.14 wt% MgO) and Mn-rich ilmenites (2-11 wt%
D

MnO; Kaminsky et al., 2001 and references therein). Trace-element analysis of

E
PT

ilmenites included in diamonds from the Siberian craton showed enrichment not

only in Mg and Cr, but also in Ni, and high Nb/Zr ratios (Sobolev et al., 1997).
CE

These compositional signatures have been used to argue for a metasomatic

AC

origin of both ilmenite and diamond (Sobolev et al., 1997). Similarly, coexistence

of ilmenite and lindsleyite-mathiashite (LIMA) group minerals as inclusions in

diamond has been interpreted as further evidence for a common low-T

metasomatic origin (Haggerty, 1983).

 ACCEPTED MANUSCRIPT

Although a cold geotherm SCLM in NE Angola (Ashchepkov et al., 2012;

Griffin et al., 2009; O’Reilly et al., 2009) may envisage favourable conditions for

this low-T metasomatism, the trace-element compositions of the ilmenites

studied here differ from those of the magnesian ilmenite found as diamond

inclusions in other kimberlites worldwide (fig.6). As a consequence, additional

PT
systematic petrographic studies coupled with chemical analyses are still required

RI
to define any genetic link between the metasomatism that led to ilmenite

SC
crystallisation and that responsible for diamond formation in Angola.
NU
7.4. Ilmenite as diamond indicator mineral
MA

Because of the ubiquity of ilmenite in kimberlites, much effort has been

D

expended on the characterisation of ilmenite during exploration stages. Its

E

composition has been used to constrain conditions in the mantle sampled by the
PT

kimberlitic magma and thus assess the diamond grade of targeted pipes. Gurney
CE

and Zweistra (1995) suggested that the Fe2O3 - MgO diagram could be used to

evaluate the diamond grade of a given kimberlite. Their results are based on the
AC

premise that ilmenite showing high Fe3+ contents would have crystallised in an

oxidising environment, where diamond could not be formed or preserved.

Ilmenite megacrysts from both economic and non-economic South African

kimberlites could also be consistently distinguished based on this diagram

(Griffin et al., 1997). However, analysis of ilmenite xenocrysts in high-grade

 ACCEPTED MANUSCRIPT

kimberlites worldwide has revealed that this model is not universally applicable

(Schulze and references therein, 1995; Robles-Cruz et al., 2009; Carmody et al.,

2014).

Recent improvement of in-situ techniques such as LA-ICP-MS has spurred the

PT
use of trace element chemistry as a tool for assessing the diamond grade of the

kimberlites. The Zr/Nb ratio of ilmenite recently has been proposed as an

RI
additional discriminant of the diamond grade of kimberlites (Carmody et al.,

SC
2014). These authors defined a threshold value of Zr/Nb ~ 0.37; only ilmenites
NU
recovered from high-grade kimberlites have higher Zr/Nb values. Similarly, they

observed that ilmenites from barren or very low-grade kimberlites usually have
MA

Nb/Ta < 10.

In this work, for the sake of simplicity, only the ilmenite populations that
E D

could be acquired during the exploration stage (>50 microns) have been
PT

included in the Fe2O3- MgO and Zr/Nb-Nb/Ta discriminant diagrams shown in

CE

fig.8. All of the kimberlites are considered to be moderately diamondiferous. The

Fe2O3 - MgO diagram shows no correlation between the diamond preservation of

AC

the kimberlites and the composition of their ilmenite xenocrysts. On the contrary,

ilmenites from the Cat115 kimberlite mostly fall in the marginal-preservation

field. Similarly, neither the composition of ilmenite inclusions in diamond nor the

data from this study support the use of the Zr/Nb vs Nb/Ta diagram to assess the

diamond grade of the kimberlites. Moreover, symplectitic ilmenites typically

 ACCEPTED MANUSCRIPT

show high Zr/Nb ratios and thus the threshold value of Zr/Nb ~ 0.37 in this case

would not be linked to diamond formation, but rather would indicate the degree

of interaction with the kimberlitic magma.

In striking contrast with the composition of most ilmenite kimberlitic

PT
xenocrysts worldwide, manganoan (0.63-2.49 wt% MnO) ilmenite xenocrysts

were found in Juina (Brazil) (Kaminsky and Belousova, 2009). Their trace-

RI
element composition clearly differs from that of the common ilmenite xenocrysts,

SC
being significantly poorer in Cr, Ca, Ni, Nb and Ta. Such compositions, however,
NU
are comparable to those of ilmenite inclusions in superdeep diamonds from the

same area (Kaminsky et al., 2001; Kaminsky and Belousova, 2009). As a

MA

consequence, these Mn-rich ilmenite nodules were considered as primary phases

crystallised from a very deep (>670 km) magma, in a Ti-rich environment and
D

genetically related to diamond (Kaminsky et al., 2001). These conclusions lead

E
PT

these researchers to suggest that the compositional range of the “classic”

composition of DIMs should be enlarged, especially in the areas with minor

CE

amounts of the usual indicator minerals. However, the Mn-rich ilmenite found in
AC

the Angolan kimberlites is a secondary/late phase, being either a secondary

replacement of the original ilmenite xenocrysts or crystallising as small plates

from a Mn-rich fluid. Admittedly, given the small size of tabular ilmenite, it would

be already filtered out at early stages of diamond exploration and it would not

have an impact on diamond exploration. However, the secondary replacement of

 ACCEPTED MANUSCRIPT

large ilmenite xenocrysts by Mn-rich ilmenite has been proven significant and

may have a misleading effect at prospection stages. Therefore, although from the

geotectonic perspective a link between the Angolan and the Brazilian kimberlite

fields would be reasonable, manganoan ilmenite cannot be used here as a new

diamond indicator unless an early origin for the Mn-rich ilmenite could be

PT
demonstrated in other contexts.

RI
SC
8. Conclusions
NU
The occurrence and composition of ilmenite can only be understood through
MA

studies encompassing both textural characterisation and chemical analysis.

Ferric ilmenite nodules in the Angolan kimberlites are suggested to result from
D

disaggregation of a metasomatised rock in crustal domains. In contrast, Mg-rich

E

ilmenite nodules could result either from disaggregation of mantle-

PT

metasomatised rocks; or could be megacrysts crystallised from a proto-

CE

kimberlite magma at mantle depths. A multi-stage model is proposed to explain

ilmenite petrogenesis in the Angolan kimberlites, including: a) crystallisation of

AC

ilmenite in crustal domains or in mantle-metasomatised sites under different

environmental conditions; b) xenolith disaggregation and/or megacryst

entrainment to form the corresponding populations of ilmenite xenocrysts; c)

widespread replacement of the original xenocrysts by Mg-rich ilmenite due to

interaction with the kimberlitic magma; d) secondary replacement with a

 ACCEPTED MANUSCRIPT

manganoan ilmenite either by interaction with a late fluid of carbonatitic affinity

or infiltration of hydrothermal Mn-rich fluids; and e) subsolidus alteration of the

previously crystallised phases in an oxidising environment.

