Major and Trace Element Geochemistry of Dayakou Vanadium-Dominant Emerald from Malipo (Yunnan, China): Genetic Model and Geographic Origin Determination

: Emerald from the deposit at Dayakou is classified as a vanadium-dominant emerald together with Lened, Muzo, Mohmand, and Eidsvoll emeralds. Although previous studies defined these V-dominant emeralds and traced the genesis of the Dayakou deposit, there has not been any systematic comparison or discrimination on V-dominant emeralds from these deposits. The origin of the parental fluid and the crystallization process of the Dayakou emerald remain controversial. In this study, both major and trace element signatures of 34 V-dominant samples from Dayakou are analyzed through electron microprobe analysis (EMPA) and laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS). Dayakou emeralds are characterized by high ratios of V/Cr and the enrichment of Li, Cs, W, Sn, and As. These geochemical fingerprints indicate a parental fluid of an Early Cretaceous early-stage granitic fluid associated with Laojunshan granite. The considerable concentration variation of Rb, Cs, Ga and the presence of V-rich oxy-schorl-dravite inclusions in a color zoned sample suggest two generations of emerald precipitation. Thus, a more detailed idealized mineralization model for the Dayakou deposit is proposed. A series of plots, such as Rb vs. Cs, V vs. V/Cr, LILE vs. CTE, and Li vs. Sc, are constructed to discriminate the provenance of V-dominant emeralds. adjacent metamorphic local zoning with outer enriched in intermediate emerald-bearing quartz an emerald-tourmaline-bearing (tourmaline-Tur) quartz


Introduction
Emerald, the green gem variety of beryl, is considered one of the most precious gems. With the ideal formula of Be3Al2(Si6O18), emerald crystallizes in a hexagonal structure which consists of six-membered rings (Si6O18) 12− bonded by tetrahedrally coordinated Be 2+ (T1 site) and octahedrally coordinated Al 3+ (Y site). Y sites, preferentially occupied by Al 3+ , can accommodate Mg, Mn, Fe, Cr, V, Ti, Sc, Co, and Ni ions, while T1 sites can accommodate Li [1][2][3][4][5]. The six-membered rings filled by Si 4+ (T2 site) are stacked one above the other, forming channels of approximately 5.1 Å in diameter along the c-axis. These cavities are large enough to incorporate single molecules of H2O, D2O, HDO, CO2, etc. In addition, many large size cations, such as alkali metal ions, are incorporated into the channel to compensate for the deficiency of positive charges as a result of divalent cation substitutions at Y sites [2,[6][7][8][9][10][11][12].
Most previous works have discussed the element concentration differences among the emerald deposits worldwide. Those studies emphasized the channel cations (Na, K) and isomorphic substituents (Fe, Mg, and chromophoric Cr and V) at the Y, T1, and T2 sites [5,[13][14][15][16][17][18][19]. References [2] and [10] summarized emerald deposits worldwide and collected their chemical compositions. They defined Dayakou (China), Lened (Canada), Muzo (Colombia), Mohmand (Pakistan), and Eidsvoll (Norway) emeralds as typical V-dominant emeralds based on the greater content of V2O3 than Cr2O3. Those V-dominant emeralds can be easily distinguished from Cr-dominant emeralds. However, there has not been a systematic comparison or discrimination of those V-dominant emeralds from various deposits. With regards to Dayakou emeralds, their chemical compositions have been recently reported [20], but not analyzed in detail.
Trace elements of emerald are powerful indicators for determining the geographic provenance but the number of related studies is limited. For the first time, using Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry (LA-ICP-MS) and Secondary Ion Mass Spectrometry (SIMS) techniques, [21] and [16] determined the origin of emeralds from several deposits by trace elements data, such as Li, B, Rb, Cs, V, Cr, Sc, Ga, etc. So far, only [20] has reported trace elements of Dayakou emerald, but their work focused on the Al 3+ replacement degree and Cr/V ratio. The geochemical fingerprints of the Dayakou emerald are not yet determined.
Besides, the trace element geochemistry of emeralds was used for building the genetic model. B, Ga, Ba, Sr, and alkali elements Li, Cs, and Rb are regarded as indicators of parental fluid and the fractionation and evolution for pegmatite [3,4,22,23]. Reference [24] reported the formation conditions of Dayakou emerald, but the crystallization process and the sources of emerald-forming fluids have yet to be determined.
In this study, thirty-four single Dayakou emerald crystals are investigated. Electron microprobe analysis and LA-ICP-MS method were used for measuring major and trace elements of the Dayakou emerald. These analyses can determine the geochemical fingerprints of the Dayakou emerald and help discriminate the geographic origin of V-dominant emeralds. In addition, the geochemical features are linked to the genesis of the Dayakou emerald, revealing the origin of parent fluid and the crystallization process of emerald. The results of this study improve the genetic model of the Dayakou deposit, and determine the geographic origin of V-dominant emeralds.

