STABLE ISOTOPE COMPOSITION AND MODEL OF TIN ORE FORMATION IN THE ĐÀ LẠT ZONE, SOUTH VIỆT NAM

NGUYỄN THỊ BÍCH THỦY

Institute of Geosciences and Mineral Resources, Thanh Xuân, Hà Nội

Abstract: Numerous bodies of granitoids and volcanic rocks exposed in the Đà Lạt zone consist of subduction-related products. The granitoids belong to an Andean-type arc in the West Pacific region running from SE China through S Việt Nam to SW Borneo. They have been divided into 3 complexes: 1) The Định Quán; 2) Đèo Cả; and 3) Cà Ná ones. These rocks are of subalkaline affinity and belong to the high-K calc-alkaline series. Most of them display features typical of I-type granites.

Tin deposits in the Đà Lạt zone are structurally controlled and hosted by both Jurassic sedimentary and late Mesozoic igneous rocks. The tin deposits are genetically associated with highly differentiated, fine-grained, biotite-bearing granites of the Cà Ná Complex, and strongly altered granodiorites of the Định Quán Complex. Tin mineralization in fine-grained granites of the Cà Ná Complex occurs in 2 forms, as dissemination and greisen zone: 1) Disseminations of accessory cassiterite occur commonly in the roof zones. This form of mineralization is usually accompanied by intense albitization; 2) Lodes and fracture-controlled greisen zone contain commonly cassiterite in association with the hydrothermal alteration (greisenized). Cassiterite occurs in majority of the silicate-cassiterite type that is characterized by the development of cassiterite, quartz with variable muscovite, tourmaline and ± fluorite. The calculated oxygen and hydrogen isotopic compositions of tin ore fluids indicate that magmatic waters were the major source of the fluid. In so far as oxygen and hydrogen are concerned, only minor amount of meteoric water has been admixed with the magmatic water in the hydrothermal circulation system. The temperatures of tin ore formation have been calculated using oxygen isotopic fractionation of mineral pairs ranging from ~ 350 to 500^oC, mainly from 350 to 400^oC.

I. INTRODUCTION

The Đà Lạt Zone, 300×150 km in size, is located in South Việt Nam (Fig. 1). Tin mineralization, forming economic deposits, has been found. The deposits are structurally controlled and hosted by both Jurassic sedimentary and late Mesozoic igneous rocks. Tin deposits around the world are known to be closely associated with granitic rocks and related effusives [10]. In the Đà Lạt Zone, tin deposits consist of mineralized granite cupolas with veining and pervasive greisenized zones in faults and fractures within granitoid plutons of the Cà Ná and Định Quán Complexes or sediments of the La Ngà Formation. Tin mineralization occurs in two types, namely cassiterite-silicate and cassiterite-sulfides. Cassiterite-silicate deposits are characterized by the presence of cassiterite, quartz with variable tourmaline, muscovite and base metal sulfides. Cassiterite-sulfide type is less recognized and mainly characterized by development of cassiterite associated with arsenopyrite, pyrite, and chalcopyrite in quartz veins. It is generally observed that, structure is the most dominant control on the localization of the tin mineralization in the Đà Lạt zone.

The close association of tin deposits with granitic rocks has prompted several studies on mineralogical composition and geochemical characteristics of granites, that are thought to be related to this mineralization [11]. The same author has investigated the mineralogical composition of greisenized zone and ore veins. Despite these efforts, there are many aspects of ore evolution that remain poorly understood. One of the most crucial of these is the sources of ore-forming elements and of ore-forming fluids, involved in mineralization and isotopic composition of magmatic fluids, as well as hydrothermal fluids in ore vein systems. Particularly, the question whether cassiterite had magmatic and/or sedimentary sources or were scavenged from crustal rocks by convective circulation of (magmatic) hydrothermal fluids has to be solved.

This paper focuses on the genesis of ore fluids, crystallization temperatures of cassiterite mineralization, mainly using the stable isotopic data obtained from co-existing minerals. The nature of the relationship of magmatic rocks to processes of tin deposition is also touched upon.

II. SAMPLE DESCRIPTION

The samples used in investigation were collected at 4 areas, namely Núi Cao, Đăk Hoa, Ma Ty and Sa Võ - Đa Tan Ky. The locations of the samples are shown in Fig. 1. Some of representative specimens and microphotographs are displayed in Fig. 2.

