Geochemical Journal, Vol. 38, pp.

363 to 381, 2004

HFSE, REE, and PGE geochemistry of three sedimentary rock-hosted disseminated gold deposits in southwestern Guizhou Province, China
ZHIWEI B AO,1,2* Z HENHUA ZHAO,1 JAYANTA GUHA2 and ANTHONY E. WILLIAMS-J ONES3
1 2

Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, Wushan, Guangzhou 510640, China Department of Applied Sciences, University of Quebec at Chicoutimi, Chicoutimi, P.Q., G7H 2B1, Canada 3 Department of Geological Sciences, University of McGill, Montreal, P.Q., H3A 2A7, Canada (Received January 14, 2003; Accepted February 23, 2004)

The sedimentary rock-hosted disseminated gold (SRHDG) deposits in southwest Guizhou Province, China, are located near the buried edge of the Yangtze Craton. Three deposits, Lannigou, Getang and Zimudang, were chosen for this study. The deposits are hosted by fine-grained clastic and calcareous sedimentary rocks and formed at low temperatures from low salinity solutions of low to intermediate pH and low fO2 . Chemical analyses of the ores and host rocks show that high field strength elements, such as Nb, Ta, Zr and Hf, were immobile during hydrothermal alteration. However, rare earth elements (REE) were mobile. Light rare earth elements (LREE) were depleted, whereas heavy rare earth elements (HREE) were enriched. A differentiation of platinum group elements (PGE) concentrations was documented. The ores are enriched in Pd and Pt, consistent with the higher solubility of these two elements in ore fluids. The Pd/Ir ratios of the ores are lower than those of the ultramafic rocks and the Emeishan basalts in the region. Thus, the ultramafic intrusives and basalts could not have been the main source of the PGEs and gold in the SRHDG deposits. Rather, these metals were derived mainly from the sedimentary host rocks as evidenced by the trace element and stable isotopic characteristics of the ores. Our data suggest that the SRHDG deposits in southwest Guizhou are most likely of amagmatic origin, which is consistent with the lack of igneous intrusives in the vicinity of the deposits, the very weak metamorphic grade of the rocks, the lack of metamorphic effects within the time range of the SRHDG deposits, and coincidence in time of gold mineralization with regional extension.
Keywords: REE, PGE, sedimentary rock-hosted disseminated gold deposit, geochemistry

INTRODUCTION Sedimentary rock-hosted disseminated gold (SRHDG) deposits, also referred to as Carlin-type gold deposits, were not recognized in China until the early 1980s (Liu et al., 1994). Since then, over 100 SRHDG deposits of various sizes have been discovered in Guizhou, Sichuan, Yunnan, Guangxi, Shanxi, and other provinces, and have become an important gold resource in China. The known SRHDG deposits are located mainly along the edge of the Precambrian Yangtze Craton, and are hosted in upper Permian to Triassic sedimentary rocks. The host rocks in southwest Guizhou Province are generally siliciclastic turbidite or carbonate formations with minor clayey basic volcanic rock debris, ranging from Cambrian to Triassic in age. Gold mineralization commonly occurs at the contact of siliciclastic and carbonate layers in siliciclastic rocks characterized by (1) enrichment in organic matter, (2) the presence of volcanic material, and
*Corresponding author (e-mail: baozw@gig.ac.cn) Copyright 2004 by The Geochemical Society of Japan.

(3) an elevated content of ferrous carbonate (ferroan dolomite-siderite). Most of the gold deposits in southwest Guizhou Province are localized by anticlines and reverse faults produced by the contractional phase of Yanshanian orogeny (190 to 65 Ma); therefore, gold mineralization probably took place during the subsequent extensional phase (Casaceli and Gemuts, 1985). Recent geochronologic studies suggest that the SRHDG deposits in southwest Guizhou formed around 80105 Ma (Su et al., 1998b; Zhang and Yang, 1992; Hu et al., 2002). The source(s) of water, gold, and other components of the ore fluids for SRHDG deposits have been extensively debated and there is still no conclusions. In the case of the mid-Tertiary Carlin-type deposits of the USA, the widely favored model is one in which the deposits form due to intrusion-driven circulation of meteoric water, plus or minus inputs of magmatic fluids (see the review of Hofstra and Cline, 2000). Meteoric fluid circulation resulting from crustal extension and/or widespread magmatism or to the ascent of metamorphic fluids to shallow levels in the crust might also be important for some
363

Fig. 1. Location of the main SRHDG mineralization areas in China relative to tectonic features and major cities. The SRHDG deposits are mainly distributed in the Yunnan-Guizhou-Guangxi area (A) at the southwestern margin of the Yangtze Craton and the west Qinling-Dabie Orogenic Belt and (B) an area on the northern margin of the Yangtze Craton (modified from Liu et al., 1994).

Fig. 2. Sketch map of the Youjiang-Nanpanjiang Basin (after Xiao et al., 1998). 1. Craton; 2. Anticline; 3. Basin; 4. Fault; 5. Gold deposit; 6. City.

and its potential application to determine the source of gold. deposits. Most of the gold is interpreted to have come from sedimentary sources (Sillitoe and Bonham, 1990; Arehart et al., 1993; Philips and Powell, 1993; Hofstra, 1995; Kuehn and Rose, 1995; Ilchik and Barton, 1997; Tosdal et al., 1998; Hofstra and Cline, 2000). In China, contradicting models have been proposed for SRHDG deposits. Many researchers consider that the gold was derived mainly from gold-rich strata during deep circulation of meteoric water and/or formation water. According to this model the water was driven either by elevated geothermal gradients during regional extension and/or accompanying magmatism (Liu and Geng, 1985; Wang et al ., 1992; He et al ., 1993; He, 1996; Hu et al., 2002). Others propose a more direct magmatic link for some deposits and a mixed source of gold involving host rocks, magmatic rocks, and even rocks at deep or intermediate crust (Zhang et al., 1996; Zhu et al., 1998; Liu et al ., 1999). Several research papers, mainly in Chinese, have been published on the deposits of southwest Guizhou Province, but they are rather general and have contradicting conclusions. Most have been carried out on the geology and tectonic control, disseminated nature of the gold, geochemical characteristics such as S, Pb, and Sr isotopic compositions, gold exploration and organic geochemical aspects of the deposits. However, much less work has been done on the trace element geochemistry of the SRHDG deposits. The objective of this research is to investigate trace elements behavior during SRHDG mineralization
364 Z. Bao et al .

GEOLOGY The SRHDG deposits in southwest Guizhou Province, China, are located near the buried southwestern edge of the Yangtze Craton and form part of the Dian (Yunnan)Qian (Guizhou)-Gui (Guangxi) ore province (often referred to as the gold triangle in the Chinese literature) which is largely underlain by the Youjiang-Nanpanjiang Basin (Figs. 1 and 2). Within the Dian-Qian-Gui ore province, southwest Guizhou hosts the majority of the SRHDG mineral resources. During the Early Paleozoic Period, the YoujiangNanpanjiang basin was part of the fringe of the Yangtze Craton and in the late Paleozoic Period, formed the relatively active South China Platform (Silurian to Devonian) and evolved into foreland basins in early to middle Triassic Period. The associated sediments comprise a series of deep-sea sequences consisting of siliceous pelitic sequence and turbidites containing volcanic detritus (Zeng et al., 1992). Volcanic rocks are present in the sequence from the Early Devonian to the mid-Triassic but are absent between the middle to the late Carboniferous. The older volcanic rocks (D1-P 1) are found only along the southern margin of the basin in north Guangxi and northeast Yunnan Province. These comprise calc-alkaline and alkaline basalts. Later volcanic activity (P2-T2) was more extensive, and is represented by late Permian Emeishan basalts on the northwest margin and early to middle

