Metamorphic rocks are the records of plate tectonic processes whose reconstruction relies on correct estimates of the pressures and temperatures (P-T) experienced by these rocks through time. Unlike chemical geothermobarometry, elastic geobarometry does not rely on chemical equilibrium between minerals, so it has the potential to provide information on overstepping of reaction boundaries and to identify other examples of non-equilibrium behavior in rocks. Here we introduce a method that exploits the anisotropy in elastic properties of minerals to determine the unique P and T of entrapment from a single inclusion in a mineral host. We apply it to preserved quartz inclusions in garnet from eclogite xenoliths hosted in Yakutian kimberlites (Russia). Our results demonstrate that quartz trapped in garnet can be preserved when the rock reaches the stability field of coesite (the high-pressure and hightemperature polymorph of quartz) at 3 GPa and 850 °C. This supports a metamorphic origin for these xenoliths and sheds light on the mechanisms of craton accretion from a subducted crustal protolith. Furthermore, we show that interpreting P and T conditions reached by a rock from the simple phase identification of key inclusion minerals can be misleading. INTRODUCTION The mechanisms attending the downward transport of crustal material into the mantle and its return back to Earth’s surface (exhumation) are still a matter of vigorous debate. Chemical information only allows the interpretation of the measurements on mineral and rock composition in terms of pressure (P) for perfectly lithostatic systems under ideal chemical equilibrium. However, significant overstepping of reaction boundaries (Spear and Pattison, 2017) as well as the presence of non-lithostatic stresses might prevent the correct interpretation of P, and in turn depths, reached by crustal rocks during subduction and metamorphism. Host-inclusion geobarometry provides an alternative and complementary method to determine pressures and temperatures (P-T) attained during the history of rocks (Zhang, 1998; Angel et al., 2014b, 2015). A mineral trapped as an inclusion within another host mineral is not free to expand or contract as would a free crystal but is constrained by the host mineral. This results in the development of stress in the inclusion that differs from the external stress or pressure applied to the host mineral, both while it is in the earth and afterwards when we examine the rock at room pressure in the laboratory. The stress state of the inclusion arises from the change in P and T from the time of its entrapment, so measurement of the stress state of the still-entrapped inclusion while in the laboratory enables the conditions of entrapment to be calculated, provided no plastic or brittle deformation occurred upon exhumation after entrapment. However, the current state of the art is based upon the assumption that both the host and the inclusion are elastically isotropic. But, no mineral is isotropic in elastic properties and this may cause errors in calculated P and T. This also means that a measurement of a single inclusion pressure while the host is at room pressure provides only one constraint on the entrapment conditions. As a consequence, one can only calculate a line in P-T space (the entrapment isomeke) which represents possible entrapment conditions of the inclusion (Rosenfeld and Chase, 1961; Angel et al., 2014b). Here we describe how the anisotropy of mineral inclusions can be exploited to determine unique P-T conditions last recorded by the rock. The basic idea behind this approach is that an anisotropic inclusion will exhibit different stresses and strain along different crystallographic directions. By measuring these strains from a single inclusion, we obtain two or three independent data which, in combination with the known P-T variation of the unit-cell parameters of the inclusion mineral and the host, enable both the P and T of entrapment or elastic equilibration of the inclusion to be determined. ECLOGITE XENOLITH FROM THE MIR PIPE (YAKUTIA) The Mir pipe (Yakutian kimberlites, Russia) is a relatively young kimberlite (360 Ma) for which eruption temperatures of ∼1000 °C and very fast ascent rates or short residence time (<0.1 m.y.) have been estimated (Korsakov et al., 2009; Zhukov and Korsakov, 2015). The Mir pipe kimberlite carried to the surface eclogite xenoliths made of omphacite, garnet, and rutile (Taylor et al., 2003; Tomilenko et al., 2005; Zhukov and Korsakov, 2015). The major and trace element bulk composition of these eclogite xenoliths suggests that they may be derived from subducted oceanic crust (Shimizu and Sobolev, 1995; Taylor et al., 2003). The studied xenolith contains coarse homogeneous garnet hosting relatively large primary quartz *E-mail: matteo.alvaro@unipv.it Published online XX Month 2019 Downloaded from https://pubs.geoscienceworld.org/gsa/geology/article-pdf/4858978/g46617.pdf by guest on 31 October 2019 2 www.gsapubs.org | Volume XX | Number XX | GEOLOGY | Geological Society of America inclusions (Fig. 1). Previous models show that the short eruption time prevents any significant resetting of inclusion pressures by plastic flow of the garnet during upward transport from the mantle (Zhong et al., 2018). This is confirmed by the estimated residual pressures of 1.0–1.2 GPa for the quartz inclusions, which are significantly higher than 0.1–0.6 GPa reported on other quartz inclusions in garnets from coesite-grade and diamond-grade ultrahigh-pressure (UHP) rocks (see Korsakov et al., 2009). The relatively large sizes and high pressures of the inclusions and the homogeneous composition of the garnet host, coupled with the fast kimberlite ascent which may have been insufficient to reset the rock-forming minerals, make these eclogites an ideal case to test and prove the potential of anisotropic elastic geobarometry. DETERMINATION OF INCLUSION STRAINS The birefringence haloes around the four selected inclusions (see Fig. 1) indicate the presence of significant residual stresses in the inclusions (Korsakov et al., 2007, 2009; Howell et al., 2010; Campomenosi et al., 2018;). The upshift of the Raman bands (Fig. 2A; for further details, see the GSA Data Repository1) at various positions across the inclusion confirms that the inclusion is under significant residual strain. Quantitative values of the strains have been determined from the wavenumber shifts of the Raman bands at 128, 206, 464, and 696 cm−1 by using the mode Grüneisen tensors of quartz (Murri et al., 2018, 2019; Angel et al., 2019; Bonazzi et al., 2019). The strains determined in this way from several inclusions in the same garnet are identical within estimated standard deviations, and agree with the measurements by single-crystal X-ray diffraction (Table 1). The residual strains as determined by X-ray diffraction and by micro-Raman spectroscopy (MRS) cannot be directly used to back-calculate the residual pressure at entrapment conditions because they are the product of two processes: the contrast in the elastic properties of the host and inclusion over P and T that leads to the inclusions exhibiting an excess pressure, and the mutual elastic relaxation of the system driven by this excess pressure (Angel et al., 2014b). Before calculating entrapment conditions by using the equations of state of the host and inclusion minerals, the residual strains must be corrected for elastic relaxation (Angel et al., 2017b). Correction for elastic relaxation of a spherical inclusion in an elastically isotropic system depends only on the elastic properties of the host and inclusion (Zhang, 1998; Angel et al., 2017b). For faceted or complex-shaped inclusions, and for all elastically anisotropic inclusions, the final stress state depends on both the geometry of the system and the elastic anisotropy of the inclusion-host pair (Eshelby, 1957; Zhukov and Korsakov, 2015; Campomenosi et al., 2018; Mazzucchelli et al., 2018, 2019). Therefore, for all of these cases, the elastic relaxation must be calculated numerically (Mazzucchelli et al., 2018, 2019) using the exact shape of the inclusion and its full elastic properties together with those for the host. We determined the three-dimensional (3D) model of the entire sample (see Fig. 1E) from X-ray microtomography measurements at the TOMCAT (Tomographic Microscopy and Coherent Radiology Experiments) beamline of the Swiss Light Source (SLS; Paul Scherrer Institut) (Stampanoni et al., 2006). From these images, we created a meshed 3-D model in order to perform finite element (FE) analyses that allow us to calculate the amount of elastic relaxation that has occurred (Mazzucchelli et al., 2018). Because of the anisotropy of quartz, the amount of elastic relaxation is different in different directions. Only after correction for the elastic relaxation can we demonstrate that the measured quartz inclusions were subject to isotropic strain (Table 1; Fig. 2B) as a consequence of the change of the dimensions of the cubic host mineral with P and T. The corrected strains of all four inclusions are identical within the estimated uncertainties, consistent with them having been trapped under the same conditions and having experienced the same post-entrapment history. 1GSA Data Repository item 2020006, theoretical backgound for the calculations, is available online at http://www.geosociety.org/datarepository/2020/, or on request from editing@geosociety.org. A B

