LA-ICP-MS U–Pb Dating of Cenozoic Rutile Inclusions in the Yuanjiang Marble-Hosted Ruby Deposit, Ailao Shan Complex, Southwest China

Among the marble-hosted ruby deposits in the Himalayan tectonic belt, which yields the highest-quality rubies in the world, the Yuanjiang deposit is the only economically viable one located in China. More attempts are necessary to put constraints on the ore-forming age of these marble-hosted ruby deposits. Here, we dated rutile inclusions in the Yuanjiang rubies using the LA-ICP-MS U–Pb method, which yielded a lower intercept 206Pb/238U age of 20.2 ± 1.2 Ma on the Tera-Wasserburg plot, close to the 22.5–22.2 Ma 40Ar/39Ar ages of phlogopite from the ruby host matrix assemblage. Our U–Pb rutile age put a constraint on the cooling history of the Yuanjiang rubies deposit. The new rutile age is consistent with our previous model that shows the ca. 28–22 Ma left lateral shearing plays an important role in transporting the ruby deposit toward the surface. This study provides the first example of in-situ U–Pb dating of rutile in the Himalayan tectonic belt, demonstrating the great potential of U–Pb rutile geochronology for Cenozoic mineral deposits.


Introduction
The marble-hosted ruby deposits in Central and Southeast Asia produce rubies of the highest quality [1]. Characterized by deep-red color and high transparency, these ruby deposits were developed during the Cenozoic India-Eurasia continental collision and outcrop in sutures or extrusion shearing zones ( Figure 1A, [1,2]). The Yuanjiang marblehosted ruby deposit in the Ailao Shan Complex is the only economic ruby deposit in China up to now [3,4].
The simple formula Al 2 O 3 of ruby carries limited geochronological information, which is a premise to understand the relationship between ruby mineralization and tectonic evolution. Only limited geochronological tools have been used to date the ore-forming age of the ruby deposit, such as 40 Ar-39 Ar dating of micas from the ruby host matrix assemblage and U-Pb dating of zircon inclusions in rubies. For the Yuanjiang ruby deposit, a previous study obtained the 40 Ar-39 Ar ages of phlogopites and U-Pb age of zircon inclusions, which are 22.5-22.2 Ma and 36.2 Ma, respectively [5]. However, the closure temperature of micas 40 Ar- 39 Ar system (425 • C, [6]) is significantly lower than the marblehosted ruby formation temperature (610-790 • C, [1]), and integrating with the fact that micas can be easily overprinted by later tectonic disturbance often render the interpretations of mica 40 Ar-39 Ar ages ambiguous [7]. The U-Pb system closure temperature of zircon is close to ruby formation, and zircon inclusions are protected by hosted rubies from later interaction of permeating fluids and U-Pb resetting, making zircon inclusions as potential candidates for constraining the crystallization age of gem ruby and sapphire deposits [1,5,[8][9][10][11][12]. However, the potential residual of inherited cores may result in mixed, even meaningless ages, considering that the analytical spots are usually more than 20 µm in LA-ICP-MS or SIMS analyses [13]. Therefore, to better define the ore-forming age of the marble-hosted ruby deposits, geochronological constraints from other minerals syngenetic to rubies are warranted.  [1]). (B) Geological map of the Yuanjiang marble-hosted ruby deposit (modified after [5]). (C) The cross-section profile map of Shaku anticline (Its location corresponds to A-A′ in Figure 1B).
The simple formula Al2O3 of ruby carries limited geochronological information, which is a premise to understand the relationship between ruby mineralization and tectonic evolution. Only limited geochronological tools have been used to date the oreforming age of the ruby deposit, such as 40 Ar-39 Ar dating of micas from the ruby host matrix assemblage and U-Pb dating of zircon inclusions in rubies. For the Yuanjiang ruby deposit, a previous study obtained the 40 Ar-39 Ar ages of phlogopites and U-Pb age of zircon inclusions, which are 22.5-22.2 Ma and 36.2 Ma, respectively [5]. However, the closure temperature of micas 40 Ar-39 Ar system (425 °C, [6]) is significantly lower than the marble-hosted ruby formation temperature (610-790 °C, [1]), and integrating with the fact that micas can be easily overprinted by later tectonic disturbance often render the interpretations of mica 40 Ar-39 Ar ages ambiguous [7]. The U-Pb system closure temperature of zircon is close to ruby formation, and zircon inclusions are protected by hosted rubies from later interaction of permeating fluids and U-Pb resetting, making zircon inclusions as potential candidates for constraining the crystallization age of gem ruby and sapphire deposits [1,5,[8][9][10][11][12]. However, the potential residual of inherited cores may result in mixed, even meaningless ages, considering that the analytical spots are usually more than 20 μm in LA-ICP-MS or SIMS analyses [13]. Therefore, to better define the ore-forming age of the marble-hosted ruby deposits, geochronological constraints from other minerals syngenetic to rubies are warranted.
Rutile, another common mineral inclusion in the Central and Southeast Asia marblehosted rubies [1], is potentially suitable candidate for U-Pb dating if sufficient U was incorporated [14]. In contrast to the ambiguous interpretation of U-Pb age of zircon (e.g., inherited, metamorphic or even mixed ages), the U-Pb age of rutile records the latest cooling age to the Pb diffusion closure temperature of rutile as the result of faster Pb diffusion rate [13][14][15][16]. Therefore, dating the rutile inclusions has the potential to provide new constraint on the ore-forming age of ruby deposits that can be compared with geochronology produced by other methods. LA-ICP-MS U-Pb dating of rutile inclusions in corundum (ruby and sapphire) was firstly reported by [13], defining a minimum age of 499-533 Ma for corundum growth in the Mozambique belt.  [1]). (B) Geological map of the Yuanjiang marble-hosted ruby deposit (modified after [5]). (C) The cross-section profile map of Shaku anticline (Its location corresponds to A-A in Figure 1B).
Rutile, another common mineral inclusion in the Central and Southeast Asia marblehosted rubies [1], is potentially suitable candidate for U-Pb dating if sufficient U was incorporated [14]. In contrast to the ambiguous interpretation of U-Pb age of zircon (e.g., inherited, metamorphic or even mixed ages), the U-Pb age of rutile records the latest cooling age to the Pb diffusion closure temperature of rutile as the result of faster Pb diffusion rate [13][14][15][16]. Therefore, dating the rutile inclusions has the potential to provide new constraint on the ore-forming age of ruby deposits that can be compared with geochronology produced by other methods. LA-ICP-MS U-Pb dating of rutile inclusions in corundum (ruby and sapphire) was firstly reported by [13], defining a minimum age of 499-533 Ma for corundum growth in the Mozambique belt.
Here, we present compositional and chronological data of rutile inclusions in rubies from the Yuanjiang marble-hosted ruby deposit, which set an example of dating Cenozoic rutile. Coupled with the previously obtained U-Pb zircon inclusion age and 40 Ar/ 39 Ar phlogopite ages, the genetic link between ruby mineralization and tectonic evolution in the Ailao Shan-Red River shear zone was further discussed in the Section 5.3.