Table 1. Three types of SRHDG deposits in southwest Guizhou Province, China

Triassic basic to intermediate-acidic volcanic rocks on the southeast margin of the basin (Liu et al., 1993). Voluminous basalts, covering an area of 5.1 105 km2, erupted in the Permian to Early Triassic Period over a brief time interval (251~253 Ma) forming the large Emeishan igneous province in the western part of the Yangtze Craton (Xu, 2002). The Emeishan basalts consist of mainly alkaline continental flood basalt and minor basaltic pyroclastic rocks. The thickness of the basalts decreases eastwards from more than 5000 m in Binchuan, Yunnan Province to few hundred or even a few tens of meters in Guizhou Province. In the western part of the study area, the basalt layers are more than 100 m thick and commonly include intercalated submarine siliceous exhalite (Wu and Li, 1993). The SRHDG deposits all formed at low temperature (fluid inclusions mainly yield T h in the range of 150 250 C) from solutions of low salinity, weak to neutral acidity, and low fO 2 (e.g., Hu et al., 2002). The gold deposits in southwest Guizhou Province can be classified roughly into three types on the basis of sedimentary environment, lithology of the host rocks, mineral assemblages, and ore metal association (Tan, 1994). The three types are characterized by the associations of Au-Sb, AuHg-(Tl), and Au-As-(Sb), and are represented respectively by the Getang, Zimudang, and Lannigou gold deposits which are hosted by rocks of Late Permian, Permian-Early Triassic, and Middle Triassic, respectively (Table 1). The different geological and geochemical characteristics of the three types of deposits may result partly from differences in the chemical compositions of their host rocks (possible source beds), which formed in different geological settings. The Lannigou gold deposit is adjacent to a former realgar deposit located in Zhenfeng County. Tectonically,

Fig. 3. Fault control of gold mineralization along the Huangchanggou section of the Lannigou deposit (after Luo, 1994).

the Lannigou gold deposit occurs in the northern part of the Youjiang-Nanpanjiang basin (or rift) on the southwest fringe of the Yangtze Craton. The sedimentary environment evolved from that of a passive continental margin to a foreland basin during the Early to Middle Triassic Period. Magmatic activity was insignificant, a few small sub-alkali ultramafic Yanshannian intrusives are exposed in Beiceng, Zhenfeng County, about 27 km NW of the gold deposit. The mining area is located SW of the Banchang thrust fault and the eastern limb of the Laizishan anticline. The Lannigou deposit is hosted by Upper Triassic and Early to Middle Triassic rocks, especially the Bianyang and Xuman Formations of Middle Triassic

HFSE, REE and PGE geochemistry of gold deposits in SW China 365

Fig. 4. Cross section along the 3-3 prospecting line, Erlongkou ore block, Getang gold deposit (after Wang, Y-G. et al., 1994). 1. Maokou Formation limestone; 2. Silicified brecciated limestone (mineralized); 3. Ore body; 4. The Longtan Formation carbonaceous shale; 5. Claystone.

Ladinian stage and Anisian stage (Fig. 3). The lenticular gold ore bodies are mainly hosted by WNW trending faults that discordant to bedding and partly by the NE trending subsidiary faults, preferentially in fine-grained sandstone and siltstone with mudstone intercalations. The Getang gold deposit is located in Anlong and Xingyi County, 27 km northwest of Anlong town on the eastern limb of the Getang dome, a doubly plunging anticline about 50 km long, which trends northwestsoutheast. Gold mineralization is mainly confined to inter-stratal faults and related fracture zones, and occurs within silicified clay breccia and limestone breccia developed along an unconformity between the Upper Permian Longtan Formation (Luoping stage) and the Lower Permian Maokou Formation (Maokou stage) (Fig. 4). Gold is concentrated along a 3 to 15 m-thick breccia horizon at the base of the Longtan Formation (P2l1). The breccia lies over the erosional unconformity between the Longtan and Maokou Formation, which was controlled by the topography of palaeo-base level erosion. Regionally, this horizon is characterized by karst development, but in the vicinity of the Getang deposit, it may be superimposed by a bedding-plane thrust fault (Cunningham et al., 1988). The nature and thickness of the breccia layer (0.5~43 m) vary considerably as a result of variations in the hydrodynamic and palaeoerosion surface conditions. The irregular breccias consist mainly of the Lower Permian Maokou Formation limestone and Upper Permian Longtan Formation clay rock and carbonaceous mudstone. Lenticular ore-bodies crop out discontinuously along the breccia horizon over a distance of about 1000 m and consist of silicified, argillaceous, and carbonate breccia fragments in a chalcedonic quartz matrix. The Zimudang gold deposit is localized by a fault zone in the western part of the NW trending Huijiabao anticline, which is a symmetrical fold about 20 km long and 7 km wide. This fault zone, which is sub-parallel to the fold-axis of the anticline, is characterized by breccia and
366 Z. Bao et al .

cataclasite. It controls the distribution and occurrences of ores; more than 10 gold showings with accompanying Hg, As, Tl and Sb mineralization have been discovered in the anticline. The ore bodies are hosted by fracture zones, which cut through impure carbonate and siltstone of the Early Triassic (Indian stage) Yelang to Upper Permian (Luoping stage) Longtan Formations. The ore bodies are mainly confined to faults F1 and F2 (Fig. 5), and to bedding parallel faults (or bed slips) that are surrounded by meters wide alteration zones. Fault F1, about 2000 m long, hosts the major part of the known gold reserve. Within this structure, ore bodies occur as plates or lenses in the center of the altered zones of fractures, and there are no clear boundaries between the ores and the host rocks. The F2 fault, which hosts some irregular lenticular ore bodies, is a steep normal fault oriented subparallel to the axis of the anticline. Some bedding parallel faults also host small lenticular gold ore bodies. The three deposits have many geological and geochemical characteristics in common. The distribution of gold mineralization in the three deposits is controlled by faults cutting the anticlines. Orebodies are larger in the horizons with high permeability, high Au background (up to ~10 ppb) and carbonaceous matter. The fact that SRHDG host rocks, especially the turbidites, generally have high background gold concentrations was quoted as evidence for the host rocks being the source beds for the mineralization (He et al., 1993; Zhang et al., 1992). The gold mineralization occurring in the Upper Permian sequences is limited to carbonaceous mudstone, shale and breccia, located in the lower part of the Longtan Formation, e.g., the Getang gold deposit. The Zimudang gold deposit mainly occurs in argillaceous, silty limestone and siltstone in the upper part of the Lower Triassic Yelang Formation. The average gold contents of the Maokou, Longtan and Changxing Formations are 1.3, 4.2, and 1.1 ppb, respectively. The average gold content of the Yelang Formation is 4.4 ppb. The Longtan and Yelang Formations, which host some of the gold ores, are more enriched

360

F1
T1y
1b

T1y

1b

T1y1a

T1y1a

T1y1c

P2l

P2l3
3 2l

T1y1b

T1y1a
P
P2l2

P2l2

P2l1
0 200m

F2

P2l1

P2l2

P2l3

T1y1a

T1y

1b

T1y

1c

Fig. 5. 32-32 exploration section of the Zimudang gold deposit (after Wang, Y-G. et al., 1994). 1. Lower Longtan Formation; 2. Middle Longtan Formation; 3. Upper Longtan Formation; 4. Lower part of Lower Yelang Formation; 5. Middle part of Lower Yelang Formation; 6. Upper part of Lower Yelang Formation; 7. Lime stone; 8. Siltstone; 9. Ore body. Ore bodies are hosted in the Permian Longtan and Triassic Yelang Formations, especially the permeable carbonaceous layers with high background Au in the siltstone and argillic limestone in the middle Longtan and upper part of Lower Yelang Formation.