Materials
Thirty four representative emerald samples (YE-1 to YE-34) from the Dayakou occurrence, Yunnan, China were selected for this study. These samples consist of crystal fractions and euhedral columnar single crystals with typical beryl crystal habit. Sized from 2 to 20 mm, these crystals are opaque to translucent and their colors cover various shades of green. All the samples were polished with double parallel sides and analyzed by LA-ICP-MS. Twenty-one individual crystals were selected for EMPA.

Laser Ablation-Inductively Coupled Plasma Mass Spectrometry
In situ trace element measurements were performed using a X-Series ICP-MS (Thermo Fisher Scientific, Waltham, MA, USA) fitted with a 343 nm femto-second laser ablation system (J-100, Applied Spectra, West Sacramento, CA, USA), housed at the National Research Center for Geoanalysis (CAGS). The radiofrequency power of ICP-MS was 1300 W. Helium gas carrying the ablated sample aerosol from the chamber was mixed with argon gas and nitrogen as an additional diatomic gas to enhance sensitivity. A baffled-type smoothing device was used in front of the ICP-MS to reduce fluctuation effects induced by laser-ablation pulses and to improve the analytic quality like [38]. Samples were ablated for 60 s at a repetition rate of 8 Hz at 8 J/cm 2 , and ablation pits were ~50 μm in diameter. Each analysis incorporated an approximate 20 s background acquisition (gas blank) followed by 50 s data acquisition from the sample. Every twelve analyses were followed by a calibration process with two analyses of NIST 610 and one analysis of NIST 612 in order to correct the time-dependent drift of sensitivity and mass discrimination. All elemental concentrations were calculated by applying 29 Si as an internal standard. Data reduction was carried out with the commercial software ICPMSDataCal 10.8 [39]. The analytical procedures and calibration methods were similar to those described by [39]. The detection limits of LA-ICP-MS range from 0.05 to 0.1 ppm for REE. The precision and accuracy are about 10% rel. at ppm level.

Results
Major and minor elemental compositions of the Dayakou emerald samples expressed in oxides and formula obtained from individual crystals by EMPA are presented in Table 1. A total of 49  electron microprobe analyses of emeralds and four analyses of tourmaline inclusions were obtained  from 21 individual crystals. Trace element compositions expressed in ppm were obtained from 34 individual emerald crystals, and results are given in Table 2. An overview of chemical compositions is displayed in boxplots ( Figure 2). Al ions (avg. 1.714 apfu) are the ideal occupant of the Y site, but show obvious deficits. Mg (avg. 0.138 apfu), Fe (avg. 0.031 apfu), and V (avg. 0.053 apfu) ions are the main substituents for Al, in the order Mg > V > Fe. The sum of Y site substituents (avg. 0.230 apfu) ranges from 6.9% to 18.4% of the total site occupancy.