1. Núi Cao area: Four samples (NC-1/1, NC-2/1, NC-2/3 and NC-2/4) were taken from the currently mined deposit. The deposit contains pervasive greisenized zones in fault, whose dip and strike are on the average of 45^o to 70^o and 310^o to 330^o, respectively. Sample NC-1/1 was collected in the middle part of the zone, where rocks are totally altered and of dark colour, containing quartz, tourmaline and cassiterite. Three samples (NC-2/1, NC-2/3 and NC-2/4) were collected in about 1 km SE of sample NC-1/1 locality. Sample NC-2/1 is from the central part of greisenized fault zone, completely altered and of dark colour, while sample NC-2/3 was taken from greisenized adjacent part, about 25 cm from NC-2/1. Sample NC-2/3 is also completely altered, but still conserves the igneous structure, similar to that of the Định Quán granitoids. Zircon is an abundant accessory mineral in this sample. Sample NC-2/4 was encountered in the developed well, at the depth of about 40 m beneath sample NC-2/1. Mineralogical composition consists of plagioclase, K-feldspar, quartz, biotite, hornblende and opaque phases. Biotite is altered to form cassiterite, ilmenite and rutile. Macroscopic and microscopic examination indicates that this sample is a granodiorite, similar to rocks of the Định Quán Complex.

2. Đar Hoa area: One sample DH was collected in the active tin deposit, in about 1-1.5 km SW of sample NC-2/1. In general, veins in this deposit are irregular, steep (55-60^o), forming a broad zone (Fig. 2, sample DH). The width of veins ranges from several up to 40 cm. These veins are hosted by the La Ngà sedimentary rocks, which are mainly composed of weathered sandstone and siltstone.

3. Ma Ty area: Sample MT-2156 was taken from quartz-cassiterite vein selvage of an inactive deposit in Ma Ty area. This selvage zone consists of fine-grained, biotite-bearing granite, which are greisenized, comprising dominantly quartz (40 to 50 %), muscovite (20 to 25 %), K-feldspar (20 to 30 %) and several percent of accessory minerals, such as fluorite, cassiterite, pyrite, chalcopyrite, and hematite. One of the most interesting features of the Ma Ty deposit is the development of muscovite as product of pervasive granite greisenization, of local vein-related wall-rock alteration. This muscovite is the most abundant mineral in the Ma Ty quartz-cassiterite veins. Tourmaline has not been found in this tin deposit.

4. Sa Võ - Đa Tan Ky area: Three samples were taken within Đa Tan Ky pluton. The pluton consists of leucocratic fine-grained granite and granitic aplite of the Cà Ná Complex. Sericitization and greisenization are commonly observed at the roof of the pluton. Cassiterite occurs in two mineralization types, that are quartz-

Figure 1. Simplified geological map of the Đà Lạt Zone showing distribution of granitoid rocks (Tiến et al, 1991) and sampling sites. Upper left inset indicates that from Middle Jurassic through Middle Cretaceous times the SE Asian margin was an Andean-type arc (Taylor and Hayes, 1983). NW subduction beneath the continent is evidenced by widespread rhyolitic volcanism and granitic intrusions along SE China (e.g. Jahn et al, 1976) and SE Việt Nam. Lower right inset shows Việt Nam, locations of study area and neighbouring countries.

tourmaline-cassiterite (sample DK-1) and quartz-cassiterite (sample DK-2). The quartz-tourmaline-cassiterite type consists of quartz and tourmaline (about 60-98 %), cassiterite, sulphide and less hematite. The latter type, which does not contain tourmaline, is characterized by big grains of cassiterite, forming a band within the quartz vein (Fig. 3, DK-2). Minor amount of pyrite, chalcopyrite, sphalerite, and arsenopyrite are also present. Pyrite is commonly altered to limonite. Sample DK-3 was collected in the granite greisenization zone, which contains cassiterite with blur oscillatory zoning pattern (Fig. 3, DK-3).

III. STRUCTURAL GEOLOGY AND MINERALIZATION

The most prominent structural features in the Đà Lạt zone are normal fault zones, which strike mainly NE-SW, NW-SE, and N-S. The NE-SW striking faults appear to control the shape of the eastern coastline of the study area. The oldest and deep NW-SE trending fault zones had a major influence on localization of ore deposits in the Đà Lạt zone. The main mineralized fault zones in the deposits of the study area also strike NW-SE and dip 45-80^o. These faults cut the Jurassic sedimentary and Cretaceous granitic rocks and thus were active during, or after, the Yanshanian orogeny. The NW-SE faults were cut and offset by NE-SW trending fault zones.

The tin mineralization is associated with leucocratic fined-grained granites of the Cà Ná and probably with altered biotite-hornblende bearing granodiorites of Định Quán Complexes. Tin mineralization in fine-grained granites of the Cà Ná Complex occurs in two types as dissemination and as greisen veins. 1) Dissemination of accessory cassiterite occurs commonly in the margin and roof zones. This form of mineralization is usually accompanied by intense albitisation; 2) Lodes and fracture-controlled greisen zones contain cassiterite probably associated with post-magmatic alteration. Greisen consists of tin-bearing veins and pervasively greisenized zones. The most of ore veins or mineralization zones range from several centimetres to tens metres in width and are ten to hundreds metres long. For the Định Quán Complex, the tin mineralization is associated with granodiorites, which are later altered.

Figure 3. CL images and EDS photograph of cassiterite separated from sample MT-2156 (Ma Ty), DK-1, DK-3 and DK-2 (Đa Tan Ky), respectively.