than other strata. For example, in the Xuman area, the average gold content of the middle Triassic Series is as high as 13.4 ppb (Zhang et al ., 1992). For the three studied deposits, silicification and pyritization are the principal significant alteration types associated with gold mineralization. Gold mineralization is associated with elevated concentrations of As, Hg, Tl. In the ores, Au occurs mainly in hydrothermal pyrite as micron to submicron-sized grains. In some cases, visible micron size gold was observed in oxidized ores (He et al., 1993; Sun et al., 1993). However, oxidized ores only represent a small portion of the known Au reserves in the three deposits. The mean grades and known tonnages of the Lannigou, Zimudang and Getang deposits are 7.0, 6.0, and 6.2 g/t and 80, 60, and 22 t respectively (Hu et al., 2002). RARE EARTH ELEMENTS AND HIGH FIELD STRENGTH E LEMENTS REE have historically been generally regarded as being insensitive to hydrothermal alteration. However, literature supporting REE mobility during hydrothermal processes has been rapidly expanding in the last two decades (Nesbitt, 1979; Alderton et al ., 1980; Humphris, 1984; Marsh, 1991; Williams-Jones and Wood, 1992; Gouveia et al ., 1993; Mongelli, 1993; Prudencio et al., 1993; Van der Weijden and Van der Weijden, 1995). Recently, the potential of REE to provide information about

ore forming processes through REE mobility or immobility during ore formation processes has been widely recognized, and detailed REE investigations have been performed on several metallic ore deposits (McLennan and Taylor, 1979; Taylor and Fryer, 1980, 1982; Campbell et al ., 1984; Giere, 1986; Whitford et al ., 1988; Lottermoser, 1992; Parr, 1992; Wood and Williams-Jones, 1994; Bierlein, 1995; Bierlein et al ., 1999). The mobility of trace elements in altered host-rocks relative to their unaltered counterparts, and their distribution in host rocks and ores can place constraints on physico-chemical characteristics of ore-bearing fluids, and contribute to understanding of ore-forming processes (Lottermoser, 1992). Wall rock alteration in SRHDG deposits, especially sulfidation, silicification, and decarbonatization, have been investigated extensively. Studies have examined stable isotope composition variation and the degree of pyritization of host rock iron around deposits, documented geochemical signatures, and have proposed potential applications to geochemical exploration (Hofstra et al ., 1991; Stenger et al ., 1998a, b; Woitsekhowskaya and Peters, 1998; Hofstra and Cline, 2000). However, studies focusing on high field strength and rare earth elements in the ores and the altered host rocks are rare. Samples have therefore been collected from underground and surface exposures of the Lannigou, Getang and Zimudang deposit, in order to assess the behavior of REE during hydrothermal alteration associated with SRHDG mineralization. Samples included siltstone, argillaceous limestone,

HFSE, REE and PGE geochemistry of gold deposits in SW China 367

368 Z. Bao et al .

Table 2. Trace element compositions of the ores and host rocks from the SRHDG deposits in SW Guizhou Province (ppm)

25 20

80

R2=0.91

60

R2=0.93

Nb

Nb
0 100 200 300 400

15 10 5 0

40 20 0 0 100 200 300 400 500 600

Zr
2 1.5 5

Zr

R2=0.90
4 3

R2=0.95

Ta

1 0.5 0 0 25 20 100 200 300 400

Ta
2 1 0 0 80 100 200 300 400 500 600

Zr

Zr
70 60 50 40 30 20 10 0

R2=0.90

R2=0.93

Nb

10 5 0 0 100 200 300 400

Nb

15

100

200

300

400

500

600

Zr
5 4 6

Zr

R2=0.94

5 4

R2=0.77

Yb

Yb

3 2 1
0 0 100 200 300 400

3 2 1 0 0 100 200 300 400 500 600

Zr
80 60 100

Zr

R =0.87

80

R2=0.84

La

40 20 0
0 100 200 300 400

La

60 40 20 0
0 100 200 300 400 500 600

Zr

Zr

Ores

Sedimentary rocks

Fig. 6. Binary trace element plots for samples of the SRHDG deposits of southwest Guizhou. A, B, C, D, Esamples from the Lannigou deposit; F, G, H, I, Jsamples from the three investigated SRHDG deposits of southwest Guizhou.

breccia, and ore from different stratigraphic layers sharing similar source; brief descriptions of these samples are given in Table 2. Analytical procedure Trace element concentrations of the samples were analyzed using ICP-MS method. The analytical procedure for the ICP-MS analysis is fully described by Qi et al. (2000). Briefly, one hundred milligrams of each powdered sample was dissolved by 1 ml of HF (38%) and 0.5 ml of

HNO3 (68%) in a separate top screwed PTFE bomb, and 1 ml of 1 g/ml Rh solution was added as an internal standard. The final residue was retrieved by 8 ml of 40% HNO3, diluted up to 100 ml in volume. Trace element concentrations of the samples were analyzed using the Finnigan MAT ICP-MS at the State Key Laboratory of Ore Geochemistry, Guiyang Institute of Geochemistry, Chinese Academy of Sciences. The detection limits for Sc, Cr, Co, Ni, Cu, Zn, Sr, Zr, Mo, Cs, Ba, Pb range from 0.010.2 g/l; and for Y, Nb, Hf, Ta, W, Th, U and REE

HFSE, REE and PGE geochemistry of gold deposits in SW China 369

370 Z. Bao et al .

Table 3.1. Trace element correlations for the ores and host rocks from the Lannigou deposit

Table 3.2. Trace element correlations for the ores and host rocks of three SRHDG deposits in Southwest Guizhou Province

HFSE, REE and PGE geochemistry of gold deposits in SW China 371

[A]
2 1
LN-4 LN-8

Sandstone
Bioclastic limestone

Sample/NASC

LN-12 Siltstone LN-13 Oxidized ore LN-14 Semi-oxidized ore LN-15 Primary ore LN-45 Reef limestone LN-46 Primary ore

0.1

0.01 10

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Sample/NASC

[B]

ZMD-1 Oxidized ore ZMD-3 Silicified argillaceous limestone

1 La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

ZMD-4 Carbonaceous matter

10

Sample/NASC

[C]
1

GT-2 Primary ore GT-4 Oxidized ore GT-5 Silicified carbonaceous shale GT-6

Coal bed

0.1

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Fig. 7. NASC normalized REE patterns of ores and host rocks from the SRHDG deposits in southwest Guizhou. [A] The Lannigou deposit; [B] The Zimudang deposit; [C] The Getang deposit (NASCNorth American shale composite composition by Gromet et al., 1984).

are from 0.0010.005 g/l. The international standard, GSR-1, was used to calibrate element concentrations. The analytical precision is about 5% for all the trace elements listed in Table 2. Major elements, such as Al, Ca, Mg, K, and Na, were not detected due to the low precision and poor reproducibility of this method. Results and discussion Trace element concentrations in sedimentary rocks are generally rather heterogeneous and difficult to compare directly. However, owing to the efficient mixing of source lithologies during sedimentary processes, sedimentary rocks sharing the same sources are commonly homogeneous enough to share similar trace element patterns (McLennan, 1989). This makes it possible to investigate trace element variations due to alteration. A common practice in evaluating the compositional changes which accompanying hydrothermal alteration is to compare the concentrations of elements of interest to
372 Z. Bao et al .

that of an immobile element. If a trace element is immobile during alteration, the addition of material (as in silicification) dilutes its concentration, whereas extraction of material (as by dissolution) concentrates it. Consequently, a pair of immobile elements will plot on a common regression line that passes through the origin, such that elements retain a constant inter-element ratio. If the sedimentary rocks in a particular lithological horizon were deposited from the same source material, the trace components might be homogenous enough for geochemical comparison (e.g., MacLean and Kranidiotis, 1987; Hofstra and Cline, 2000). In the case of chemically incompatible elements, the correlation coefficients for regression lines that pass through the bulk composition and origin can be used to select the most immobile element pair (MacLean and Kranidiotis, 1987). High field strength elements like Nb, Ta, Zr, and Hf, which are widely considered to be immobile, correlate highly with each other in the Lannigou