Major and Minor Elements from EMPA
The chromophore elements (Cr, V and Fe) in emerald are plotted as oxides in Figure 3a,b. Cr and V are generally considered to be the primary chromophores. In most cases, Cr is dominant as the higher Cr2O3 contents than V2O3. However, samples from Dayakou (China), Lened (Canada), Muzo (Colombia), Mohmand (Pakistan), Eidsvoll (Norway), Jos (Nigeria), and Poona (Australia) are the main exceptions. V2O3 content is greater than Cr2O3 content in most or partial samples from those deposits. The Panjshir Valley (Afghanistan) emerald is also suspected of being V-dominant emerald on account of the high V2O3 content. Thus, these eight deposits are plotted together in Figure 3b Figure 4a shows Al versus the sum of its substituents at the Y site of emeralds from Dayakou in comparison with worldwide samples from the literature. An expected negative correlation is observed, due to the common isomorphic substitution in octahedral site. In most cases, the Al occupancy in the octahedra is more than 75%, except for Madagascar deposits. Among V-dominant deposits (Figure 4b), the Muzo, Poona, and Eidsvoll compositions are at the Al-rich end, the Dayakou samples show a wide range of Al occupancy, and the Al occupancy of Mohmand is close to 75%. To compensate for the charge deficit introduced by the substitution of divalent cations (R 2+ = avg. 0.174 apfu) for Al ions at Y site, the structure channel incorporates monovalent alkali cations (A + = avg. 0.165 apfu). Figure 5a displays a nearly 1:1 correlation between R 2+ (including Mg, Fe, Mn, Co, Ni) and A + (including Na, K, Rb, Cs). Compositions of some samples lie below the 1:1 line, denoting that some of the Fe is present as Fe 3+ . Points of partial Dayakou samples lie above the 1:1 line, suggesting that Li + substitutes Be 2+ at T1 sites. In Figure 5b, five colored areas represent the discriminated clusters of V-dominant deposits. Clusters of Jos fall below the 1:1 line obviously, and those of Dayakou, Lened, Muzo, and Panjshir Valley overlap completely. In view of the content of R 2+ and A + , the V-dominant deposits are distributed in the order Eidsvoll < Poona < Dayakou, Muzo, Lened, and Panjshir Valley < Mohmand. The main substituents for Al at Y site are plotted as oxides in Figure 6. Mg is the main substituent in most cases, as the majority of points is close to the MgO corner (Figure 6a). Only samples from Australia (except the Poona occurrence) show a relatively low content of MgO. Clusters of Davdar and Zimbabwe suggest that Cr2O3 content is higher than FeO content. With regard to V-dominant emerald, vanadium is the more significant substituent at the octahedral site.

Trace Elements from LA-ICP-MS Analysis
The bulk continental crust normalized [45] multi-element spider diagram (Figure 7) was constructed to illustrate the variation of trace elements in Dayakou emerald samples and to identify the differences among V-dominant emerald deposits worldwide. Trace elements of Dayakou samples in the diagram include light elements (LE: Li, B), incompatible elements (IE: Cs−Ti), and transition metal chromophoric elements (CTE: Cr and V). Other trace elements results from worldwide V-dominant emeralds are summarized from the previous work reported by [16,20,21,42].  Figure 7. The contents of Li, Cs, Ga, Ti, and V are variable between relatively narrow limits, while those of W, U, Pb, and Cr between a wide range. In contrast to other V-dominant emerald deposits, the Dayakou emerald shows the highest contents of B, Cs, and Ba, and incorporates significant W, Sn, and As. To explore the discrimination further, a series of binary diagrams were constructed (Figures 8-12). In the Rb versus Cs binary diagram (Figure 8), the Dayakou emerald shows an obvious positive linear correlation between the Rb and Cs, which can be expressed by the following equation: The Rb content ranges from 7.97 to 73.11 ppm. The Cs content ranges from 737 to 5034 ppm, which is considerably higher than that in other V-dominant deposits. Cs content of Lened emeralds ranges from 621 to 1397 ppm, whereas Rb content is extremely low and generally below the detection limit. Among the rest of V-dominant deposits where Cs content is less than 500 ppm, Muzo emeralds show the lowest Rb and Cs contents (Rb < 5 ppm, Cs < 50 ppm), while Poona emeralds show the relatively highest contents (Rb > 60 ppm, Cs > 250 ppm). Clusters of Eidsvoll emeralds (20 < Rb < 65 ppm, 30 < Cs < 150 ppm) completely overlap with those of Panjshir emeralds.
The V/Cr ratio is a common indicator of emerald. A plot of V/Cr ratio versus V (Figure 9) content can easily separate Muzo, Poona, Jos, and Panjshir Valley emeralds. These four deposits are plotted at the area with a V/Cr ratio of less than 4 and a V content of less than 5000 ppm. Emeralds from Muzo are clearly distinguished based on the higher content of V and V/Cr ratios (V ~5000 ppm, V/Cr > 2), and those from Poona, Jos, and Panjshir Valley can be separated by V content. The V/Cr ratios of the Eidsvoll emeralds are lower than 25, and those of Lened emeralds are approximately 100. Only some samples from Dayakou show a V/Cr ratio higher than 110.    Table 3. REE studies of emeralds are very limited, as the REE contents of emeralds are always negligible. The Dayakou emerald does not incorporate REEs in large amount, and the concentrations of Eu, Sm, and HREE are always close to or below the detection limit. The total REE (TREE) content of the Dayakou emerald has a relatively wide range (5.03 to 44.84 ppm). The ∑Ce/∑Y ratio ranges from 1.05 to 4.28, and the (La/Yb)N is generally greater than 1, which indicates the relative enrichment of the light rare earth element (LREE). Chondrite-normalized [47] REE patterns (Figure 12) of the Dayakou emerald show overall flat-trending profiles, strong negative Ce anomalies (σ Ce = 0.14 to 0.70), and moderate negative Eu anomalies (σ Eu = 0.33 to 0.72).

