Lithos 118 (2010) 145155

Contents lists available at ScienceDirect

Lithos
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / l i t h o s

Trace-element modeling of the magmatic evolution of rare-earth-rich carbonatite from the Miaoya deposit, Central China
Cheng Xu a,, Jindrich Kynicky b, Anton R. Chakhmouradian c, Ian H. Campbell d, Charlotte M. Allen d
a

Laboratory of Orogenic Belts and Crustal Evolution, Peking University, Beijing 100871, China Department of Geology and Pedology, Mendel University, Brno, Czech Republic Department of Geological Sciences, University of Manitoba, Winnipeg, Manitoba, Canada d Research School of Earth Sciences, Australian National University, Canberra, Australia
b c

a r t i c l e

i n f o

a b s t r a c t
Carbonatites are known to contain the highest concentrations of rare-earth elements (REE) among all igneous rocks. The REE distribution of carbonatites is commonly believed to be controlled by that of the rockforming Ca minerals (i.e., calcite, dolomite, and ankerite) and apatite because of their high modal content and tolerance for the substitution of Ca by light REE (LREE). Contrary to this conjecture, calcite from the Miaoya carbonatite (China), analyzed in situ by laser-ablation inductively-coupled-plasma mass-spectrometry, is characterized by low REE contents (100260 ppm) and relatively at chondrite-normalized REE distribution patterns [average (La/Yb)CN = 1.6]. The carbonatite contains abundant REE-rich minerals, including monazite and uorapatite, both precipitated earlier than the REE-poor calcite, and REEuorocarbonates that postdated the calcite. Hydrothermal REE-bearing uorite and barite veins are not observed at Miaoya. The textural and analytical evidence indicates that the initially high concentrations of REE and P in the carbonatitic magma facilitated early precipitation of REE-rich phosphates. Subsequent crystallization of REE-poor calcite led to enrichment of the residual liquid in REE, particularly LREE. This implies that REE are generally incompatible with respect to calcite and the calcite/melt partition coefcients for heavy REE (HREE) are signicantly greater than those for LREE. Precipitation of REE-uorocarbonates late in the evolutionary history resulted in depletion of the residual liquid in LREE, as manifested by the development of HREE-enriched late-stage calcite [(La/Yb)CN 0.7] in syenites associated with the carbonatite. The observed variations of REE distribution between calcite and whole rocks are interpreted to arise from multistage fractional crystallization (phosphates calcite REE-uorocarbonates) from an initially REE-rich carbonatitic liquid. 2010 Elsevier B.V. All rights reserved.

Article history: Received 3 November 2009 Accepted 1 April 2010 Available online 13 April 2010 Keywords: Carbonatite Calcite REE partitioning Fractional crystallization REE deposits Miaoya China

1. Introduction The utility of rare-earth elements (REE) as petrogenetic tracers in igneous processes has long been recognized. A large amount of experimental data on hydrothermal REE transport and partitioning is available (e.g., Wood, 1990; Haas et al., 1995; Migdisov et al., 2008). In conjunction with voluminous empirical data (e.g., Morogan, 1989; Williams-Jones et al., 2000; Wall and Zaitsev, 2004; Bhn, 2008), this experimental evidence has led to an improved understanding of the behavior of REE during waterrock interaction. However, the effects of crystal fractionation involving multiple phases separating at different stages during magma evolution have not been studied in adequate detail for carbonatites. Carbonatites are known to contain the highest concentrations of REE (500 to N10 000 ppm) among all igneous rocks,

Corresponding author. E-mail address: xucheng1999@hotmail.com (C. Xu). 0024-4937/$ see front matter 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.lithos.2010.04.003

and are a major source of economic REE mineralization. However, carbonatite-related REE deposits usually have very limited reserves, and only a few economically important deposits have been documented since the discovery of the world-class Mountain Pass deposit in California (Castor, 2008). In most carbonatites, REE are primarily dispersed in rock-forming and accessory Ca minerals such as calcite, dolomite, apatite and silicates (e.g., allanite; Hornig-Kjarsgaard, 1998). Rare-earth minerals (particularly, carbonates and phosphates) are commonly enriched in magnesiocarbonatite and ferrocarbonatite dikes that may be either igneous or hydrothermal rocks (Le Bas, 1989; Wall and Mariano, 1996; Ruberti et al., 2008). The Miaoya carbonatite stocks in China consist predominantly of calcite and contain abundant REE minerals (up to 5 vol.%), making this deposit an economic REE resource (Qian and Li, 1996). Importantly, the Miaoya carbonatite is not associated with obvious hydrothermal REE mineralization, which distinguishes it from most other REE deposits (e.g., Samson et al., 2004; Smith, 2007; Xu et al., 2008). The main objective of the present work is to establish genetic relations between calcite and

146

C. Xu et al. / Lithos 118 (2010) 145155

major REE hosts in the Miaoya carbonatite and to assess the parameters that control the distribution of REE in minerals precipitating from carbonatitic liquids. 2. Geological setting The Miaoya REE deposit is located in the southwestern margin of the Wudang Terrane at the southern margin of the Qinling orogenic belt (Fig. 1). The Qinling belt, which separates the North China block from the South China block, is divided into two parts (North and South Qinling) along the Shangdan suture (Fig. 1). The North Qinling belt is

regarded as a Mid-Paleozoic orogen with widespread Paleozoic island-arc type magmatism and metamorphism (Sun et al., 2002). The South Qinling belt is interpreted as a Late-Paleozoic to EarlyMesozoic orogen characterized by abundant Triassic granitic magmatism and metamorphism (Mattauer et al., 1985; Sun et al., 2000). There is a general consensus that the North Qinling belt was an active continental margin, whereas the South Qinling structure was a passive margin forming the northernmost part of the South China block (Lu et al., 2006). Detailed descriptions of the geological framework and tectonic evolution of the Qinling region are given in Meng and Zhang (2000), Ratschbacher et al. (2003) and Chen et al. (2009).

Fig. 1. Schematic geological map of the Miaoya REE deposit (modied after Li, 1980; Meng and Zhang, 2000).

C. Xu et al. / Lithos 118 (2010) 145155

147

Fig. 2. Selected characteristic textures observed at the Miaoya REE deposit. (A) and (B) Photomicrographs of carbonatite and contaminated syenite, respectively, in cross-polarized transmitted light; (C) cathodoluminescence in primary calcite; (DH) back-scattered electron images of REE minerals in carbonatites. Symbols: Cal, calcite; Ls, laser spot; Mnz, monazite-(Ce); Ap, uorapatite; Rt, rutile; Bt, bastnsite-(Ce); Sy, synchysite-(Ce). Note alteration and calcite veining in the syenite (B), development of monazite and uorapatite prior to primary calcite (D, E), crystallization of bastnsite-(Ce) after primary calcite (F), and syntaxial intergrowths of bastnsite-(Ce) and synchysite-(Ce) (G, H).