Fig. 8. Chondrite normalized REE patterns of the ore and host rocks of the SRHDG deposits from southwest Guizhou. [A] The Lannigou deposit; [B] The Zimudang deposit; [C] The Getang deposit. Values for chondrite normalization from Boynton (1984).

deposit, and in the three deposits as a whole. The correlations are consistent with the detrital components of the sediments having similar source areas. Sedimentary environment and alteration can shift the immobile element concentrations but have little affect on interelement ratios, which are controlled largely by the source of detritus in the rocks. Within the analytical errors, the chemically incompatible elements Nb, Ta against Zr plot on regression lines through the origin (Fig. 6). This means that these elements are geochemically coherent and immobile during the ore-forming alterations. The rare earth elements are less well correlated with the above high field strength elements and the regression lines shift from the origin (Fig. 6), which suggests that the rare earth elements were somewhat mobile during alteration. The interelement correlation among the rare earth elements is significant (generally with coefficients larger than 0.8), which is consistent with their coherent geochemical behavior. However, the interelement correlation among LREEs and HREEs treated separately are considerably stronger

(>0.97), which indicates fractionation between the two subgroups. W, Th, U, Mo, Sr, Ba and Zn are mobile relative to Zr during hydrothermal alteration as shown by their poor correlations with the latter (Table 3). The NASC (North American Shale Composite)- and chondrite-normalized REE patterns of the ores and host rocks are shown in Figs. 7 and 8, respectively. Samples from the deposits show similar REE patterns, except for the limestone. The limestone samples from the Lannigou deposit, which are quite pure and extremely low in REE, show negative Ce anomalies (0.330.44) (Fig. 8). Moreover, although the REE patterns of the ores (i.e., the most altered host rocks) and the unaltered sedimentary host rocks are quite comparable, minor but consistent differences are still discernable. The (La/Yb)N ratios of the ores (7.58.9) are apparently lower than those of the unaltered host rocks (10.514.4), indicating a fractionation of REEs during the hydrothermal alteration. Relative to NASC, the samples from the Zimudang deposit are enriched in LREE and Eu, which is probably a result of the volcanic

HFSE, REE and PGE geochemistry of gold deposits in SW China 373

Silicified and pyritized siltstone (primary ore, CA)

450

(A)
400 350 300 250 200 150
Ce

Sr

Y = 1.4926X

R = 0.99
Zr Ba

100 50 Sb 0 0 50
Cu (B)

Cr Zn

100

150

200 250 300 o Unaltered siltstone (C )

350

400

A Silicified and pyritized siltstone (primary ore, C )

Fig. 10. Chondrite normalized REE composition of fluid in fluid inclusions from the Lannigou deposit (data from Su et al., 1998a). Values for chondrite normalization from Boynton (1984).

70 60 50 40
Sb

(B)
Yb*10 Y

La

Cu

Nd

Tm*50 Co Eu*10 Ho*10 Th Nb Sc Pb

30 20 10 0 0
W UGd

Tb*10

Ta

Pr Sm Ce/10 D Hf ErLu*10

10

20

30

40

50

Unaltered siltstone (Co)

Fig. 9. An isocon diagram showing trace elements mobility for samples from the Lannigou deposit ((B) is an enlargement of the small area in (A) adjacent to origin).

component in the rocks. The oxidized ores are relatively depleted in HREE because of preferential leaching by meteoric water (Wood, 1990). An isocon approach (Grant, 1986; Hofstra and Cline, 2000) on the two samples from the Lannigou deposit, LN12 and LN-15, representing unaltered siltstone and altered siltstone (the primary ore) from the Xiuman Formation turbidite, shows that the REE and Y are mobile during alteration (Fig. 9). The slope of the best-fit isocon (M o/ M A) was 1.49, which is equivalent to MA/Mo = 0.67 and corresponds to a mass decrease of 33%; M A and Mo represent the mass of an element in altered and original rock, respectively. The isocon diagram (Fig. 9) shows that the LREE are depleted, whereas the HREE were enriched during alteration. For example, the LREE La (9.5%), Ce (8.6%), Pr (5.9%) show mass losses; and the HREE Gd (19%), Tb (17%), Dy (14%), Ho (9.2%), Er (8.5%) show mass gains. However, the variations of the other REEs are rather minor such as Nd (2.3%), Tm (3.4%), Yb (7.4%), and Lu (2.2%), and within the analytical errors.
374 Z. Bao et al .

The physico-chemical characteristics of hydrothermal fluids control the mobility of REE during the alteration. Fluid inclusion data, including temperature, pressure, salinity, density, composition, etc., can be used to understand REE mobility, even though the fluid inclusions in SRHDG deposits are rather small and many are difficult to relate directly to gold mineralization. The main characteristics of the primary inclusions are summarized in Table 4 (Yang and Dong, 1994; Zhang and Zhang, 1999; Zhu and He, 1996; He et al., 1993; Liu et al., 1999; Hu et al., 2002). The fluid inclusion investigation suggests that the hydrothermal solution was a low salinity, near neutral and relatively reducing meteoric water. The fluid salinities vary in the range of 4%~14% NaCl equivalent. The major components of the fluid inclusions in the Lannigou, Getang and Zimudang deposits determined through decrepitating and gas/liquid chromatography show that the ions are dominated by (Ca2+), Na+, K+, while the anions are mainly SO42, Cl and F. The calculated pH and Eh vary in the range of 5.56~7.33, and 0.41~ 0.54 respectively (Zhang and Zhang, 1999). Ideally, the REE composition of hydrothermal fluid can be determined from fluid inclusion analysis. However, since fluid inclusions are usually very small in size, and the REE concentrations are rather low, analysis is difficult, and consequently such data are rarely reported (Ghazi et al., 1993; Banks et al., 1994). Su et al. (1998a) reported three data for the SRHDG deposits of Guizhou Province using the decrepitating-leaching method. However, their estimation of absolute concentrations is unreliable, because of the heterogeneous distribution of the inclusions in quartz, and there are multiple generations of inclusions with different compositions. This simple fact alone makes it almost impossible to obtain truly duplicate samples for a test of analytical accuracy. Thus, it is more reasonable to assume that the patterns, or rather their inter-elemental ratios, of the inclusions are more reliable

Table 4. Characteristics of fluid inclusions in ores from the three deposits

Types of the inclusions in the ores Size Homogenization temperatures Salinity Ions Anions 40 Ar/3 6 Ar Liquid inclusion, CO2 bearing inclusion, gas inclusion and organic bearing inclusion 515 m 80C~320C, mainly 200 50C 4% 14% NaCl equivalent; decreases from early to late stage of mineralization Na +, K+, Ca 2 +, Mg2 + Cl , SO4 2 , F 298328 (atmospheric argon 3 6 Ar > 90%)

