Content uploaded by Martin Peter Smith
Author content
All content in this area was uploaded by Martin Peter Smith on Mar 28, 2018
Content may be subject to copyright.
The origin of secondary heavy rare earth element enrichment in
carbonatites: Constraints from the evolution of the Huanglongpu
district, China
M. Smith
a,
⁎, J. Kynicky
b
,ChenXu
c
, Wenlei Song
c
,J.Spratt
d
,T.Jeffries
d
, M. Brtnicky
b
,
A. Kopriva
b,e
, D. Cangelosi
f
a
School of Environment and Technology, University of Brighton, Brighton, UK
b
Mendel University in Brno, Zemedelska 1, CZ-613 00 Brno, Czech Republic
c
Peking University, School of Earth & Space Science, Laboratory for Orogenic Belts & Crustal Evolution, Beijing 100871, PR China
d
Department of Earth Sciences, The Natural History Museum, London, UK
e
TESCAN Brno, s.r.o., Libusina 1, CZ-623 00 Brno, Czech Republic
f
School of Earth and Environment, University of Leeds, Leeds LS2 9JT, UK
abstractarticle info
Article history:
Received 27 October 2017
Accepted 25 February 2018
Available online 04 March 2018
The silico-carbonatite dykes of the Huanglongpu area, LesserQinling, China, are unusual in that they are quartz-
bearing, Mo-mineralised and enriched in the heavy rare earth elements (HREE) relative to typical carbonatites.
The textures of REE minerals indicate crystallisation of monazite-(Ce), bastnäsite-(Ce), parisite-(Ce)
and aeschynite-(Ce) as magmatic phases. Burbankite was also potentially an early crystallising phase.
Monazite-(Ce) was subsequently altered to produce a second generation of apatite, which was in turn replaced
and overgrown by britholite-(Ce), accompanied by the formation of allanite-(Ce). Bastnäsite and parisite where
replaced by synchysite-(Ce) and röntgenite-(Ce). Aeschynite-(Ce) was altered to uranopyrochlore and then
pyrochlore with uraninite inclusions. The mineralogical evolution reflects the evolution from magmatic
carbonatite, to more silica-rich conditions during early hydrothermal processes, to fully hydrothermal conditions
accompanied by the formation of sulphate minerals. Each alteration stage resulted in the preferential leaching of
the LREE and enrichment in the HREE.Mass balance considerations indicate hydrothermal fluids must have con-
tributedHREE to the mineralisation. The evolution of the fluorcarbonate mineralassemblage requiresan increase
in a
Ca
2+
and a
CO
3
2−
in the metasomatic fluid (where ais activity), and breakdown of HREE-enriched calcite
may have been the HREE source. Leaching in the presence of strong, LREE-selective ligands (Cl
−
) may account
for the depletion in late stage minerals in the LREE, but cannot account for subsequent preferential HREE addition.
Fluid inclusion data indicate the presence of sulphate-rich brines duringalteration, and hence sulphate complex-
ation mayhave been important forpreferential HREEtransport. Alongside HREE-enriched magmatic sources,and
enrichment during magmatic processes, late stage alteration with non-LREE-selective ligands may be critical in
forming HREE-enriched carbonatites.
© 2018 Published by Elsevier B.V.
Keywords:
Rare earth elements
Carbonatite
Hydrothermal
Sulphate
Brine
1. Introduction
The rare earth elements (REE) (here defined as the lanthanides plus
Y), have long been a focus of research because of their utility as tracers
for geochemical processes from magma source and differentiation,
through to hydrothermal system evolution (e.g. Chakhmouradian,
2006;Haas et al., 1995;Henderson, 1984;Migdisov et al., 2009). In re-
cent years they have also become a focus of intense economic interest
because of the restriction in supply caused by a focus of production
in China, coupled with increasing use in renewable energy and high
technology applications (Chakhmouradian and Wall, 2012). Rare earth
element resources are dominated by concentrations, either from mag-
matic or hydrothermal processes, associated with alkaline igneous
rocks, notably carbonatites. A significant issue with primary resources
is that carbonatite concentrations tend to be preferentially enriched in
the light REE (La to Nd), which, with the exception of Nd, are less eco-
nomically important than the heavy REE (Eu to Lu) (Chakhmouradian
and Wall, 2012). Middle to HREE enriched carbonatites do occur how-
ever, and are of great interest both because of their economic potential,
and because of the need to understand the processes leading to
this enrichment (e.g. Xu et al., 2007). Hydrothermal processes have
been shown to fractionate the REE via differential solubility during
leaching of primary phases, controls on relative solubility exerted by
Lithos 308–309 (2018) 65–82
⁎Corresponding author.
E-mail address: martin.smith@brighton.ac.uk (M. Smith).
https://doi.org/10.1016/j.lithos.2018.02.027
0024-4937/© 2018 Published by Elsevier B.V.
Contents lists available at ScienceDirect
Lithos
journal homepage: www.elsevier.com/locate/lithos
the variation in the stability of aqueous complex ions with different
ligands (Haas et al., 1995;Migdisov et al., 2009;Wood, 1990), variation
in the relative solubility of secondary phases, and all of these coupled
with dynamic flow and reaction pathways (Williams-Jones et al., 2013).
In this study we have investigated the paragenetic and chemical evolu-
tion of rare earth minerals in the Huanglongpu carbonatites, Qinling
Mountains, China (Fig. 1;Xu et al., 2007, 2010;Song et al., 2015). These
carbonatites are exceptional because of their association with economic
molybdenite mineralisation, a relatively silica-rich geochemistry and
hence an association with quartz, and a relative enrichment in the
HREE compared to typical carbonatites (Kynicky et al., 2012). The excep-
tional geochemistry of the dykes, and the preservation of a wide range
of reaction textures, make this an ideal site to assess the role of post-
magmatic processes in the genesis of HREE-enriched carbonatites glob-
ally. Heavy rare earth enrichment has been reported from a number of
carbonatites, typically associated with late stage hydrothermal phenom-
ena (e.g. Tundulu - Ngwenya, 1993;Songwe-Swinden and Hall, 2012;
Bear Mountain - Moore et al., 2017;Lofdal-Bodeving et al., 2017). Bulk
rock analyses of the Huanglongpu carbonatites show that their REE
patterns are transitional between more typical carbonatites and the
HREE-enriched hydrothermal deposits (Fig. 2). Thus the Huanglongpu
district may provide an exceptionally well exposed site, suitable for
characterisation of hydrothermal enrichment process, with wider im-
plications for HREE-enrichment in carbonatites globally.
2. Geological setting
The Huanglongpu district is composed of four carbonatite-related
orefields with a total ore reserve of N180 Kt of Mo. Almost all spatially
Fig. 1. (A) Geological mapof the eastern Qinling mountainsshowing the location of the Huanglongpu area. Inset showsthe overall location withinChina in relationto major crustal blocks.
(B) Geological map of the Huanglongpu Mo-district (highlighted in bold in A).Maps adapted from Xu et al. (2010). Samples labelled HLP are fromthe Dashigou site, sampleslabelled YT are
from Yuantao.
Fig. 2. Comparison of ra tios characteris ing the REE pattern for the Huanglongpu
carbonatites with Chinese REE mineralised carbonatites (Kynicky et al., 2012), typical
carbonatites (H ornig-Kjarsgaa rd, 1998), and carbonatites with reported high HREE
contents (Bodeving et al., 2017;Moore et al., 2017;Ngwenya, 1993;Swinden and Hall,
2012) Data from Huanglongpu from Xu et al. (2010),Kynicky et al. (2012) and Song
et al. (2015). Chondrite values from Sun and McDonough (1989) here and throughout.
66 M. Smith et al. / Lithos 308–309 (2018) 65–82
Table 1
Localities and descriptions of samples used in this study.
Locality⁎Deposit Sample description Analysed REE and associated minerals#
Longitude Latitude
HLP1 410,530 3,803,439 Dashigou Sulphide mineralised calcite-quartz carbonatite. Mnz, Bst, Par, All
HLP2 410,530 3,803,439 Dashigou Sulphide mineralised calcite-quartz carbonatite. Mnz, Ap, Brh, All, Par, Syn
HLP3 410,530 3,803,439 Dashigou Sulphide mineralised calcite-quartz carbonatite. NA
HLP4 410,530 3,803,439 Dashigou Sulphide mineralised calcite-quartz carbonatite. NA
HLP5 410,530 3,803,439 Dashigou Sulphide mineralised carbonatite. Par, Syn, Rnt, Xen
HLP6 410,530 3,803,439 Dashigou Barite altered calcite carbonatite. Mnz, Bst, Par, Xen
HLP7 410,530 3,803,439 Dashigou Sulphide mineralised calcite-quartz carbonatite. Mnz
HLP10 410,530 3,803,439 Dashigou Sulphide-calcite veinlet cutting calcite-quartz carbonatite. Mnz, Ap, Brh, All, Par, Syn
HLP15 410,530 3,803,439 Dashigou Sulphide mineralised calcite-quartz carbonatite. Mnz, Pyc, Aesch, Par, Syn, Rnt
HLP A 409,886 3,804,124 Dashigou Calcite-quartz carbonatite. Ap
HLP B 409,886 3,804,124 Dashigou Molybdenite mineralised calcite-quartz carbonatite. Mnz
HLP C 409,886 3,804,124 Dashigou Contact zone between calcite carbonatite and quartz-core. Mnz
YT1 410,732 3,806,281 Yuantao Calcite-quartz carbonatite with minor sulphide veining. Mnz
YT2 410,732 3,806,281 Yuantao Sulphate (barite-celestine) rich carbonatite. Pyc, Aesch, Mnz, Bst, Par, Syn, Rnt
YT6 410,732 3,806,281 Yuantao Calcite-quartz carbonatite with minor sulphide veining. Par, Syn, Rnt, Xen
#Mineral abbreviations –Mnz –Monazite; Bst –Bastnäsite; Par –Parisite; Syn –Synchysite; Rnt –Röntgenite; All –Allanite; Brh –Britholite; Xen –Xenotime; Ap –apatite; Pyc –
Pyrochlore; Aesch –Aeschynite.
⁎Grid reference format WGS84 sector 49S.
Fig. 3. Field relationships from Dashigou open pit. (A) Highly parallel carbonatite dykes and veins. Field of view ~30 m. Dykes are upto 3 m diameter, with narrowest examples 2–3cm.
(B) Internal structure of a single dyke with carbonate concentrated on the margins and quartz in the centre. Sulphide mineralisation concentrated at the margin between quartz and
carbonate. Margins are fenitised. (C) Composite dyke showing multiple opening stages. (D) Intersecting dykes in conjugate orientation, but with offset of one dyke by the other.
(E) Intersection of carbonate-quartz-molybdenite dyke with cross cutting barite-calcite vein/dyke. (F) Individual dyke section showing transition from carbonate-rich margins
associated with K-feldspar, through a sulphide (pyrite-molybdenite-galena) zone to a central quartz-rich zone. Cc –calcite; Qtz –quartz; Mly –molybdenite; Brt –barite.
67M. Smith et al. / Lithos 308–309 (2018) 65–82
associated porphyry and porphyry-skarn Mo deposits in the Qinling belt
formed in the Late Jurassic-Early Cretaceous, as indicated by molybde-
nite Re-Os dates ranging from 148 to 112 Ma (Huang et al., 1994;Mao
et al., 2008), whereas molybdenite from the Huanglongpu deposits
is much older, yielding Re-Os ages from 209 to 221 Ma and monazite
U-Pb and Th-Pb ages from 208.9 ± 4.6 Ma and 213.6 ± 4.0 Ma (Huang
et al., 1994;Song et al., 2016;Stein et al., 1997). Alkali granites and
syenites have been identified in the area, but these are peraluminous,
inferred to be derived from similarsources to the Mo-bearing granitoids
(typically granodiorite) and have zircon U-Pb ages of 131 ± 1 Ma
(Zhao et al., 2010), and are therefore unrelated to the carbonatites and
REE-mineralisation. All these data indicate that the mineralisation of
the dykes is not a result of overprint by subsequent magmatism.