148

C. Xu et al. / Lithos 118 (2010) 145155

C. Xu et al. / Lithos 118 (2010) 145155

149

At Miaoya, REE-bearing carbonatites and associated syenites, emplaced in Mesoproterozoic low-grade metavolcanic rocks (mainly quartz keratophyre) and Early Paleozoic carbonaceous shale, cover an area of 6.5 km2 at the current level of erosion. The only available geochronological study of these carbonatites gives a biotite KAr age of 278 Ma (Li, 1980). The carbonatites intrude the syenites as stocks and minor dikes and are composed predominantly of medium- to ne-grained calcite (N85 vol.%). Cathodoluminescence (CL) imaging of calcite shows a consistent zoning pattern consisting of a yellow core and an orange rim (Fig. 2). Minor and accessory phases include biotite, alkali feldspar, uorapatite, magnetite, ilmenite, perovskite, zircon, ferrocolumbite, Nb-rich rutile, monazite-(Ce), bastnsite-(Ce) and synchysite-(Ce). A few carbonatite samples contain coarse-grained graphite and ankerite; the latter mineral is developed as networks of veinlets postdating the graphitization of carbonatite (Li, 1980). The principal REE minerals in the carbonatite are uorapatite, monazite(Ce), bastnsite-(Ce) and synchysite-(Ce) (Fig. 2). The uorapatite and monazite-(Ce) are early phases formed prior to calcite. Fluorapatite typically makes up 15% of the rock by volume. It occurs as elongate and partially resorbed grains, as well as layered segregations (Fig. 2) commonly associated with small (b1.0 mm) euhedral grains of Nb-rich rutile. Monazite-(Ce) commonly occurs as equant grains. Both bastnsite-(Ce) and synchysite-(Ce) crystallized later than the calcite. In some cases, the two uorocarbonate minerals form syntaxial intergrowths. The Miaoya deposit has an estimated ore reserve of 1.21 million tons of ore at an average grade of 1.72 wt.% REE2O3 (Qian and Li, 1996). Note that late-stage hydrothermal mineralization commonly found in association with carbonatites and consisting primarily of uorite and barite, is not observed at the Miaoya deposit or anywhere in the country rocks. The latter observation is at variance with the suggestion of Williams-Jones et al. (2000) that the occurrence of uorite is an important guide in the exploration for economic REE deposits. 3. Analytical methods The whole-rock abundances of trace elements in samples from Miaoya were determined in solution by inductively-coupled-plasma mass-spectrometry (ICPMS) using a Thermo VG PQ ExCell instrument at the University of Hong Kong and following the methodology described in Qi et al. (2000). Replicate analyses of the samples and reference materials showed that the accuracy of the REE determinations was better than 10%. In-situ laser-ablation ICPMS (LA-ICPMS) analyses of carbonate minerals were performed on polished thin sections and grain mounts at the Australian National University. The diameter of the ablation spot was varied between 54 and 86 m depending on the size and compositional homogeneity of individual grains. The NIST 610 and 612 glasses were used as a calibration standard assuming the element concentrations reported by Eggins et al. (1997). The Ca content determined independently by wavelength-dispersive X-ray spectrometry (WDS) was used as an internal standard. The lower detection limits for individual elements were calculated following the approach of Longerich et al. (1996). The analytical precision is estimated to be better than 5% at the ppm level. In-run signal intensity for indicative trace elements was monitored during the analysis to make sure that inclusions and adjacent mineral grains were not accidentally sampled by ablation. ICPMS sensitivity in dry-plasma (laser) mode is such that the instrument can be tuned for low oxide production, which results in very low molecular interferences, such as ThO/Th 0.5%; other

interferences are less signicant. These conditions are difcult to achieve for wet plasmas, and oxide production for solution ICPMS is commonly a few percent. In particular, light-REE oxides interfere with heavier REE producing spurious positive anomalies for some elements (particularly in the case of Gd; see Fig. 3C). The major-element compositions of calcite, uorapatite and REE minerals were measured by WDS using a Cameca SX 100 electron microprobe at the Joint Electron Microscopy and Microanalysis Laboratory (Institute of Geological Sciences, Masaryk University and Czech Geological Survey). The instrument was operated at a beam current of 10 nA and an accelerating voltage of 15 kV. The beam was defocused to a spot size of 510 m to minimize beam-induced damage and thermal decomposition of the minerals. The following standards and X-ray emission lines were used in the analysis of calcite: dolomite (Mg and K), calcite (Ca and K), willemite (Mn and K), olivine (Fe and K) and celestine (Sr and L). For the analysis of REE minerals, andradite (Si and Fe, K line for both), wollastonite (Ca and K), apatite (P and K), topaz (F and K), synthetic ThO2 (Th and M), Y-Al garnet (Y and K), REE glasses and orthophosphates (La, Ce and Sm with L line; Pr, Nd, Gd and Dy with L) were used. For the uorapatite, uorapatite (Ca, P, F, and K), andradite (Si and K), celestine (Sr and L) and the aforementioned REE standards were used. For the REE-bearing minerals, the raw WDS data were corrected using empirical interference values for REE and other elements potentially interfering with the REE signals determined for the wellcharacterized synthetic standards. Following this procedure, the lower limits of detection for the individual REE were calculated to be as follows (wt.%): Y (0.031), La (0.040), Ce (0.058), Pr (0.100), Nd (0.093), Sm (0.051), Gd (0.110), Dy (0.114). The data were reduced and corrected using the PAP routine (Pouchou and Pichoir, 1984). 4. Results The results of the solution ICPMS and LA-ICPMS trace element analyses of carbonatites, calcite and uorapatite are presented in Table 1. The compositions of calcite and REE minerals determined by electron-microprobe analysis are given in Tables 2 and 3. The complete set of analytical data is available as Supplementary online data appended to this article. 4.1. Carbonatites and carbonate minerals The Miaoya carbonatites are characterized by consistently high abundances of Sr and light REE (LREE), but variable levels of Rb, Ba, Th, U, Nb and Ta, which reects local variations in the content of biotite and Nb-bearing accessory phases (particularly, perovskite). In traceelement distribution diagrams normalized to the primitive mantle composition (Fig. 3), all samples show negative Pb, Zr and Hf anomalies typical of calcite carbonatites worldwide (Nelson et al., 1988; Woolley and Kempe, 1989; and references therein). The degree of LREE enrichment varies signicantly among the studied samples [(La/Yb)N = 2754] and appears to correlate with the Ba, Th and Pb contents (Fig. 3A). The chondrite-normalized REE proles are smooth and lack obvious Ce or Eu anomalies (Fig. 3C). In contrast, the syenite sample altered and impregnated by calcite veinlets (MY-17) shows higher levels of Rb and Ba, but lower levels of Sr and REE (Fig. 3A). All examined carbonatites are composed predominantly of Sr-rich calcite (0.22.8 wt.% SrO) with low to moderate Fe, Mn and Mg contents (02, 11.2 and 0.20.7 wt.% respective oxides). The REE abundances, measured by LA-ICPMS, range from 95 to 258 ppm and

Fig. 3. (A, B) Trace-element abundances in Miaoya carbonatites, contaminated syenite (WR), calcite (C) and uorapatite (Ap) normalized to the primitive mantle compositions. (C, D) Chondrite-normalized REE and Y abundances in the Miaoya carbonatites, contaminated syenite, calcite and uorapatite; MY-17 and MY-17C are contaminated syenite and latestage calcite from veinlets in sample MY-17, respectively. Normalization values for both primitive mantle and chondrite are from McDonough and Sun (1995).