than their contents; the absolute values are probably meaningless. The weak negative Ce anomalies and positive or mild negative Eu anomalies (Fig. 10) show that the oreforming environment was rather reducing (Bau, 1991). Relative to the ores and the host rocks, the fluids are LREE enriched; the entrapped fluids reflect the compositions of the relict solutions after rock-fluid interaction. The LREE enrichment of the fluids is consistent with the relative LREE depletion of the altered rocks and the ores. The mobility (gains and losses) of REE during alteration is controlled largely by: (1) the REE concentrations of the reacting minerals; (2) the relative stability of these minerals towards the fluid; (3) the availability of sites within the secondary minerals to accommodate REE released during alteration; (4) the REE concentrations of the hydrothermal fluid; and (5) the ability of the fluid phase to remove REE from the system. Accordingly, the final chemical composition of the rocks is a function of the precursor rock composition, alteration assemblage, physico-chemical properties of the ore-forming solution, and fluid/rock ratios. The mineral assemblage of the related alteration includes clay minerals (illite, kaolinite, and minor smectite), quartz, calcite and sulfides. As the detritus in the sedimentary rocks are from roughly the same sources, the physico-chemical properties and chemical evolution of the fluids will be the most important factors controlling the differentiation of the REE. As shown by the fluid inclusion results the hydrothermal solution was mild acidic to neutral and Cl dominated. The possible REE complexes in the fluids include REE-Cl, REE-F, REE-CO 3, and REE-SO 4 complexes. As temperatures below ~200 C, the association constants of REE-Cl complexes are relatively constant across the group, however, the REE-F and REE-CO3 complexes show steady increase in stability with increasing atomic number (decreasing ionic radius) (Wood, 1990). Thus, F and CO 3 2 complexing could result in the fractionation of the REE in aqueous solutions, if all other factors are held constant (Wood, 1990; Haas et al., 1995; Smith et al., 2000). During argillization and carbonatization the F and CO32 concentrations of the fluid will decrease and result in preferential HREE precipitation. As no correlation between Au enrichment and REE variations was observed, which

probably due to that gold is transported mainly by complex agents different from those for REE. So REE mobility alone could resolve the source of gold of SRHDG deposits. PLATINUM GROUP ELEMENTS The platinum group elements can be divided on the basis of their associations into two sub-groupsthe Irgroup (IPGEOs, Ir, and Ru) and the Pd-group (PPGE Rh, Pt and Pd) (Rollison, 1993). Gold is associated with the Pd-group. The PGE were considered to be refractory inert elements and were often used as fingerprints in investigating magmatic system evolution (Campbell et al., 1983). They were also considered to be immobile under geologically reasonable hydrothermal conditions and to be transported and concentrated primarily by magmatic processes (Barnes et al ., 1985; Cawthorn and Barry, 1992). However, in the last two decades, more and more studies have shown that PGE are mobile during alteration, even at low temperatures in supergene environments (Barnes et al ., 1985; Crocket and Kabir, 1987; Bowles et al ., 1994; Hannigan and Peucker-Ehrenbrink, 1998; Hattori and Hulbert, 1998; Helmy and Mogessie, 2001). An important feature of the PGE in the context of SRHDG deposits is that they can be transported together with Au in aqueous solutions as bisulfides, chlorides, and organic complexes (McKibben and Williams, 1990; Pan and Wood, 1994; Gammons et al., 1992; Wood, 1996). PGE and Au have been observed associated with each other in many geological circumstances. Platinum group element minerals, such as platinum-gold, occur as inclusions in pyrite and arsenopyrite, and have been observed in SRHDG deposits in southwestern Guizhou Province (Wang, K-R. et al., 1994) and elsewhere (Mazdab et al., 1999). In some SRHDG deposits the PGE concentrations are high enough to contribute significantly to the economic value of the deposit*. Since basalt and alkali ultramafic rocks near areas of gold mineralization in
*Hou et al. (1991). Metallogeny and expectation of the fine grain disseminated gold deposits in southeastern Yunnan Province. Tianjin geological academy of the ministry of metallurgic industry, Tianjin. Unpublished report: I-88-92-01-4-1.

HFSE, REE and PGE geochemistry of gold deposits in SW China 375

Table 5. PGE and Au concentrations of the ores, host rocks, basalt and alkali ultramafic rocks in southwest Guizhou (ppb)

*Su (personal communication, 2000); **Average value of the 14 Ermeishan basalt samples (Zhang and Li, 1998).

southwest Guizhou have been considered to be the main source of Au (Liu et al., 1999), the PGE contents of these rocks and sedimentary rocks were analyzed to evaluate their possible contribution to gold mineralization in the deposits studied. Analytical procedure The PGE concentrations of the samples were analyzed in the ICP-MS laboratory of the Institute of Geochemistry, Chinese Academy of Sciences using a Finnigan MAT ICPMS following the procedure described by Qi and Hu (1999). Samples, 5 g each, were decomposed with Na2O2 in corundum crucibles. After dissolving the fused disc with aqua regia and evaporating the solution to dryness to remove SiO 2, the PGE were concentrated by the Te coprecipitation method. The concentrations of Ru, Pd, Ir and Pt were determined by isotope dilution-inductively coupled plasma mass spectrometry (ID-ICP-MS). The Rh and Au contents were determined by internal standardICP-MS. The detection limits for these elements are less than 0.3 ng/g. The following standards, WGB-1, UMT-1, and WPR-1, were analyzed to control and the quality of the analyses. The precision of the PGE analyses was between 1.5% and 6.5% RSD. Results and discussion The PGE and gold concentrations of the ores and rocks are listed in Table 5. As shown by the chondrite normalized PGE patterns, the PGE contents of the ores are generally higher than those of the unaltered host rocks (Fig. 11). The variation of the PGE concentrations of the ores, altered host rocks, and unaltered rocks may be due partly
376 Z. Bao et al .

to mass loss or dilution during alteration. However, the PGE fractionation of altered rocks and ores relative to unaltered host rocks is obvious. For instance, the primary ore (LN-15) from the Lannigou deposit, which underwent pervasive silicification and sulfidation, is enriched in Pt and Pd relative to its unaltered counterpart (LN-12) (Table 5). The relative enrichment of Pt and Pd in altered rocks and ores suggests that the hydrothermal solution was enriched in Pt and Pd. This enrichment is consistent with their higher solubilities and mobility of these metals in hydrothermal systems (Barnes et al ., 1985; Mountain and Wood, 1988; Gammons et al., 1992). It is interesting to note that the carbonaceous beds of the Zimudang deposit are significantly enriched in Pd. The Pd/Ir ratio of the carbonaceous bed is 11.2, which is higher than that of the primary ore (Pd/Ir = 6.54). The PGE distribution pattern of the carbonaceous matter may therefore indicate that the hydrothermal solution was enriched in Pd. As shown in Fig. 11, the basaltic and alkali ultramafic rocks outside the mining area are enriched in Pt and Pd. Their PGE profiles are different from those of samples from the three deposits. The Pd/Ir ratio of the alkali ultramafic rocks is relatively high (40.1~41.3) and the average Pd/Ir ratio of the Emeishan basaltic rocks is as high as 78.5. Thus, if these rocks were the main source of PGE for the SRHDG deposits, the Pd/Ir ratios of the ores and the pyrite would be much higher. It is clear that the properties of the different PGE in the ores are similar to those of the host rocks, rather than those of the basaltic and ultramafic rocks. Therefore, the hypothesis that the ore-forming material was mainly derived from the

0.01

0.01

Sample/Chondrite

0.001
LN-4 Sandstone LN-8 Limestone LN-12 Siltstone LN-13 Oxidized ore LN-14 Semioxidized ore LN-15 Primary ore

Sample/chondrite

0.001 GT-1 GT-2 GT-4 GT-6 Ir Ru Rh Pt

Limestone Primary ore Oxidized ore Coal bed

0.0001 0.1

0.0001

Pd

Ir

Ru

Rh

Pt

Pd

D
0.1
ZMD-1 Oxidized ZMD-3 Primary Ore ZMD-4 Carbonaceous matter ZMD-8 Limestone

0.01

0.01

Sample/Chondrite

0.001 0.0001
CJ-05 CJ-01

Sample/chondrite

0.001

Ultramafic rock

0.0001

Ermeishan(14 average) (shadow area showing the range of Emeishan basalts) Ir Ru Rh Pt Pd

0.00001

Ir

Ru

Rh

Pt

Pd

Fig. 11. Chondrite normalized PGE profiles for the host rocks, ores, and pyrite from the SRHDG deposits in southwestern Guizhou Province. The values for chondrite normalization are from Naldrett and Duke (1980). A: The host rocks and the ores from the Lannigou gold deposit; B: The ores and host rocks from the Zimudang gold deposit; C: The ores and host rocks from the Getang gold deposit; D: The Emeishan basalt and the alkali ultramafic rocks.