The Huanglongpu deposits belong spatially and genetically to the
Qinling orogenic belt which is subdivided into two main parts, North
Qinling and South Qinling, separated by the Shangdan suture (Fig. 1).
The northern border of the North Qinling is marked by a relatively
narrow, straight, steep north-dipping fault zone, the Machaoying
fault zone, which is strongly temporally and spatially associated with
the Cenozoic rift basin in the north. The southern border of the South
Qinling is separated from the South China block by the Mianlue suture.
The detailed geological framework and tectonic evolution of the
Qinling region have been described by Xue et al. (1996),Meng and
Zhang (2000),andRatschbacher et al. (2003), and also in the context
of carbonatite evolution by Xu et al. (2010, 2014) and Song et al.
(2015).
The ore bodies occur discontinuously over a total distance of 6 km
and an area of 23 km
2
. The distribution of carbonatites is predomi-
nantly controlled by a northwest-striking, extensional fault zone. The
Huanglongpu Mo deposits consist of separate mineralised bodies at
Yuantou, Wengongling, Dashigou, Shijiawan I, II, Taoyuan and Erdaohe
(Fig. 1b). With the exception of Shijiawan I, which is hosted by granite
porphyry, the rest of the deposits are associated with carbonatite
dykes (Xu et al., 2007, 2010). The samples studied here come from the
Yuantao (YT) and Dashigou (HLP) deposits (Table 1).
The carbonatite dykes are highly parallel, predominantly dipping
N to NNW, at steep angles (strike/dip ~260/50–88°N) and consist of
calcite, kutnahorite, quartz, potassium feldspar, barite, pyrite, galena,
sphalerite, molybdenite, monazite, Ca-REE-fluorcarbonates, apatite,
britholite, pyrochlore, uraninite, REE fluorides, burbankite, celestine,
strontianite and brannerite. Minor fluorite is found at Shijiawan and
Yuantao. The dykes range in thickness from ~10 m to ~0.1 m (Fig. 3A-C)
with lateral extents ranging from 10 m to N1 km. Rarer dykes occur in
orientations conjugate to the main set (strike/dip ~350/50–80°E;
Fig. 3D). Minor offsets suggest this set may be slightly later than the
main set, although the conjugate orientation and presence of conjugate
veins merging with the main set (Fig. 3E) suggest they werevery nearly
contemporaneous. Where cross-cutting relationships can be clearly
observed, barite-celestine bearing dykes and veins cut the earlier
carbonate-quartz-sulphide bearing dykes (Fig. 3E). Alteration enve-
lopes consisting of biotite, epidote, pyrite and anhydrite are developed
at the selvages of the dykes (Fig. 3B). The carbonatite dykes are domi-
nated by calcite and other carbonates along dyke margins, with central
infills of quartz. Quartz is dominantly restricted to cores of dykes and
may be mainly hydrothermal in origin (Fig. 3B), although some dykes
are composite and the result of repeated opening events (Fig. 3C).
This suggests the primary dykes may have been calico-carbonatite in
composition although the bulk dyke composition is silico-carbonatite.
In the carbonatites, molybdenite occurs mainly as disseminated grains
and intergranular and fracture-hosted films (Fig. 3F), sometimes
associated with pyrite, galena and sphalerite suggesting a subsolidus,
hydrothermal origin for at least some of the sulphide assemblage.
Disseminated molybdenite is also found along fractures in fenitized
gneiss near its contact with the dykes.
Fig. 4. Generalised summary of the major mineral paragenesis in the Huanglongpu deposits based on this study and observations in Xu et al. (2010, 2014),andSong et al. (2015).
68 M. Smith et al. / Lithos 308–309 (2018) 65–82
3. Methods
3.1. Scanning electron microscopy
Scanning electron microscopic (SEM) observation of samples
was carried out at Mendel University, Brno, Czech Republic, and the
University of Brighton, UK. In Brno a FEG-SEM TESCAN MIRA 3 XMU
was used. X-ray spectra were collected using either a Bruker Quantax
800 or Oxford Instruments XMax 80 EDXspectrometer using an acceler-
ating voltage of 25 kV and a beam current of 5 nA. At Brighton a Zeiss
EVO LS 15 SEM equipped with an Oxford Instruments XMax 80 EDX
spectrometer was used, at 20 kV accelerating voltage and 1.2 nA beam
current.
3.2. Electron microprobe
The major-element compositions of selected major and accessory
mineral phases from the Huanglongpu carbonatites were measured
by wavelength-dispersive X-ray spectrometry (WDS) using Cameca
SX100 electron microprobes at the Joint Electron Microscopy and
Microanalysis Laboratory, Institute of Geological Sciences, Masaryk
University and Czech Geological Survey, and the Natural History
Museum (NHM), London. At Masaryk University the instrument was
operated at a beam current of 10 nA and an accelerating voltage of
15 kV, and at the NHM at 20 nA and 20 kV. Where possible the beam
was defocused to a spot size of 5–10 μm to minimize beam-induced
damage and thermal decomposition of fluorcarbonate phases. The
Fig. 5. Back scattered electron images of early magmatic phases. (A) Monazite within calcite carbonatite, with overgrowth and fracture fill of molybdenite. (B) Apatite in calcite-quartz
carbonatite. (C) Zircon with xenotime fracture fill in calcite. (D) Isolated, patchily zoned xenotime in calcite. (E) Possibly primary parisite in calcite carbonatite. (F) Pyrochlore with
uraninite inclusions overgrown by molybdenite, galena and REE fluorcarbonates, all in calcite carbonatite. (G) Backscattered electron image of intergrown barite and ancyltite in calcite
carbonatite. The barite-ancylite intergrowth is inferred to represent a pseudomorph after early burbankite. (H) Association of parisite-(Ce) with galena and molybdenite. Mnz –
monazite; Cc –calcite; Ap –apatite; Qtz –quartz; Phl –phlogopite; Zrn –zircon; Xnt –xenotime; Par –parisite; Pyr –pyrochlore; Urn –uraninite; PbS –galena; MoS –molybdenite;
Fc –REE fluorcarbonate. Brt –barite; Anc - ancylite.
69M. Smith et al. / Lithos 308–309 (2018) 65–82
standards and emission lines used are summarised in E-Appendix 1,
alongside typical detection limits for fluorcarbonate phases. The data
were reduced and corrected using the PAP routine (Pouchou and
Pichoir, 1984). Inter-element peak interferences, particularly for the
REE, were corrected for by analysis of the pure standards following
methods outlined by Williams (1996).
3.3. La-ICPMS
Trace-element analysis of selected carbonates, oxides and phosphates
by laser-ablation inductively-coupled-plasma mass-spectrometry (LA-
ICP-MS) was performed at the Laboratory of Atomic Spectrochemistry,
Masaryk University, and the NHM, London using New-Wave UP-213
Frequency 348 quintupled Nd:YAG laser systems operated at a wave-
length of 213 nm and pulse duration of 4.2 ns. At Masaryk an Agilent
7500ce spectrometer was used, and at the NHM an ICP-MS 349 –
ThermoElemental PlasmaQuad III. Helium was used as a carrier gas
with a flow rate of 1 l/min. The samples were analysed using a spot
diameter of 25 or 30 μm, dwell time of 60 s, repetition rate of 10 Hz and
fluence of 5 J/cm
2
. The calcium content determined by WDS was used
as an internal standard for all carbonates and phosphates, and external
calibration was performed using glass standards NIST 610 and 612. For
low Ca or Ca-free phases the internal standard used was Ce. Columbia
River Basalt glass BCR2G was analysed as a quality check. Within run
reproducibility of NIST612 analyses was typically within 1% of the pub-
lished values, and of BCR2 within 5% of published values.
4. Results
4.1. Mineralogical evolution of the Huanglongpu system
The paragenetic evolution of the Huanglongpu system is summarised
in Fig. 4 and the REE mineral assemblage is illustrated in Figs. 5 and 6.
Early REE mineral crystallisation is characterised by the occurrence
of sub- to euhedral monazite-(Ce), parisite-(Ce), bastnäsite-(Ce) and
REE-bearing apatite (Fig. 5A, B). The REE are also hosted in zircon and
xenotime (Fig. 5C, D). Xenotime is developed both as individual grains
(Fig. 5D) and hosted in fractures in zircon (Fig. 5C). The patchy zonation
Fig. 6. Back scattered electron images of reaction textures showingthe development of thesecondary REE mineralassemblage. (A) Apatite corona developed on monazite associated with
britholite in calcitecarbonatite. (B) Britholite overgrowth and replacement ofapatite with relict monazite in apatite core. In calcite-quartz carbonatite. (C) Allanite overgrowth on apatite,
with subsequent overgrowth and replacement by parisite. (D) Synchysite overgrowth on monazite. (E) Bastnäsite replacement by parisite-synchysite syntaxial intergrowth. All are then
replaced by röntgenite. (F) Alteration of aeschynite to uranopyrochlore and pyrochlore Abbreviations as Fig. 5,plusBrh–britholite; Fld –K-feldspar; Snc –synchysite; Rnt –röntgenite;
Aesch –aeschynite; Upyc –uranopyrochlore; Pyc –pyrochlore; Urn –uraninite.
70 M. Smith et al. / Lithos 308–309 (2018) 65–82
of xenotime (Fig. 5D) suggests it may have been subject to alteration and
dissolution-reprecipitation processes. Xenotime was not studied in de-
tail here as it does not constitute a major contributor to the HREE budget.
Potentially primary parisite-(Ce) (indicated by the lack of alteration to
other Ca-REE fluorcarbonates) occurs within the carbonate-dominant
zone of the dykes (Fig. 5E) in association with sulphides. Early magmatic
crystallisation also included the formation of Nb minerals, dominated by
Table 2
Representative bulk rock analyses of Huanglongpu carbonatites. Data from Xu et al. (2010),
Kynicky et al. (2012) and Song et al. (2015).LREE–La to Sm; HREE –Eu to L u + Y.