150

Table 1 ICPMS and LA-ICPMS analyses of carbonatites, contaminated syenite (MY-17) and calcite, apatite crystals, respectively (ppm). Sample MY-7 number WR Calcite Core Core RimCore Rb Sr Y Zr Nb Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U 27.1 3724 158 129 388 328 629 1076 82.5 288 44.2 12.7 43.0 5.76 32.4 6.23 16.2 1.98 10.8 1.40 1.46 21.0 34.8 104 36.0 0.05 8301 88.2 BDL 0.20 83.9 18.3 49.8 7.25 35.6 10.8 4.32 13.0 2.28 14.9 2.98 8.05 NA 7.89 1.09 NA 0.01 13.9 BDL 0.01 0.09 9228 85.1 BDL 0.01 71.6 22.5 62.9 8.92 42.0 10.8 4.14 12.3 2.17 13.7 2.88 7.75 NA 7.59 1.05 NA BDL 12.9 0.01 0.02 0.16 6143 96.0 0.75 0.01 34.1 12.3 34.1 5.09 27.0 9.26 3.97 12.0 2.26 15.0 3.15 8.60 NA 9.02 1.30 NA BDL 34.2 BDL 0.05 BDL 8653 105 0.26 0.13 154 33.7 81.9 10.9 51.1 14.2 5.55 16.8 2.91 18.2 3.56 9.66 NA 8.76 1.24 NA BDL 10.6 0.37 0.21 Core 0.09 10,533 88.7 0.07 0.03 199 31.1 75.0 10.8 47.9 13.1 4.96 14.7 2.45 15.3 3.06 8.0 NA 7.80 1.14 NA BDL 8.10 0.03 0.03 MY-10 Apatite Core Core Core Core Core Core Core BDL 5713 84.8 BDL BDL 115 26.6 66.0 8.89 42.0 11.2 4.34 13.2 2.27 14.2 2.92 8.11 NA 8.12 1.22 NA BDL 7.0 BDL BDL 0.09 6278 85.2 0.22 0.17 54.3 15.6 44.6 6.48 32.3 9.33 4.0 11.4 2.10 13.7 2.81 7.89 NA 8.48 1.20 NA 0.01 18.9 0.16 0.12 BDL 5126 87.3 BDL 0.01 77.9 19.6 54.0 7.89 39.2 10.9 4.39 12.6 2.18 14.0 2.81 7.94 NA 9.14 1.34 NA BDL 15.1 0.02 0.01 0.06 8897 293 2.23 0.69 21.8 1393 2957 340 1300 194 53.0 137 16.5 76.2 11.7 24.0 NA 12.8 1.33 NA 0.01 2.64 28.5 0.70 0.08 8193 337 6.03 1.88 24.7 1369 3005 353 1373 216 57.6 152 18.5 86.6 13.4 28.1 NA 14.8 1.61 NA 0.03 3.48 46.8 2.79 0.06 9482 254 0.60 0.20 21.3 1195 2474 276 1031 157 43.7 114 13.8 64.7 10.1 20.6 NA 11.0 1.16 NA BDL 2.62 12.7 0.45 0.11 7294 472 10.4 3.16 25.8 1320 3006 362 1438 241 66.9 183 23.7 116 18.3 37.9 NA 21.2 2.32 NA 0.06 5.31 74.1 3.95 3.54 6883 122 79.4 54.0 179 514 852 63.4 221 34.0 10.5 33.1 4.03 20.7 4.02 10.8 1.43 8.42 1.17 0.56 0.78 19.6 4.37 2.35 WR Calcite Core Core Core Core Core Core Core Core BDL 5865 114 BDL 0.01 29.8 10.9 26.1 3.71 17.9 6.15 3.37 9.96 2.09 15.6 3.45 10.5 NA 11.9 1.72 NA BDL 58.2 0.02 0.04 BDL 5955 121 BDL BDL 27.8 18.0 36.2 4.48 19.8 6.78 3.57 10.1 2.09 15.6 3.80 12.2 NA 14.6 2.29 NA BDL 57.4 BDL 0.03 BDL 6137 105 BDL BDL 21.3 12.0 28.3 3.72 17.9 5.87 2.85 8.48 1.79 13.8 3.25 10.0 NA 12.0 1.76 NA BDL 55.5 BDL 0.01 BDL 5344 108 BDL 0.01 24.5 13.3 30.3 4.05 18.9 5.90 3.20 9.51 1.98 14.5 3.40 10.1 NA 11.6 1.78 NA BDL 50.7 BDL 0.06 BDL 4207 102 0.19 BDL 32.2 14.6 34.1 4.58 20.0 6.40 3.30 9.01 1.92 14.3 3.32 10.0 NA 12.8 2.01 NA BDL 55.4 BDL 0.06 BDL 4950 111 BDL BDL 26.3 15.0 33.1 4.30 20.1 6.27 3.43 10.1 2.09 15.3 3.53 10.7 NA 12.3 1.70 NA BDL 49.7 BDL 0.07 BDL 5549 104 BDL 0.01 40.8 14.0 29.6 3.93 17.8 5.88 2.69 9.22 2.01 14.2 3.27 9.78 NA 12.0 1.77 NA BDL 29.9 0.02 0.10 BDL 5470 114 0.02 BDL 26.0 12.0 29.7 4.27 21.2 7.33 3.64 11.0 2.22 15.6 3.61 10.9 NA 13.1 1.84 NA BDL 56.1 BDL 0.04 2.15 5091 105 11.2 122 433 589 1054 80.2 278 41.9 12.2 38.1 4.33 20.6 3.76 10.3 1.37 8.56 1.26 0.23 0.98 8.14 1.71 0.96 MY-13 WR Calcite RimCore BDL 1586 58.3 0.01 BDL 31.9 24.1 58.5 7.81 34.7 8.77 3.27 9.48 1.49 9.16 2.01 6.05 NA 7.82 1.40 NA BDL 4.73 BDL BDL BDL 1659 58.2 BDL BDL 32.6 24.7 61.5 8.24 37.0 9.33 3.44 9.69 1.56 9.49 1.88 5.41 NA 6.44 1.0 NA BDL 4.61 BDL BDL RimCore BDL 1773 70.4 0.01 0.13 31.5 14.4 40.6 6.15 30.4 8.66 3.49 9.76 1.68 10.5 2.28 6.81 NA 7.59 1.21 NA BDL 7.37 BDL BDL BDL 1286 58.7 BDL 0.01 30.0 23.7 58.4 7.84 35.3 8.81 3.28 9.14 1.56 9.48 2.11 6.11 NA 8.0 1.43 NA BDL 4.16 0.03 0.05 Core Core Core Core RimCore 0.03 1549 67.0 BDL BDL 41.0 21.0 59.9 8.74 40.2 10.5 3.81 10.7 1.71 10.7 2.20 6.25 NA 7.36 1.15 NA BDL 13.8 BDL BDL BDL 1355 60.7 BDL 0.01 43.6 23.6 60.1 8.17 36.6 9.07 3.31 9.51 1.59 9.20 1.96 5.69 NA 6.52 1.02 NA BDL 9.29 0.01 BDL BDL 58.4 BDL BDL 37.4 16.2 49.2 7.36 33.3 8.65 3.23 8.74 1.43 8.76 1.87 5.65 NA 7.05 1.13 NA BDL 9.97 0.07 BDL BDL 1700 73.0 BDL BDL 46.1 15.7 48.3 7.42 35.3 10.3 3.92 11.2 1.91 11.9 2.41 6.86 NA 7.38 1.08 NA BDL 21.6 0.09 BDL BDL 2009 78.7 BDL BDL 50.6 18.3 51.9 7.77 37.1 11.0 4.10 11.8 2.05 13.0 2.69 7.44 NA 7.87 1.13 NA BDL 25.0 BDL BDL 0.05 C. Xu et al. / Lithos 118 (2010) 145155 77.0 0.01 0.02 63.0 21.8 59.0 8.25 37.1 10.4 3.86 11.2 1.94 12.1 2.39 7.17 NA 7.51 1.11 NA BDL 28.3 BDL BDL