Major and Trace Elements of Color Zoned Sample and Tourmaline Inclusions
Sample YE-6 is a representative color-zoned emerald crystal, with a greenish-white core and a medium-green rim. EMP and LA-ICP-MS analyses across the emerald crystal shown in Figure 13a were conducted traversing color zones, and the analyses of eight spots are given in Tables 1 and 2. Needle-like euhedral fuscous inclusions whose cross-sections are convex triangles are only observed in the rim zone and crystal surface of this sample (Figure 13b). These inclusions proved to be tourmalines through and EMPA (Table 1). They show variable orientations but the directions of the c axis of most crystals are consistent with that of the emerald, suggesting that tourmaline mineralization formed simultaneously or earlier compared with the crystallization of rim zone.
Tourmaline has the general chemical formula XY3Z6(T6O18)(BO3)3V3W. Based on EMP analyses of these fuscous tourmalines, T sites in the crystal structure are mainly occupied by Si, octahedral Z sites are dominated by Al, followed by V, Cr, Mg, and the Y sites mainly host Fe (Mg), followed by Mg (Fe), V, and Ti. The X sites are mainly occupied by Na, but also contain Ca, K, and vacancies. The W sites are dominated by O. According to the ions at X, Y and W sites, the fuscous tourmaline is considered to be V-rich intermediate oxy-schorl-dravite [48][49][50]. Compositional variations of trace elements in sample YE-6 are shown in Figure 13c. Most of the elements show stable content in the greenish-white core zone, excluding V. However, from the core to rim, remarkable variations are observed. The contents of Fe (1375 to 2198 ppm), Ga (4 to 14.7 ppm), Rb (9.6 to 27.4 ppm), and Cs (535 to 3108 ppm) increase considerably. The contents of V (970 to 10,077 ppm) show a continuous increase from the center to edge and are constantly higher than the contents of Fe and Cr.

The Origin of Dayakou Emerald Parental Fluid
Emerald deposits are classified as tectonic magmatic-related Type I deposits and tectonic metamorphic-related Type II deposits with several sub-types based on geological environment, host-rock types, and formation conditions. The Dayakou deposit was classified as tectonic magmatic-related emerald deposit hosted in meta-sedimentary rocks (Type IB) [10]. Considering the presence of syntectonic intrusions of emerald-bearing pegmatites and quartz veins and the highest concentration of Cs ever reported for emeralds, with average content of 1754 ppm, and with average Li content of 353 ppm, a magmatic origin is preferred for the parental fluids of Dayakou emeralds [1,4,10,14,16,22,51]. In addition, the enrichment of Cs, W, Sn, and As in emeralds and the presence of scheelite, tourmaline in pegmatite veins indicate a supply of alkali and incompatible elements such as Be, Rb, Cs, W, Sn, As, F, and B from the parental fluids. Nevertheless, pegmatites in the Dayakou deposit do not show a direct relationship with a granitic source. The Laojunshan granite, outcropping approximate 5 km west of the Dayakou occurrence, shows a similar enrichment in W, Sn, and Zn with depletion in Ba, Sr, and Ti [32], and thus is considered to be the source of the Be-bearing parent fluids of pegmatites. However, the formation of Laojunshan granite dated back to 86−118 Ma [31][32][33], which indicates the granites slightly post-date the emerald formation (124 Ma) [24]. Considering the protracted Cretaceous granitoid magmatism in the region [27,29,32,33], Laojunshan granite and an Early Cretaceous granitoid intrusion genetically related to the emerald mineralization are presumably associated with the same magmatic event. The Rb content, an indicator of fractionation and evolution of pegmatite and granite, can support this hypothesis. The significant enrichment of Rb with values between 326 and 502 ppm in Laojunshan granites suggests a highly evolved late-stage fluid after considerable period of fractionation [23,32], whereas the contents of Rb in emerald-bearing pegmatite (218 ppm), emerald-bearing quartz vein (78.7 ppm) [24], and Dayakou emerald (avg. 22.8 ppm) indicate a less evolved early-stage fluid. In conclusion, the evidence mentioned above supports a hypothesis for the genesis of Dayakou emerald that the parental fluid is the Early Cretaceous less evolved granitic fluid associated with Laojunshan granite.