This complex petrogenesis could thus result in misleading information about

PT
the diamond potential of a kimberlitic pipe if textural and trace-element

evidences are disregarded. Our work shows that secondary Mg enrichment of the

RI
ilmenite xenocrysts is unrelated to reducing conditions that could favour

SC
diamond formation/preservation in the mantle. In such cases, it is essential that
NU
ilmenite data are combined with information from other DIMs to avoid making

decisions with a high economic cost, since potentially diamondiferous

MA

kimberlites could be discarded as diamondiferous, and vice versa. Similarly, Mn-

rich ilmenite should be disregarded as a diamond indicator mineral, unless

D

textural studies can prove its primary origin.

E
PT
CE

ACKNOWLEDGEMENTS
AC

This research was supported by the projects CGL2006-12973 and CGL2009-

13758 of the Ministerio de Ciencia e Innovación of Spain; the SGR 589, SGR 444

and 2014 SGR 1661 projects of the AGAUR-Generalitat de Catalunya. Logistic

assistance for the field trips was provided by the mining company CATOCA SL.

The Ph.D. studies of M. Castillo-Oliver in Barcelona were supported by an FPU

grant given by Ministerio de Ciencia e Innovación of Spain. The research

 ACCEPTED MANUSCRIPT

activities were also funded by the Fundació Pedro Pons (UB), a research bursary

given by the Faculty of Geology from the Universitat de Barcelona and a Cotutelle

International Macquarie University Research Excellence Scholarship (iMQRES–

No. 2014210). The authors would also like to acknowledge Dr. Xavier Llovet and

Eva Prats for their assistance with EMP and SEM analysis at the Serveis

PT
Científico-Tècnics (UB); as well as Dr. José Ignacio Gil Ibarguchi, Dr. Sonia García

RI
de Madinabeitia, Dr. Aratz Beranoaguirre and Dr. Maria Eugenia Sanchez-Lorda

SC
for their help with the LA-ICP-MS analysis at the IBERCRON facilities (UPV/EHU).

This is contribution XXX from the ARC National Key Centre for Geochemical
NU
Evolution and Metallogeny of Continents (www.GEMOC.mq.edu.au) and paper
MA

XXX from the ARC Centre of Excellence for Core to Crust Fluid Systems

(www.CCFS.mq.edu.au). The authors would like to acknowledge the constructive

D

and pertinent comments of Dr Andy Moore and Dr Sebastian Tappe, which

E

significantly improved the manuscript.

PT
CE
AC
 ACCEPTED MANUSCRIPT

FIGURE CAPTIONS

Figure 1. Simplified geological map of the Lunda Norte kimberlite province, with the

location of the studied intrusions: Cat115 (1) and Tchiuzo (2). Geological and structural

maps modified from Perevalov et al.,(1992), Guiraud et al. (2005), Egorov et al. (2007)

and Robles-Cruz et al. (2012).

PT
RI
Figure 2. Textural types of ilmenite found in the Angolan kimberlites (BSE images). (a-

SC
c) ilmenite (ilm) in xenoliths. (a) metasomatic ilmenite associated with a secondary
NU
clinopyroxene (cpx2) replacing an early clinopyroxene (cpx1) and olivine; (b) anhedral

ilmenite, partially altered to titanite (ttn), with apatite (ap), clinopyroxene, calcite (cal)
MA

and kaersutite (krs); (c) ilmenite with thin hematite (hem) exsolutions, surrounded by

spongy garnet+orthopyroxene intergrowths (grt) and zircon (zrn) (d) ilmenite nodule
D

showing recrystallization set in the kimberlite matrix.

E
PT

Figure 2 (continuation). Textural types of ilmenite found in the Angolan kimberlites

CE

(BSE images). (e) Fe3±-ilmenite with fine hematite exsolutions replaced by symplectitic,
AC

darker Mg-rich ilmenite, which crystallises along fractures and grain boundaries.

Detailed image (f) shows the complex replacement sequence of the symplectitic

ilmenite: ulvöspinel (usp), perovskite (prv) and titanite (ttn); (g) nodular ilmenite

replaced by Mg- and Mn-rich ilmenite along grain boundaries of the recrystallised

grains, together with tabular Mn-rich ilmenite set in the kimberlite matrix; (h) late Mn-
 ACCEPTED MANUSCRIPT

rich ilmenite, replacing previous rutile (rut) and ulvöspinel (usp), set in a matrix of

serpentine (srp), calcite (cal) and phlogopite (phl).

Figure 3. Major element chemistry of the different textural types of ilmenite of the

Angolan kimberlites. (a) TiO2-MgO, including the kimberlitic ilmenite field as defined by

Wyatt et al. (2004). (b) Cr2O3-MgO. (c) classification of the analysed ilmenites in terms

PT
of their geikielite (MgTiO3), ilmenite (FeTiO3), hematite (Fe2O3) and pyrophanite

RI
(MnTiO3) components, after Mitchell (1979) and Tompkins and Haggerty (1985).

SC
Different colours represent different compositional groups: ilmenite sensu strictu

(purple), ferric ilmenite (yellow), Mg-rich ilmenite (red/orange) and Mn-rich ilmenite
NU
(green). The compositional domain of ilmenite xenocrysts of the Catoca kimberlite

(Robles-Cruz et al., 2009) is represented in grey colour. A Mg-enrichment trend is

MA

observed in both Fe3+ ilmenites (Ia, yellow arrow) and magnesian ilmenites (Ib, black

arrow). The green arrow (II) shows the late Mn-ilmenite replacement trend.
E D
PT

Figure 4. Bivariate diagrams showing the correlation between the major and minor
CE

elements in the different textural groups of ilmenite, expressed in atoms per formula

unit (apfu). Different colours represent different compositional groups: ilmenite sensu
AC

strictu (purple), ferric ilmenite (yellow), Mg-rich ilmenite (red/orange) and Mn-rich

ilmenite (green). A Mg-enrichment trend is observed in both Fe3+ ilmenites (Ia, yellow

arrow) and magnesian ilmenites (Ib, black arrow). The green arrow (II) shows the late

Mn-ilmenite replacement trend

 ACCEPTED MANUSCRIPT

Figure 5. BSE images and element profile of zoned symplectitic ilmenite of the Tchiuzo

kimberlite, showing that the main changes in the brightness of the BSE images are

correlated with its variation in Mg.

PT
Figure 6. Trace-element compositional plots for the different textural groups of ilmenite

found in the kimberlites of the Lunda Norte province. There is a positive correlation

RI
between Nb and the other HFSE (Ta, Zr, Hf) as well as with V. Notice the enrichment in

SC
HFSE of the Mg enrichment in ilmenite nodules and symplectitic ilmenites. Data from

ilmenite inclusions in diamond (Sobolev, 1977; Kaminsky et al., 2000, 2001), as well as
NU
trace-element compositions of ilmenite xenocrysts from other kimberlites worldwide
MA

(Kaminsky and Belousova, 2009; Kryvdik, 2014; Moore et al., 1992) have been included

for comparison (open diamonds and yellow field, respectively).

E D

Figure 7. Schematic model of the multi-stage petrogenesis of ilmenite in the studied

PT

kimberlites, including the average composition of ilmenite at each stage. Mineral

CE

abbreviations: clinopyroxene (cpx), ilmenite (ilm), garnet (grt), serpentine (srp),

apatite (ap), kaersutite (krs) and olivine (ol).