Table 1 ICPMS1 (continued) analyses of carbonatites, contaminated syenite (MY-17) and calcite, apatite crystals, respectively (ppm). Table and LA-ICPMS Sample MY-14 number WR Calcite Core Rb Sr Y Zr Nb Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf Ta Pb Th U 5.30 2446 90.2 32.7 758 512 280 548 43.6 156 25.3 7.78 23.5 2.95 15.2 2.91 8.4 1.18 7.66 1.17 0.53 1.91 16.1 3.60 0.77 BDL 2021 48.3 BDL 0.01 32.1 17.2 45.5 6.67 31.9 7.81 2.72 7.65 1.14 6.96 1.46 4.47 NA 5.49 0.87 NA BDL 9.76 BDL BDL Core BDL 1732 39.5 BDL BDL 36.4 17.4 42.4 5.58 24.8 6.01 2.05 5.55 0.88 5.48 1.25 3.77 NA 5.34 0.93 NA BDL 8.66 BDL BDL RimCore BDL 1881 45.2 BDL BDL 34.2 18.0 47.1 6.60 31.1 7.55 2.54 7.06 1.07 6.62 1.35 4.03 NA 4.98 0.79 NA BDL 8.85 BDL BDL BDL 2000 49.3 BDL BDL 32.3 19.1 49.7 7.06 33.1 8.18 2.83 7.78 1.19 7.04 1.50 4.42 NA 5.56 0.90 NA BDL 9.55 BDL BDL Core BDL 1881 43.7 BDL BDL 34.8 14.1 40.2 5.99 28.5 7.22 2.41 6.93 1.03 6.29 1.30 4.06 NA 5.10 0.83 NA BDL 8.93 BDL BDL Core BDL 1679 44.7 BDL 0.01 40.3 14.6 39.7 5.67 27.1 6.76 2.36 6.67 1.05 6.40 1.41 4.26 NA 5.48 0.87 NA BDL 9.28 BDL BDL RimCore BDL 1937 43.6 BDL 1.24 51.8 10.3 28.3 4.41 22.3 6.54 2.46 6.56 1.06 6.30 1.31 4.06 NA 5.49 0.92 NA 0.02 9.39 0.03 0.03 BDL 1697 44.7 BDL BDL 34.5 17.2 41.3 5.63 26.2 6.50 2.33 6.64 1.04 6.57 1.35 4.21 NA 5.26 0.86 NA BDL 9.59 BDL BDL Core BDL 2482 59.9 0.02 BDL 30.5 17.9 48.1 6.95 34.3 9.88 3.58 10.4 1.67 10.3 2.10 6.0 NA 7.0 1.06 NA BDL 10.9 BDL BDL RimCore BDL 2463 57.9 BDL 0.01 31.7 12.2 35.5 5.48 27.9 9.01 3.43 9.40 1.58 9.78 2.0 5.74 NA 6.57 1.03 NA BDL 10.8 0.02 0.03 BDL 2288 57.2 BDL BDL 39.2 19.8 50.9 7.30 34.3 9.11 3.35 9.47 1.59 9.80 2.0 5.76 NA 6.51 1.07 NA BDL 9.74 0.01 0.02 RimCore BDL 2288 55.0 1.47 0.57 55.7 10.7 31.9 5.06 26.4 8.27 3.11 9.05 1.51 9.01 1.89 5.46 NA 6.24 0.95 NA BDL 9.81 0.03 0.01 BDL 2277 55.5 BDL BDL 31.4 11.8 35.2 5.48 27.1 8.60 3.23 9.37 1.54 9.53 1.88 5.51 NA 6.31 0.98 NA BDL 9.86 BDL BDL Rim BDL 2293 55.7 BDL 0.01 36.3 12.2 35.7 5.50 28.7 8.53 3.29 9.39 1.57 9.35 1.90 5.56 NA 6.22 0.93 NA BDL 9.85 BDL BDL Core BDL 2373 57.6 BDL 1.25 42.5 14.4 38.7 5.97 30.2 9.22 3.44 9.66 1.63 9.92 2.06 5.72 NA 6.54 0.93 NA BDL 10.3 0.04 0.01 RimCoreRim BDL 2578 59.2 BDL BDL 32.5 16.4 45.6 6.87 34.3 9.70 3.57 10.0 1.69 10.1 2.12 6.04 NA 6.88 1.02 NA BDL 11.7 BDL BDL BDL 2490 58.0 BDL BDL 37.3 19.9 51.6 7.44 36.0 9.62 3.61 10.3 1.63 10.1 2.07 5.92 NA 6.71 1.03 NA BDL 11.2 BDL BDL BDL 2552 58.4 BDL 0.01 29.2 16.3 45.3 6.77 33.3 9.87 3.56 9.92 1.65 10.1 2.01 5.96 NA 6.86 1.05 NA BDL 11.8 0.02 0.05
208

MY-15 WR Calcite Core 27.3 4907 88.3 102 110 3503 463 861 68.3 245 35.7 10.3 32.5 3.58 17.1 3.08 8.13 1.0 5.83 0.83 1.12 5.33 34.3 8.28 13.8 Pb. 0.02 1905 34.0 BDL BDL 26.0 17.9 43.0 5.68 25.4 5.53 2.04 5.90 0.84 5.51 1.22 4.12 NA 5.40 0.94 NA BDL 16.3 BDL BDL Core BDL 2738 47.7 IE IE 37.3 9.59 26.4 3.85 18.9 5.0 2.18 7.01 1.16 7.83 1.69 5.21 NA 5.76 0.89 NA IE 5.78 BDL IE

MY-16 WR Calcite Core 3.27 11,000 112 18.7 9.38 202 543 891 65.2 224 33.8 10.1 33.0 3.83 19.1 3.63 10.0 1.30 7.63 1.06 0.22 0.27 17.1 4.69 1.53 BDL 3438 70.0 IE 0.01 79.0 18.4 61.1 8.95 40.5 11.7 4.25 12.4 1.91 11.8 2.30 6.32 NA 5.76 0.83 NA IE 9.71 IE IE Core BDL 3381 67.6 IE 0.03 67.9 15.6 55.8 8.37 39.1 11.8 4.50 12.9 1.95 12.3 2.42 6.83 NA 6.17 0.95 NA IE 8.46 IE IE

MY-17 WR Calcite Core 113 702 25.0 65.7 1151 15,169 154 360 29.5 97 12.7 3.60 12.1 1.06 4.5 0.84 2.40 0.31 1.94 0.29 1.93 2.95 25.2 10.9 2.24 BDL 2703 143 0.08 0.01 33.1 9.42 28.4 4.54 24.7 11.1 5.78 17.2 3.23 20.1 3.87 10.2 NA 9.04 1.20 NA BDL 41.4 0.02 BDL Core 0.06 2798 146 BDL BDL 33.7 6.69 21.7 3.56 20.4 10.2 5.51 16.5 3.17 20.1 3.87 10.3 NA 9.18 1.25 NA BDL 41.3 0.10 BDL Core BDL 3343 149 0.03 0.01 29.9 7.29 23.0 3.80 21.5 10.6 5.61 17.2 3.24 20.2 3.90 10.4 NA 9.43 1.25 NA BDL 46.1 0.02 BDL Core BDL 2628 141 0.04 0.02 28.4 6.80 21.9 3.61 20.3 9.85 5.25 15.9 3.0 18.8 3.66 9.69 NA 8.78 1.22 NA BDL 49.9 BDL BDL Core BDL 2574 125 0.05 0.01 32.3 5.70 16.3 2.71 15.7 8.44 4.53 13.9 2.52 15.7 2.98 7.73 NA 7.85 1.09 NA BDL 32.2 1.57 0.54 Core BDL 3008 146 0.61 0.12 52.9 8.43 20.2 3.28 20.0 10.15 5.37 16.3 2.96 18.3 3.61 9.65 NA 9.69 1.37 NA BDL 46.8 1.55 0.17

C. Xu et al. / Lithos 118 (2010) 145155

WR, whole-rock; BDL, below detection limit; NA, not analyzed; IE, inclusion encountered; Pb concentrations in calcites are modeled from

151

152 Table 2 Main element compositions of carbonates (wt.%). Sample In carbonatite MeaN FeO MnO MgO CaO SrO CO2a Total
a

C. Xu et al. / Lithos 118 (2010) 145155

n = 15 2 1.40 0.54 0.26 2.42 1.62 0.76 1.58

In veinlets cross-cutting the syenites Mean 1.24 3.90 0.33 49.32 1.66 42.94 99.38

n=5 2 0.75 1.31 0.16 0.70 0.52 0.64 1.42

0.75 0.48 0.40 53.37 1.24 43.61 99.84

Calculated by charge balance.