Geological Setting
As one of the marble-hosted ruby deposits in Central and Southeast Asia, the Yuanjiang ruby deposit is located in the middle of the Ailao Shan Complex, the longest part of Ailao Shan-Red River shear zone (ASRR, Figure 1A, [1]). The Yuanjiang ruby deposit outcrops in the NE limb of Shaku anticline and is trapped between the Tangfang Fault and the Red River Fault ( Figure 1B,C). In general, the mineral assemblage of the Yuanjiang ruby deposits is very similar to that of other ruby deposits in Central and Southeast Asia [1,5]. The detailed regional and ore geological setting has been described in our previous paper [5].

Materials and Methods
Ruby-bearing marble samples are collected from three layers of orebody at the Shaku village (102 • 05 31.41 E, 23 • 27 52.03 N). Rubies are found as spotted crystals in coarsegrained marble or associated with phlogopite, graphite and pyrite in foliations [5]. Rutile inclusions are identified in the hosted ruby crystals under microscope transmitted light by the characters of high relief and dark-brown color ( Figure 2A).

Geological Setting
As one of the marble-hosted ruby deposits in Central and Southeast Asia, the Yuanjiang ruby deposit is located in the middle of the Ailao Shan Complex, the longest part of Ailao Shan-Red River shear zone (ASRR, Figure 1A, [1]). The Yuanjiang ruby deposit outcrops in the NE limb of Shaku anticline and is trapped between the Tangfang Fault and the Red River Fault ( Figure 1B,C). In general, the mineral assemblage of the Yuanjiang ruby deposits is very similar to that of other ruby deposits in Central and Southeast Asia [1,5]. The detailed regional and ore geological setting has been described in our previous paper [5].