Multi-Stage Crystallization of the Dayakou Emerald
Most of the emerald-bearing pegmatite veins in the Dayakou occurrence show a local zoning with an outer zone enriched in feldspars, an intermediate zone of scheelite-emerald-bearing quartz vein, and an inner zone of emerald-tourmaline-bearing quartz vein [27]. The moderate negative Eu anomalies in REE patterns of the Dayakou emerald are likely the result of the incorporation of Eu into feldspars in the outer zone. And the decreased negative Eu anomalies, which are not as strong as those in Laojunshan granites, might be attributed to the enrichment of ligands such as chlorine and especially fluorine in the fluids which can elevate the partition coefficient of Eu [23].
In the inner zone of pegmatites, tourmalines coexist with emeralds or occur as inclusions which sometimes show the same orientation with the hosted emerald. It indicates that the tourmaline mineralization formed simultaneously or earlier compared with emerald mineralization, which is similar to the Lened emerald [14]. The presence of tourmaline inclusions and the zoned nature of the emerald ( Figure 13) suggest that there may be at least two generations of emerald precipitation at Dayakou [18,52]. First, the greenish-white emerald core is inferred to have formed in the early-stage pegmatitic fluid with a low concentration of Rb, Cs, and Ga and the absence of tourmaline inclusions. V and Cr with low concentrations are released by the circulation and infiltration of fluids. The medium-green emerald rim is supposed to crystallize in a more evolved late-stage pegmatitic-hydrothermal fluid with the presence of tourmaline inclusions and a considerable increase of Rb, Cs, Ga, and replacement degree of Y sites. Crystallization of tourmalines and the increase of incompatible Rb indicate the late stages of pegmatite evolution when incompatible elements and volatiles (e.g., H3BO3) have become sufficiently enriched in the melts to crystallize mineral phases where these elements are essential structural components. The pegmatite melts tend to reach fluid saturation [23].
Nevertheless, an unexpected high V2O3 content (2.92-3.87 wt.%) is noticeable in the oxy-schorl-dravite inclusions, which is consistent with previous study of the Dayakou deposit [53] but different from tourmalines from other deposits. [13] reported Cr2O3 contents in coexisting tourmalines from Regal Ridge emerald deposit in Canada are between 0.01 and 0.20 wt.% with a maximum of 3.28 wt.%. [54] reported a maximum Cr2O3 content of 0.14 wt.% in coexisting tourmalines from the Habachtal emerald deposit in Austria. And [14] reported V2O3 contents in coexisting tourmalines from the Lened emerald deposit in Canada are between 0.09 wt.% and 0.89 wt.%. A possible interpretation for this extremely high V content in tourmalines from the Dayakou deposit is the affection of fluid of metamorphic origin. The oxygen isotope data (δ 18 O = 10.6-12.4‰) for the Dayakou emeralds were consistent with both magmatic and metamorphic origins for the source of the mineralizing fluid, and the fluctuating salinities from fluid inclusion data suggested the mixture of a brine related to pegmatites and a local metamorphic fluid within regional host rocks [24]. It supports that the late-stage pegmatitic-hydrothermal fluids mixed with the local metamorphic fluids.

Mineralization Model for the Dayakou Emerald
An idealized mineralization model for the Dayakou deposit is shown in Figure 14. The deposit is genetically linked to the Early Cretaceous early-stage granite intruding into Neoproterozoic Saxi metamorphic unit. Geochemical evidence suggests that the magmatic-pegmatitic fluids enriched in elements such as Be, Li, Cs, W, Sn, As, B, and F while the granofel of Saxi Unit enriched in V [24,27]. The emerald formation was controlled by the availability of V in the granofels and Be present in the emerald-bearing pegmatite and quartz veins. Incompatible beryllium was transported as Cl and F complex in magmatic-hydrothermal environments [14,15,55]. The circulation and infiltration of magmatic-pegmatitic fluids were channeled by shear zones and structural fractures and released V from V-rich granofels by the metasomatism. The first generation of emerald precipitation occurred in the early-stage pegmatite and quartz veins. The second generation of emerald was precipitated in the more evolved pegmatitic-hydrothermal fluids that might be mixed by the local metamorphic fluids.