produce an essentially at and smooth chondrite-normalized prole (Fig. 3D). In some calcite crystals, the LREE contents gradually decrease from the core toward the rim, but no systematic variation in either Sr or heavy REE (HREE) was observed (Table 1). In contrast to the wholerock patterns, the primary calcite does not show any signicant enrichment in LREE relative to HREE [(La/Yb)CN = 0.62.7]. Also in contrast to the whole-rock trace-element distribution, the calcite is signicantly enriched in Sr and Pb relative to LREE [(Sr/La)PM = 213; (Pb/La)PM = 214]. With the exception of Ba, U and, in some samples, Th, other trace elements are at or below their lower limit of detection by LA-ICPMS (Fig. 3B). Late-stage calcite in veinlets cross-cutting the syenites (Fig. 2) has Sr and Mg contents (1.21.9 and 0.30.5 wt.% respective oxides) similar to those in the carbonatitic calcite, but signicantly higher Mn abundances (up to 4.7 wt.% MnO). The traceelement composition of the late-stage calcite is also noticeably different, being relatively depleted in LREE [(La/Nd)CN = 0.60.8]. The (La/Yb)CN ratio is highest in the whole-rock analyses (on average, 41), as compared with an average value of 1.6 in the primary calcite and 0.6 in the late-stage calcite veins. 4.2. REE minerals Fluorapatite in the carbonatite contains high F and SrO contents (on average, 4.1 and 1.31 wt.%, respectively) typical of carbonatitic apatite (Chakhmouradian et al., 2008). In comparison with the calcite, uorapatite is signicantly enriched in REE and especially LREE [(La/ Yb)CN = 4274]. Note that compositions with N3.8 wt.% F give nonstoichiometric formulae, in which the number of F atoms exceeds one.

The highest recorded F content (4.4 wt.%) corresponds to 1.2 atoms of F per formula unit. Similar F-rich compositions (up to 1.35 atoms of F per formula unit) have been described in the literature from various rock types, including carbonatites (e.g., Wall and Mariano, 1996; Bhn et al., 2001). Although it is sometimes argued that the excess of F is an analytical artifact arising from beam-induced ion diffusion in the sample, there are numerous examples of non-stoichiometric uorapatites analyzed by techniques that would not produce such artifacts (Binder and Troll, 1989; McClellan and Van Kauwenbergh, 1991). It is also noteworthy that synthetic apatite-type compounds with a signicant excess of tunnel anions (up to 1.25 atoms) are well known in materials science (e.g., Guillot et al., 2009). Monazite-(Ce) shows relatively little compositional variation and is strongly enriched in LREE relative to the heavier REE, of which only Gd, Dy and Y were quantied [average (La/Y)CN = 271]. Comparison of the Miaoya compositions with those for the Kola carbonatites (Wall and Zaitsev, 2004) shows that the latter have comparable levels of Th (typically b 0.5 wt.% ThO2), but generally higher Sr and lower Y contents (0.9 and 0.2 wt.% respective oxides). To our knowledge, HREE-enriched compositions comparable to those obtained in the present study have been previously reported only from the Lofdal carbonatite in Namibia (Wall et al., 2008). Bastnsite-(Ce) and synchysite-(Ce) form syntaxial intergrowths that formed in the nal stages of carbonatite evolution (Fig. 2). From bastnsite-(Ce) to synchysite-(Ce), the Ca and HREE contents increase, while the LREE and F contents decrease. Both minerals show strong enrichment in LREE relative to the heavier REE; the (La/Y)CN ratio varies from an average value of 110 in synchysite-(Ce) to ca. 500 in basnsite-(Ce) (Fig. 4). 5. Discussion In order to understand the behavior of REE during the evolution of carbonatitic magma, the whole-rock REE budget needs to be considered in the context of trace-element data for individual minerals and REE partition coefcients. Unfortunately, partitioning of trace elements between carbonate melts and rock-forming carbonates has not been studied experimentally. The data available in the literature are limited to qualitative assessments of traceelement (including REE) partitioning between calcite and melt or coprecipitating phases (e.g., Bhn et al., 2001; Dawson and Hinton,

Table 3 Representative analysis of REE minerals in Miaoya carbonatites (wt.%). Fluorapatite Mean CaO FeO SiO2 P2O5 SrO ThO2 Y2O3 La2O3 Ce2O3 Pr2O3 Nd2O3 Sm2O3 Gd2O3 Dy2O3 CO2 H2O F F=O Total 54.44 NA 0.13 37.77 1.29 NA BDL 0.10 0.22 BDL 0.14 BDL BDL BDL 2.71 0 3.87 1.63 99.03 n = 10 2 1.68 0.32 2.0 0.50 Monazite-(Ce) Mean 0.46 BDL 0.12 31.10 NA 0.29 0.56 18.03 32.95 3.09 10.38 1.65 0.95 0.29 0 0 BDL 99.88 n = 10 2 1.29 0.06 0.36 0.27 0.44 2.16 1.45 0.58 0.96 0.36 0.22 0.10 0 0 Synchysite-(Ce) Mean 16.10 0.18 0.14 BDL NA 0.10 0.84 13.00 24.79 2.23 9.88 1.29 0.77 0.29 27.11 0.46 4.98 2.10 100.01 n = 10 2 0.97 0.30 0.15 Bastnsite-(Ce) Mean 0.25 0.20 0.40 BDL NA 0.12 0.31 20.83 37.40 2.81 9.66 1.03 0.37 BDL 20.05 0.63 7.33 3.09 98.31 n = 10 2 0.20 0.40 0.66

0.17 0.38 0.28

1.35 0 0.86 0.37 1.52

1.08

0.05 0.36 1.18 1.11 0.30 2.45 0.35 0.22 0.09 0.89 0.64 1.32 0.55 2.73

0.12 0.18 1.65 1.41 0.42 2.00 0.85 0.37 0.27 0.68 1.46 0.62 1.06

Abbreviations as in Tables 1 and 2. CO2 in uroapatite is calculated by stoichiometry and in other REE minerals by charge balance. H2O is calculated by charge balance.

C. Xu et al. / Lithos 118 (2010) 145155

153

Fig. 4. Chondrite-normalized REE + Y abundances in synchysite-(Ce), bastnsite-(Ce) and monazite-(Ce) in the Miaoya carbonatites. Normalization values for chondrite are from McDonough and Sun (1995).

2003). Carbonatitic calcite generally contains lower levels of REE than its host rock (e.g., Hornig-Kjarsgaard, 1998; Xu et al., 2008), implying that REE partition coefcients of calcite (Cal/LDREE) are b1. Other minerals precipitated early in the evolution of most carbonatites include apatite, magnetite, olivine and clinopyroxene (e.g., Le Bas, 1989; Chakhmouradian et al., 2008), of which apatite is capable of sequestering signicant amounts of REE. Thus, Hornig-Kjarsgaard (1998) argued that rock-forming carbonates together with apatite control the overall budget and relative enrichment of specic REE in carbonatites. Bhn et al. (2001) modeled REE fractionation in a carbonatitic liquid at various proportions of precipitating calcite, uorapatite and clinopyroxene, and Cal/LDREE ranging from 0.05 for La to 0.01 for Lu. Their results (LaCN = 2179335; La/YbCN = 67361) are consistent with the published data for carbonatitic calcite in both the degree of REE enrichment and slope of chondrite-normalized REE prole (e.g., Eby, 1975; Hornig-Kjarsgaard, 1998; Xu et al., 2008). By contrast, our data for the Miaoya carbonatitic calcite are in marked contrast with the published data for primary calcite from carbonatites worldwide and the ndings of Bhn et al. (2001). It is noteworthy that REE minerals are typically regarded as late-stage or secondary phases and, as such, are omitted from consideration. The primary calcite in the Miaoya carbonatites is characterized by low REE contents and at REE proles, suggesting that REE were fractionated from the parental carbonatite liquid prior to the precipitation of the calcite, possibly in a transient magma chamber below the present erosion level. The initial carbonatite liquid was rich in P and F, as manifested by the common occurrence of uorapatite in these carbonatites. The association of uorapatite with monazite-(Ce) indicates initially high REE abundances in the magma. Fractional crystallization of these two minerals resulted in depletion of the residual liquid in REE and especially LREE. This conclusion is supported by the experimental data of Wyllie et al. (1996) that initially low REE concentrations in the magma are dispersed among major rock-forming constituents (predominantly carbonates), and REE mineralization does not develop. This camouage effect explains why calcite in most carbonatites is characterized by relatively high REE contents and signicant enrichment in LREE, i.e. positive slope of chondrite-normalized REE proles (e.g., Eby, 1975; Hornig-Kjarsgaard, 1998; Xu et al., 2008). Hornig-Kjarsgaard (1998) suggested that the LREE-enriched signature of carbonatites reects the chemistry of their parental magmas and cannot be reasonably explained by crystal fractionation. Ionov and Harmer (2002), on the contrary, found that the REE abundances in carbonate melts entrapped in mantle xenoliths are very low. They suggested that mantle-derived carbonate liquids are REE-poor, and extreme REE enrichment observed in many carbona-