The silicate-cassiterite mineralization type encountered in the Đa Tan Ky and Núi Cao areas (Fig. 3, sample DK-2, NC-1/1) is characterized by early development of cassiterite with quartz, tourmaline and ± fluorite, with later development of arsenopyrite and base metal sulphides. Cassiterite is the only ore mineral extracted from the ore samples, occurring as euhedral crystals or crystal groups, generally in bipyramidal form. Cassiterite crystals from Ma Ty and Đa Tan Ky areas are dark brown (samples MT-2156; DK-1) and exhibit finely homogeneous zoning, which is not observed in cassiterite at Núi Cao (samples NC2/1, NC2/3 and NC2/4) location. The zoning patterns of cassiterite grains in samples MT-2156 and DK-1 are clearly shown in CL photographs (Fig. 3). The same observation, but blur, is made for cassiterite crystals in sample DK-3. The EDS photograph of cassiterite of sample DK-2 is from a cassiterite-quartz vein, which cross-cuts the granite of the Đa Tan Ky pluton.

IV. HYDROTHERMAL ALTERATION OF THE GRANITOIDS

The δ¹⁸O values of quartz from all three granitoid complexes as well as of cassiterite and tourmaline from greisenized granite zones and ore vein are illustrated in Fig. 4 and Table 1. It is clear that cassiterite from vein has lower d¹⁸O than those from altered granodiorites. Quartz separated from quartz-cassiterite vein and greisenized granite has higher d¹⁸O values than those separated from granitoids of three complexes. Small isotopic fractionation (δ¹⁸O<0.3 ‰) at magmatic temperatures (about 700-750^oC) allow the isotopic composition of pristine granite to be used as a guide to the isotopic composition of any aqueous phase that might have unmixed from the magma [26]. Determination on mineral separates of quartz, K-feldspar, and biotite yields δ¹⁸O fractionations of 1.6 to 3.2 ‰ for quartz-feldspar and 4.6 to 6.2 ‰ for quartz-biotite (Table 4.5). For minerals crystallized from granitic melts, the δ¹⁸O values, e.g. for feldspar and biotite are expected, respectively, to be less than 1.5 and 3.0 to 4.5 ‰ lower than those for coexisting quartz [7]. Therefore, this δ¹⁸O fractionation range for quartz-feldspar and quartz-biotite pairs in the Đà Lạt granitoids, in combination with medium- to coarse-grained granitoids may indicate subsolidus oxygen isotopic exchange of feldspar and biotite with water depleted in ¹⁸O. The later hydrothermal alteration stage of the Đà Lạt granitoids is characterized by two types of greisenized rocks, which represent complete and incomplete greisenization processes.

The incomplete greisenization is clearly observed in the fine-grained Cà Ná granite, and characterized by the destruction of igneous textures and the replacement of plagioclase and biotite by stable hydrothermal mineral assemblage of quartz, muscovite, cassiterite, K-feldspar, sulphides, and ± fluorite, ± tourmaline. Rocks associated with the complete greisenization underwent argillitic alteration, as encountered in Núi Cao (sample NC-1/1, NC-2/1) location. The primary rocks and textures of the completely greisenized rocks are difficult to identify. However, less-altered samples NC-2/3 and NC-2/4 contain large amount of mafic minerals (biotite and hornblende) and preserved igneous texture characteristic similar to the Định Quán rocks. Thus, it is possible to suggest that the completely greisenized rocks as well as the samples NC-2/3 and NC-2/4 belong to the Định Quán Complex.

The completely greisenized rocks contain quartz, cassiterite, tourmaline, and ± K-feldspar ± muscovite. In the polished section of the samples NC-2/3, NC-2/4 and DK-3 the mineral assemblage of cassiterite + ilmenite + rutile + titanite is commonly observed. Tin was probably liberated by destruction of biotite during hydrothermal activity and deposited as cassiterite in the mineralized zones and/or in veins and form economic tin deposit.

V. GEOTHERMOMETRY AND OXYGEN ISOTOPE COMPOSITION OF HYDROTHERMAL FLUIDS IN ORE VEIN SYSTEMS

Quartz, cassiterite and tourmaline in samples NC-1/1, NC-2/1, NC-2/4, DH, and DK-1 were selected in order to apply oxygen isotope thermometry for the tin mineralization stage. There are considerable variation of δ¹⁸O values of quartz, cassiterite, and tourmaline among these samples (Fig. 4, Table 1). This suggests that the variations in temperatures and/or oxygen isotopic composition of fluids from which those minerals precipitated are relatively large. Three quartz-cassiterite pairs have oxygen isotope fractionations of from 8.7 to 7.0 ‰. Calculated temperatures range from 430 to 500^oC for the quartz-cassiterite pairs, using the calibration of Zhang et al [32], are in good agreement with those obtained by using the fractionation expression of Zheng [33]. Two quartz-tourmaline mineral pairs from samples NC-1/1 and DK-1 indicate temperatures for isotopic equilibrium of 330^o and 462^oC, respectively, applying the calibration of Zheng [33].