AC

Figure 8. Discrimination diagrams used to define the diamond grade of a kimberlite

from the composition of its ilmenite xenocrysts. (a) Fe2O3 vs MgO diagram, proposed by

Gurney and Zweistra (1995). The diamond grade of the pyroclastic kimberlites of the

Cat115 and Tchiuzo pipes are 39% and 46%, respectively, of the grade of the Catoca
 ACCEPTED MANUSCRIPT

kimberlite (6th largest diamond producer worldwide), which classifies them as

kimberlites of moderate diamond grade. The field in grey includes data of ilmenite

grains from the Catoca kimberlite (Robles-Cruz et al., 2009). (b) Zr/Nb vs Nb/Ta

diagram, modified after Carmody et al. (2014). The dashed line indicates the threshold

Zr/Nb ~ 0.37 that the authors proposed to assess the diamond grade of the kimberlite.

PT
The fields represented in grey were drawn after a collection of ilmenite xenocrysts of

Siberian kimberlites with different diamond grade (Carmody et al., 2014). Data from

RI
ilmenite inclusions found in diamond are also represented for comparative purposes (

SC
Sobolev et al., 1997; Kaminsky et al., 2001).
NU
MA
E D
PT
CE
AC
 ACCEPTED MANUSCRIPT

TABLE CAPTIONS

Table 1. Major element analyses of ilmenite (wt %) and the calculated structural

formula in atoms per formula unit (apfu) normalised to 3 oxygens. Abbreviations:

Compositional groups: 1. Ilmenite sensu strictu; 2. Fe3+-rich ilmenite; 3. Mg-rich

ilmenite and 4. Mn-rich ilmenite. Textures: met.1 = metasomatic ilmenite (with apatite,

PT
clinopyroxene and amphibole); met.2 = metasomatic ilmenite (veinlets in peridotites);

RI
exsol = ilmenite with hematite exsolutions; sympl = symplectitic ilmenite; Mg-rich =

SC
Mg enrichment in ilmenites,usually already rich in Mg (darker areas in BSE images);

recrist. = ilmenite nodules showing recristallisation.

NU
MA

Table 2. Trace element analyses of ilmenite (ppm). Bdl stands for below detection limit

analyses. Abbreviations as in table 1.

E D
PT
CE
AC
 ACCEPTED MANUSCRIPT

REFERENCES

Araújo, D.P., Griffin, W.L., O&apos;Reilly, S.Y., 2009. Mantle melts, metasomatism and diamond
formation: Insights from melt inclusions in xenoliths from Diavik, Slave Craton. Lithos 112, 675–
682. doi:10.1016/j.lithos.2009.06.005
Ashchepkov, I.V., Rotman, a. Y., Somov, S.V., Afanasiev, V.P., Downes, H., Logvinova, a. M.,
Nossyko, S., Shimupi, J., Palessky, S.V., Khmelnikova, O.S., Vladykin, N.V., 2012. Composition
and thermal structure of the lithospheric mantle beneath kimberlite pipes from the Catoca
cluster, Angola. Tectonophysics 530–531, 128–151. doi:10.1016/j.tecto.2011.12.007

PT
Boctor, N.Z., Boyd, F.R., 1981. Oxide minerals in a layered kimberlite–carbonatite sill from
Benfontein, South Africa. Contrib. to Mineral. Petrol. 76, 253–259.

RI
Boyd & Danchin 1980 Lherzolites, eclogites, and megacrysts from some kimberlites of Angola.pdf,
1980.

SC
Carmody, L., Taylor, L. a., Thaisen, K.G., Tychkov, N.S., Bodnar, R.J., Sobolev, N. V., Pokhilenko, L.N.,
Pokhilenko, N.P., 2014. Ilmenite as diamond indicator mineral in the Siberian Craton: a tool to
predict diamond potential. Econ. Geol. 109, 775–783.
NU
Cas, R., Porritt, L., Pittari, A., Hayman, P., 2008. A new approach to kimberlite facies terminology
using a revised general approach to the nomenclature of all volcanic rocks and deposits:
Descriptive to genetic. J. Volcanol. Geotherm. Res. 174, 226–240.
MA

doi:10.1016/j.jvolgeores.2007.12.018
Castillo-Oliver, M., Galí, S., Melgarejo, J.C., Griffin, W.L., Belousova, E., Pearson, N.J., Watangua, M.,
O’Reilly, S.Y., 2016. Trace-element geochemistry and U–Pb dating of perovskite in kimberlites
of the Lunda Norte province (NE Angola): Petrogenetic and tectonic implications. Chem. Geol.
D

426, 118–134. doi:10.1016/j.chemgeo.2015.12.014

E

Chakhmouradian, A.R., Mitchell, R.H., 2001. Three compositional varieties of perovskite from
PT

kimberlites of the Lac de Gras field ( Northwest Territories , Canada ). Mineral. Mag. 65, 133–
148.
Chakhmouradian, A.R., Mitchell, R.H., 1999. Niobian ilmenite, hydroxylapatite and sulfatian
CE

alternative hosts for incompatible in calcite elements. Can. Mineral. 37, 1177–1189.
Dare, S. a S., Barnes, S.J., Beaudoin, G., 2012. Variation in trace element content of magnetite
crystallized from a fractionating sulfide liquid, Sudbury, Canada: Implications for provenance
AC

discrimination. Geochim. Cosmochim. Acta 88, 27–50. doi:10.1016/j.gca.2012.04.032

Dawson, J.B., 2002. Metasomatism and Partial Melting in Upper-Mantle Peridotite Xenoliths from
the Lashaine Volcano, Northern Tanzania. J. Pet. 43, 1749–1777.
doi:10.1093/petrology/43.9.1749
Dawson, J.B., Smith, J. V., 1977. The MARID (mica-amphibole-rutile-ilmenite-diopside) suite of
xenoliths in kimberlite. Geochim. Cosmochim. Acta 41, 309–323. doi:10.1016/0016-
7037(77)90239-3
Donohue, P.H., Simonetti, A., Neal, C.R., 2012. Chemical Characterisation of Natural Ilmenite: A
Possible New Reference Material. Geostand. Geoanalytical Res. 36, 61–73. doi:10.1111/j.1751-
908X.2011.00124.x
 ACCEPTED MANUSCRIPT

Eggler, D.H., Mccallum, M.E., Smith, C.B., 1979. Assemblages in kimberlite form Colorado and
Southern Wyoming: petrology, geothermometry-barometry, and areal distribution, in: The
Mantle Sample: Inclusion in Kimberlites and Other Volcanics. pp. 213–226.
Egorov, K.N., Roman, E.F., Podvysotsky, V.T., Sablukov, S.M., Garanin, V.K., D’yakonov, D.B., 2007.
New data on kimberlite magmatism in southwestern Angola. Russ. Geol. Geophys. 48, 323–
336. doi:10.1016/j.rgg.200
Faure, S., 2010. World Kimberlites CONSOREM Database (Version 3), Consortium de Recherche en
Exploration Minérale CONSOREM, Université du Québec à Montréal [WWW Document].
www.consorem.ca.