Materials and Methods
Ruby-bearing marble samples are collected from three layers of orebody at the Shaku village (102°05′31.41″ E, 23°27′52.03″ N). Rubies are found as spotted crystals in coarsegrained marble or associated with phlogopite, graphite and pyrite in foliations [5]. Rutile inclusions are identified in the hosted ruby crystals under microscope transmitted light by the characters of high relief and dark-brown color ( Figure 2A). Raman spectroscopy of rutile inclusions was conducted by using Renishaw inVia confocal micro-Raman spectrometer at Instrumental Analysis and Research Center, Sun Yat-Sen University (SYSU). Ar + laser with 514.5 nm excitation was used and the Raman signals were collected over 50 to 2000 cm −1 .
Rutile inclusions were carefully polished until it totally exposed on the surface of hosted ruby crystals for LA-ICP-MS analyses. Back-scattered electron images (BSE) of rutile inclusions were conducted by ∑SIGMA scanning electron microscope (SEM) to check its inner structure at the School of Earth Sciences and Engineering, SYSU. The trace Raman spectroscopy of rutile inclusions was conducted by using Renishaw inVia confocal micro-Raman spectrometer at Instrumental Analysis and Research Center, Sun Yat-Sen University (SYSU). Ar + laser with 514.5 nm excitation was used and the Raman signals were collected over 50 to 2000 cm −1 .
Rutile inclusions were carefully polished until it totally exposed on the surface of hosted ruby crystals for LA-ICP-MS analyses. Back-scattered electron images (BSE) of rutile inclusions were conducted by ∑SIGMA scanning electron microscope (SEM) to check its inner structure at the School of Earth Sciences and Engineering, SYSU. The trace elements and U-Pb isotopic composition analyses of rutile inclusions were conducted at the Key Laboratory of Marine Resources and Coastal Engineering, SYSU. 22 dated spots were conducted on 8 rutile inclusions from 8 selected ruby crystals. Calibration rutile reference material R10 (~30 ppm U, 1090 ± 5 Ma, [17]) was served as matrix-matched internal standard to calibrate the U-Pb isotope fractionation of rutile inclusions [13], and the quality control material R19 yielded a concordant weighted average age of 493 ± 8 Ma (2σ, MSWD = 0.2) during analyses, which is in line with its thermal infrared mass spectrometry (TIMS) 206 Pb/ 238 U age within the analytical uncertainties (~15 ppm U, 489.5 ± 0.9 Ma, [16]). During U-Pb dating, trace elements of rutile inclusions were simultaneously determined by monitoring 24 Mg, 27 Al, 29 49 Ti was taken as an internal standard for unknown rutile after normalized TiO 2 to 100% [13]. The averaged results of quality control material BCR-2G, rutile R10, and R19 are in concordance within 5% reported values [16][17][18].
An ArF excimer laser ablation system with Ar-He stream transportation system attached to an Agilent 7700× ICP-MS was used to acquired ion-signal intensities. A 32 µm spot is generally used with an energy density of 5 J/cm 2 and a repetition rate of 5 Hz. Each analysis consists of 20 s background acquisition followed by 45 s sample analysis and 25 s washout. Off-line selection, integration of background and analyzed signals, time-drift correction, quantitative calibration for trace elemental analyses and U-Pb isotopic results were conducted by ICPMSDataCal 10.2 software [19]. Concordia diagrams were performed using Isoplot3 software [20].
An ArF excimer laser ablation system with Ar-He stream transportation system attached to an Agilent 7700× ICP-MS was used to acquired ion-signal intensities. A 32 μm spot is generally used with an energy density of 5 J/cm 2 and a repetition rate of 5 Hz. Each analysis consists of 20 s background acquisition followed by 45 s sample analysis and 25 s washout. Off-line selection, integration of background and analyzed signals, time-drift correction, quantitative calibration for trace elemental analyses and U-Pb isotopic results were conducted by ICPMSDataCal 10.2 software [19]. Concordia diagrams were performed using Isoplot3 software [20].