ultramafic rocks (Liu et al ., 1999) is not supported by the distribution of the PGE. Moreover, the ultramafic rocks occur as small dikes that are far from gold mineralization. The conclusion that the source of the metals is not igneous is also supported by sulfur isotopic data on the SRHDG deposits, which show that the H2S in ore fluids was largely derived from the sedimentary host rocks (Hu et al., 2002). On the contrary, the PGE features of the ores are similar to those of the sedimentary host rocks (Fig. 11). Thus, given that gold was transported mainly by reduced sulfur complexes, then gold too must have been derived from the host rocks. CONCLUSION Trace element geochemical investigation of the sediment-hosted disseminated gold deposits, represented by the Lannigou, Getang and Zimudang deposits, shows

that: (1) High field strength elements, such as Nb, Ta, Zr, and Hf, were immobile during hydrothermal alteration and gold mineralization. (2) REEs were mobile during the alteration. LREE were depleted, whereas HREE were enriched. (3) PGE investigation of the ores, host rocks, basalt, and the ultramafic dikes show that the altered host rocks and ores are relatively enriched in Pd and Pt, and that the PGE profiles of the ores are similar to those of the host rock rather than those of the ultramafic rocks and the Emeishan basalts. Thus, the late Yanshanian ultramafic dikes and the basalts were not a major source of PGE or gold. (4) REE and PGE signatures of the three deposits show that the ore-forming fluid was relatively reduced and suggest that the gold was probably derived from the sedimentary host rocks, which is consistent with the conclu-

HFSE, REE and PGE geochemistry of gold deposits in SW China 377

sion based on sulfur isotopic data. In addition, the lack of igneous intrusives in the vicinity of the deposits, the weak metamorphic grade of the rocks, the lack of metamorphic effects within the time range of the SRHDG deposits, and coincidence in time of gold mineralization with regional extension support the hypothesis that the SRHDG deposits in southwest Guizhou are most likely of amagmatic origin.
Acknowledgments This study was partly supported by an AUCC-CIDA SULCP program supervised by professor Jayanta Guha (UQAC), the National Climbing Program of China (Grant No. 95-Yu-25), the innovation program of Chinese Academy of Sciences (CAS) (KZCXZ-SW-117-05), and the Guangzhou Institute of Geochemistry, CAS (GIGCX-03-04). We would like also to thank Mr. Xunfeng Liu, the engineer-in-chief of Geological Bureau of Guizhou Province, for his helpful suggestions and information, and the Zimudang gold mine, Getang gold mine, and the Gold Industry Bureau of Zhenfeng County, for access to the mining area and additional assistance during the field-based part of this study.

R EFERENCE
Alderton, D. H. M., Pearce, J. A. and Potts, P. J. (1980) Rare earth element mobility during granite alteration: Evidence from southwest England. Earth and Planetary Science Letters 49, 149165. Arehart, G. B., Foland, K. A., Naeser, C. W. and Kestler, S. E. (1993) 40Ar/39Ar, K/Ar, and fission track geochronology of sediment-hosted disseminated gold deposits at Post/Betze, Carlin Trend, northeastern Nevada. Economic Geology 88, 622646. Banks, D. A., Yardley, B. W. D., Campbell, A. R. and Jarvis, K. E. (1994) REE composition of an aqueous magmatic fluid: A fluid inclusion study from the Capitan Pluton, New Mexico, USA. Chemical Geology 113 (34), 259272. Barnes, S. J., Naldrett, A. J. and Gorton, M. P. (1985) The origin of the fractionation of platinum-group elements in terrestrial magmas. Chemical Geology 53, 303323. Bau, M. (1991) Rare-earth element mobility during hydrothermal and metamorphic fluid-rock interaction and the significance of the oxidation state of europium. Chemical Geology 93, 219230. Bierlein, F. P. (1995) Rare-earth element geochemistry of clastic and chemical metasedimentary rocks associated with hydrothermal sulphide mineralisation in the Olary Block, South Australia. Chemical Geology 122, 7798. Bierlein, F. P., Waldron, H. M. and Arne, D. C. (1999) Behaviour of rare earth and high strength elements during hydrothermal alteration of meta-turbidities associated with mesothermal gold mineralization in central Victoria, Australia. Journal of Geochemical Exploration 67, 109125. Bowles, J. F. W., Gize, A. P. and Cowden, A. (1994) The mobility of the platinum-group elements in the soils of the Freetown Peninsula, Sierra Leone. The Canadian Mineralogist 32, 957967. Boynton, W. V. (1984) Geochemistry of the rare earth elements:

Meteorite studies. Rare Earth Element Geochemistry (Henderson, P., ed.), 63114, Elsevier. Campbell, I. H., Naldrett, A. J. and Barnes, S. J. (1983) A model for the origin of the platinum-rich sulfide horizons in the Bushveld and Stillwater complexes. Journal of Petrology 24, 133165. Campbell, I. H., Lesher, C. M., Coad, P., Franklin, J. M., Gorton, M. P. and Thurston, P. C. (1984) Rare earth element mobility in alteration pipes below massive Cu-Zn-sulfide deposits. Chemical Geology 45, 181202. Casaceli, R. J. and Gemuts, I. (1985) Precious metal deposits related to continental extension. Global Tectonics and Metallogeny 3 (1), 78. Cawthorn, R. G. and Barry, S. D. (1992) The role of intercumulus residua in the formation of pegmatoid associated with the UG2 chromite, Bushveld Complex. Australia Journal of Earth Sciences 39, 263276. Crocket, J. H. and Kabir, A. (1987) PGE in Hawaiian basalt: Implications of hydrothermal alteration on PGE mobility in volcanic fluids. Geo-Platinum (Prichard, H. M., Potts, P. J., Bowles, J. F. W. and Cribb, S. J., eds.), 87, 256. Cunningham, C. G., Ashely, R. P., Chou, I-M., Huang, H., Wang, C-Y. and Li, W-K. (1988) Newly discovered sedimentaryhosted disseminated gold deposits in the Peoples Republic of China. Economic Geology 83, 14621467. Gammons, C. H., Bloom, M. S. and Yu, Y. (1992) Experimental investigation of the hydrothermal geochemistry of platinum and palladium. I. Solubility of platinum and palladium sulfide minerals in NaCl/H 2 SO 4 solutions at 300 C. Geochimica et Cosmochimica Acta 56, 38813894. Ghazi, A. M., Vanko, D. A., Roedder, E. and Seeley, R. C. (1993) Determination of rare earth elements in fluid inclusions by inductively coupled plasma-mass spectrometry (ICP-MS). Geochimica et Cosmochimica Acta 57 (18), 45134516. Giere, R. (1986) Zirconolite, allanite and hoegbomite in a marble skarn from the Bergell contact aureole: Implications for mobility of Ti, Zr and REE. Contributions to Mineralogy and Petrology 93, 459470. Gouveia, M. A., Prudencio, M. I., Figtueiredo, M. O., Pereira, L. C. J., Waerrnborgh, J. C., Morgado, I., Pena, T. and Lopes, A. (1993) Behavior of REE and other trace and major elements during weathering of granitic rocks, Evora, Portugal. Chemical Geology 107 , 293296. Grant, J. A. (1986) The isocon diagramA simple solution to Gresens equation for Metasomatic alteration. Economic Geology 81, 19761982. Gromet, L. P., Dymek, R. F., Haskin, L. A. and Korotev, R. L. (1984) The North American Shale Composite: Its compilation, major and trace element characteristics. Geochimica et Cosmochimica Acta 48, 24692482. Haas, J. R., Shock, E. L. and Sassani, D. C. (1995) Rare earth elements in hydrothermal systems: Estimates of standard partial molar thermodynamic properties of aqueous complexes of the rare earth elements at high pressures and temperatures. Geochimica et Cosmochimica Acta 59(21), 4329 4350. Hannigan, R. and Peucker-Ehrenbrink, B. (1998) Weathering of black shales as sources of radiogenic Os in seawater. Geological Society of America, 1998 Annual Meeting Ab-