Sample HLP-1 HLP-3 HLP-4 HLP-5 HLP-6
Major elements (weight %)
SiO2 1.18 1.56 1.07 1.27 4.26
TiO2 0.08 0.10 0.09 0.07 0.18
Al2O3 2.10 2.24 1.54 1.25 6.36
Fe2O3 0.37 0.51 0.41 0.48 0.35
MnO 1.00 2.59 1.47 1.29 1.20
MgO 0.40 0.44 0.42 0.41 0.33
CaO 53.35 51.39 53.01 53.03 48.41
Na2O 0.08 0.15 0.08 0.09 0.18
K2O 0.18 0.06 0.04 0.02 1.03
P2O5 0.04 0.05 0.03 0.03 0.05
CO2 40.94 40.65 41.48 41.67 37.37
Total 99.71 99.74 99.63 99.61 99.72
Trace elements (ppm)
Rb 3.0 0.9 0.4 0.3 5.8
Ba 912.0 154.0 184.0 197.0 510.0
Th 0.8 0.2 0.0 0.2 0.6
U 0.6 0.9 0.3 0.9 0.2
Nb 1.2 0.5 0.4 0.9 0.4
Ta 0.2 0.4 0.3 0.3 0.2
La 220.0 130.0 279.0 140.0 186.0
Ce 516.0 445.0 764.0 516.0 527.0
Pr 47.9 46.1 76.9 53.2 52.4
Sr 7753.0 6938.0 8441.0 7096.0 7957.0
Nd 200.0 210.0 336.0 240.0 230.0
Sm 41.7 58.2 71.9 60.1 50.9
Zr 0.4 0.4 0.3 0.2 29.3
Hf 0.2 0.5 0.3 0.4 0.8
Eu 11.3 17.9 18.6 17.2 13.5
Gd 38.0 61.2 60.7 56.8 45.5
Tb 5.7 11.3 8.8 9.5 6.8
Dy 34.3 77.4 49.6 59.7 41.6
Y 365.0 841.0 426.0 589.0 421.0
Ho 8.3 19.2 10.9 13.6 9.7
Er 28.6 67.3 34.2 44.9 32.9
Tm 4.9 11.8 5.2 7.5 5.5
Yb 35.3 86.0 34.7 53.0 37.9
Lu 5.6 13.3 4.8 8.1 5.6
ΣREE 1563 2096 2181 1869 1666
HREE/LREE 1.91 0.74 2.34 1.17 1.69
Table 3
Representative major element (WDS and EDS) and trace element (LA-ICPMS) analyses of calcite.
LA-ICPMS DL# HLPB HLPB HLP5 HLP5 YT6 YT6 HLP10 HLP7 HLP7
Major element concentration in weight %.
Technique EDS EDS EDS EDS EDS EDS EDS WDS WDS
CaO 51.84 54.23 52.88 53.61 53.30 52.26 53.28 53.38 52.86
MgO 0.51 0.00 0.31 0.31 0.12 0.00 0.00 0.40 0.28
MnO 4.14 3.29 3.44 2.92 2.77 4.02 3.34 2.67 2.02
FeO 0.53 0.00 0.29 0.00 0.81 1.17 0.00 0.29 0.28
SrO 0.61 0.00 0.70 0.72 0.64 0.26 1.11 1.29 1.27
LA-ICPMS concentration in ppm.
La 0.03 29 59 4 95 2 47 14 30 152
Ce 0.03 107 132 9 169 6 96 39 53 179
Pr 0.03 17 19 2 20 1 13 6 7 18
Nd 0.31 97 83 8 73 3 53 29 29 43
Sm 0.17 29 21 3 13 2 9 6 7 8
Eu 0.06 9 7 1 4 1 3 2 2 3
Gd 0.24 37 30 3 25 3 10 7 6 8
Tb 0.03 6 5 1 2 1 2 1 1 1
Dy 0.14 44 40 6 16 14 19 7 8 5
Y 0.03 424 373 76 161 83 129 86 83 62
Ho 0.03 11 10 2 4 4 5 2 2 1
Er 0.15 42 38 9 15 14 17 9 5 8
Tm 0.03 7 7 2 3 3 3 2 2 2
Yb 0.14 55 55 25 23 22 19 16 15 13
Lu 0.02 9 9 6 4 3 3 3 4 3
Th 0.03 BD
a
BD
a
BD
a
BD
a
BD
a
BD
a
BD
a
BD
a
BD
a
U 0.03 BD
a
BD
a
BD
a
BD
a
BD
a
BD
a
BD
a
BD
a
BD
a
#DL - Detection Limit
a
Below detection.
Fig. 7. (A) Chondrite normalised REE content of calcite from LA-ICPMS analyses. (B) MnO
content of calcite from EPMA.
71M. Smith et al. / Lithos 308–309 (2018) 65–82
aeschynite-(Ce), but also including pyrochlore (Fig. 5F). Early stage mag-
matic crystallisation is also inferred to have included the formation of
primary Na-Ba-Sr-REE carbonates of the burbankite group. Potential
burbankite pseudomorphs are preserved as coarse grained, hexagonal
pods (up to 2 cm in diameter) in-filled by barite, or a solid solution of
barite-celestine, strontianite and ancylite (Fig. 5G). Burbankite has not
been directly observed, but this texture and assemblage is typical of
burbankite replacement (Zaitsev et al., 1998, 2002).
The subsequent development of the REE mineral paragenesis is re-
corded by reaction textures in phosphates and silicates, the colloform
banded infill of niobate pseudomorphs, and the development of syn-
to epitaxial alteration and overgrowth in the fluorcarbonates. The
fluorcarbonate assemblage appears to be related to the formation of
sulphides (notably galena and molybdenite; Fig. 5H). The earliest
of these textures involves the partial replacement and overgrowth of
monazite-(Ce) by REE-enriched apatite (Fig. 6A). The apatite is zoned,
with more REE-depleted areas developed towards the outer rim of the
coronae. The apatite is replaced in some instances by an outer layer of
britholite-(Ce), sometimes giving multi-layered structures (Fig. 6B).
Britholite-(Ce) is sometimes at the contact with quartz within the
dyke cores. These outer layers are frequently associated with molybde-
nite. Potentially synchronous with the development of britholite-(Ce)
is the formation of allanite-(Ce), which in some instances overgrows
apatite. In at least one case this is associated with the development of
later stage parisite-(Ce) (Fig. 6C), and Ca-REE-fluorcarbonates may over-
grow all the preceding phases (e.g. Fig. 6D). The syntaxial intergrowth
and epitaxial overgrowth patterns of the Ca-REE-fluorcarbonates allow
a sequence of their formation to be determined, with early parisite-
(Ce), sometimes intergrown with, and sometimes overgrown and partly
replaced by, bastnäsite-(Ce). Parisite-(Ce) also occurs intergrown and
overgrown by synchysite-(Ce), and ultimately all are overgrown and
partially replaced by röntgenite-(Ce) (Fig. 6E). Synchysite-(Ce) and
röntgenite-(Ce) are always enriched in Y relative to bastnäsite-(Ce)
and parisite-(Ce), and in rare examples occur as the Y-dominant phase
(synchysite-(Y)). Textures involving fluorcarbonate overgrowth are
typically developed within calcite.
Fig. 8. Chondrite normalised REE + U + Th content of REE minerals. Grey shading indicates the range of analyses from corresponding EPMA spots. (a) Monazite. (b) Pa risite and
Bastnäsite. (c) Apatite subdivided by discrete grains (Primary) and apatite involved in reaction textures. (d) Britholite. (e) Allanite. (f) Synchysite and röntgenite.
72 M. Smith et al. / Lithos 308–309 (2018) 65–82
The textural evolution of the niobium mineral paragenesis parallels
that of the REE minerals. Euhedral pseudomorphs are commonly associ-
ated with monazite (Fig. 6F). The initial Nb phase seems to have been
aeschynite-(Ce), on the basis of well crystalline, subhedral relicts within
larger pseudomorphs (Fig. 6F). These early Nb-phases formed either
synchronously with monazite, or pre-monazite crystallisation as
indicated by the overgrowth of pseudomorphous aggregates of Nb-
minerals by monazite-(Ce). Aeschynite-(Ce) is subsequently partially
replaced by uranopyrochlore, and then both are replaced and over-
grown by pyrochlore. The replacement of the primary aeschynite-(Ce)
is accompanied by the development of uraninite inclusions (Fig. 6F).
4.2. REE geochemistry
The full results of LA-ICPMS and EPMA are available in Electronic
Appendices 2 and 3. The results of LA-ICPMS and EPMA analyses are
generally in good agreement for the LREE. Agreement is less good for
the M-HREE and low levels of U and Th where concentrations approach
the detection limit for EPMA for the conditions used. In these cases the
LA-ICPMS data are the most accurate values. The values are compared
in E-Appendix 3. Variation between EPMA and LA-ICPMS for corre-
sponding points is a function of the relative volume analysed in each
case, and theabsolute precision of each technique. Low analytical totals
for the fluorcarbonate minerals andniobate minerals occur in the EPMA
data. In the case of the fluorcarbonates this is because a narrow beam
diameter was used in order to analyse the fine intergrowth in some
reaction textures, resulting in beam damage during analysis. In the
case of the niobates low analytical totals are typical for EPMA analyses
of pyrochlore and aeschynite because of their susceptibility to hydration
and metamictisation (Lumpkin and Ewing, 1992, 1995).
The whole rock geochemistry of the Huanglongpu carbonatites
is depleted in the LREE and enriched in the HREE compared to most
carbonatite occurrences (Xu et al., 2010;Kynicky et al., 2012;Song
et al., 2015;Table 2;Fig. 2). This is also reflected in the trace element
chemistry of calcite (Table 3), which in the majority of the dykes
shows relatively flat REE patterns, but may vary from LREE-enriched
with flat HREE segments to strongly HREE-enriched (Fig. 7). This varia-
tion reflects variation in calcite generations, as HREE enriched patterns
occur in calcite adjacent to the late stage Ca-REE fluorcarbonates.
The REE patterns of REE-minerals and apatite show dramatic changes
with evolution of the mineral paragenesis (Fig. 8;Tables 4-6). Both
monazite-(Ce) (Table 4) and primary bastnäsite-(Ce) and parisite-
(Ce) (Table 5) show strongly LREE-enriched patterns with very minor
negative anomalies at Eu relative to Sm and Gd, and Y relative to Dy
and Ho (Fig. 8A, B). Primary apatite (Table 4) shows relatively flat to up-
wards convex patterns with variably negative or positive Eu anomalies.
In contrast secondary apatite in coronas around monazite is LREE-
enriched, with flat HREE patterns from Dy to Lu and a consistently
negativeEu anomaly (Fig. 8C). Similar patterns, albeit at higher absolute
concentration, are observed in britholite and allanite (Fig. 8D, E;
Table 6). Secondary REE-fluorcarbonate minerals either show patterns
comparable to primary bastnäsite-(Ce) and parisite-(Ce), or show very
strong HREE-enrichment, most notably in synchysite-(Y) replacing ear-
lier phases, and in röntgenite (Fig. 8F; Table 5).
For economic considerations it is also important to quantify the
behaviour of U and Th in the REE phases and during the development
of the paragenesis. This is important as enrichment of the HREE may
be characteristically accompanied by enrichment in U, and particularly
in Th (e.g. Lofdal, Namibia; Wall et al., 2008; and Tundulu, Malawi;
Broom-Fendley et al., 2016). For this reason chondrite normalised
Table 4
Representative analyses of monazite (M) and apatite (A) by electron microprobe (EPMA) and LA-ICPMS.
EPMA DL
wt%
LA-ICPMS DL HLP1
M-Grain
HLP10
M-Corona
HLP-B
M-Grain
HLP15
M-Grain
HLP7
M-Corona
YT2
M-Grain
YT1
M-Grain
HLP10
A-Corona
HLP10
A-Corona Rim
HLP-A
A-Grain
CaO 0.03 0.16 0.87 0.19 0.21 0.18 1.10 0.38 49.12 52.79 55.13
SrO 0.05 B.D. B.D. B.D. B.D. B.D. B.D. B.D. 0.71 0.59 B.D.
MnO 0.03 B.D. B.D. B.D. B.D. B.D. B.D. B.D. 0.23 0.10 0.06
La2O3 0.08 19.17 20.77 22.04 22.43 21.17 20.85 21.87 1.27 0.69 B.D.
Ce2O3 0.10 32.74 33.48 34.22 34.85 34.99 34.19 33.25 2.64 1.34 B.D.
Pr2O3 0.16 3.06 2.54 2.80 2.79 3.137 2.85 2.79 0.27 0.09 B.D.