PT
Gaspar, J.C., Wyllie, P.J., 1984. The alleged kimberlite-carbonatite relationship: evidence from
ilmenite and spinel from Premier and Wesselton mines and the Benfontein sill, South Africa.

RI
Contrib. to Mineral. Petrol. 85, 133–140. doi:10.1007/BF00371703
Gaspar, J.C., Wyllie, P.J., 1983. Ilmenite ( high Mg , Mn , Nb ) in the carbonatites from the

SC
Jacupiranga Complex , Brazil. Am. Mineral. 68, 960–971.
Giuliani, A., Phillips, D., Woodhead, J.D., Kamenetsky, V.S., Fiorentini, M.L., Maas, R., Soltys, A.,
Armstrong, R.A., 2015. Did diamond-bearing orangeites originate from MARID-veined
NU
peridotites in the lithospheric mantle? Nat. Commun. 6, 1–10. doi:10.1038/ncomms7837
Golubkova, A.B., Nosova, A.A., Larionova, Y.O., 2013. Mg-ilmenite megacrysts from the Arkhangelsk
kimberlites, Russia: Genesis and interaction with kimberlite melt and postkimberlite fluid.
MA

Geochemistry Int. 51, 353–381. doi:10.1134/S0016702913030038

Grégoire, M., Bell, D., Le Roex, a., 2002. Trace element geochemistry of phlogopite-rich mafic
mantle xenoliths: their classification and their relationship to phlogopite-bearing peridotites
and kimberlites revisited. Contrib. to Mineral. Petrol. 142, 603–625. doi:10.1007/s00410-001-
D

0315-8
E

Griffin, W.L., Moore, R.O., Ryan, C.G., Gurney, J.J., Win, T.T., 1997. Geochemistry of magnesian
PT

ilmenite megacrysts from Southern African kimberlites. Russ. Geol. Geophys. 38, 421–443.
Griffin, W.L., O’Reilly, S.Y., Afonso, J.C., Begg, G.C., 2009. The composition and evolution of
lithospheric mantle: A re-evaluation and its tectonic implications. J. Petrol. 50, 1185–1204.
CE

doi:10.1093/petrology/egn033
Griffin, W.L., Ryan, C.G., 1995. Trace elements in indicator minerals: area selection and target
evaluation in diamond exploration. J. Geochemical Explor. 53, 311–337. doi:10.1016/0375-
AC

6742(94)00015-4
Griffin, W.L., Smith, D., Boyd, F.R., Cousens, D.R.I., Ryan, C.G., Sie, S.H., Suter, G.F., 1989. Trace-
element zoning in garnets from sheared mantle xenoliths. Geochim. Cosmochim. Acta 53, 561–
567.
Guiraud, R., Bosworth, W., Thierry, J., Delplanque, A., 2005. Phanerozoic geological evolution of
Northern and Central Africa: An overview. J. African Earth Sci. 43, 83–143.
Gurney, J.J., Moore, R.O., 1993. Geochemical correlations between kimberlitic indicator minerals and
diamonds, in: Diamonds: Exploration, Sampling and Evaluation. pp. 149–171.
Gurney, J.J., Zweistra, P., 1995. The interpretation of the major element compositions of mantle
 ACCEPTED MANUSCRIPT

minerals in diamond exploration. J. Geochemical Explor. 53, 293–309. doi:10.1016/0375-

6742(94)00021-3
Haggerty, S.E., 1991a. Oxide mineralogy of the upper mantle, in: Oxide Minerals. Mineralogical
Society of America Reviews in Mineralogy, vol. 25, pp. 355– 416.
Haggerty, S.E., 1991b. Oxide textures; a mini-atlas. Rev. Mineral. Geochemistry 25, 129–219.
Haggerty, S.E., 1983. Lindseyite (Ba) and mathiasite (K): two new chromium-titanates in the
crichtonite series from the upper mantle. Am. Mineral. 68, 494–505.
Harris, J.W., Hawthorne, J.B., Oosterveld, M.M., 1984. A comparison of diamond characteristics from

PT
the DeBeers pool mines, Kimberley, South Africa. Ann. Sci. Univ. Clermont Ferrand 74, 1–13.
Harte, B., 1983. Mantle peridotites and processes - the kimberlite sample, in: Hawkesworth, C.J.

RI
(Ed.), Continental Basalts and Mantle Xenoliths. Nantwich, pp. 46–91.
Harte, B., Gurney, J.J., 1981. The mode of formation of chromium-poor megacryst suites from

SC
kimberlites. J. Geol. 89, 749–753.
Harte, B., Gurney, J.J., 1980. The mode of formation of chromium-poor megacryst suites from
kimberlites. J. Geol. 89, 749–753.
NU
Jacques, A.L., Hall, A.E., Sheraton, J.W., Smith, C.B., Sun, S.S., Drew, R.M., Foudoulis, C., Ellingsen, K.,
1989. Composition of crystalline inclusions and C isotopic composition of Argyle and Ellendale
diamonds, in: Ross, J., Jaques, A.L., Ferguson, J., Green, D.H., O’Reilly, S.Y., Danchin, R.V., Janse,
MA

A.J.A. (Eds.), Kimberlites and Related Rocks, Proceedings of the Fourth International Kimberlite
Conference. Geological Society of Australia Special Publication n14, vol.2, pp. 966–989.
Jang, Y.D., Naslund, H.., 2003. Major and trace element variation in ilmenite in the Skaergaard
D

Intrusion: petrologic implications. Chem. Geol. 193, 109–125. doi:10.1016/S0009-

2541(02)00224-3
E

Jelsma, H.A., Wit, M.J. de, Thiart, C., Dirks, P.H.G.., Viola, G., Basson, I.J., Anckar, E., 2004.
PT

Preferential distribution along transcontinental corridors of kimberlites and related rocks of

Southern Africa. South African J. Geol. 107, 301–324.
Jelsma, H., Barnett, W., Richards, S., Lister, G., 2009. Tectonic setting of kimberlites. Lithos 112, 155–
CE

165. doi:10.1016/j.lithos.2009.06.030
Jelsma, H., Krishnan, U., Perritt, S., Preston, R., Lemotlo, L., Linde, G. Van Der, Armstrong, R., Joy, S.,
AC

Costa, J., Facatino, M., Posser, A., Kumar, M., Wallace, C., Chinn, I., Henning, A., 2013.
Kimberlites from Central Angola: a case study of exploration findings, in: Proceedings of 10th
International Kimberlite Conference. Springer, New Delhi, pp. 173–190. doi:10.1007/978-81-
322-1173-0
Jones, R.A., 1987. Strontium and neodymium isotopic and rare earth element evidence for the
genesis of megacrysts in kimberlites of southern Africa, in: Mantle Xenoliths. John Wiley and
Sons, Chichester, UK, pp. 711–724.
Kamenetsky, V.S., Golovin, A. V., Maas, R., Giuliani, A., Kamenetsky, M.B., Weiss, Y., 2014. Towards a
new model for kimberlite petrogenesis: Evidence from unaltered kimberlites and mantle
minerals. Earth-Science Rev. 139, 145–167. doi:10.1016/j.earscirev.2014.09.004
Kaminsky, F., Zakharchenko, O., Davies, R., Griffin, W., Khachatryan-Blinova, G., Shiryaev, a., 2001.
 ACCEPTED MANUSCRIPT