378 Z. Bao et al .

stracts with ProgramsGeological Society of America 30(7), 223. Hattori, K. and Hulbert, L. J. (1998) PGE patterns and osmium isotope compositions of uranium veins in the Nicholson deposit, Athabasca Basin. Geological Society of America, 1998 Annual Meeting Abstracts with ProgramsGeological Society of America 30 (7), 371. He, L-X. (1996) A genetic model for gold deposits in southwestern Guizhou: A co-source of heat, water and orematerials. Guizhou Geology 13(2), 154160 (in Chinese). He, L-X., Zeng, R-L. and Lin, L-Q. (1993) Gold Geology of Guizhou Province. Geology Publishing House, Beijing, 156 pp. (in Chinese). Helmy, H. M. and Mogessie, A. (2001) Grabbro Alkarem, Eastern Desert, Egypt, Cu-Ni-PGE mineralization in a concentrically zoned mafic-ultramafic complex. Mineralium Deposita 36, 5871. Hofstra, A. H. (1995) Timing and duration of Carlin-type gold deposits in Nevada and UtahRelation to back-arc extension and magmatism. Geological Society of America Abstracts with Programs 27(6), A-329. Hofstra, A. H. and Cline, J. S. (2000) Characteristics and models from Carlin-type gold deposits (Chapter 5). SEG Reviews 13, 163220. Hofstra, A. H., Leventhal, J. S., Northrop, H. R., Landis, G. P., Rye, R. O., Birak, D. J. and Dahl, A. R. (1991) Genesis of sediment-hosted disseminated-gold deposits by fluid mixing and sulfidation: Chemical-reaction-path modeling of ore-depositional processes documented in the Jerritt Canyon district, Nevada. Geology 19, 3640. Hu, Rui-Zhong, Su, Wen-Chao, Bi, Xian-Wu, Tu, Guang-Zhi and Hofstra, A. H. (2002) Geology and geochemistry of Carlin-type gold deposits in China. Mineralium Deposita 37, 378392. Humphris, S. E. (1984) The mobility of the rare earth elements in crust. Rare Earth Element Geochemistry (Developments in Geochemistry, 2) (Henderson, P., ed.), 317342 (Chapter 9), Elsevier, Amsterdam. Ilchik, R. P. and Barton, M. D. (1997) An amagmatic origin of Carlin-type gold deposits. Economic Geology 92 (3), 269 288. Kuehn, C. A. and Rose, A. W. (1995) Carlin gold deposits, Nevada: Origin in a deep zone of mixing between normally pressured and over pressured fluids. Economic Geology 90, 1736. Liu, D-S. and Geng, W-H. (1985) On the mineral association and mineralization conditions of the Carlin-type gold deposits in China. Geochimica No. 3, 277282 (in Chinese). Liu, D-S., Tan, Y-J., Wang, J-Y. and Wei, L-M. (1994) Carlintype gold deposits in China. Chinese Carlin-type Gold Deposits (Liu, Dongsheng, Tan, Yunjian, Wang, Jianye and Jiang Shufang, eds.), 1 36, Nanjing University Press, Nanjing (in Chinese). Liu, W-J., Zeng, Y-F., Zhang, J-Q., Chen, H-D. and Zhang, RC. (1993) The geochemical characteristics of volcanic rocks and structural setting in Youjiang basin, Guangxi. Geology of Guangxi 6 (2), 114 (in Chinese with English abstract). Liu, X-F., Su, W-C. and Zhu, L-M. (1999) An approach on mechanism of juvenile fluid mineralization for Carlin-type

gold deposits in Yunnan-Guizhou-Guangxi. Geology and Prospecting 35(1), 1419 (in Chinese). Lottermoser, B. G. (1992) Rare earth elements and hydrothermal ore formation processes. Ore Geology Review 7(1), 25 41. Luo, X-H. (1994) Geological characteristics, genesis and exploration of the Lannigou gold deposit, Zhenfeng, Guizhou Province. Carlin-type Gold Deposits in China (Liu Dongsheng, ed.), 100 115, Nanjing University Press (in Chinese). MacLean, W. H. and Kranidiotis, P. (1987) Immobile elements as monitors of mass transfer in hydrothermal alteration: Phelps Dodge massive sulfide deposit, Matagami, Quebec. Economic Geology 82, 951962. Marsh, J. S. (1991) REE fractionation and Ce anomalies in weathered Karoo dolerite. Chemical Geology 90, 189194. Mazdab, F. K., Barton, M. D. and Hervig, R. L. (1999) Trace element distributions in iron sulfides and associated oxides, with a focus on Fe-oxide-rich hydrothermal systems. Geological Society of America, 1999 Annual Meeting Abstracts with ProgramsGeological Society of America 31 (7), 32. McKibben, M. A. and Williams, A. E. (1990) Solubility and transport of platinum-group elements and Au in saline hydrothermal fluids: Constraints from geothermal brine data. Economic Geology 85, 19261934. McLennan, S. M. (1989) Rare earth elements in sedimentary rocks: influence of provenance and sedimentary processes. Geochemistry and Mineralogy of Rare Earth Elements (Lipin, B. R. and McKay, G. A., eds.), Reviews in Mineralogy 21, 169200. McLennan, S. M. and Taylor, S. R. (1979) Rare earth element mobility with associated with uranium mineralization. Nature (London) 282, 247250. Mongelli, G. (1993) REE and other trace elements in a granitic weathering profile from Serre, Southern Italy. Chemical Geology 103 , 1725. Mountain, B. W. and Wood, S. A. (1988) Chemical controls on the solubility, transport, and deposition of platinum and palladium in hydrothermal solutions: A thermodynamic approach. Economic Geology 83, 492510. Naldrett, A. J. and Duke, J. M. (1980) Platinum metals in magmatic sulfide ores. Science 208 , 14171424. Nesbitt, H. W. (1979) Mobility and fractionation of rare earth elements during weathering of a granodiorite. Nature (London) 279 , 206210. Pan, P. and Wood, S. A. (1994) Solubility of Pt and Pd sulfides and Au metal in aqueous bisulfide solutions. II. Results at 200 C to 350 C and saturated vapor pressure. Mineralium Deposita 29, 373390. Parr, J. M. (1992) Rare-earth element distribution in exhalites associated with Broken Hill-type mineralisation at the Pinnacles deposit, New South Wales, Australia. Chemical Geology 100 , 7391. Philips, G. N. and Powell, R. (1993) Link between gold provinces. Economic Geology 88, 10841098. Prudencio, M. I., Braga, M. S. A. and Gouveia, M. A. (1993) REE mobilization, fractionation and precipitation during weathering of basalts. Chemical Geology 107 , 251254. Qi, L. and Hu, J. (1999) Fast determination of platinum group