Nd2O3 0.08 9.82 7.77 8.01 8.08 8.45 8.44 8.06 0.84 0.32 B.D.
Sm2O3 0.04 1.04 0.52 0.55 0.66 0.72 0.65 0.66 0.10 B.D. B.D.
ThO2 0.09 1.00 0.77 0.62 0.27 0.45 0.35 0.43 B.D. B.D. B.D.
SiO2 0.03 0.19 0.80 0.57 BD 0.53 0.56 0.55 2.07 0.43 0.34
P2O5 0.03 28.63 26.59 28.10 29.60 29.32 28.14 28.28 34.79 38.76 40.74
F 0.08 B.D. B.D. B.D. B.D. B.D. B.D. B.D. 4.40 5.20 5.38
Total 95.81 94.11 97.10 98.89 98.95 97.13 96.27 96.80 100.30 101.91
O=F 1.83 2.14 2.27
Total 95.81 94.11 97.10 98.89 98.95 97.13 96.27 94.97 98.16 99.65
ppm ppm EPMA EPMA LA-ICPMS EPMA EPMA EPMA EPMA LA-ICPMS LA-ICPMS LA-ICPMS
La 701 0.01 160,600 177,100 220,300 191,300 180,500 177,800 186,500 13,200 9250 51
Ce 867 0.01 279,500 285,800 291,700 297,500 298,700 291,900 283,800 27,400 19,300 161
Pr 1325 0.01 26,100 21,700 26,600 23,800 26,800 24,400 23,900 2800 1950 22
Nd 723 0.07 84,200 66,600 77,900 69,200 72,400 72,400 69,100 8550 6140 94
Sm 368 0.04 8980 4500 5920 5670 6200 5560 5700 862 755 23
Eu 151 0.01 BD
a
BD 920 BD 560 BD BD 159 165 27
Gd 66 0.04 4900 BD 2560 BD 1750 BD 2350 448 579 22
Tb 0.01 NA NA 141 NA 600 NA NA 38 67 4
Dy 1223 0.03 BD BD 334 BD BD BD BD 159 448 26
Y 1071 0.01 5800 BD 883 BD BD BD BD 1190 2502 164
Ho 0.01 NA
b
NA 34 NA NA NA NA 30 85 6
Er 1309 0.03 BD BD 49 BD BD BD BD 95 252 21
Tm 0.01 NA NA 3 NA NA NA NA 14 39 3
Yb 843 0.04 BD BD 9 BD BD BD BD 114 267 30
Lu 679 0.01 BD BD 1 BD BD BD BD 18 35 5
Th 786 0.01 8770 6755 5980 BD BD BD BD 16,054 5470 10
U 1492 0.01 BD BD 110 BD BD BD BD 49 135 114
# DL - Detection Limit.
a
Below detection.
b
Not analysed.
73M. Smith et al. / Lithos 308–309 (2018) 65–82
U and Th concentrations are plotted alongside the REE in Fig. 8. In all
cases Th is significantly enriched relative to U and the HREE, except
for in the secondary Ca-REE fluorcarbonates. In late stage synchysite
and röntgentite the pattern is reversed, and Th is relatively depleted.
The absolute variation in U and Th content is shown in Tables 4-7.
For apatite and britholite-(Ce) the principle substitution leading
to REE incorporation is of the form [Ca
2+
]
−1
[P
5+
]
−1
[REE
3+
][Si
4+
].
In allanite ΣREE is consistently close to 1 a.p.f.u. (atoms per formula
unit), but variation of the Fe:Al ratio indicates development of a ferri-
allanite substitution, and Fe
3+
dominant within the allanite structure.
In some instance more Mn-rich allanites are developed (Table 6).
Formulae calculations (E-Appendix 2) suggest there is an excess of F
over stoichiometric amounts in some apatite analyses. This is accompa-
nied by a deficiency in P, suggesting a francolite (carbonate apatite)
component, although there may be issues with F mobility under the
electron beam. Theoretical carbonate contents based on theP deficiency
are calculated in E-Appendix 2, but are only indicative concentrations as
they assume a full phosphate site occupancy.
The niobates within the Huanglongpu district are consistently
Nb-rich with minimal Ta content (Fig. 9A, B; Table 7). Compositions
show a continuous transition between aeschynite-(Ce) and pyrochlore,
with limited development of an A-site vacancy (Fig. 9C). Primary
aeschynite-(Ce) has relatively low U contents, whilst pyrochlore
shows a continuous variation in U content (Fig. 9D). In terms of REE pat-
tern aeschynite-(Ce)is consistently LREEenriched, and the REE patterns
is comparable to monazite-(Ce) and bastnäsite-(Ce) (Fig. 10A). The very
fine nature of the banding in the replacement products of aeschynite-
(Ce) restricted the applicability of LA-ICPMS, hence the analyses com-
pleted are dominantly by electron microprobe. The occurrence of a
limited number of coarse uranopyrochlore grains allowed some analyses
(Fig. 10B). These show that alteration of uranopyrochlore to pyrochlore
resulted in a relative enrichment in the HREE. The transition from
aeschynite-(Ce) to uranopyrochlore shows a slope of ~0.5 on a plot of
ΣREE vs U, and is consistent with REE leaching and the addition of U
during alteration according to the substitution [REE
3+
]
−2
[U
4+
][Ca
2+
]
(Fig. 10C).
5. Discussion
5.1. The evolution of the Huanglongpu REE mineralisation
The mineralogical evolution of the Huanglongpu REE system reflects
that of the host carbonatite. The primary REE mineralogy consists
of monazite and bastnäsite, with phases with no essential REE con-
tributing to the HREE budget (apatite and zircon). A major contribu-
tion to the primary REE budget may have come from the presence
of burbankite, now pseudomorphed by ancylite and barite. Early
Nb mineralisation appears to have been dominated by aeschynite,
followed by replacement by uranopyrochlore and finally pyrochlore.
The formation of a REE-rich A-B oxide at this stage relates to similar
melt compositions to those during monazite deposition. The saturation
of carbonatite melts in monazite and basnästite is typically related
to passive enrichment on fractional crystallisation (Le Bas, 1987;
Mitchell, 2005) which is a likely factor in this case (Xu et al., 2007).
Aeschynite-(Ce) saturation in melts has not been specifically addressed
in the literature, but the formation of Nb-minerals in general has been
related to saturation following crystal fractionation, magma mixing,
and redistribution of Nb-minerals by density currents (Mitchell,
2015). The reduction in niobate REE content through the paragenesis
mirrors the reduction seen in phosphates and may reflect a similar pro-
cess of fractional crystallisation and potentially REE-loss to a fluid phase.
The further development of the REE paragenesis features the
overgrowth of monazite by REE-bearing apatite. Assuming a fixed phos-
phate framework (an assumption supported by the overall mineralogi-
cal change and data on phosphate mineral solubility –Tropper et al.,
2011) this implies an increase in a
Ca
2+
and a
HF
in the surrounding fluids
Table 5
Representative analyses of REE fluorcarbonates by electron microprobe (EPMA) and LA-ICPMS. Mineral abbreviations as in Table 1.
EPMA DL# LA-ICPMS DL YT6 HLP5 HLP6 HLP15 YT2
wt% Par Rnt Bst Par Rnt Syn Bst Syn Par Rnt Syn Bst
CaO 0.03 10.15 16.41 0.23 9.92 13.01 16.50 0.05 17.76 9.72 15.75 21.23 0.10
FeO 0.03 BD BD 0.00 BD 0.00 1.44 BD BD BD BD BD BD
SrO 0.07 0.75 0.60 0.07 0.09 0.06 0.06 bdt 0.05 BD BD BD BD
Y2O3 0.04 0.32 0.13 0.41 1.42 1.30 6.17 BD 3.29 0.76 0.50 0.94 0.32
La2O3 0.05 18.78 13.83 14.91 11.00 12.73 7.22 23.66 9.29 13.02 8.82 7.52 18.87
Ce2O3 0.07 30.10 26.52 34.28 26.11 23.32 18.70 33.66 20.88 27.21 20.31 18.43 30.10
Pr2O3 0.10 2.67 2.54 3.49 3.09 2.65 2.28 2.67 2.30 2.80 2.33 2.34 2.67
Nd2O3 0.05 9.23 8.15 12.01 11.80 10.04 8.60 8.19 9.31 10.74 9.77 8.80 9.23
Sm2O3 0.03 0.98 0.73 1.44 2.06 2.06 1.66 0.55 2.10 1.25 1.15 1.47 0.98
SiO2 0.03 0.16 0.14 0.131 0.11 13.28 0.13 0.13 0.08 0.20 0.33 0.30 0.10
F 0.04 5.43 4.70 6.982 4.91 3.39 4.03 7.12 3.93 4.45 4.69 4.67 6.86
ppm ppm LA LA EPMA LA EPMA LA EPMA EPMA EPMA EPMA EPMA EPMA
La 591 0.03 85,000 202,500 127,200 75,900 108,500 63,900 201,800 79,200 111,000 75,200 64,100 190,000
Ce 875 0.02 156,200 338,600 292,700 144,600 199,000 159,800 287,300 179,200 232,200 173,300 157,300 310,500
Pr 1187 0.01 14,200 32,100 29,800 17,700 22,600 20,000 22,800 19,700 23,900 19,900 20,000 25,300
Nd 669 0.26 46,900 104,300 102,900 71,600 86,200 87,900 70,200 79,800 92,100 83,800 75,500 75,700
Sm 437 0.22 4120 9270 12,450 11,600 17,700 19,200 4710 1810 10,800 9900 12,700 4500
Eu 151 0.05 717 1550 3500 2460 13,000 4700 BD
a
17,600 2300 BD 750 1100
Gd 29 0.10 1830 4040 1300 7640 3300 17,400 1300 6600 4800 5000 5990 1550
Tb 0.02 131 245 NA
b
915 BD 2750 BD BD BD BD BD BD
Dy 1255 0.07 488 795 BD 5040 BD 15,900 BD BD BD BD BD BD
Y 328 0.02 1500 2360 3200 27,300 10,200 113,300 BD 25,900 6000 3900 7400 840
Ho 0.01 62 91 NA 978 NA 3920 NA NA NA NA NA NA
Er 1247 0.09 98 150 BD 2650 NA 13,200 BD BD BD BD BD BD
Tm 0.02 6 11 NA 429 NA 2320 NA NA NA NA NA NA
Yb 663 0.14 21 47 342 2760 NA 19,500 BD 817 BD BD BD BD
Lu 626 0.03 2 4 BD 353 NA 2260 BD BD BD BD BD BD
Th 763 0.02 5370 13,500 176 673 NA 1570 4570 4570 BD 13,867 475 BD
U 1394 0.02 16 38 BD 1030 NA 6520 BD BD BD BD BD BD
# DL - Detection Limit.
a
Below detection.
b
Not analysed.
74 M. Smith et al. / Lithos 308–309 (2018) 65–82
(where a is activity), and a decrease in a
REE
3+
. That this shift resulted in a
decrease in a
REE
3+
probably reflects decline in melt REE content with
early crystallisation of monazite and REE-rich niobates. The a
HF
never
reached the point of widespread fluorite saturation, and may have
been buffered to low levels by the formation of fluorapatite.