Superdeep diamonds from the Juina area, Mato Grosso State, Brazil. Contrib. to Mineral.
Petrol. 140, 734–753. doi:10.1007/s004100000221
Kaminsky, F. V., Belousova, E. a., 2009. Manganoan ilmenite as kimberlite/diamond indicator
mineral. Russ. Geol. Geophys. 50, 1212–1220. doi:10.1016/j.rgg.2009.11.019
Kaminsky, F. V., Zakharchenko, O.D., Griffin, W.L., Channer, D.M.D., Khachatryan-Blinova, G.K., 2000.
Diamond from the Guaniamo area, Venezuela. Can. Mineral. 38, 1347–1370.
doi:10.2113/gscanmin.38.6.1347
Konzett, J., Armstrong, R.A., Sweeney, R.J., Compston, W., 1998. The timing of MARID metasomatism

PT
in the Kaapvaal mantle: An ion probe study of zircons from MARID xenoliths. Earth Planet. Sci.
Lett. 160, 133–145.
Kopylova, M.G., Nowell, G.M., Pearson, D.G., Markovic, G., 2009. Crystallization of megacrysts from

RI
protokimberlitic fluids: Geochemical evidence from high-Cr megacrysts in the Jericho
kimberlite. Lithos 112S, 284–295. doi:10.1016/j.lithos.2009.06.008

SC
Kostrovitsky, S.I., Malkovets, V.G., Verichev, E.M., Garanin, V.K., Suvorova, L. V., 2004. Megacrysts
from the Grib kimberlite pipe (Arkhangelsk Province, Russia). Lithos 77, 511–523.
doi:10.1016/j.lithos.2004.03.014
NU
Kotel’nikov, D.D., Zinchuk, N.N., Zhukhlistov, A.P., 2005. Stages of serpentine and phlogopite
transformation in the Catoca kimberlite pipe, Angola. Dokl. Earth Sci. 403, 866–869.
MA

Kryvdik, S.G., 2014. Geochemical features of ilmenites from the alkaline complexes of the Ukrainian
Shield: LA-ICP MS data. Geochemistry Int. 52, 287–295. doi:10.1134/S0016702914040041
Malkovets, V.G., Griffin, W.L., O’Reilly, S.Y., Wood, B.J., 2007. Diamond, subcalcic garnet, and mantle
metasomatism: Kimberlite sampling patterns define the link. Geology 35, 339.
D

doi:10.1130/G23092A.1
E

Meyer, H.O.A., Svisero, D.P., 1975. Mineral inclusions in Brazilian diamonds. Phys. Chem. Earth 9,
785–795.
PT

Mitchell, R.H., 1986. Kimberlites, Mineralogy, Geochemistry and Petrology.

Mitchell, R.H., 1979. The alleged kimberlite-carbonatite relationship: additional contrary
CE

mineralogical evidence. Am. J. Sci. 279, 570–589.

Mitchell, R.H., 1978. Manganoan magnesian ilmenite and titanian clinohumite from the Jacupiranga
AC

carbonatite, São Paulo, Brazil. Am. Mineral. 63, 544–547.

Moore, A., Belousova, E., 2005. Crystallization of Cr-poor and Cr-rich megacryst suites from the host
kimberlite magma: implications for mantle structure and the generation of kimberlite magmas.
Contrib. to Mineral. Petrol. 149, 462–481. doi:10.1007/s00410-005-0663-x
Moore, A., Blenkinsop, T., Woody, F., 2008. Controls on post-Gondwana alkaline volcanism in
Southern Africa. Earth Planet. Sci. Lett. 268, 151–164. doi:10.1016/j.epsl.2008.01.007
Moore, A., Lock, N.P., 2001. The origin of mantle-derived megacrysts and sheared peridotites-
evidence from kimberlites in the northern Lesotho-Orange Free State (South Africa) and
Botswana pipe clusters. South African J. Geol. 104, 23–38.
Moore, R.O., Griffin, W.L., Gurney, J.J., Ryan, C.G., Cousens, D.R., Sie, S.H., Suter, G.F., 1992. Trace
 ACCEPTED MANUSCRIPT

element geochemistry of ilmenite megacrysts from the Monastery kimberlite, South Africa.
Lithos 29, 1–18. doi:10.1016/0024-4937(92)90031-S
Mvuemba Ntanda, F., Moreau, J., Meyer, H.O.A., 1982. Particularites Des Inclusions Cristallines
Primaires Des Diamants Du Kasai , Zaire. Can. Mineral. 20, 217–230.
Nikitina, L.P., Korolev, N.M., Zinchenko, V.N., Felix, J.T., 2014. Eclogites from the upper mantle
beneath the Kasai Craton (Western Africa): Petrography, whole-rock geochemistry and UPb
zircon age. Precambrian Res. 249, 13–32. doi:10.1016/j.precamres.2014.04.014
Nixon, P.H., Boyd, F.R., 1973. The discrete nodule (megacryst) association in kimberlites from

PT
northern Lesotho, in: Nixon, P.H. (Ed.), Lesotho Kimberlites. Cape and Transvaal Printers, Cape
Town, pp. 67–75.
Nowell, G.M., Pearson, D.G., Bell, D.R., Carlson, R.W., Smith, C.B., Kempton, P.D., Noble, S.R., 2004.

RI
Hf Isotope systematics of kimberlites and their megacrysts: new constraints on their source
regions. J. Petrol. 45, 1583–1612. doi:10.1093/petrology/egh024

SC
O’Reilly, S.Y., Griffin, W.L., 2013. Mantle metasomatism, in: Harlov, D.E., Austrheim, H. (Eds.),
Metasomatism and the Chemical Transformation of Rock, Lecture Notes in Earth System
Sciences. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 471–534. doi:10.1007/978-3-642-
NU
28394-9
O’Reilly, S.Y., Zhang, M., Griffin, W.L., Begg, G., Hronsky, J., 2009. Ultradeep continental roots and
their oceanic remnants: A solution to the geochemical “mantle reservoir” problem? Lithos 112,
MA

1043–1054. doi:10.1016/j.lithos.2009.04.028
Ogilvie-Harris, R.C., Field, M., Sparks, R.S.J., Walter, M.J., 2009. Perovskite from the Dutoitspan
kimberlite, Kimberley, South Africa: implications for magmatic processes. Mineral. Mag. 73,
915–928. doi:10.1180/minmag.2009.073.6.915
D

Pasteris, J.D., 1980. The significance of groundmass ilmentite and megacryst ilmentite in kimberlites.
E

Contrib. to Mineral. Petrol. 75, 315–325.