Oxygen isotopic composition of quartz, cassiterite, and tourmaline-forming fluids can be calculated using the isotope fractionation equations between minerals and water. The temperature estimates from quartz-cassiterite and quartz-tourmaline pairs are used for this calculation. Quartz-water, cassiterite-water and tourmaline-water fractionations were calculated using the expressions of Friedman and O’Neil [7, 32] and Zheng [33], respectively:

Table 1. Isotope analysis and geothermometry of coexisting minerals in tin mineralization zones and calculated values of hydrothermal fluids

Mineral (sample)	δ^18O (±0.2‰)	δD (±1‰)	Oxygen isotope thermometry (T^oC)	Calculated δ^18Owater (‰)	Calculated δDwater (‰)
Quartz (NC-1/1)	11.4		330	5.0
Tourmaline (NC-1/1)	7.9	−68.5	330	5.8	−72
Cassiterite (NC-1/1)	2.6		330	6.2
Quartz (NC-2/1)	10.4		437	6.6
Cassiterite (NC-2/1)	2.5		437	6.6
Quartz (NC-2/4)	10.0		500	7.2
Cassiterite (NC-2/4)	3.0		500	7.2
Quartz (DH)	11.1		376	6.0
Cassiterite (DH)	2.4		376	6.0
Quartz (DK-1)	10.0		462	6.6
Tourmaline (DK-1)	7.4	−80.1	462	7.0	−79

Calculated δ¹⁸O values of water coexisting with hydrothermal minerals are shown in Table 1. The δ¹⁸O_water values for the quartz-forming fluids are quite similar to those of the cassiterite and tourmaline-forming water during the stage of tin mineralization. This indicates that the minerals deposited from the same fluids and are in isotopic equilibrium. The small range of calculated δ¹⁸O_water values of fluids in equilibrium with quartz, cassiterite and tourmaline for each sample reflects the range of temperatures that appears to be associated with given quartz, cassiterite and tourmaline values.

VI. GEOCHEMISTRY IN RELATION TO TIN-BEARING POTENTIAL

The geochemical composition, particularly trace element relationship and ratios, can be used as reliable indicators of the ore-bearing potential of granite rocks [3, 22]. This application is based on the difference in the migration characteristics of related elements during magmatic and postmagmatic processes. As discussed in the previous section, granites of the Cà Ná Complex are highly fractionated rocks and their geochemical features are similar to most tin-bearing granite rocks [8, 30, 31]. Rubidium and lithium concentration, which have been widely considered as good evidence of tin-bearing granites [3, 25], are higher in the Cà Ná rocks in comparing to the Định Quán and Đèo Cả rocks [29]. The significantly enhanced lithium values may in part reflect the genetic association of the granites with Li-rich greisen. Rb is extremely enriched in the granites, whereas, the reverse is true for Ba and Sr. Consequently, Ba/Rb and Sr/Rb ratios of the Cà Ná rocks are low. In addition, low K/Rb ratios are also observed for the Cà Ná granites [29]. These relationships suggest that Rb is significant enriched relative to K, Sr and Ba during post-magmatic processes related to mineralization.

It has also been observed that Na is strongly depleted in several greisen samples from the Cà Ná granite, probably as result of hydrothermal leaching associated with incipient greisenization, and consequently higher K/Na ratios are obtained for greisenized granites (K/Na = 6.5, sample CN2.1). High Rb/Zr ratios (≥ 2.0, up to 15.4) exist for most analyzed samples of the Cà Ná Complex [29]. These relationships are diagnostic characteristics of tin-bearing granites as reported from many workers [20].

Table 2. Electron microprobe analyses (wt %) of cassiterite from Ma Ty
and Sa Võ - Đa Tan Ky areas

Sample	MT-2156
Crystal	1	1	1	1	1	1	1	1	1	1	1
Zones	dark	light	dark	light	light	light	dark	light	dark	dark	dark
SnO2	97.76	98.99	96.10	100.32	99.55	97.82	96.67	97.84	97.06	96.03	97.10
Ta2O5	0.01	0.03	0.00	0.00	0.00	0.00	0.05	0.15	0.03	0.00	0.15
TiO2	0.99	0.03	1.25	0.07	0.21	0.16	0.56	0.15	0.09	1.13	0.36
Nb2O5	0.79	0.05	0.00	0.05	0.00	0.00	0.46	0.04	0.01	0.10	0.38
FeO	0.22	0.21	1.48	0.17	0.24	0.01	0.08	0.25	0.19	0.54	0.15
Total	99.2	99.3	98.8	100.6	100	98.0	97.8	98.4	97.4	97.8	98.1

Table 2. (Continued)