Geographic Origin Determination
It is obvious that almost all points of those V-dominant deposits fall into the V-dominant area in Figure 15, but partial points of Poona, Muzo, Panjshir Valley and Jos do not fall into the gray area in Figure 3. This difference can be explained by the low contents and the high precision of data expressed in ppm from LA-ICP-MS or SIMS analysis. Based on Figures 3 and 15, the Dayakou (China), Lened (Canada), Muzo (Colombia), Mohmand (Pakistan), and Eidsvoll (Norway) deposits are defined as completely V-dominant emerald deposits, as the majority of emeralds from those deposits shows the predominance of V among chromophores. The Jos (Nigeria), Panjshir Valley (Afghanistan), and Poona (Australia) deposits are defined as partially V-dominant emerald deposits, as only partial samples show V-dominance. Various binary diagrams generated from the trace element data highlight peculiar differences inside these V-dominant emerald deposits and show great potential in differentiating the provenance. Figure 15. Plot of vanadium (V) versus chromium (Cr) concentrations (expressed in ppm) in emeralds from seven V-dominant deposits. Gray triangular area indicates the V-dominance. Sources of data: [46] (average of five analyses), [16,42] (average of three analyses).
The Dayakou emerald is characterized by significant concentration in W, Sn, and As. V/Cr ratios of some samples are greater than 110. The highest Li, Cs, and LILE contents indicate a magmatic origin of fluid for emerald mineralization which is comparable to the Lened in Canada and Eidsvoll in Norway [14,43]. The Lened emerald shows a similarly high concentration of Cs but depletes in Rb (below the detection limit). However, the Eidsvoll emerald unconventionally are depleted in Li, Cs, and LILE, but are enriched in Al and CTE. Samples from Eidsvoll show a high V/Cr ratio (4 to 25) and a great concentration of V, as the source of vanadium is black shale. The Muzo emerald is characterized by the enrichment of Mg and negligible LILE content, especially for Rb and Cs, which is related to the high salinity brines from evaporates that are poor in Rb and Cs [10]. Those Muzo samples hosted in V-bearing black shale also display a V/Cr ratio of 2 to 4 and high CTE content. The Jos emerald displays a low CTE, Li, and Sc content and low V/Cr ratio, but the concentration of Fe is extremely high, as the Fe-rich metasomatic fluids related to emerald mineralization are produced by the emplacement of younger granites [10,56]. The Panjshir Valley emerald shows a low V/Cr ratio (<2) and high CTE content, indicating the similar content of Cr and V in surrounding metamorphic schists [10]. Furthermore, the enrichment of Sc in the Panjshir Valley emerald is a definitive criterion for identification. The Poona emerald shows a low CTE content and V/Cr ratios, but displays a relatively high concentration of Li, Rb, and Cs, which is likely to suggest that the precipitation of these emerald samples occurred during syn-tectonic pegmatite emplacement [18]. The Mohmand V-dominant emerald is not discussed due to the lack of data.
One of the feasible schemes to discriminate the origin of V-dominant emerald deposits is proposed in Figure 16. The plot of Rb vs. Cs (Figure 8) displays five distinct clusters which allow discriminating four deposits (Dayakou, Lened, Poona, and Muzo). The other indiscriminate cluster includes the Eidsvoll, Jos, and Panjshir Valley deposits. The plot of V vs. V/Cr (Figure 9) helps distinguish the Eidsvoll deposits from the other two. Plots of LILE vs. CTE ( Figure 10) and Li vs. Sc ( Figure 11) can easily differentiate the Jos and Panjshir Valley deposits. Figure 16. The feasible scheme to discriminate the provenance of vanadium-dominant emerald deposits. Rubidium-Rb; caesium-Cs; vanadium-V; chromium-Cr; lithium-Li; scandium-Sc; large ion lithophile element-LILE; chromophoric transition elements-CTE; bdl = below detection limit.

Conclusions
The major and trace element compositions of the Dayakou emerald were investigated using electron microprobe and LA-ICP-MS. Results are compared with the characteristics of emeralds from seven typical V-dominant deposits worldwide. The results demonstrate that: 1) The Dayakou emerald is characterized by the enrichment of Cs, W, Sn, and As, with high ratios of V/Cr. 2) Eight deposits are defined as V-dominant emerald deposits by the content of Cr and V. They can be discriminated by a series of plots such as Rb vs. Cs, V vs. V/Cr, LILE vs. CTE, and Li vs. Sc. 3) The parental fluid of Dayakou emeralds is considered to be an Early Cretaceous early-stage granitic fluid associated with Laojunshan granite. Two generations of emerald precipitation are proposed: (i) Emeralds crystallized in early-stage pegmatitic fluid; (ii) Emeralds crystallized in more evolved late-stage pegmatitic-hydrothermal fluid that might be mixed by the local metamorphic fluids.