Trace Elemental and U-Pb Dating Results of Rutile Inclusions
Rutile inclusions in Yuanjiang ruby crystals are anhedral and oval shapes (80-200 µm in length and 50-100 µm in width), showing homogeneous structure without visible zoning and inherited cores in BSE images ( Figure 2B-D).
LA-ICP-MS trace elemental contents data for rutile inclusions are listed in Table 1

The Relationship between Rutile and Ruby
Rutile inclusions in Yuanjiang rubies are intact and maintain distance from fractures, meeting the criteria of syngenetic inclusions [5,23]. During LA-ICP-MS analyses, the occurrence of abnormal Al signal peak in the LA-ICP-MS time-resolved spectra (e.g., Ru-2, Ru-18; Table 1; Figure 5) with the unusually higher Al contents of rutile inclusions (Table 1) indicate that rubies (Al2O3) also occasionally occur as inclusions in the rutile inclusions though not directly proved by petrographic observation, this mutual encapsulated relationship indicates that the rubies and rutiles syngenetically grew under the same conditions.

The Relationship between Rutile and Ruby
Rutile inclusions in Yuanjiang rubies are intact and maintain distance from fractures, meeting the criteria of syngenetic inclusions [5,23]. During LA-ICP-MS analyses, the occurrence of abnormal Al signal peak in the LA-ICP-MS time-resolved spectra (e.g., Ru-2, Ru-18; Table 1; Figure 5) with the unusually higher Al contents of rutile inclusions (Table 1) indicate that rubies (Al 2 O 3 ) also occasionally occur as inclusions in the rutile inclusions though not directly proved by petrographic observation, this mutual encapsulated relationship indicates that the rubies and rutiles syngenetically grew under the same conditions. Geochemically, these rutile inclusions are distinguished by their high V, Nb, Cr, Zr content (with the average of 4682 ppm, 3454 ppm, 3136 ppm, 2249 ppm, respectively) and low Fe (with the average of 73 ppm, Table 1) content compared with metamafic and metapelitic rutiles. These characters distinct our rutile inclusions from the more commonly reported metamafic and metapelitic rutiles ( Figure 6; [14,[24][25][26][27]). This difference can be attributed to that our rutile-rubies were crystalized from exceptional V-rich, Cr-rich and Fe-poor ruby-bearing carbonate protolith [28]. Moreover, the V, Cu, W, Sn, Sb contents (with the average of 12, 151, 50 and 1 ppm, respectively) of rutile inclusions are significantly lower than those of hydrothermal rutile [22,29], which suggest that rutile inclusions were protected by their host ruby from the physical interference of later hydrothermal fluids. Thus, the syngenetic, high U, undisturbed rutile inclusions have great potentials to provide more constraints on the timing of ruby growth in the Yuanjiang area [13].

The Relationship between Rutile and Ruby
Rutile inclusions in Yuanjiang rubies are intact and maintain distance from fractures, meeting the criteria of syngenetic inclusions [5,23]. During LA-ICP-MS analyses, the occurrence of abnormal Al signal peak in the LA-ICP-MS time-resolved spectra (e.g., Ru-2, Ru-18; Table 1; Figure 5) with the unusually higher Al contents of rutile inclusions (Table 1) indicate that rubies (Al2O3) also occasionally occur as inclusions in the rutile inclusions though not directly proved by petrographic observation, this mutual encapsulated relationship indicates that the rubies and rutiles syngenetically grew under the same conditions. Geochemically, these rutile inclusions are distinguished by their high V, Nb, Cr, Zr content (with the average of 4682 ppm, 3454ppm, 3136 ppm, 2249 ppm, respectively) and low Fe (with the average of 73 ppm, Table 1) content compared with metamafic and metapelitic rutiles. These characters distinct our rutile inclusions from the more commonly reported metamafic and metapelitic rutiles ( Figure 6; [14,[24][25][26][27]). This difference can be attributed to that our rutile-rubies were crystalized from exceptional Vrich, Cr-rich and Fe-poor ruby-bearing carbonate protolith [28]. Moreover, the V, Cu, W, Sn, Sb contents (with the average of 12, 151, 50 and 1 ppm, respectively) of rutile inclusions are significantly lower than those of hydrothermal rutile [22,29], which suggest that rutile inclusions were protected by their host ruby from the physical interference of later hydrothermal fluids. Thus, the syngenetic, high U, undisturbed rutile inclusions have great potentials to provide more constraints on the timing of ruby growth in the Yuanjiang area [13].  [25,26]. Log (Cr/Nb) = 0 according to [27].