Sample	DK-1
Crystal	1	1	1	2	2	2	3	3	3	3
SnO2	97.10	96.33	97.48	97.65	97.07	97.91	96.9	95.88	98.09	97.42
Ta2O5	0.06	0.00	0.08	0.00	0.02	0.05	0.09	0.00	0.00	0.02
TiO2	0.26	0.54	0.15	0.00	0.25	0.08	0.06	0.27	0.31	0.09
Nb2O5	0.00	0.00	0.01	0.00	0.00	0.02	0.03	0.00	0.02	0.00
FeO	0.25	0.61	0.12	0.02	0.12	0.10	0.53	1.33	0.27	0.23
Total	97.7	97.5	97.8	97.7	97.5	98.3	97.6	97.5	98.7	97.8

Sample	DK-2
Crystal	1	1	1	2	2	3	3
SnO2	96.1	94.46	95.07	95.48	94.00	94.28	94.77
Ta2O5	0.00	0.08	0.00	0.02	0.00	0.04	0.00
TiO2	0.06	0.00	0.14	0.25	0.00	0.03	0.00
Nb2O5	0.00	0.00	0.02	0.03	0.05	0.03	0.01
FeO	1.56	3.00	2.11	2.10	3.68	3.11	2.15
Total	98.7	97.5	97.34	97.9	97.7	97.5	97.9

Sample	DK-3
Crystal	1	1	1	1	1	1
SnO2	96.91	97.68	96.55	96.85	97.54	96.92
Ta2O5	0.00	0.07	0.04	0.00	0.00	0.31
TiO2	0.44	0.38	0.49	0.31	0.53	0.05
Nb2O5	0.08	0.19	0.08	0.01	0.00	0.00
FeO	0.07	0.16	0.19	0.18	0.64	0.35
Total	97.5	98.5	97.4	97.4	98.7	97.6

Sample	NC-2/1
Crystal	1	1	1	1	1	1	2	2	2
SnO2	95.76	98.98	98.68	98.33	98.06	97.89	98.29	98.37	98.35
Ta2O5	0.00	0.00	0.01	0.04	0.00	0.00	0.03	0.00	0.00
TiO2	1.58	1.76	0.61	0.48	0.62	0.76	0.71	0.65	0.43
Nb2O5	0.02	0.00	0.00	0.00	0.00	0.03	0.16	0.00	0.00
FeO	0.63	0.16	0.18	0.04	0.12	0.14	0.05	0.15	0.14
Total	98.0	99.9	99.5	98.9	98.8	98.8	99.2	99.2	98.9

Five cassiterite samples were analyzed by electron microprobe and the analytical results are summarized in Table 2. Cassiterite collected from different locations has similar concentration of SnO₂ and Nb₂O₅, but slightly higher TiO₂ content is observed for cassiterite in the Núi Cao deposit. The striking feature of cassiterite from a quartz-cassiterite vein style in Đa Tan Ky is that it contains significantly higher FeO concentration (sample DK-3) compared to cassiterite from quartz-tourmaline-cassiterite vein style. The cassiterite crystals from Ma Ty display oscillatory zoning pattern, each zone was analyzed and the results show that the brighter zones in the CL image slightly have higher SnO₂, but lower FeO and Ta₂O₅ contents than the darker zones.

VII. DISCUSSION

1. Origin of ore fluids

The origin of tin fluids could be constrained on the basis of the stable isotopic data. The oxygen and hydrogen isotopic composition of granitoids from three complexes cannot be well defined, since there is evidence for subsolidus exchange of the granitoids with water of lighter isotopic composition. However, the least altered δ¹⁸O values for the granitoids can be estimated to be close to the values of fresh feldspar [27] or 1.0 to 1.5 ‰ lower than the values of quartz. The later method is preferred since quartz has a much lower exchange rate than feldspar and biotite [4], and it is therefore more likely to have preserved its original oxygen isotope composition. Quartz from the least altered available samples (DQ-8, CN-10, and DC-15) were selected for this estimation, and the data suggest δ¹⁸O values of about 8.0, 8.5, and 7.5 ‰ for the Định Quán, Cà Ná and Đèo Cả granitoids, respectively. If the biotite with the highest δ¹⁸O values from these 3 samples had equilibrated with hydrothermal waters at 500^oC (the estimated temperature of subsolidus exchange), δD values of -58, -52, and -54 ‰ (Table 3) for the water are indicated. Therefore, the magmatic waters derived from the Định Quán, Cà Ná, and Đèo Cả granitoids are expected to have δD values within this range, or heavier.

Table 3. Stable isotope analyses of biotite from representative granitoid samples and calculated values of magmatic water. The calculated δ¹⁸O and δD were calculated using the calibrations of Friedman and O’Neil [7] and Suzouki and Epstein [24], respectively.