The subsequent paragenetic stage is the formation of britholite-(Ce),
and by inference allanite-(Ce), reflecting increasing a
SiO
2
in the
mineralising fluid. This is most clearly marked in the overall paragenetic
evolution of the system by the formation of the quartz-rich cores to
the dykes (Kynicky et al., 2012;Song et al., 2015;Xu et al., 2010). This
stage can be unequivocally related to the initiation of hydrothermal
conditions as fluid inclusion studies (Cangelosi, 2016;Song et al.,
2015) indicate the presence of primary, sulphate-rich brines. Early
aqueous-carbonic fluid inclusions with sulphate daughter minerals con-
tain nearly pure CO
2
, and have salinities in the aqueous phase deter-
mined from clathrate melting of 5–19 wt% NaCl equivalent (Wt% NaCl
eq.). These inclusion decrepitate on heating, and so homogenisation
temperatures cannot be determined. However, they occur in assem-
blages with liquid CO
2
inclusions, suggesting aqueous-carbonic fluid
immiscibility during trapping (Cangelosi, 2016). In a 20 wt% salt fluid,
below 100 MPa XCO
2
in the water-rich fluid duringimmiscibility is neg-
ligible, and below 50 MPa in a 5 wt% salt fluid. For both salinities there
should be significant water in the CO
2
-rich fluid above 200–300 °C
(Bowers and Helgeson, 1983). This suggests that these inclusions were
trapped between 300 and 200 °C and 100-50 MPa. Fluid inclusions in
fracture-fill quartz and calcite are typically 2 phase aqueous liquid plus
vapour inclusions. These homogenise between 265 and 120 °C, and
have salinities from 6 to 15 wt% NaCl eq. Thus hydrothermal quartz depo-
sition and the alteration of REE minerals took place from 300 to 120 °C,
at pressures below 100 MPa, from saline brines that evolved from
aqueous-carbonic compositions to aqueous-dominated compositions.
The final stage in the REE mineral evolution is the formation of
Ca-REE fluorcarbonates overgrowing all REE-phases and replacing
bastnäsite. This reflects increasing a
Ca
2+
and a
CO
3
2−
in the hydrothermal
fluid, and has been linked to decreasing T in many systems (Smith et al.,
2000;Williams-Jones and Wood, 1992). However, Gysi and Williams-
Jones (2015) have demonstrated that the compositional field of
bastnäsite stability increases with cooling at constant a
Ca
2+
in the system
REE
2
O
3
-CaF
2
-CaCO
3
,soachangeinfluid chemistry is necessary for the re-
placement reactions to develop (Fig. 11). This is most likely achieved as
the fluid cools, with the resulting dissociation of acid species (e.g. HCl,
H
2
SO
4
) and a corresponding drop in pH resulting in calcite dissolution.
5.2. Element mobility during sub-solidus metasomatism
At each alteration stage identified above there is a marked change
in REE pattern. Fig. 12 shows the composition of alteration phases nor-
malised to the precursor phase, modified by factors to allow for likely
reaction stoichiometry assuming a fixed framework of either phos-
phate, carbonate or oxide.
Table 6
Representative analyses of REE silicates by electron microprobe (EPMA) and LA-ICPMS.
EPMA DL LA-ICPMS DL HLP 10 HLP11
wt% ppm Britholite Britholite Allanite Allanite Allanite Allanite
Major elements by EPMA
CaO 0.03 11.59 12.06 8.69 8.66 9.99 11.05
MgO 0.02 BD
a
BD 4.65 4.76 0.85 1.07
MnO 0.04 0.68 0.58 6.72 6.42 1.49 1.46
Fe2O3 0.03 BD BD 5.03 4.63 23.01 21.49
Al2O3 0.02 BD BD 12.51 12.68 10.45 11.83
SiO2 0.03 19.42 18.50 30.28 30.30 29.96 30.70
La2O3 0.08 18.07 15.91 13.51 13.24 6.56 4.25
Ce2O3 0.10 31.71 29.31 12.62 12.50 10.79 9.94
Pr2O3 0.15 2.67 2.72 0.7 0.7 1.14 1.23
Nd2O3 0.08 8.15 8.17 1.30 1.30 4.27 4.27
Sm2O3 0.02 0.63 0.80 BD BD 0.65 0.65
Eu2O3 0.02 BD BD BD BD BD BD
Gd2O3 0.01 BD 0.24 BD BD 0.34 0.38
Dy2O3 0.14 BD 0.25 BD BD 0.05 0.08
TiO2 0.03 BD BD 0.51 0.49 0.20 0.41
P2O5 0.02 1.44 2.14 BD BD BD BD
F 0.05 2.70 2.75 1.45 1.43 0.23 0.29
Total 98.23 96.77 97.96 97.11 100.43 100.39
O = F 2.09 2.09 2.09 2.09 2.09 2.09
Total 96.15 94.68 95.87 95.02 98.34 98.30
Trace elements by LA-ICPMS
La 0.02 158,400 147,500 94,000 107,600 102,000 111,000
Ce 0.04 101,500 260,600 91,300 100,500 112,400 105,800
Pr 0.03 10,900 26,600 5740 5940 7430 6450
Nd 0.38 35,700 91,000 11,700 11,900 15,200 12,300
Sm 0.33 4170 9290 719 709 565 461
Eu 0.04 956 2050 160 140 77.4 61
Gd 0.13 2810 5710 386 373 317 293
Tb 0.03 326 682 40 32.8 14.1 14
Dy 0.15 1840 3800 199 153 42.7 50
Y 0.03 12,200 25,500 1270 869 337 378
Ho 0.03 405 820 40 31 9.09 11
Er 0.16 1220 2540 143 105 33.4 41
Tm 0.03 183 379 26 20 7.21 9
Yb 0.20 1350 2760 212 160 75.7 89
Lu 0.03 182 332 28 22.8 13.6 16
Th 0.10 3040 5360 250 248 164 367
U 0.06 1250 5160 4 5 10.8 14
# DL - Detection Limit.
a
Below detection.
75M. Smith et al. / Lithos 308–309 (2018) 65–82
The early alteration of monazite to apatite can be represented by
Eq. (1):
3REEPO4þ5Ca2þþF−¼Ca5PO4
ðÞ
3Fþ3REE3þ
Monazite Apatite
ð1Þ
Consideration of both reaction stoichiometry and the mineral chem-
istry indicates that the REE must have been significantly mobilised
during this phase of alteration, and Fig. 12A indicates the leaching of
the L-MREE (La-Ho), but addition of the lower abundance HREE (Er to
Lu). Yttrium behaves anomalously, being enriched relative to Dy and
Ho (Fig. 12A). Further alteration is visible on the rimsof apatite coronas
around monazite, and again indicates selective leaching of the LREE and
immobility or passive enrichment of the HREE (Fig. 12B).
It is likely that the alteration of early aeschynite to uranopyrochlore
and ultimately pyrochlore (Eqs. (2)-(3)) plus uraninite also occurred at
this stage, as it also marks progressive depletion of a primary phase in
the REE. It also implies an increase in fluid F-activities, as for alteration
of monazite to apatite. Throughout this process alteration primarily af-
fects the more mobile A-site cations (Ecrit, 2005;Lumpkin and Ewing,
1992, 1995, 1996), leaving a relatively unaltered B-site oxide frame-
work. As with monazite alteration, REE leaching is restricted to the
LREE (La-Eu), with relative immobility of the HREE (Fig. 12C).
REE;CaÞðTi;NbðÞ
2O6þF−þU4þ¼U;Ca;REEðÞ
2Nb;TiðÞ
2O6FðÞþREE3þ
Aeschynite Uranopyrochlore
ð2Þ
U;Ca;REEðÞ
2Nb;TaðÞ
2O6FðÞ¼Ca2Nb2O6FðÞþUO2þREE3þ
Uranopyrochlore Pyrochlore
ð3Þ
In both these cases the initial metasomatism is likely to be at the late
magmatic stage. Fluorite is not common in the paragenesis, as would be
expected for the interaction of an aqueous F-rich fluid with pre-existing
calcite, which would also buffer F-species activities to low levels in any
hydrothermal fluid. The alteration of uranopyrochlore to pyrochlore
plus uraninite is likely to reflect hydrothermal alteration, however,
with the relative immobility of U reflecting low fO
2
as implied by the
sulphide mineral paragenesis (Fig. 12D).
The alteration of apatite to britholite requires mobile framework
ions (Eq. (4)). Phosphate mobility from apatite via alteration to
britholite requires low pH, and high SiO
4
4−
activities.
Ca5PO4
ðÞ
3Fþ3SiO4−
4þ3REE3þ¼REE3Ca2SiO4
ðÞ
3Fþ3Ca2þþ3PO43−
Apatite Britholite
ð4Þ
This is indicative of a shift to full hydrothermal conditions, and
accompanies the formation of quartz cores to dykes. The product to pre-
cursor normalised REE patterns of britholite-(Ce) vary (Fig. 12E), typi-
cally indicating either no fractionation of the REE during alteration of
apatite, or preferential addition of the HREE. The assumption of a fixed
crystallographic framework with substitution of SiO
4
4−
for PO
4
3−
implies
no significant volume change on replacement. The actual mechanism of
apatite replacement is likely to involve some component of dissolution-
reprecipitation, which would generate some microporosity (Putnis,
2002) evident in apatite, but the factor of 10–20 increase in REE would
require a much more significant volume reduction without a metaso-
matic REE flux. This is also supported by the formation of neoformed
allanite around apatite, with comparable REE patterns to britholite. The
textural evidence therefore favours metasomatic addition of the REE.
Table 7
Representative analyses of Niobiumminerals by electronmicroprobe (EPMA)and LA-ICPMS. Tantalum was included in all analyses but was belowdetection limits andso is not shown in
the table.
EPMA DL LA-ICPMS DL HLP 15 Uranopyc Pyc YT2
Aesch
Aesch Aesch HLP11 Uranopyc Uranopyc
wt% EPMA EPMA EPMA EPMA EPMA LA LA
CaO 0.03 3.11 1.85 1.82 2.24 3.20 2.8
c
2.8
c
SrO 0.11 BD
a
BD 0.76 0.53 0.60 1.36 1.22
MnO 0.05 0.38 0.21 0.20 BD BD 0.35 0.36
FeO 0.04 1.23 12.70 1.35 0.47 0.27 1.54 1.48
TiO2 0.03 22.22 29.61 14.32 12.07 6.27 22.52 21.46
Nb2O5 0.13 28.66 42.23 24.35 16.39 13.61 22.41 21.85
SiO2 0.03 2.65 0.75 0.47 0.60 0.52 1.47 1.28
UO2 0.14 30.14 1.95 20.09 18.02 11.78 26.22 18.44
P2O5 0.02 BD
a
0.04 7.16 4.50 16.00 NA
b
NA
b
F 0.04 0.11 BD
a
0.27 2.18 0.50 NA
b
NA
b
Total 90.42 91.39 93.59 89.49 95.83 78.67 68.89
ppm ppm EPMA EPMA EPMA EPMA EPMA LA-ICPMS LA-ICPMS
La 656 0.02 BD BD 54,000 71,000 11,000 3492 3247
Ce 743 0.04 3280 2910 98,690 141,000 187,800 13,365 12,258
Pr 1290 0.03 BD BD 8850 13,100 15,900 1854 1668
Nd 764 0.38 3900 BD 28,060 41,000 48,100 10,914 9566
Sm 487 0.33 730 BD 3000 4900 4100 1750 1486
Eu 168 0.04 BD BD BD BD BD 401 371
Gd 53 0.13 1300 BD BD 950 BD 1303 1039
Tb 0.03 NA
b
NA
b
NA
b
NA
b
NA
b
154 117
Dy 1183 0.15 BD BD BD 1600 BD 1618 1193
Y 863 0.03 BD 2090 1600 3800 BD 3529 3300
Ho 0.03 NA
b
NA
b
NA
b
NA
b
NA
b
239 167
Er 1286 0.16 BD BD BD BD BD 1151 813
Tm 0.03 NA
b
NA
b
NA
b
NA
b
NA
b
118 86
Yb 824 0.20 BD BD BD BD BD 1254 902
Lu 654 0.03 BD BD BD BD BD 85 57
Th 803 0.10 BD BD BD BD 3322 1017 847
# DL - Detection Limit.
a
Below detection.
b
Not analysed.
c
Analyses from HLP 11 by LA-ICPMS and normalised to a mean Cacontent of 2000 ppm.