tites requires that these primary liquids undergo complex fractionation processes during their ascent. However, it is noteworthy that some mantle-derived primary carbonate liquids contain high levels of REE (ca. 1000 ppm) and show strong LREE enrichment (Moine et al., 2004). Thus, at least some pristine carbonatitic magmas contain high REE levels that may be further enhanced and modied by fractionation processes. Both monazite and apatite in carbonatites show a strong afnity for LREE (Hornig-Kjarsgaard, 1998; Wall and Zaitsev, 2004), which is consistent with the trace-element data obtained in the present work. Thus, early precipitation of these minerals will lower the proportion of LREE in the residual liquid prior to the crystallization of carbonates. Although there appears to be a consensus that the Cal/LDREE coefcients are low, neither their absolute values nor the magnitude of variation across the REE series have been determined. In Bhn et al. (2001)'s model discussed above, calcite is assumed to have a greater capacity for LREE relative to HREE. If this is the case, i.e. the Cal/LDREE values decrease from La toward Lu, precipitation of voluminous quantities of calcite should produce an evolved derivative liquid enriched in HREE relative to LREE, particularly if calcite was preceded on the liquidus by LREE-rich phosphates (as is the case at Miaoya), and most LREE were removed. Clearly, such HREE enrichment did not occur at Miaoya, where the latest phases to crystallize are REEuorocarbonates showing the highest (La/Y)CN values of all minerals studied here. This implies that the residual liquids that precipitated REE-uorocarbonates had higher LREE/HREE ratios than the primary parental liquids. The Cal/LDREE values may be affected by (1) crystalchemical effects arising from charge/size interrelations of substituent elements, and (2) complexing properties of the melt, where both F and Cl are likely to become enriched with fractional crystallization and, hence, change the partitioning behavior of REE, Y and, perhaps, other high-charge cations (Ellison and Hess, 1989; Wood, 1990). To reconcile the observed discrepancy, we propose that Cal/LDREE values actually increase with decreasing ionic radius (i.e., with increasing atomic number in the REE series). Combining the apatitecarbonatite melt partition coefcients (Ap/LDREE) determined experimentally by Klemme and Dalp (2003) with the calcite uorapatite coefcients estimated by Dawson and Hinton (2003) gives Cal/LDREE values ranging from 0.08 for La, 0.14 for Sm to 0.21 for Y and 0.27 for Lu. This overall trend is in agreement with our interpretation that LREE are less compatible with respect to calcite than HREE. Although Klemme and Dalp (2003) conducted their experiments at different temperature (T) and pressure (P) relative to those at which apatite and calcite equilibrate in plutonic carbonatites (Dawson and Hinton, 2003), it is extremely unlikely that the relative compatibility of large and small REE cations will reverse. For example, the Cal/LDREE values calculated from the data of Guzmics et al. (2008) for high-P (2.23.8 GPa) partitioning of REE between apatite and a carbonatitic melt in mantle xenoliths, conrm that La (0.33) is about two times less compatible than Y (0.77) and three times less compatible than Lu (0.98) in calcite. Because the conditions of Klemme and Dalp (2003) 's experiment (1 GPa and 1250 C) are closer to the conditions of early magmatic crystallization at Miaoya than those of apatitecarbonatite equilibration in the mantle (Guzmics et al., 2008), and also because empirically determined Ap/L DREE values may not be very accurate, we will use the coefcients calculated on the basis of Klemme and Dalp (2003)'s data to model changes in the REE budget of the parental carbonatitic liquid in response to precipitation of REE-rich phosphates (monazite and apatite) and calcite based on the Rayleigh equation: Cs = Ds C0 F
Ds 1

where CS and C0 are the concentrations of an element in the crystallizing mineral and in the melt, respectively, F is the fraction of melt remaining, and DS is the mineral-melt distribution coefcient.

154

C. Xu et al. / Lithos 118 (2010) 145155

The changes in composition of the parental melt in response to mineral crystallization were monitored using the following equation: C0 =
Cs

=D s F

Ds 1

The monazite and apatite are considered to have crystallized nearly contemporaneously, and the DS values for apatite were taken from Klemme and Dalp (2003). Our model shows that the carbonatite liquid undergoes a dramatic depletion in LREE with increasing fractionation of uorapatite and, subsequently, calcite (Fig. 5). Partitioning of LREE into the phosphates early in the evolutionary history leads to depletion of the evolved liquid in LREE relative to HREE with (La/Y)CN ratios decreasing from an average value of 37 to 3.8 (Fig. 5). More than 90% of the total LREE budget of the parental liquid is sequestered in the phosphates prior to calcite crystallization. It is thus clear that, if the parental carbonatitic liquid is saturated with respect to P2O5, the bulk of REE will be removed early in the evolutionary history either in apatite-group minerals or in monazite. However, fractionation of voluminous quantities of calcite characterized by low Cal/LDREE values will produce residual liquids strongly enriched in REE and especially LREE, potentially to the level where REE-uorocarbonates become a stable liquidus phase. The late REE-uorocarbonates in most carbonatites show euhedral morphology and are not associated with uorite or any other gangue mineral commonly accompanying bastnsite and related minerals in hydrothermal deposits (e.g., Williams-Jones et al., 2000; Samson et al., 2004; Xu et al., 2008). Syntaxial intergrowth of Ca-rich and Ca-poor REE-uorocarbonates (Fig. 2) is considered to reect changes in the chemical composition of the parental liquid (Ni et al., 1993). All these features indicate that the REE-uorocarbonates in the Miaoya carbonatites are primary phases rather than products of interaction of postmagmatic uorine-rich aqueous uids causing local remobilization and re-deposition of REE (cf. Ngwenya, 1994; Schmitt et al., 2002; Bhn, 2008). In carbonatites, REE minerals commonly follow a trend of progressive Nd enrichment and decreasing La/NdCN ratio from early, relatively high-temperature mineralization types to lowtemperature replacement assemblages; this trend sometimes culminates in enrichment of the most evolved mineral assemblages in HREE (Zaitsev et al., 1998; Smith et al., 2000). This evolutionary path is consistent with a change of uid regime from an early, orthomagmatic, CO2-rich uid to a late-stage aqueous uid, and with the