Sample	δ^18O	δD	T^oC using for calculatation of δ^18Owater	Calculated δ^18Owater	Calculated δDwater
DQ-8	3.5	-91	500	8.0	-58
CN-10	4.9	-86	500	8.5	-52
DC-15	4.4	-81	500	7.5	-54

Compared to the granitoids the calculated δ¹⁸O values of 5.0 to 7.2 ‰ for hydrothermal fluids (Table 1) that precipitated cassiterite and tourmaline are slightly lighter isotopically, suggesting that a small component of meteoric water mixed with magmatic water in the hydrothermal circulation system. The δD_water values for tourmaline-forming fluids were calculated using the empirical tourmaline-water isotope fractionation factor of Kotzer et al. [16] and are given in Table 1. These values are lower than those of the magmatic water indicating that an influx of meteoric water caused fluid mixing and dropped in the isotopic composition. The constricted range of δ¹⁸O values for cassiterite and tourmaline (Table 1) suggests that these minerals were precipitated over a narrow range of mixing of the hydrothermal fluids.

2. Model for processes of tin deposition

The geochemical evidence and the field relationship show that tin mineralization is associated with the fine-grained granites of the Cà Ná Complex and probably the hornblende, biotite-bearing granodiorite of the Định Quán Complex. This then raises a question as to a manner of the relationship of magmatic rocks to processes of ore deposition. This can be discussed below.

For the Cà Ná Complex, the fine-grained, biotite-bearing granite, which is rocks of the late phase and greisenized, are genetically as well as spatially related to tin mineralization, whereas the medium and coarse-grained rocks of early phase do not contain cassiterite. It is reported that the behaviour of Sn in granitic complexes is dependent mainly on the ƒ(_O2) of the melt. In oxidized melts, tin, in the 4+ valence, substitutes in minerals, such as magnetite, titanite and micas [16, 17, 27) and is not concentrated in melt by fractional crystallization. By contrast, in reduced granites, tin is dominantly divalent, and its abundance increases with fractionation [6, 18]. Linnen et al. [19] also reported that the strong increase of SnO₂ solubility with decreasing ƒ(_O2) indicates a significant dissolution of tin into the melt as Sn²⁺ at more reduced condition. The similar conclusion is reached by Taylor and Wall [28], though their experiments indicate a solubility, that is one order of magnitude lower than that in the study of Linnen et al. [19]. Granite-related tin deposits are commonly considered to be crystallized under reducing condition and classified as ilmenite series granitoids [28]. A crustal source contained very low Sn content [2-10 ppm, 18] and therefore, first batch partial melting processes will clearly lead to a strong depletion of tin regardless of the ƒ(_O2) conditions. It follows that for a granitic liquid to be subsequently saturated with cassiterite, Sn must be strongly incompatible and a high degree of fractional crystallization is required. Evidence for such processes are naturally occurring is provided by ongonites, which contain up to 500 ppm Sn [12] and the Macusani volcanics up to 194 ppm Sn [1]. In the Đà Lạt zone, the leucocratic and fine-grained, biotite-bearing granites and granitic aplites of the Cà Ná Complex, which are the late and highly fractionated phases. The fine-grained granites containing 42-95 ppm Sn and granitic aplite up to 1067 ppm Sn, whereas the coarse-grained granite of this complex contain 8-35 ppm Sn [7]. This indicates that tin concentration of the Cà Ná granite increases with fractional crystallization.

Tin as Sn²⁺ is ultimately, under reducing condition, partitioned to a volatile-rich fluid phase [2, 21, 23] that exsolved while the fine-grained granites crystallized. This exsolving magmatic fluid phase which carried Sn seems to have been locally trapped in cupolas of previously crystallized fine-grained granites and caused muscovite and greisen alteration of granites and cassiterite deposited. Tin mineralization may be localized by faults or in fractures that are developed as the late Mesozoic granitoid intruded, cooled and contracted. The oxygen and hydrogen data of hydrothermal fluid, in which cassiterite deposited, presented above do not support the involvement of quantities of meteoric water in cassiterite-stage mineralization. The combining between the stable isotope data and oscillatory zoning of cassiterite (Fig. 3) may suggest an orthomagmatic origin for cassiterite or at least for disseminated cassiterite in roof zones and pervasive greisen granites of the Cà Ná Complex. But the question “How and under what conditions Sn can be transported from the Cà Ná granitic melt to aqueous phase and precipitation of cassiterite, as well as chemical composition of mineralized-fluid?” is still a remaining problem, due to the failure of fluid inclusion data, as well as the chemical composition of magmatic-derived biotites including halogen elemental group. However, from the data in this study a suggestion for transport of tin can be made. As mentioned above, the residual melt of the Cà Ná granites are enriched in tin. Tin was transported from this melt to the aqueous phase as chloride complexes under low f(_O2) condition – a possibility consistent with its general polymetallic nature [2, 6, 9, 14, 19, 23]. Rarely seen fluorite may support this suggestion. The common occurrences of muscovite and/or quartz-tourmaline in tin deposits may suggest that the tin-bearing aqueous fluid has high HCl/KCl ratio and is rich in B content, respectively.