Interpretation of Rutile Inclusion U-Pb Dating Results
Our results reveal that rutile inclusions contain high U (with the average of 62 ppm, Table 2). Given the most of dating data are far away from the Y-axis ( 207 Pb/ 206 Pb), the lower intercept of 20.2 ± 1.2 Ma on the discordia can be taken as reliable age of rutile inclusions (Figure 4, [30]).   [25,26]. Log (Cr/Nb) = 0 according to [27].

Interpretation of Rutile Inclusion U-Pb Dating Results
Our results reveal that rutile inclusions contain high U (with the average of 62 ppm, Table 2). Given the most of dating data are far away from the Y-axis ( 207 Pb/ 206 Pb), the lower intercept of 20.2 ± 1.2 Ma on the discordia can be taken as reliable age of rutile inclusions (Figure 4, [30]). The marked gap between the U-Pb age of rutile (20.2 ± 1.2 Ma) and zircon inclusions (36.2 ± 1.1 Ma, [5]) suggests that the 'armoring effect' of rutile inclusions by the hosted rubies has been ineffective, consistent with the reported ineffective garnet 'shielding' on rutile inclusions under multi-thermal conditions in North Dabie eclogite [31]. Considering that the Pb diffusion rate of rutile is significantly faster than that of zircon, the new obtained 206 Pb/ 238 U 20.2 ± 1.2 Ma age of rutile inclusions, which falls close to the range of 40 Ar-39 Ar ages of phlogopite at 22.5-22.2 Ma [5], is interpreted as the cooling age of the Yuanjiang ruby deposit.

The Spatial-Temporal Relationship of the Ruby Mineralization and Tectonic Evolution
The widespread distribution of multi-stage foliation, mylonitic rocks, migmatites, leucocratic dikes indicate the complicated Oligocene-Miocene tectonic evolution of the Ailao Shan-Red River shear zone in response to progressive India-Eurasia collision. The U-Pb dating of monazite and zircon inclusions in the pre-shearing corundum as well as zircons from the syn-kinematic leucocratic dikes indicate that the initiation of the largescale left lateral shearing is around~28 Ma [32,33]. Zircon inclusions in the rubies were dated at 36.2 and 38.1 Ma for the Yuanjiang deposit [5] and the Luc-Yen deposit [7,34], respectively, suggesting that the trigger of the marble-hosted ruby mineralization was prior to the left lateral shearing and coeval with the local crustal thickening related continental subduction-collision [5]. The obtained 206 Pb/ 238 U age of rutile inclusions in Yuanjiang rubies, 20.2 ± 1.2 Ma, is approximate to the range of phlogopite 40 Ar-39 Ar ages of Yuanjiang deposit (22.5-22.2 Ma) [5] and Yen Bai deposit (24.4-23.2 Ma) [7], collectively reflecting that the left lateral shearing plays an important role in transporting ruby deposit toward the surface [5].

Conclusions
Rutile inclusions in Yuanjiang rubies possess the characteristic of high V, Nb, Cr, Zr and low Fe content, which clearly discriminated our rubies-hosted rutile inclusions from the metamafic, metapelitic and hydrothermal rutile. In-situ LA-ICP-MS U-Pb dating of rutile inclusions in the Yuanjiang rubies, which yielded a reliable 206 Pb/ 238 U age of 20.2 ± 1.2 Ma, is approximate to the 40 Ar-39 Ar ages of phlogopite, helps to constrain the cooling age of Yuanjiang ruby deposit. Our study provides the first example of LA-ICP-MS U-Pb dating of rutile in the Himalayan tectonic belt, suggesting that U-Pb dating of rutile is a geochronological tool of great potential for young Cenozoic deposits.
Author Contributions: T.Y. and X.S. conceived and designed the experiments; T.Y., Y.L., G.S. and T.Y. did sample preparation and analyzed the data; T.Y. wrote the manuscript; X.S. substantially revised the original manuscript. All authors have read and agreed to the published version of the manuscript.