76 M. Smith et al. / Lithos 308–309 (2018) 65–82
The textural record of the fluorcarbonate mineralisation indicates
initially relatively uniform bastnäsite and more abundant parisite,
inferred to represent magmatic phases, followed by the progressive
development of syntaxial intergrowth with synchysite and finally over-
growth by synchysite and röntgenite. Calcium-REE fluorcarbonates
also overgrow monazite and allanite, suggesting some corrosion and
REE mobility from these phases during the hydrothermal alteration
stage. Because of the likely low F activity in any hydrothermal phase
(Williams-Jones et al., 2012), the initial development of fluorcarbonates
from earlier phases may relate to reaction with late magmatic fluids.
However, subsequent alteration reactions can take place without signif-
icant addition of F. Although reactions constructed on this basis imply
minimal REE mobility (Eqs. (5)–(8)), precursor normalised REE pat-
terns again imply significant leaching of the LREE and addition of the
HREE (Fig. 12F).
2REE PO4
ðÞþCa2þþ3CO2−
3þ2F−¼Ca REEðÞ
2CO3
ðÞ
3F2þ2PO3−
4
Monazite Parisite
ð5Þ
2REE CO3
ðÞFþCa2þþCO2−
3¼Ca REEðÞ
2CO3
ðÞ
3F2
Basna
·· site Parisite
ð6Þ
Ca REE
ðÞ
2CO3
ðÞ
3F2þCa2þþCO2−
3¼2CaREE CO3
ðÞ
2F
Parisite Synchysite
ð7Þ
3Ca REEðÞ
2CO3
ðÞ
3F2þCa2þþCO2−
3¼2Ca2REEðÞ
3CO3
ðÞ
5F3
Parisite Ro
··ntgenite
ð8Þ
Equations for the replacement of monazite and allanite by parisite
cannot be developed on the basis of an immobile crystallographic
framework. However, consideration of the mineral analyses again
suggests relative losses in LREE and gains in HREE from reactants to
products.
Uranium and thorium have been compared as products vs precur-
sors in Fig. 12 because both these elements may be concentrated
in REE phases, and are a significant issue in the assessment of
an REE resource. Thorium is typically immobile in all the reactions
examined here, consistent with its low solubility in aqueous fluids
(Bailey and Ragnarsdottir, 1994). Uranium shows variable behaviour
throughout the alteration sequence. In the progressive alteration of
aeschynite to pyrochlore U is relative immobile until the alteration of
uranopyrochlore to pyrochlore, at which point it is still potentially con-
served at the 10 s of micron scale, nucleating as uraninite in altered areas
as small (1–50 μm) inclusions in pyrochlore. These inclusions are associ-
ated with microporosity, consistent with a dissolution-reprecipitation
mechanism for niobate alteration. During the main alteration sequence
the most notable change in U is in the overgrowth and replacement of
the fluorcarbonates. The association of the hydrothermal stage with
the formation of barite and celestine, and the presence of sulphate-rich
brines within fluid inclusions, suggests that the later hydrothermal
fluids where relatively oxidising, consistent with the mobilisation of U
as uranyl complexes (Ragnarsdottir and Charlet, 2000).
5.3. Origins of HREE enrichment at Huanglongpu
The overall HREE enrichment of the Huanglongpu carbonatites
has been previously related to a recycled ocean crust component in
the lithospheric mantle indicated by Nd-Sr radioisotope studies and
Fig. 9. Ternary diagrams in atomic % showing the major element composition of niobium minerals. A-site vacancy calculated on the basis of 6 oxygens. (a) B-sitecations.(b)Allmajor
cations. (c) Major A-site cations. (d) Uranium and A-site cations.
77M. Smith et al. / Lithos 308–309 (2018) 65–82
Mg stable isotope analyses of calcite (Table 8;Xu etal., 2010, 2011;Song
et al., 2016), accounting for the district scale enrichment in Mo, accom-
panied by small degrees of melting from a garnet-poor source(Xu et al.,
2007). This primary source control has been proposed to be enhanced
by fractional crystallisation of calcite, with the dykes at the current
level of exposure representing relatively HREE-enriched carbonate
left as residuum on the walls of flowing carbonatite dykes from an
initially more LREE-enriched melt (Xu et al., 2007). Both stable (O, C
and S: Huang et al., 1984;Xu et al., 2010;Songet al., 2015) and radioiso-
tope data (Xu et al., 2007, 2011), indicate that components within
mineralisation originate with a primary carbonatite magma with only
minor input from external fluids. The shift from carbonate dominated
crystallisation to quartz and then sulphides and sulphates records the
magmatic to hydrothermal transition, with variation in redox condi-
tions during the hydrothermal stage (Huang et al., 1984), driven by
either mixing with a minor external fluid component, interaction with
relatively oxidised wall rocks, or phase separation. The chemical de-
velopment of the mineral paragenesis during this process clearly con-
tributed to the HREE enrichment of the final dyke rocks.
The trends developed in the REE patterns at Huanglongpu can be ex-
plained by consideration of experimental data on solubility of REE min-
erals and aqueous complex stability for REE species. Solubility products
for fluorides and phosphates are higher for Y and HREE than for LREE
(Migdisov et al., 2009;Tropper et al., 2011;Fig. 13A). The preferentially
solubility of the HREE end-members in low ligand strength fluids can-
not therefore result in passive enrichment in the HREE. However, ligand
concentrations can significantly modify this behaviour. Experimental
data at 1GPa and 800 °C) have shown Ce-monazite becomes more
soluble than Y-xenotime at high X
NaCl
and low X
NaF
(where X is mole
fraction; Tropper et al., 2011, 2012). Determinations of LnCl
2+
aqueous
complex ion stability constants (Logβ
1
) show that chloride preferen-
tially forms complex ions with the LREE, and hence would promote
LREE-mineral solubility at high concentrations (Fig. 13B). Low solubility
Fig. 10. (A) Chondrite normalised REE contents of aeschynite, uranopyrochlore and pyrochlore from EMPA. (B) Chondrite normalised REE contents of pyrochlore and uraninite from
LA-ICPMS analyses. (C) Correlation of U and REE contents of aechynite and pyrochlore.
Fig. 11. Activity diagram for the sys tem REE
2
O
3
-CaF
2
-CaCO
3
. Black lines show phase
boundaries at 300 °C, logaCa
2+
=−6, logaREE
3+
=−12. Solid red lines show phase
boundaries for the rel ative stability of bastnäsite and parisite at 400 °C and the same
fluid composition. Dashed red lines show phase boundaries for the relative stability of
bastnäsite and parisite at 300 °C, logaCa
2+
=−4, logaREE
3+
=−8. Adapted from Gysi
and Williams-Jones (2015). (For interpretation of the references to colour in this figure
legend, the reader is referred to the web version of this article.)
78 M. Smith et al. / Lithos 308–309 (2018) 65–82
constants for fluocerite and fluorcarbonate minerals (Williams-Jones
et al., 2012) mean high F activities are unlikely to have a major role. A
process of preferential LREE mineral solubility enhanced by chloride
complex formation is likely to have resulted in the LREE depletion
of monazite on the formation of apatite, the subsequent depletion of
apatite rims, and the progressive leaching of the REE during alteration
from aeschynite to pyrochlore. It cannot, however, account for the sig-
nificant addition of the HREE inferred in the alteration of apatite to
britholite, and the alteration of parisite and bastnäsite to synchysite
and röntgentite.
Late enrichment in the HREE is observable in calcite, and indicates
that the Huanglongpu system evolved to HREE-rich conditions, presum-
ably because of higher initial HREE contents in the original magma (Xu
et al., 2007). For hydrothermal fluids at this later stage to have devel-
oped high HREE contents requires either direct reaction with HREE-
rich calcite, or a mechanism allowing fluids to either partition HREE
preferentially from late stage melts. Very few commonly occurring li-
gands are HREE-selective, although the contrast in formation constants
between the LREE and HREE declines at lower T. However, theoretical
extrapolation of stability constants from low T data (Haas et al., 1995;
Wood, 1990) and experimental data (Migdisov and Williams-Jones,
2008) all indicate SO
4
2−
is a possible ligand with minimal difference
in stability between LREE and HREE (Fig. 13B). Carbonate is also a
potentially less LREE selective ligand at hydrothermal temperatures
(Haas et al., 1995). Fluid inclusion data from quartz cores to the
carbonatite dykes (Cangelosi, 2016;Song et al., 2016) indicate hydro-
thermal fluids in the Huanglongpu district were saturated with re-
spect to sulphate minerals, and arcanite (K
2
SO
4
), anhydrite (CaSO
4
),
glauberite (Na
2
Ca(SO
4
)
2
), and gorgeyite (K
2
Ca
5
(SO
4
)
6
·H
2
O) have all
Fig. 12. Product to precursor normalised REE plots inrelation to reactiontextures, taking into account reaction stoichiometry assuming immobile frameworkions. Error bars arebased on
2σof the mean of analyses on individual reaction textures where N= 2 or more. (a) Reaction of monazite to apatite. (b) Apatite core normalised to leached apatite rim. (c) Reaction of
apatite to britholite. (d) Reaction o f parasite to synchysite. (e) Reaction o f aeschynite to pyrochlore, and pyrochlore to uranopyrochlore (EPMA). (f) Reaction of Pyrochlor e to
urnaopyrchlore (LA-ICPMS).
Table 8
Summary of previousradio-and stable isotopic data from Huanglongpu. Data are consistent
with an enriched mantle source, with a possible subducted carbonate component. Narrow
ranges for stable isotope data suggest limited mixing of fluid sources, and are consistent
with a man tle-deri ved carbonatite orig in for fluids and solution components including S.
Reference Isotope System Min Max Estimated fluid
composition
Song et al. (2016) δ
25
Mg (‰)−0.68 −0.56 n=9
δ
26
Mg (‰)−1.28 −1.07 n = 9
δ
34
Ssulphides(‰)−10.5 −6.5 n=18 ~1‰
δ
34
S Barite (‰) 4.6 5.1 n=6
Xu et al. (2011)
87
Sr/
86
Sr
i
0.705 0.706 n=10
ε
Nd(t, CHUR)
−10.1 −4.3 n = 10
207
Pb/
206
Pb
i
0.883 0.887 n = 10
208
Pb/
206
Pb
i
2.146 2.159 n = 10
Song et al. (2015) δ
18
O quartz 8.1 10.2 n=5
Xu et al. (2010) δ
18
O calcite 7.2 9.2 n=8
δ
13
C calcite −6.9 −6.5 n = 8
Huang et al.
(1984)
δ
34
S sulphides −14.7 −4n=25 ~1‰
δ
34
S Barite 4.7 7.9 n = 8
δ
18
O calcite 8.5 9.5 n = 8
δ
13
C calcite −7−6.6 n = 8
79M. Smith et al. / Lithos 308–309 (2018) 65–82
been identified as either daughteror heterogeneously trapped phases in
fluid inclusions (Cangelosi, 2016). Migdisov et al. (2016) demonstrated
that sulphate complexes of the form REE(SO
4
)
2
−
can predominate in
fluoride-containing, Cl-dominated brines at intermediate pH (~3–8),
from 200 to 400 °C in the presence of cerianite. These conditions are
entirely consistent with the example studied here. Dissolution of calcite
as pH falls with decreasing T, in the presence of sulphate and carbonate
ligands has the potential to release HREE from calcite resulting in the
formation of late HREE enriched minerals.