HFSE, REE and PGE geochemistry of gold deposits in SW China 379

elements and gold in geological samples by ICP-MS. Rock and Mineral Analysis 18 (4), 267279 (in Chinese). Qi, L., Hu, J. and Gregoire, D. C. (2000) Determination of trace elements in granites by inductively coupled plasma mass spectrometry. Talanta 51, 507513. Rollison, H. R. (1993) Using Geochemical Data: Evaluation, Presentation, Interpretation. John Wily & Sons, New York, 352 pp. Sillitoe, R. H. and Bonham, H. F. (1990) Sediment-hosted gold depositsDistal products of magmatic-hydrothermal systems. Geology 18, 157161. Smith, M. P., Henderson, P. and Campbell, L. S. (2000) Fractionation of the REE during hydrothermal processes: Constraints from the Bayan Obo Fe-REE-Nb deposit, Inner Mongolia, China. Geochimica et Cosmochimica Acta 64(18), 31413160. Stenger, D. P., Kesler, S. E., Peltonen, D. R. and Tapper, C. J. (1998a) Deposition of gold in Carlin-type deposits: The role of sulfidation and decarbonation at Twin Creeks, Nevada. Economic Geology 93 , 201215. Stenger, D. P., Kesler, S. E. and Vennemann, T. (1998b) Carbon and oxygen isotope zoning around Carlin-type gold deposits: A reconnaissance survey at Twin Creeks, Nevada. Journal of Geochemical Exploration 63 (2), 105121. Su, W-C., Qi, L., Hu, R-Z. and Zhang, G-P. (1998a) Analysis of rare-earth elements in fluid inclusions by inductivity coupled plasma-mass spectrometry (ICP-MS). Chinese Science Bulletin 43(22), 19221927. Su, W-C., Yang, K-Y., Hu, R-Z. and Chen, F. (1998b) Fluid inclusion chronological study of the Carlin-type gold deposits in southwestern China: As exemplified by the Lannigou gold deposit, Guizhou Province. Acta Mineralogica Sinica 18(3), 359362 (in Chinese). Sun, Z-Y., Yie, X-X., Liu, Y-K. and Wuan, G-Q. (1993) Submicron gold analysis by AEM combined EPMA. Chinese Science Bulletin 38(9), 803806 (in Chinese). Tan, Y-J. (1994) Geochemical types of the micro- and finegrained disseminated gold deposits in Yunnan-GuizhouGuangxi region. Mineral Deposits 13 (4), 308 321 (in Chinese). Taylor, R. P. and Fryer, B. J. (1980) Multi-stage hydrothermal alteration in porphyry copper systems in northern Turkey: The temporal interplay of potassic, prophylitic and phyllic fluids. Canadian Journal of Earth Sciences 17, 901926. Taylor, R. P. and Fryer, B. J. (1982) Rare earth element geochemistry as an aid to interpreting hydrothermal ore deposits. Metallization Associated with Acid Magmatism (Evans, A. M., ed.), 357365, Wiley, New York. Tosdal, R. M., Clin, J. S., Hofstra, A. H., Peters, S. G., Wooden, J. L. and Young-Mitchell, M. N. (1998) Mixed sources of Pb in sedimentary-rock-hosted Au deposits, Northern Nevada. Contributions to the Gold Metallogeny of Northern Nevada . U.S.G.S. Open-File Report, 98-338 (Tosdal, R. M., ed.), 223233, U.S. Geological Survey, Menlo Park, CA. Van der Weijden, C. H. and Van der Weijden, R. D. (1995) Mobility of major, minor and some redox-sensitive trace elements and rare earth elements during weathering of four granitoids in central Portugal. Chemical Geology 125(3/4),

149168. Wang, K-R., Zhou, Y-Q., Sun, L-G. and Ren, Z-G. (1994) Study on the Gold Occurrence from Several Typical Carlin-type Gold Deposits in China. Publish House of the Chinese University of Science and Technology, 103 pp. (in Chinese). Wang, X-Z., Cheng, J-P. and Zhang, B-G. (1992) Geochemistry of the Reworked Type Gold Deposits in China . Science Press, Beijing, 176 pp. (in Chinese). Wang, Y-G., Suo, S-T. and Zhang, M-F. (1994) Tectonics and Carlin-type Gold Deposits in Southwestern Guizhou, China. Geological Publishing House, Beijing 115 pp. (in Chinese). Whitford, D. J., Korsch, M. L., Porritt, P. M. and Craven, S. J. (1988) Rare-earth element mobility around the volcanogenenic polymetallic massive sulfide deposit at Que River, Tasmania, Australia. Chemical Geology 68 , 105119. Williams-Jones, A. E. and Wood, S. A. (1992) A preliminary petrogenetic grid for REE fluorocarbonates and associated minerals. Geochimica et Cosmochimica Acta 56, 725738. Woitsekhowskaya, M. B. and Peters, S. G. (1998) Geochemical modeling of alteration and gold deposition at the Betze deposit, Eureka County, Nevada. U.S.G.S. Open-File Report, 98-338 (Tosdal, M., ed.), 211222, U.S. Geological Survey, Menlo Park, CA. 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 natural waters. Chemical Geology 82, 159 186. Wood, S. A. (1996) The role of humic substances in the transport and fixation of metals of economic interest (Au, Pt, Pd, U, V). Organic and Ore Deposits (Girodano, T. H., ed.), Ore Geology Reviews 11 (13), 131. Wood, S. A. and Williams-Jones, A. E. (1994) The aqueous geochemistry of the rare-earth elements and yttrium. 4. Monazite solubility and REE mobility in exhalative massive sulfide-depositing environments. Chemical Geology 115, 4760. Wu, J. and Li, S-T. (1993) Application of thermal-process parameters in ore-forming process analysis: A case study of the gold-triangle area between Yunnan, Guizhou and Guangxi. Earth Science Journal of China University of Geosciences 18 (6), 725734 (in Chinese). Xiao, R-G., Chen, H-Q. and Fan, J. (1998) Geological condition controlling Carlin-type gold deposits of YunnanGuizhou-Guangxi area. Acta Mineralogica Sinica 18 (3), 344349 (in Chinese). Xu, Yigang (2002) Mantle plumes, large igneous provinces and their geological consequences. Earth Science Frontier 9(4), 341 352, China University of Geosciences, Beijing (in Chinese). Yang, K-Y. and Dong, Z-S. (1994) Carlin-type gold deposits in Yunnan, Guizhou, Guangxi and northwestern Sichuan Provinces. Advances of Gold Deposit Researches in China (Gold Resource Section of Chinese Academy of Sciences, ed.), 284347, Volume I (2), Dizhen Publish House (in Chinese). Zeng, Y-F., Liu, W-J. and Chen, H-D. (1992) Sedimentation characteristics and evolution of the Youjiang complex basin. Guangxi Geology 5 (4), 114. Zhang, C-J. and Li, X-L. (1998) Geochemical characteristics

380 Z. Bao et al .

of platinum group elements in Emeishan basalts. Acta Petrologica Sinica 14(3), 299304 (in Chinese). Zhang, F. and Yang, K-Y. (1992) Fission track chronological study of the fine disseminated gold deposits in southwestern Guizhou. Chinese Science Bulletin 37(17), 15931595 (in Chinese). Zhang, J-R., Lu, J-J., Zhou, K-J. and Zhang, X-H. (1992) Metallogenic characteristics in micro-fine disseminated gold deposit belt of east southern of Yangtze plate. Journal of Nanjiing University (Earth Sciences) 4 (3), 76 85 (in Chinese). Zhang, Y-X., Cun, G. and Liu, L-D. (1996) Gold Deposits in China: Advances and Considerations. Geological Publishing House, Beijing, 205 pp. (in Chinese).

Zhang, Z-J. and Zhang, W-H. (1999) Investigation into metallogenic fluid feature in Carlin-type gold deposits and its relation to mineralization in southwest Guizhou Province. Earth ScienceJournal of China University of Geosciences 24(1), 7478 (in Chinese). Zhu, L-M. and He, M-Y. (1996) Geochemistry of the fluid inclusions in disseminated gold deposits of southwestern Guizhou Province. Volcanology and Mineral Resources 17 (1/2), 6575 (in Chinese). Zhu, L-M., Liu, X-F., Jin, J-F. and He, M-Y. (1998) The study of the time-space distribution and source of ore-forming fluid for the fine-disseminated gold deposits in the YunnanGuizhou-Gangxi area. Scientia Geologica Sinica 33 (4), 463473 (in Chinese).

HFSE, REE and PGE geochemistry of gold deposits in SW China 381