PT

Paton, C., Hellstrom, J., Paul, B., Woodhead, J., Hergt, J., 2011. Iolite: Freeware for the visualisation
and processing of mass spectrometric data. J. Anal. At. Spectrom. 26, 2508.
doi:10.1039/c1ja10172b
CE

Pereira, E., Rodrigues, J., Reis, B., 2003. Synopsis of Lunda geology , NE Angola : Implications for
diamond exploration. Comun. Inst. Geol. e Min. 90, 189–212.
AC

Perevalov, O.V., Voinovsky, A.S., Tselikovsky, A.F., Agueev, Y.L., Polskoi, F.R., Khódirev, V.L.,
Kondrátiev, A.I., 1992. Geology of Angola: Explanatory Notes of the Geological Map at a Scale
of 1:1.000.000; Geological Survey of Angola: Luanda, Angola.
Pivin, M., Féménias, O., Demaiffe, D., 2009. Metasomatic mantle origin for Mbuji-Mayi and
Kundelungu garnet and clinopyroxene megacrysts (Democratic Republic of Congo). Lithos 112,
951–960. doi:10.1016/j.lithos.2009.03.050
Pouchou, J.L., Pichoir, F., 1984. A new model for quantitative X-ray microanalysis. Part 1.
Applications to the analysis of homogeneous samples. Rech. Aerosp. (English Ed. 3, 11–38.
Reis, B., 1972. Preliminary note on the distribution and tectonic controls on kimberlites in Angola.
24th Int. Geollgical Congr. Montr. Section 4, 276–281.
 ACCEPTED MANUSCRIPT

Robles-Cruz, S.E., Escayola, M., Jackson, S., Galí, S., Pervov, V., Watangua, M., Gonçalves, A.,
Melgarejo, J.C., 2012. U–Pb SHRIMP geochronology of zircon from the Catoca kimberlite,
Angola: Implications for diamond exploration. Chem. Geol. 310–311, 137–147.
doi:10.1016/j.chemgeo.2012.04.001
Robles-Cruz, S.E., Watangua, M., Isidoro, L., Melgarejo, J.C., Galí, S., Olimpio, A., 2009. Contrasting
compositions and textures of ilmenite in the Catoca kimberlite, Angola, and implications in
exploration for diamond. Lithos 112, 966–975. doi:10.1016/j.lithos.2009.05.040
Schulze, D.J., 1987. Megacrysts from alkalic volcanic rocks, in: Mantle Xenoliths. Wiley, Chichester,
pp. 433–451.

PT
Schulze, D.J., 1984. Cr-poor megacrysts in the Hamilton Branch kimberlite, Kentucky, in: Kornprobst,
J. (Ed.), Proceedings 3rd International Kimberlite Conference. Elsevier, Amsterdam, pp. 97–108.

RI
Schulze, D.J., Anderson, P.F.N., Hearn, B.C., Hetman, C.M., 1995. Origin and Significance of Ilmenite
Megacrysts and Macrocrysts from Kimberlite. Int. Geol. Rev. 37, 780–812.

SC
doi:10.1080/00206819509465427
Shee, S.R., Gurney, J.J., 1979. The mineralogy of xenoliths from Orapa, Botswana, in: The Mantle
Sample: Inclusion in Kimberlites and Other Volcanics. pp. 37–49.
NU
Smart, K.A., Cartigny, P., Tappe, S., Brien, H.O., Klemme, S., 2017. Lithospheric diamond formation as
a consequence of methane-rich volatile flooding: An example from diamondiferous eclogite
xenoliths of the Karelian craton (Finland). Geochim. Cosmochim. Acta 206, 312–342.
MA

doi:10.1016/j.gca.2017.03.014
Sobolev, N. V., 1977. Deep-seated inclusions in kimberlites and the problem of the composition of
the upper mantle. American Geophysical Union, Washington, D. C.
D

Sobolev, N. V., Kaminsky, F. V., Griffin, W.L., Yefimova, E.S., Win, T.T., Ryan, C.G., Botkunov, A.I.,
1997. Mineral inclusions in diamonds from the Sputnik kimberlite pipe, Yakutia. Lithos 39, 135–
E

157. doi:10.1016/S0024-4937(96)00022-9
PT

Stachel, T., Aulbach, S., Brey, G.P., Harris, J.W., Leost, I., Tappert, R., Viljoen, K.S., 2004. The trace
element composition of silicate inclusions in diamonds: A review. Lithos 77, 1–19.
doi:10.1016/j.lithos.2004.03.027
CE

Tappe, S., Foley, S.F., Jenner, G. a., Heaman, L.M., Kjarsgaard, B. a., Romer, R.L., Stracke, A., Joyce,
N., Hoefs, J., 2006. Genesis of ultramafic lamprophyres and carbonatites at Aillik Bay, Labrador:
A consequence of incipient lithospheric thinning beneath the North Atlantic Craton. J. Petrol.
AC

47, 1261–1315. doi:10.1093/petrology/egl008

Tappe, S., Jenner, G.A., Foley, S.F., Heaman, L., Besserer, D., Kjarsgaard, B.A., Ryan, B., 2004. Torngat
ultramafic lamprophyres and their relation to the North Atlantic Alkaline Province. Lithos 76,
491–518. doi:10.1016/j.lithos.2004.03.040
Tappe, S., Kjarsgaard, B. a., Kurszlaukis, S., Nowell, G.M., Phillips, D., 2014. Petrology and Nd-Hf
Isotope Geochemistry of the Neoproterozoic Amon Kimberlite Sills, Baffin Island (Canada):
Evidence for Deep Mantle Magmatic Activity Linked to Supercontinent Cycles, Journal of
Petrology. doi:10.1093/petrology/egu048
Tappe, S., Pearson, D.G., Nowell, G., Nielsen, T., Milstead, P., Muehlenbachs, K., 2011. A fresh
isotopic look at Greenland kimberlites: Cratonic mantle lithosphere imprint on deep source
 ACCEPTED MANUSCRIPT

signal. Earth Planet. Sci. Lett. 305, 235–248. doi:10.1016/j.epsl.2011.03.005

Tappe, S., Romer, R.L., Stracke, A., Steenfelt, A., Smart, K.A., Muehlenbachs, K., Torsvik, T.H., 2017.
Sources and mobility of carbonate melts beneath cratons, with implications for deep carbon
cycling, metasomatism and rift initiation. Earth Planet. Sci. Lett. 466, 152–167.
doi:10.1016/j.epsl.2017.03.011
Tompkins, L. a., Haggerty, S.E., 1985. Groundmass oxide minerals in the Koidu kimberlite dikes,
Sierra Leone, West Africa. Contrib. to Mineral. Petrol. 91, 245–263. doi:10.1007/BF00413351
Turcotte, D.L., Oxburgh, E.R., 1973. Mid-plate tectonics. Nature 244, 337–339.