In contrast to the Cà Ná granites, rocks of the Định Quán Complex have the typical features of I-type granitoids, which are considered to form under relatively oxidized condition, and hence tin is not concentrated in the residual melt by fractionation. Consequently, the residual melt is insufficient concentration of tin to form ore deposit. However, the tin deposit, which is considered to be related to Định Quán granodiorites, is observed in Núi Cao area. Thus, a speculative mechanism for forming the Định Quán rock-related tin deposit is considered as a process of remobilization and redeposition of tin and will be discussed below.

The Định Quán granitoids derived by partial melting of lower continental crust can probably be rich in tin (whole-rock samples contain ca. 12 ppm Sn, [29]). This tin as Sn⁴⁺, nevertheless, may be dispersed in ferromagnesian, rock-forming minerals by substituting for Ti⁴⁺ and Fe³⁺ in biotite and hornblende [15, 16, 27, 28], which are major mafic minerals present in Định Quán granodiorites. Intensive alteration of these granodiorites may lead to liberation of Sn and other elements from these mafic minerals, mainly by destruction of biotite. Since a mineral assemblage of cassiterite, rutile, titanite and ilmenite is commonly observed in altered biotite from all polished sections (NC-2/1, NC-2/3 and NC-2/4). Consequently, the leached Sn was redeposited as cassiterite and made available for tin deposit. A similar mechanism was reported for cassiterite in hydrothermal ore deposit in Southeast Asia [13]. The presence of tourmaline in the tin deposit indicates that the hydrothermal fluid in which cassiterite deposited also contains B. Magmatic water is the major source of the hydrothermal fluid on the basis of the oxygen and hydrogen data represented in Table 2.

VIII. CONCLUSIONS

The calculated oxygen and hydrogen isotopic compositions of tin ore fluids show that magmatic water was the major source of the fluid. In so far as oxygen and hydrogen are concerned, only minor amounts of meteoric water have been admixed with the magmatic water in the hydrothermal circulation system. The temperatures of tin ore formation were in the range of ~ 350^o to 500^oC, mainly 350^o to 400^oC.

Hydrothermal veins and greisen tin deposits are genetically and spatially associated with the fine-grained, biotite-bearing granites that are the late and highly evolved phases of the Cà Ná Complex, and altered granodiorites of the Định Quán Complex induced by the intrusion of Cà Ná granites. The petrographical and geochemical characteristics of the fine-grained granites are like those of granites located elsewhere in the world that are associated with deposits of tin ore.

Evolution of the Cà Ná Complex produces its association with tin deposits. Intense alteration of fine-grained, biotite-bearing granites formed a cassiterite-bearing greisen. The alteration was achieved by late-stage, hydrothermal fluid that was locally trapped in cupolas beneath the impermeable fine-grained granitic and La Ngà sedimentary rocks. The effect of the fluid was to convert the rocks to an assemblage of quartz, muscovite, tourmaline, cassiterite, ± fluorite, and sulphides. Hydrothermal alteration involving greisenization of the Định Quán granodiorites due to intrusion of the Cà Ná granites leads to leaching of Sn and other elements from biotite. The leached materials were probably partially redeposited in the veins of the mineralised zones by migrating solutions.

REFERENCES

1. Antipin V.S., Kovalenko V.I., Kuznetsova A.I., Persikova L.A., 1981. Distribution coefficients for tin and tungsten in ore-bearing acid igneous rocks. Geochem. Intern., 18: 92-106.

2. Bailey J.C., 1977. Fluorine in granitic rocks and melts: A review. Geochem. Geol., 19 : 1-42.

3. Beus A.A., Gregoria S.V., 1975. Geochemical exploration methods for mineral deposits. Wilmette Illinois Applied Pub. Ltd : 53-71.

4. Criss R.E., Taylor H.P., 1983. An ¹⁸O/¹⁶O and D/H study of Tertiary hydrothermal systems in the southern half of the Idaho batholith. Geol. Soc. Amer. Bull., 94 : 640-663.

5. Eadington P.J., Kinealy K., 1983. Some aspects of the hydrothermal reaction of tin during skarn formation. Geol. Soc. Australia, 30 : 461-471.

6. Eugster H.P., 1985. Granites and hydrothermal ore deposits: A geochemical framework. Miner. Mag., 49 : 7-23.

7. Friedman I., O’Neil J.R., 1977. Data of geochemistry, 6^th ed.. Chap K, complication of stable isotope fractionation factors of geochemical interest. U.S. Geol. Surv. Prof. Paper : pp. 440-KK.

8. Haapala I., 1977. Petrography and geochemistry of the Eurajoki stock, a Rapakivi-granite Complex with greisen-type mineralization in SW Finland. Finland Geol. Surv. Bull., 286: pp. 128.

9. Heinrich C.A., 1995. Geochemical evolution and hydrothermal mineral deposition in Sn (-W-base metal) and other granite-related ore systems: Some conclusions from Australia examples. Geol. Assoc. of Canada short course handbook, 23 : 203-220.