5.4. Wider implications
Sulphate is a component in many carbonatite melts (e.g.
Doroshkevich et al., 2010;Xie et al., 2015), and sulphate melt generated
by immiscibility in the late stages of magmatic crystallisation has
been identified in numerous melt inclusion studies (e.g. Andreeva
et al., 1998;Panina, 2005;Panina and Motorina, 2008;Panina and
Stoppa, 2009). Equally, sulphate-rich fluids within carbonatite-related
hydrothermal systems are widespread, with barite and celestine as
common accessories in carbonatite systems (e.g. Nikiforov et al., 2014;
Trofanenko et al., 2016;Xie et al., 2015). The key factor for the genesis
of sulphide- or sulphate-bearing carbonatite systems is the redox state
of the melt or fluid. In the magmatic stage oxidised carbonatite melts
have been proposed to be generated by oxidised mantle source regions
(Foley, 2011) and magma mingling (Moore et al., 2009), whilst in hy-
drothermal systems the redox state of the fluid is controlled by interac-
tion with the host-rocks, mineral deposition and phase separation
(Robb, 2005). There is widespread evidence for sulphate complexation
of the REE from mesothermal to epithermal systems (e.g. Inguaggiato
et al., 2015;Lewis et al., 1998). The inference that sulphate-rich
brines are responsible for the late HREE-enrichment of the REE mineral
assemblage means that REE-rich carbonatites with strong evidence
for sulphate-bearing magmas and late stage sulphate alteration may
be prospective targets for relative HREE-enrichment, although initial
HREE-enrichment in the mantle source is still a necessity for significant
HREE mineralisation. Previous research and mineral exploration has
identified HREE-enriched zones in a number of potentially economically
mineralised carbonatites, reviewed briefly in the introduction. In most of
these HREE-enrichment is related to sub-solidus hydrothermal processes
associated with barite and celestine deposition (Broom-Fendley et al.,
2017;Moore et al., 2017;Ngwenya, 1993). At Lofdal the main HREE en-
richment was inferred to be hydrothermal, associated with carbonate
matrices to breccias (Bodeving et al., 2017) with less evidence for high
sulphare activities. Thus carbonatite HREE-mineralisation is very com-
monly hydrothermal in origin, and frequently associated with sulphate-
rich fluids. Both SO
4
2−
and CO
3
2–
are potentially less LREE-selective ligands
which may contribute to HREE leaching and transport during secondary
processes, and may be effectiviely destabilised by processes such as boil-
ing, aqueous-carbonic immiscibility and water-rock interaction, whilst
Cl
−
acts to retain LREE in the fluid. Although the HREE-enrichment at
Huanglongpu is a function of both magma source regions and hydro-
thermal enrichment, the processes identified here clearly have the po-
tential to operate in many carbonatite-related hydrothermal systems.
6. Conclusions
The carbonatites of the Huanglongpu district, Qinling mountains,
China, are REE-mineralised, and HREE-enriched compared to typical
carbonatites and carbonatite-related REE deposits. The dykes and
veins evolved from early calcite dominated crystallisation to multiple
stages of late magmatic and hydrothermal activity generating multiple
generations of carbonate, K-feldspar and quartz-rich cores to the
dykes. Textural and geochemical data show that the REE and niobate
mineral assemblage evolved via fractional crystallisation from an
early magmatic stage dominated by monazite-(Ce), parisite-(Ce) and
aeschynite-(Ce), to more REE depleted conditions resulting in the over-
growth and replacement of monazite by apatite, and the alteration
of aeschynite to uranopyrochlore. Subsequent alteration reactions
record the shift to silica-rich hydrothermal conditions at tempera-
tures from 300 to 200 °C, 0.5-1 kbar, and involving aqueous-carbonic
brines, resulting in the alteration of apatite to britholite-(Ce) and the
crystallisation of allanite-(Ce). Cooling of the hydrothermal system to
N265 °C moved REE mineralisation to fluorcarbonate mineral domi-
nated, and reaction with host carbonatite resulted in raised a
Ca
2+
and a
CO
3
2−
, causing alteration of early fluorcarbonates to synchysite
and röntgenite. At every stage of alteration hydrothermal fluids leached
the LREE relative to the HREE. Solubility products for LREE minerals are
typically lower thanfor HREE except in thepresence of strongligands, in
this case probably Cl
−
,CO
3
2–
and SO
4
2−
. For some replacement reactions
the increase in the HREE concentration cannot be accounted for by
passive enrichment without significant volume loss, and so there must
have been a metasomatic flux of the HREE. The host calcite is the most
likely source of HREE as the overall carbonatite system is anomalously
HREE-enriched compared to typical systems. The sulphate-rich nature
Fig. 13. (A)Solubilityproducts(K
sp
) for monazite from 150 to 250 °C from Migdisov et al.
(2009). (B) Experimentally determined first association constants for Ln-Cl (Migdisov
et al., 2009) and Ln-SO
4
(Migdisov and Williams-Jones, 2008) aqueous complexes from
100 to 250 °C and 10 MPa. First association constants for REE-SO
4
complexes calculated
from the data of Haas et al. (1995) are shown for comparison (open circles).
80 M. Smith et al. / Lithos 308–309 (2018) 65–82
of metasomatic brines provides a mechanism via which this may occur,
as sulphate complexes show no major difference between LREE
and HREE complex ion stability. There is no evidence for a major
role for fluoride in this process, except as a depositional ligand in
the precipitation of fluorcarbonate minerals. Globally sulphate-rich
carbonatites may therefore offer significant potential for more HREE
enriched carbonatite-related resources.
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.lithos.2018.02.027.
Acknowledgements
We would like to thank Andrew Flint and Mike Heliasfor additional
help with SEM imaging at the University of Brighton. The manuscript
was improved by comments from Nelson Eby and two anonymous
referees.
Funding
This work was supported by the NERC SOS:RARE project (NE/
M011267/1), the EU H2020 (grant agreement no. 689909), grant
project (HiTech AlkCarb) and project CEITEC 2020 (LQ1601).
References
Andreeva, I.A., Naumov, V.B., Kovalenko, V.I., Kononkova, N.N., 1998. Fluoride-sulfate and
chloride-sulfate salt melts of the carbonatite-bearing complex Mushugai-Khuduk,
southern Mongolia. Petrology 6, 284–292.
Bailey, E.H., Ragnarsdottir, V., 1994. Uranium and thorium solubilities in subduction zone
fluids. Earth and Planetary Science Letters 124, 119–129.
Bodeving, S., Williams-Jones,A.E., Swinden,S., 2017. Carbonate–silicatemelt immiscibility,
REE mineralising fluids,and the evolutionof the Lofdal intrusive suite, Namibia. Lithos
268–271, 383–398.
Bowers, T.S., Helgeson, H.C., 1983. Calculation of the thermodynamic and geochemical
consequences of non-ideal mixing in the system H
2
O-CO
2
-NaCl on phase relations
in geologic systems: equation of state for H
2
O-CO
2
-NaCl fluids at high pressures
and temperatures. Geochimica et Cosmochimica Acta 47, 1247–1275.
Broom-Fendley, S., Styles, M.T., Appleton, J.D., Gunn, G., Wall, F., 2016. Eviden ce for
dissolution-reprecipitation of apatite and preferential LREE mobility in carbonatite-
derived late-stage hydrothermal processes. American Mineralogist 101, 596–611.
Broom-Fendley, S., Brady, A.E., Wall, F., Gunn, G., Dawes, W., 2017. REE minerals at the
Songwe Hillcarbonatite, Malawi: HREE-enrichment in late-stage apatite. Ore Geology
Reviews 81, 23–41.
Cangelosi, D., 2016. The Heavy Rare Ea rth Element Enrichment of the Huanglongpu
deposit, China. (MRes Thesis). Leeds Univ., UK.
Chakhmouradian, A.R., 2006. High-field-strength el ements in carbonat itic rocks:
geochemistry, crystal chemistry and significance fo r constraining th e sources of
carbonatites. Chemical Geology 235, 138–160.
Chakhmouradian, A., Wall, F., 2012. Rare earth elements: minerals, mines, magnets (and
more). Elements 8, 333–340.
Doroshkevich, A.G., Ripp, G.S., Moore, K.R., 2010. Genesis of the Khaluta alkaline-basic
Ba-Sr carbonatite complex (West Transbaikala, Russia). Mineralogy and Petrology
98, 245–268.
Ecrit, T.S., 2005. Identification and alteration trends of granitic-pegmatite-hosted (Y, REE,
U, Th)-(Nb, Ta, Ti) oxide minerals: a statistical approach. Canadian Mineralogist 43,
1291–1303.
Foley, S.F., 2011. A reappraisal of redox melting in the Earth's mantle as a function of
tectonic setting and time. Journal of Petrology 52, 1363–1391.
Gysi, A.P., Williams-Jones, A.E., 2015. The thermodynamic properties of bastnäsite-(Ce)
and parisite-(Ce). Chemical Geology 392, 87–101.
Haas, J.R., Shock, E.L., Sassani, D.C., 1995. Rare earth elements in hydrothermal systems:
estimatesof standard partial molal thermodynamic properties of aqueous complexes
of the rare earth ele ments at high pressures and temperatures. Geochimica et
Cosmochimica Acta 59, 4329–4350.
Henderson, P., 1984. Rare earthelement geochemistry. Developments in Geochemistry. 2.
Elsevier.
Hornig-Kjarsgaard, I., 1998. Rare earth elements in Sövitic carbonatites and their mineral
phases. Journal of Petrology 39, 2105–2121.
Huang, D.H., Wang, Y., Nie, F., Jiang, X., 1984. Isotopic composition of sulphur, carbon and
oxygenand source materialof the Huanglongpu carbonatite vein-type of molybdenum
(lead) deposits. Acta Geologica Sinica 63, 252–264 (in Chinese with English abstract).
Huang, D.H., Wu, C.Y., Du, A.D., He, H.L., 1994. Re–Os isotope age of molybdenum deposits
in east Qinling and their significance. Mineral Deposits 13, 221–230 (in Chinese with
English abstract).
Inguaggiato, C., Censi, P., Zuddas, P., Londono, J.M., Chacon, Z., Alzate, D., Brusca, L.,
D'Alessandro, W., 2015. Geochemistry of REE, Zr and Hf in a wide range of pH and
water composition: the Nevado del Ruiz volcano-hydrothermal system (Colombia).
Chemical Geology 417, 125–133.
Kynicky, J., Smith, M.P., Xu, C., 2012. Diversity of rare earth deposits: the key example of
China. Elements 8, 361–367.
Le Bas, M.J., 1987. Ne phelinites and Carbonatites. In: Fitton, J.G., Upton, B.G. (Eds.),
Proceedings of Rare Earths'04 MS. International Journal of Mass Spectrometry
vol. 253, pp. 87–97.
Lewis, A.J., Komninou, A., Yardley, B.W.D., Palmer, M.R., 1998. Rare earth element
speciation in geothermal fluids from Yellowsto ne National Park, Wyo ming, USA.
Geochimica et Cosmochimica Acta 62, 657–663.