PT
Wyatt, B.A., 1979. Manganoan ilmenite from the Premier kimberlite, in: Kimberlite Symposium II.
Wyatt, B. a., Baumgartner, M., Anckar, E., Grutter, H., 2004. Compositional classification of

RI
“kimberlitic” and “non-kimberlitic” ilmenite. Lithos 77, 819–840.
doi:10.1016/j.lithos.2004.04.025

SC
NU
MA
E D
PT
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
DE
PT
CE

Figure 1
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
DE
PT

Figure 2
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
DE
PT

Figure 2b
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
DE
PT

Figure 3
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
DE
PT
CE

Figure 4
AC
 ACCEPTED MANUSCRIPT

Figure 5

PT
RI
SC
NU
MA
DE
PT
CE
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
DE
PT
CE

Figure 6
AC
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
DE
PT
CE

Figure 7
AC
 ACCEPTED MANUSCRIPT

PT
Figure 8

RI
SC
NU
MA
DE
PT
CE
AC
 ACCEPTED MANUSCRIPT

Table 1. Major element analyses of ilmenite (wt %) and the calculated structural formula in atoms per formula unit (apfu) normalised to 3 oxygens
ilm_ ilm_ ilm_ ilm_1 ilm_1 ilm_1 ilm_1 ilm_1 ilm_1 ilm_1
ilm_1 ilm_4 ilm_5 ilm_8 ilm_9 ilm_17
Analysis 2 3 6 0 1 2 3 4 5 6
compositio 1 1 1 2 2 3 3 3 3 3 3 3 3 4 4 4
nal group
grt-
met. met. met. nodul nodul nodul nodul nodul symp symp tabul
P T Mn-
rich
Mn-rich
Texture pyroxeni
te
1 1
exsol exsol
2 e e e e e

R I l l ar nodul
e
seconda
ry
analysed
area
core core core core core core core core rim

S C
core core core core core core core

oxides %
SiO2 0.01 0.12 0.05 0.00 0.10 0.08 0.02 0.04
N 0.04 U0.09 0.06 0.06 0.01 0.02 0.03 0.01

TiO2
Al2O3
43.45

0.13
43.7
1
0.19
44.5
9
0.14
37.16

0.13
30.94

0.22
47.0
0
0.48
45.92

0.52
M A
50.98

0.54
52.70

0.48
48.81

0.54
48.42

0.58
49.96

0.79
42.88

2.00
50.55

0.05
53.17

0.56
51.11

0.05
Nb2O5
ZrO2
0.00
0.11
0.00
0.00
0.00
0.00
0.44
0.41
0.42
0.32
E D
0.30
0.13
0.15
0.07
0.01
0.10
0.17
0.00
0.22
0.03
0.11
0.13
0.22
0.34
0.69
0.73
0.17
0.00
0.11
0.16
0.23
0.00

PT
Cr2O3 0.18 0.24 0.94 0.41 0.46 0.33 0.85 0.85 0.51 1.04 2.13 1.02 1.72 0.12 0.50 0.08
17.1 15.5 14.6
Fe2O3
17.88

36.49
2
36.9
5
35.5
C E
30.06

28.52
41.51

23.69
6
29.1
15.80

28.18
9.78

24.33
9.26

22.09
11.44

29.19
11.22

27.40
11.01

24.61
19.59

21.29
2.05

15.92
5.88

19.37
2.31

29.39
FeO
MnO
MgO
0.08
1.33
1

A
0.31
1.13
C 8
0.27
2.35
0.12
3.03
0.08
2.63
6
0.27
7.42
0.24
7.25
1.65
11.11
1.28
13.52
0.19
8.27
0.27
8.95
0.70
11.15
0.43
10.04
29.10
0.08
9.50
10.50
16.42
0.10
CaO 0.00 0.07 0.01 0.00 0.07 0.01 0.02 0.09 0.04 0.00 0.00 0.08 0.08 0.04 0.11 0.02
ZnO 0.17 0.06 0.06 0.11 0.00 0.03 0.01 0.00 0.00 0.06 0.01 0.06 0.00 0.03 0.05 0.01
V2O3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.23 0.00 0.00 0.00
NiO 0.07 0.09 0.07 0.00 0.05 0.08 0.16 0.06 0.10 0.10 0.19 0.10 0.03 0.07 0.19 0.00
sum 99.90 99.9 99.6 100.3 100.4 99.9 99.19 99.54 100.2 99.98 99.48 100.1 99.72 98.21 100.1 99.73
 ACCEPTED MANUSCRIPT

5 1 9 9 5 0 0 3
apfu (O=3)
0.00 0.00 0.00
0.000 0.000 0.002 0.000 0.001 0.001 0.002 0.001 0.001 0.000 0.000 0.001 0.000
Si 3 1 2
0.79 0.80 0.85
0.787 0.673 0.560 0.832 0.923 0.955 0.884 0.877 0.905 0.777 0.916 0.963 0.926
Ti 2
0.00
8
0.00
1
0.01
P T
Al
0.004

0.000
5
0.00
4
0.00
0.004

0.005
0.006

0.005
4
0.00
0.015

0.002
0.015

0.000
0.014

0.002
0.015

0.002
0.016

R
0.001
I 0.022

0.002
0.057

0.008
0.001

0.002
0.016

0.001
0.001

0.003
Nb

Zr
0.001
0
0.00
0
0
0.00
0
0.005 0.004
3
0.00
2
0.001 0.001 0.000
S C
0.000 0.002 0.004 0.009 0.000 0.002 0.000

Cr
0.003
0.00
5
0.01
8
0.008 0.009
0.00
6
0.016 0.016
N 0.010 U
0.020 0.041 0.019 0.033 0.002 0.010 0.002

Fe3+
0.324

0.735
0.31
0
0.74
0.28
2
0.71
0.545

0.574
0.752

0.477
0.26
6
0.58
0.286

0.567 M A
0.177

0.490
0.168

0.445
0.207

0.588
0.203

0.552
0.199

0.496
0.355

0.429
0.037

0.321
0.106

0.390
0.042

0.592
Fe2+
0.002
3
0.00
6
0.00
0.002 0.002
E
7
D
0.00
0.005 0.034 0.026 0.004 0.006 0.014 0.009 0.593 0.194 0.335
Mn

Mg
0.048
6
0.04
1
6
0.08
4
0.109
P T
0.094
6
0.26
6
0.260 0.399 0.485 0.297 0.321 0.400 0.360 0.003 0.377 0.004

Ca
0.000
0.00
2
0.00

C
0
0.000E 0.002
0.00
0
0.001 0.002 0.001 0.000 0.000 0.002 0.002 0.001 0.003 0.001

Zn
0.003

0.000
0.00

A
1
0.00
C
0.00
1
0.00
0.002

0.000
0.000

0.000
0.00
1
0.00
0.000

0.000
0.000

0.000
0.000

0.000
0.001

0.000
0.000

0.000
0.001

0.000
0.000

0.002
0.001

0.000
0.001

0.000
0.000

0.000
V 0 0 0
0.00 0.00 0.00
0.001 0.000 0.001 0.003 0.001 0.002 0.002 0.004 0.002 0.001 0.001 0.004 0.000
Ni 2 1 2
% geikielite 5.0 4.3 8.9 11.4 9.9 26.9 26.7 39.4 46.7 29.9 32.8 39.6 36.9 0.3 37.2 0.4
% hematite 17.1 16.4 14.9 28.4 39.6 13.4 14.7 8.8 8.1 10.4 10.4 9.9 18.2 2.0 5.3 2.2
% ilmenite 77.7 78.6 75.6 60.0 50.3 59.2 58.2 48.5 42.8 59.2 56.3 49.1 44.0 34.3 38.5 62.2
 ACCEPTED MANUSCRIPT