10. Hosking K.F.G., 1967. The relationship between primary deposits and granitic rocks, in Fox.W (Eds). Technical conference on tin. 1^st in London. Intern. Tin Council. 1: 267-311.

11. Hung N.V., 1999. Tin potential of high-aluminum granites (Đa Tan Ky and Ankroet plutons) from the Đà Lạt zone. Master thesis, Nat. Univ. of Hồ Chí Minh City (in Vietnamese).

12. Ishara S., 1981. The granitoid series and mineralization. Econ. Geol., 75^th Anniversary vol., pp. 458-484.

13. Jackson K.J., Helgeson H.C., 1985. Chemical and thermodynamic constraints on hydro-thermal transport and deposition of tin: II. Interpretation of phase relations in the SE Asian tin belt. Econ. Geol., 80 : 1365-1378.

14. Keppler H., Wyllie P.J., 1989. Partitioning of Mo, W, Cu, Cu and Sn between granitic melt and fluid phase. EOS Trans. Amer. Geophys. Union, 70 : 1403.

15. Keppler H., Peter J.W., 1991. Partitioning of Cu, Sn, Mo, W, U, and Thu between melt and aqueous fluid in the system haplogranite-H₂O-HCl and haplogranite-H₂O-HF. Contrib. Miner. Petrol., 109 : 139-150.

16. Kotzer T.G., Kyser T.K., King R.W., Kerrich R., 1993. An empirical oxygen-and hydrogen-isotope geothermometer for quartz-tourmaline and tourmaline-water. Geochim. Cosmochim. Acta, 57 : 3421-3426.

17. Lehmann B., 1982. Metallogeny of tin: Magmatic differentiation versus geochemical heritage. Econ. Geol., 77 : 50-59.

18. Lehmann B., 1990. The Bolivian tin province and regional tin distribution in the central Andes: A reassessment. Econ. Geol., 85 : 1044-1058.

19. Linnen R.L., 1995. Genesis of a magmatic metamorphic hydrothermal system: The Sn-W polymetallic deposits at Pilok, Thailand. Econ. Geol., 90 : 1148-1166.

20. Olade M.A., 1980. Geochemical characteristics of tin-bearing and tin-barren granites, northern Nigeria. Econ. Geol., 75 : 71-82.

21. Pichavant M., 1988. The Miocene-Pliocene Macusani volcanics, SE Peru. Contrib. Mineral. Petrol., 100 : 325-338.

22. Smith TE., Turek A., 1976. Tin-bearing potential of some Devonian granitic rocks in SW Norva Scotia. Miner. Dep., 11 : 234-245.

23. Solomon M.S., Groves D.I., 1994. The geological setting and nature of Australia’s mineral deposits. Oxford Univ. Press. Oxford, UK.

24. Suzuoki T., Epstein, S., 1976. Hydrogen isotope fractionation between OH-bearing minerals and water. Geochim. Cosmochim. Acta. 40: 1229-1240.

25. Tauson L.V., Kozlov V.D., 1973. Distribution functions and ratios of trace elements, concentration as estimates of the ore-bearing potential of granites. In Geological exploration. London Inst. Min. Metal., pp. 37-44.

26. Taylor H.P., 1974. The application of oxygen and hydrogen isotope studies to problems of hydrothermal alteration and ore deposits. Econ. Geol., 69 : 843-883.

27. Taylor H.P., 1979. Oxygen and hydrogen isotope relation in hydrothermal mineral deposits. In Barnes H.L. (Eds). Geochemistry of hydrothermal ore deposits. Wiley Intersci., New York, pp. 236-277.

28. Taylor J.R., Wall V.J., 1992. The behaviour of tin in granitoid magmas. Econ. Geol., 87 : 403-420.

29. Thuy T.B. Nguyen, Satir M., Siebel W., Vennenmann T., Long T.V., 2003. Geochemical and isotopic constraints on the petrogenesis of granitoids from the Đà Lạt zone, southern Vietnam. J. of Asian Earth Sci., Elsevier.

30. Tischendorf G., 1973. The metallogenetic basis of tin exploration in the Erzgebirge. Inst. Min. Metal. Trans., 82B : 9-24.

31. Tischendorf G., 1977. Geochemical and petrographical characteristics of silicic magmatic rocks associated with rare-element mineralization. In M. Stemprok (Eds) Metallization associated with acid magmatism. Geol. Surv.. Prague, Czech., 2 : 41-46.

32. Zhang L.G., Liu J.X., Chen Z.S., Zhou H.B., 1994. Experimental investigation of oxygen isotope fractionation in cassiterite and wolframite. Econ. Geol., 89 : 150-157.

33. Zheng Y.F., 1991. Calculation of oxygen isotope fractionation in metal oxides. Geochim. Cosmochim. Acta, 55 : 2299-2307.

34. Zheng Y.F., 1993. Calculation of oxygen isotope fractionation in hydroxyl-bearing silicates. Earth Planetary Sci. Letters, 120 : 247-263.