Lumpkin, G.R., Ewing, R.C., 1992. Geochemical alteration of pyrochlore group minerals:
microlite subgroup. American Mineralogist 77, 179–188.
Lumpkin, G.R., Ewing, R.C., 1995. Geochemical alteration of pyrochlore group minerals:
Pyrochlore subgroup. American Mineralogist 80, 732–743.
Lumpkin, G.R., Ewing, R.C., 1996. Geochemical alteration of pyrochlore group minerals:
Betafite subgroup. Ameri can Mineralogist 81, 1237–1248.
Mao, J.W., Xie, G.Q., Bierlein, F., Qu, W.J., Du, A.D., Ye, H.S., Pirajno, F., Li, H.M.,Guo, B.J., Li,
Y.F., Yang, Z.Q., 2008. Tectonic implications from Re-Os dating of Mesozoic molybde-
num deposits in the East Qinling-Dabie orogenicbelt. Geochimica et Cosmochimica
Acta 72, 4607–4626.
Meng, Q.R., Zhang, G.W., 2000. Geologic framework and tectonic evolution of the Qinling
orogen, Central China. Tectonophysics 323, 183–196.
Migdisov, A.A., Williams-Jones, A.E., 2008. A spectrophotometric study of Nd(III), Sm(III)
and Er(III) complexation in sulfate-bearing solutions at elevated temperatures.
Geochimica et Cosmochimica Acta 72, 5291–5303.
Migdisov, A.A., Williams-Jones, A.E., Wagner, T., 2009. An experimental study of the solu-
bility and speciation of the rare earth elements (III) in fluoride- and chloride-bearing
aqueous solutions at temperatures up to 300°C. Geochimica et Cosmochimica Acta
73, 7087–7109.
Migdisov, A., Williams-Jones, A.E., Brugger, J., Caporuscio, F.A., 2016. Hydrothermal trans-
port, deposition, and fractionation of the REE:experimental data andthermodynamic
calculations. Chemical Geology 439, 13–42.
Mitchell, R., 2005. Carbonatites and carbonatites and carbonatites. Canadian Mineralogist
43, 2049–2068.
Mitchell, R., 2015. Primary and secondary niobium mineral deposits associated with
carbonatites. Ore Geology Reviews 64, 626–641.
Moore, K.R., Wall, F., Divaev, F.K., Savatenkov, V.M., 2009. Mingling of carbonate and silicate
magmas under turbulent flow conditions: evidence from rock textures and mineral chem-
istry in subvolcanic carbonatite dykes, Chagatai, Uzbekistan. Lithos 110, 65–82.
Moore, M., Chakhmouradian, A., Mariano, A.N., Sidhu, R., 2017. Evolution of rare-earth
mineralization in the bear lodge carbonatite, Wyoming: mineralogical and isotopic
evidence. Ore Geology Reviews 64, 499–521.
Ngwenya,B., 1993. Hydrothermal rare earth mineralisationin carbonatites of the Tundulu
complex, Malawi: processes at the fluid/rock interface. Geochimica et Cosmochimica
Acta 58, 2061–2072.
Nikiforov, A.V., Ozturk, H., Altu ncu, S., Lebedev, V.A., 2014. Kizi lcaoren ore-bearing
complex with carbonatites (northwestern Anatolia, Turkey): formation time and
mineralogy of rocks. Geology of Ore Deposits 56, 35–60.
Panina, L.I., 2005. Multiphase carbonate-salt immiscibility in carbonatite melts: data on
melt inclusions from the Krestovskiy massif minerals (polar Siberia). Contributions
to Mineralogy and Petrology 150, 19–36.
Panina, L.I., Motorina, I.V., 2008. Liquid immiscibility in deep-seat ed magmas and the
generation of carbonatite melts. Geochemistry International 46, 448–464.
Panina, L.I., Stoppa, F., 2009. Silicate-carbonate-salt liquid immiscibility and origin of the
sodalite-haüyne rocks: study of melt inclusions in o livine foidite from vulture
volcano, S. Italy. Central European Journal of Geosciences 1, 377–392.
Pouchou, J.L., Pichoir, F., 1984. A New Model for Quantitative Analyses. I. Application to
the Analysis of Homogeneous Samples. 3. La Recherche Aérospatiale, pp. 13–38.
Putnis, A.,2002. Mineral replacement reactions: from macroscopic observationsto micro-
scopic mechan isms. Mineralogical Magazine 66, 6 89–708.
Ragnarsdottir, K.V., Charlet, L., 2000. Uranium behaviour in natural environments. In:
Cotter-Howells, J.,Batchelder, M.,Valsami-Jones,E. (Eds.), Environmental Mineralogy.
Mineralogical Society of Great Britain and Ireland, London, pp. 333–377.
Ratschbacher, L., Hacker, B.R., Calvert, A., Webb, L.E., Grimmer, J.C., McWilliams, M.O., Ireland,
T., Dong, S., Hu, J., 2003. Tectonics of the Qinling (Central China): tectonostratigraphy,
geochronology, and deformation history. Tectonophysics 366, 1–53.
Robb, L.,2005. Introduction to Ore Forming Processes. Blackwell Science,Oxford (373 pp.).
Smith, M.P., Henderson, P., Campbell, L.S., 2000. Fractionation of the REE during hydro-
thermal processes: constraints from the Bayan obo Fe –REE–Nb deposit, Inne r
Mongolia, China. Geochimica et Cosmochimica Acta 64, 3141–3160.
Song, W., Xu, C., Qi, L., Zhou, L., Wang, L., Kynicky, J., 2015. Genesis of Si-rich carbonatites
in Huanglongpu Mo deposit, lesser Qinling orogen, China and significance for Mo
mineralization. Ore Geology Reviews 64, 756–765.
Song, W.,Xu, C., Smith, M.P.,Kynicky, J., Huang,K., Wei, C., Li, Zhou, Shu, Q., 2016. Origin of
Unusual HREE-Mo-rich Carbonatites in the Qinling Orogen, China. Scientific Reports
6:37377. https://doi.org/10.1038/srep37377.
Stein, H.J., Markey, R.J., Morgan, M.J., Du, A., Sun, Y.,1997. Highly preciseand accurate Re–
Os ages for molybdenite from the east Qinling–Dabie molybdenum belt, Shaanxi
province, China. Economic Geology 92, 827–835.
Sun, S.-S., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanic basalts:
implications for mantle composition and processes. In: Saunders, A.D., Norry, M.J.
(Eds.), Magmatism in Ocean Basins. Geol. Soc. Spec. Publ., London, pp. 313–345.
Swinden,S., Hall, M., 2012. NI 43-101 Technical Report andMineral Resource Estimate for
the Songwe Hill Rare Earth Element (REE) Project, Phalombe District, Republic of
Malawi (Report prepared for Mkango Resources by MSA group. 168 pp.).
Trofanenko, J., Williams-Jones, A.E., Simandl, G.J., Migdisov, A.A., 2016. The nature and
origin of the REE mineralization in the Wicheeda carbonatite, British Columb ia,
Canada. Economic Geology 111, 199–223.
81M. Smith et al. / Lithos 308–309 (2018) 65–82
Tropper, P., Manning, C.E., Harlov, D.E., 2011. Solubility of CePO
4
monazite and YPO
4
xenotime in H
2
O and H
2
O–NaCl at 800 °C and 1 GPa: impl ications for REE and Y
transport during high-grade metamorphism. Chemical Geology 282, 58–66.
Tropper, P., Manning, C.E., Harlov, D.E., 2012. Experimental determination of CePO
4
and
YPO
4
solubilities in H
2
O–NaF at 800 °C and 1 GPa: implications forrare earth element
transport i n high-grad e metamorphic fluids. Geofluids 13, 372–380.
Wall, F., Niku-Paavola, V.N., Storey, C., Muller, A., Jeffries, T., 2008. Xenotime-(Y) from
carbonatite dykes at Lofdal, Namibia: unusually low LREE:HREE ratio in carbonatite,
and the first dating of xenotime overgrowths on zircon. Canadian Mineralogist 46,
861–877.
Williams, C.T., 1996. Analysis of rare earth minerals. In: Jones, A.P., Wall, F., Williams, C.T.
(Eds.), Rare Earth Minerals: Chemistry, Origin and Ore Deposits. Chapman and Hall,
pp. 327–348.
Williams-Jones, A.E., Wood, S.A., 1992. A preliminary petrogenetic grid for REE
fluorocarbonates and associated minerals. Geochimic a et Cosmochimica Acta 56,
725–738.
Williams-Jones, A.E,Migdisov, A.A., Samson, I.M., 2012. Hydrothermal mobilization of the
rare earth elements –ataleof‘ceria’and ‘yttria’.Elements8,355–360.
Williams-Jones, A.E., Migdisov, A.A., Samson,I.M., 2013. Hydrothermal mobilisation of the
rare earth elements - a tale of “ceria”and “Yttria”. Elements 8, 355–360.
Wood, S.A., 1990. The aqueous geochemistry of the rare earth elements and yttriu m.
Part I. Review of available low temperature data for inorganic complexes and the
inorganic REE speciation of natural waters. Chemical Geology 82, 159–186.
Xie, Y.L., Li, Y.X., Hou, Z.Q., Cooke, D.R., Danyushevsky, L., Dominy, S.C., Yin, S.P., 2015. A
model for carbonatite hosted REE mineralisation - the Mianning-Dechang REE belt,
western Sichuan Province, China. Ore Geology Reviews 70, 595–612.
Xu, C., Campbell, I.H., Allen, C.M., Huang, Z.L., Qi, L., Zhang, H., Zhang, G.S., 2007. Flat rare
earth element patterns as an indicator of cumulate processes in the lesser Qinling
carbonatites, China. Lithos 95, 267–278.
Xu, C., Kynicky, J., Chakhmouradian, A., Qi, L., Song, W., 201 0. AuniqueModeposit
associated with carbonatites in the Qinling orogenic belt, central China. Lithos 118,
50–60.
Xu, C., Taylor, R.N., Kynicky, J., Chakhmouradian,A., Song, W., Wang, L.,2011. The origin of
enriched mantle beneath North China block: evidence from young carbonatites.
Lithos 127, 1–9.
Xu, C., Chakhmouradian, A., Taylor, R.N., Kynicky, J., Li, W., Song, W., Fletcher, I.R., 2014.
Origin of carbonatites in the south Qinling orogen: implications for crustal recycling
and timing of collision between the south and North China blocks. Geochimica et
Cosmochimica Acta 143, 189–206.
Xue, F., Lerch, M.F., Kroner, A., Reischmann, T., 1996. Tectonic evoluti on of the east
Qinling Mountains, China, in the Palaeozoic: a review and new tectonic model.
Tectonophysics 253, 271–284.
Zaitsev,A.N., Wall, F., Le Bas,M.J., 1998. REE-Sr-Ba minerals fromthe Khibina carbonatites,
Kola peninsula, Russia: their mineralogy, paragensis and evolution. Mineralogical
Magazine 62, 225–250.
Zaitsev, A.N., Demény, A., Sinder, S., Wall, F., 2002. Burbankite group minerals and their
alteration in rare earth carbonatites –source of elements and fluids (evidence from
C-O and Sr-Nd isotopic data). Lithos 62, 15–33.
Zhao, H., Mao, J., Ye, H., Xie, G., Yang, Z., 2010. Geochronology and geochemistry of the al-
kaline granite porphyry and diabase dikes in Huanglongpu area of Shaanxi Province:
petrogenesis and implications for tectonic environment. Geology in China 37 (1),
12–27 (in Chinese with English abstract).
82 M. Smith et al. / Lithos 308–309 (2018) 65–82