%
pyrophanit 0.2 0.7 0.6 0.3 0.2 0.6 0.5 3.3 2.5 0.4 0.6 1.4 0.9 63.4 19.1 35.2
e
Abbreviations: Compositional groups: 1. Ilmenite sensu strictu; 2. Fe3+-rich ilmenite; 3. Mg-rich ilmenite and 4. Mn-rich ilmenite. Textures: met.1 =
metasomatic ilmenite (with apatite, clinopyroxene and amphibole); met.2 = metasomatic ilmenite (veinlets in peridotites); exsol = ilmenite with

T
hematite exsolutions; sympl = symplectitic ilmenite; Mg-rich = Mg enrichment in ilmenites,usually already rich in Mg (darker areas in BSE images)

P
R I
S C
N U
M A
E D
P T
C E
A C
 ACCEPTED MANUSCRIPT

Table 2. Trace element analyses of ilmenite (ppm). Bdl stands for below detection limit analyses
42_ 50_ 32A 6A_ 44_ 35_i 26_j 31_ 32A 618- 22_ 60A 22_ 59- 34_ 34_ 618- 618-
35_c
Label _040 b_1 2a_ _g_0 d_0 b_1 _06 _03 c_0 _l_0 267a_ g_0 _4a_ c_0 A2_c b_0 b_0 47_g_ 47_g
55 107 87 55 25 1 2 81 98 b_139 81 052 70 _057 13 13b 112 _115
Kimb Cat Cat Tchi Cat Cat Cat Cat Cat Cat Cat Tchiuz Cat Tchi Cat
T
Tchi

P
Cat Cat Tchiuz Tchiu

I
erlite 115 115 uzo 115 115 115 115 115 115 115 o 115 uzo 115 uzo 115 115 o zo
Comp
ositio
nal
1 1 1 2 2 3 3 3 3 3 3
C
3 3 R3 3 3 3 4 4
group
Textu
grt-
pyrox
met nod exso exs met nod nod nod nod
nodule U Snod nodu nod nodul sym sym
nodule
Tabul
re enite
.1 ule l* ol* .2 ule

Mg-
ule

Mg-
ule ule

A N ule le ule e pl

Mg
pl
Mg-
ar

extra
info
ilm
rim
ilm
rim M recr
yst.
recry
st.
-
rich
Mg-
rich
Mn-
ilm
Mn-
rich

12.9 17. 24. 30. 27. 18. 25.

E D
30.
18. 18. 22. 15. 18. 15.0 130.
rim

T
Sc
4 69 1 0 1 41 3 2
30 20 19.80 6 26 17 5 bdl bdl 63.6 0
V
351 20 12 193 19
0 90 92 6 95
12
E
51P 136
3
162
13 13
9
36 28 1374 62
13 128 12 110 17 18
1 05 8 19 70 722
261
0
Cr
55 45 169 43
861 60 80 00 80
C C 18
59
493
0
834
55 58
0
10 60 2300 90
68 233 61 272 69 66 1280
0 90 0 50 60 0 840
Mn
161 24 14 139 96
4 80 86 9 2 A 20
30
183
8
281
20 18
0
40 15 2030 60
21 218 20 200 20 22 1800 248
0 50 0 50 40 0 000
166. 19 11 102 11 14 139 115
15 13 15 162 15 152. 14 15 160. 140.
Co
0 5.0 4.2 .1 4.5 7.4 .1 .6
8.0 5.5 137.0 0.0 .6 5.0 9 8.0 1.0 0 0
63 23 16 36 73 53 10 76 60 57
Ni
234 1 8 261 3.1 9 597 785 6 9 398 46 600 5 621 9 9 682 52.0
 ACCEPTED MANUSCRIPT

30 19 144 10 26 136 18 17 15 23 185. 33 27 103. 112.

Zn
810 6 1.4 .0 9.3 8 202 .1 8.0 7.0 128.0 0.8 234 5 3 3 7 0 0
17.6 4.6 13. 13. 21. 15. 19. 23. 14. 16. 23. 14. 14. 12.0 24. 22.
Ga
0 6 59 40 3 01 40 0 00 24 13.10 3 24 67 5 5 4 bdl bdl
98 137 15 51 46 60 45 45 84 88
Zr
4.69 bdl 2 1 23 8 831 735 7 8 428 6 240 0
P T
238 1 5 bdl 8.40
Nb
122. 67. 18
5 5 49
249
0
22
40
12
92
162
0
193
0
15
11
17
00 1410
13
14 774
R I
13
59
21 18
859 50 90 bdl 470
Hf
0.41
0 bdl 6
30. 45.
3
75.
5
17.
60
31.
9
32.
2
18.
70
25.
9 16.30 70C
17.
S
9.8
0
17.
60
34. 30.
8.94 6 1 bdl
0.06
00
Ta
1.8 19
5.94 80 0.1
35
281 3
14
4.2
193
.0 217
16
1.0
19
9.0
N172.0 U 15
8.0
104
.7
16
3.0
122. 27 23
8 1 8 bdl 87.0

U
0.1
09 0.1 0.0
M A
bdl bdl 2 200 bdl bdl bdl 844 bdl bdl bdl bdl bdl bdl bdl bdl bdl bdl 3.65
W
0.33 0.4 0.3 0.3
E D
0.2 0.2 0.3 0.16
7 bdl 79
Nb/T 20.6 35. 9.7
01
8.8
bdl bdl 23

P T
6.3 8.9 8.3
73
8.9
bdl
9.3
bdl
8.5
bdl 13
8.3
bdl
7.3
41
8.3
90 bdl bdl
7.9 7.9
bdl 4.10

a
Zr/N
0 90 3
0.0 0.5
6
0.5
C
5 6
E 9
0.6 0.4 0.5
1
0.3
9
0.3
4
0.3
8.20 2
0.3
9
0.3
4
0.3
7.00 3 4
0.3 0.4
- 5.40

Zr/Hf
0.04 0
11.4
4 - 10
3
32.
A
5

30
C
30.
8 0

20 40 10
1
20. 29. 26.
8
22.
80
1
25.
00
6
23.
50
0.30 5
25.
26.30 80
1
24.
50
3
25.
60
0.28 9 7
26.6 24. 29.
0 30 40
-

-
0.02
140.
00
Abbreviations: Compositional groups: 1. Ilmenite sensu strictu; 2. Fe3+-rich ilmenite; 3. Mg-rich ilmenite and 4. Mn-rich ilmenite. Textures: met.1 =
metasomatic ilmenite (with apatite, clinopyroxene and amphibole); met.2 = metasomatic ilmenite (veinlets in peridotites); exsol = ilmenite with
hematite exsolutions; sympl = symplectitic ilmenite; Mg-rich = Mg enrichment in ilmenites,usually already rich in Mg (darker areas in BSE images);
recrist. = ilmenite nodules showing recristallisation
 ACCEPTED MANUSCRIPT

PT
Graphical abstract

RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Highlights:

 Ilmenite genesis in kimberlites is a complex, multi-stage process

 Mg-rich ilmenite is produced by interaction with the kimberlitic magma
 Conventional discriminatory diagrams cannot be used to assess the diamond grade
 Combined textural and trace-element studies are essential for diamond exploration
 Mn-rich ilmenite should be disregarded as diamond indicator mineral in Angola

PT
RI
SC
NU
MA
E D
PT
CE
AC