current understanding of REE speciation in carbothermal uids (e.g., Bhn et al., 2002). However, contrary to this expected trend, the REEuorocabonates at Miaoya show strong progressive enrichment in La and other LREE relative to the earlier-precipitated calcite. This discrepancy indicates that calcite has very low partition coefcients for all REE, but particularly for LREE, leading to their enrichment in the residual liquid. This conclusion is in accord with the calculated REE distributions (Fig. 5). It is noteworthy that the late calcite in the contaminated syenites (Fig. 2) has high Sr contents typical of carbonatitic calcite from around the world (e.g., Hornig-Kjarsgaard, 1998; Yang and Le Bas, 2004; Xu et al., 2007, 2008). Severe depletion of the late carbonatitic liquid in LREE following the precipitation of REE-uorocarbonates explains the appearance of calcite with low (La/ Nd)CN values in the syenites affected by carbonatite-induced metasomatism. 6. Conclusions Carbonatites from the Miaoya REE deposit contain high levels of REE and exhibit signicant enrichment in LREE relative to HREE. Whole-rock and mineral analyses of these carbonatites show that their REE budget is controlled by minerals containing high levels of REE (i.e., uorapatite, monazite and REE-uorocarbonates). By contrast, the rock-forming calcite in the Miaoya carbonatites contains low levels of REE and shows at chondrite-normalized REE patterns. Textural evidence shows that the monazite and uorapatite crystallized prior to the REE-poor calcite, whereas the REE-uorocarbonates postdated the calcite. Our data indicate that the initial carbonatite liquids may be extremely rich in REE, but fractional crystallization of LREE-rich phosphate minerals can lower the proportion of REE and especially LREE in the residual liquid prior to calcite precipitation. Fractionation of voluminous calcite at the main magmatic stage will produce late-stage carbonatitic liquids extremely enriched in light lanthanides, assuming that calcite/melt partition coefcients are well below unity and increase from La toward Lu. In summary, the REE distribution and economic potential of carbonatites is very sensitive to crystal-fractionation processes and may involve multiple dramatic variations in REE budget driven by fractionation of a specic type of minerals (phosphates, calcite, and REE-uorocarbonates) in response to changes in the availability of (PO4)3, (CO3)2, F and other complexing ligands. Acknowledgments We thank Drs. M.F. Zhou, L. Qi for their generous help with traceelement analysis, and Drs. Anthony Mariano, Bernhard Bhn, Frances Wall and an anonymous reviewer for reviewing and improving the manuscript. Editorial comments by Nelson Eby improved sections of this manuscript. This research was nancially supported by the Chinese National Science Foundation (Nos. 40973040, 40773021) and the West Light Foundation of Chinese Academy of Sciences to Xu. ARC acknowledges the Natural Sciences and Engineering Research Council of Canada and The Anchor House for their continuing support. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.lithos.2010.04.003.

Fig. 5. Model calculations (assuming Rayleigh fractionation) for the REE concentrations in carbonatitic liquids precipitating apatite and calcite at different stages. The open and full symbols are for liquid compositions precipitating apatite and calcite, respectively. Average REE contents of apatite and calcite in sample MY-7 were used in the calculations. Ds values for apatite/carbonatite melt were taken from the experimental data (BS 19) of Klemme and Dalp (2003). Ds values for calcite/carbonatite melt were calculated from the apatite values using the inter-mineral partition coefcients of Dawson and Hinton (2003).

References
Binder, G., Troll, G., 1989. Coupled anion substitution in natural carbon-bearing apatites. Contributions to Mineralogy and Petrology 101, 394401. Bhn, B., 2008. The role of the volatile phase for REE and Y fractionation in low-silica carbonate magmas: implications from natural carbonatites, Namibia. Mineralogy and Petrology 92, 453470.

C. Xu et al. / Lithos 118 (2010) 145155 Bhn, B., Wall, F., Le Bas, M.J., 2001. Rare-earth element systematics of carbonatitc uorapatites, and their signicance for carbonatite magma evolution. Contributions to Mineralogy and Petrology 141, 572591. Bhn, B., Rankin, A.H., Dulski, P., 2002. The nature of orthomagmatic, carbonatitic uids precipitating REE, Sr-rich uorite: uid-inclusion evidence from the Okorusu uorite deposit, Namibia. Chemical Geology 186, 7598. Castor, S.B., 2008. The Mountain Pass rare-earth carbonatite and associated ultrapotassic rocks, California. Canadian Mineralogist 46, 779806. Chakhmouradian, A.R., Mumin, A.H., Demny, A., Elliott, B., 2008. Postorogenic carbonatites at Eden Lake, Trans-Hudson Orogen (northern Manitoba, Canada). Geological setting, mineralogy and geochemistry. Lithos 103, 503526. Chen, Y.J., Zhai, M.G., Jiang, S.Y., 2009. Signicant achievements and open issues in study of orogenesis and metallogenesis surrounding the North China continent. Acta Petrological Sinica 25, 26952726 (in Chinese, with English abstract). Dawson, J.B., Hinton, R.W., 2003. Trace-element content and partitioning in calcite, dolomite and apatite in carbonatite, Phalaborwa, South Africa. Mineralogical Magazine 67, 921930. Eby, G.N., 1975. Abundance and distribution of the rare-earth elements and yttrium in the rocks and minerals of the Oka carbonatite complex, Quebec. Geochimica et Cosmochimica Acta 39, 597620. Eggins, S.M., Woodhead, J.D., Kinsley, L.P.J., Mortimer, G.E., Sylvester, P., McCulloch, M. T., Hergt, J.M., Handler, M.R., 1997. A simple method for the precise determination of 40 trace elements in geological samples by ICPMS using enriched isotope internal standardization. Chemical Geology 134, 311326. Ellison, A.J.G., Hess, P.C., 1989. Solution properties of rare earth elements: inferences from immiscible liquids. Geochimica et Cosmochimica Acta 53, 19651974. Guillot, S., Beaudet-Savignat, S., Lambert, S., Vannier, R.-N., Roussel, P., Porcher, F., 2009. Evidence of local defects in the oxygen excess apatite La9.67(SiO4)6O2.5 from highresolution neutron powder diffraction. Journal of Solid State Chemistry 182, 33583364. Guzmics, T., Zajacz, Z., Kodolnyi, J., Halter, W., Szab, C., 2008. LA-ICP-MS study of apatite- and K feldspar-hosted primary carbonatite melt inclusions in clinopyroxenite xenoliths from lamprophyres, Hungary: implications for signicance of carbonatite melts in the Earth's mantle. Geochimica et Cosmochimica Acta 72, 18641886. Haas, J.R., Shock, E.L., Sassani, D.C., 1995. Rare earth elements in hydrothermal systems: estimates of standard partial modal thermodynamic properties of aqueous complexes of the rare earth elements at high pressures and temperatures. Geochimica et Cosmochimica Acta 59, 43294350. Hornig-Kjarsgaard, I., 1998. Rare earth elements in sovitic carbonatites and their mineral phases. Journal of Petrology 39, 21052121. Ionov, D., Harmer, R.E., 2002. Trace element distribution on calcitedolomite carbonatites from Spitskop: inferences for differentiation of carbonatite magmas and the origin of carbonates in mantle xenoliths. Earth and Planetary Science Letters 198, 495510. Klemme, S., Dalp, C., 2003. Trace-element partitioning between apatite and carbonatite melt. American Mineralogist 88, 639646. Le Bas, M.J., 1989. Diversication of carbonatites. In: Bell, K. (Ed.), Carbonatites: Genesis and Evolution. Unwin Hyman, London, pp. 114. Li, S., 1980. Geochemical features and petrogenesis of Miaoya carbonatites. Geochimica 4, 345355 (in Chinese, with English abstract). Longerich, H.P., Jackson, S.E., Gunter, D., 1996. Laser ablation inductively coupled plasma mass spectrometric transient signal data acquisition and analyze concentration calculation. Journal of Analytical Atomic Spectrometry 11, 899904. Lu, F.X., Zhang, B.R., Han, Y.W., Zhong, Z.Q., Ling, W.L., Zhang, H.F., Zheng, J.P., Hou, Q.Y., 2006. The Three-dimensional Lithospheric Chemical Structure in Qinling-DabieSulu Area. Geological Publish House, Beijing (in Chinese). Mattauer, M., Matte, Ph., Malavieille, L., Tapponnier, P., Maluski, H., Xu, Z.Q., Lu, Y.L., Tang, Y.Q., 1985. Tectonics of the Qinling Belt: build-up and evolution of eastern Asia. Nature 293, 212216. McClellan, G.H., Van Kauwenbergh, S.J., 1991. Mineralogical and chemical variation of francolites with geological time. Journal of the Geological Society 148, 809812. McDonough, W.F., Sun, S.S., 1995. The composition of the Earth. Chemical Geology 120, 223253. Meng, Q.R., Zhang, G.W., 2000. Geologic framework and tectonic evolution of the Qinling orogen, central China. Tectonophysics 323, 183196. Migdisov, A.A., Williams-Jones, A.E., Normand, C., Wood, S.A., 2008. A spectrophotometric study of samarium (III) speciation in chloride solutions at elevated temperatures. Geochimica et Cosmochimica Acta 72, 16111625. Moine, B.N., Grgoire, M., O'Reilly, S.Y., Delpech, G., Sheppard, S.M.F., Lorand, J.P., Renac, C., Giret, A., Cottin, J.Y., 2004. Carbonatite melt in oceanic upper mantle beneath the Kerguelen Archipelago. Lithos 75, 239252. Morogan, V., 1989. Mass transfer and REE mobility during fenitization at Aln, Sweden. Contributions to Mineralogy and Petrology 103, 2534. Nelson, D.R., Chivas, A.R., Chappell, B.W., McCulloch, M.T., 1988. Geochemical and isotopic systematics in carbonatites and implications for the evolution of ocean island sources. Geochimica et Cosmochimica Acta 52, 117.

155

Ngwenya, B.T., 1994. Hydrothermal rare earth mineralization in carbonatites of the Tundulu complex, Malawi: process at the uid/rock interface. Geochimica et Cosmochimica Acta 58, 20612072. Ni, Y., Hughes, J.M., Mariano, A.N., 1993. The atomic arrangement of bastnsite-(Ce), Ce (CO3)F, and structural elements of synchysite-(Ce), rntgenite-(Ce) and parisite(Ce). American Mineralogist 78, 415418. Pouchou, J.L., Pichoir, F., 1984. A new model for quantitative analysis. I. Application to the analysis of homogeneous samples. La Rcherche Aerosp 3, 1338. Qi, L., Hu, J., Gregoire, D.C., 2000. Determination of trace elements in granites by inductively coupled plasma mass spectrometry. Talanta 51, 507513. Qian, D.D., Li, J.Q., 1996. The Discovering History of Chinese Deposits: Hubei volume. Geological Publishing House (in Chinese), Beijing. Ratschbacher, L., Hacker, B.R., Calvert, A., Webb, L.E., Grimmer, J.C., McWilliams, M.O., Ireland, T., Dong, S.W., Hu, J.M., 2003. Tectonic of the Qinling (Central China): tectonostratigraphy, geochronology, and deformation history. Tectonophysics 366, 153. Ruberti, E., Enrich, G.E., Gomes, C.B., Comin-Chiaramonti, P., 2008. Hydrothermal rare earth mineralization in the Barra do Itapirapu (Brazil), multiple stockwork carbonatite. Canadian Mineralogist 46, 901914. Samson, I.M., Wood, S.A., Finucane, K., 2004. Fluid inclusion characteristics and genesis of the uoriteparisite mineralization in the Snowbird deposit, Montana. Economic Geology 99, 17271744. Schmitt, A.K., Trumbull, R.B., Dulski, P., Emmermann, A.R., 2002. ZrNb-REE mineralization in peralkaline granites from the Amis Complex, Brandberg (Namibia): evidence for magmatic pre-enrichment from melt inclusions. Economic Geology 97, 399413. Smith, M.P., 2007. Metasomatic silicate chemistry at the Bayan Obo Fe-REE-Nb deposit, Inner Mongolia, China: contrasting chemistry and evolution of fenitising and mineralizing uids. Lithos 93, 126148. Smith, M.P., Henderson, P., Campbell, L.S., 2000. Fractionation of the REE during hydrothermal processes: constraints from the Bayan Obo FeREENb deposit, Inner Mongolia, China. Geochimica et Cosmochimica Acta 64, 31413160. Sun, W.D., Li, S.G., Chen, Y.D., Li, Y.J., 2000. Zircon UPb dating of granitoids from South Qinling, Central China and their geological signicance. Geochimica 29, 209216 (in Chinese with English abstract). Sun, W.D., Li, S.G., Sun, Y., Zhang, G.W., Li, Q.L., 2002. Mid-paleozoic collision in the north Qinling: SmNd, RbSr and 40Ar/39Ar ages and their tectonic implications. Journal of Asian Earth Sciences 21, 6976. Wall, F., Mariano, A.N., 1996. Rare earth minerals in carbonatites: a discussion centred in the Kangankunde carbonatite, Malawi. In: Jones, A.P., Wall, F., Williams, C.T. (Eds.), Rare Earth Minerals: Chemistry, Origin and Ore Deposits. London, Mineralogical Society Series, vol. 7. Chapman and Hall, pp. 193225. Wall, F., Zaitsev, A.N., 2004. Rare earth minerals in Kola carbonatites. In: Wall, F., Zaitsev, A.N. (Eds.), Phoscorites and Carbonatites from Mantle to Mine: the Key Example of the Kola Alkaline Province. London, Mineralogical Society Series, vol. 10. Mineralogical Society, pp. 341373. Wall, F., Niku-Paavola, V.N., Storey, C., Mller, A., Jeffries, T., 2008. Xenotime-(Y) from carbonatite dykes at Lofdal, Namibia: unusually low LREE:HREE ratio in carbonatite, and the rst dating of xenotime overgrowths on zircon. Canadian Mineralogist 46, 861877. Williams-Jones, A.E., Samson, I.M., Olivo, G.R., 2000. The genesis of hydrothermal uorite-REE deposits in the Gallinas Mountains, New Mexico. Economic Geology 95, 327342. Wood, S.A., 1990. The aqueous geochemistry of the rare-earth elements and yttrium, 1. Review of available low-temperature data for inorganic complexes and the inorganic REE speciation of nature waters. Chemical Geology 82, 159186. Woolley, A.R., Kempe, D.R.C., 1989. Carbonatites: nomenclature, average chemical composition. In: Bell, K. (Ed.), Carbonatites: Genesis and Evolution. Unwin Hyman, London, pp. 114. Wyllie, P.J., Jones, A.P., Deng, J., 1996. Rare earth elements in carbonate-rich melts from mantel to crust. In: Jones, A.P., Wall, F., Williams, C.T. (Eds.), Rare Earth Minerals: Chemistry, Origin and Ore Deposits. The Mineralogical Society Series, vol. 7. Chapman and Hall, London, pp. 77103. 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, 267278. Xu, C., Campbell, I.H., Kynicky, J., Allen, C.M., Chen, Y.J., Huang, Z.L., Qi, L., 2008. Comparison of the Daluxiang and Maoniuping carbonatitic REE deposits with Bayan Obo REE deposit, China. Lithos 106, 1224. Yang, X.M., Le Bas, M.J., 2004. Chemical compositions of carbonate minerals from Bayan Obo, Inner Mongolia, China: implications for petrogenesis. Lithos 72, 97116. Zaitsev, A., Wall, F., Le Bas, M.J., 1998. REESrBa minerals from the Khibina carbonatites, Kola Peninsula, Russia: their mineralogy, paragenesis and evolution. Mineralogical Magazine 62, 225250.