中文翻译：

---

石英从柯石英稳定性场记录的化石俯冲

变质岩是板块构造过程的记录，其重建依赖于对这些岩石随时间经历的压力和温度 (PT) 的正确估计。与化学地球测温法不同，弹性地球测压法不依赖于矿物之间的化学平衡，因此它有可能提供有关越过反应边界的信息并识别岩石中非平衡行为的其他例子。在这里，我们介绍了一种方法，该方法利用矿物弹性特性的各向异性来确定来自矿物主体中单个包裹体的独特 P 和 T 包封。我们将其应用于来自雅库特金伯利岩（俄罗斯）的榴辉岩包体的石榴石中保存的石英包裹体。我们的研究结果表明，当岩石在 3 GPa 和 850 °C 下到达柯石英（石英的高压高温多晶型物）的稳定场时，可以保存困在石榴石中的石英。这支持了这些捕虏体的变质起源，并揭示了从俯冲地壳原岩吸积克拉通的机制。此外，我们表明，从关键包裹体矿物的简单相识别来解释岩石达到的 P 和 T 条件可能会产生误导。介绍 地壳物质向下输送到地幔并返回地球表面（发掘）的机制仍然是一个激烈争论的问题。对于理想化学平衡下的完美岩石静力学系统，化学信息仅允许根据压力 (P) 解释矿物和岩石成分的测量结果。然而，显着超出反应边界（Spear 和 Pattison，2017 年）以及非岩石静力应力的存在可能会阻止对 P 的正确解释，进而阻止地壳岩石在俯冲和变质作用期间达到的深度。宿主包裹体地理气压测量法提供了一种替代和补充方法来确定岩石历史过程中获得的压力和温度 (PT)（Zhang，1998 年；Angel 等人，2014b，2015 年）。作为夹杂物被困在另一主体矿物中的矿物不像自由晶体那样自由膨胀或收缩，而是受到主体矿物的约束。这导致包裹体中应力的发展与施加到宿主矿物上的外部应力或压力不同，无论是在地球中还是之后，当我们在实验室中在室温下检查岩石时。夹杂物的应力状态由夹杂时 P 和 T 的变化引起，因此在实验室中测量仍然夹杂的夹杂物的应力状态可以计算夹杂条件，前提是没有塑性或截留后挖掘时发生脆性变形。然而，当前的技术状态是基于主体和内含物都是弹性各向同性的假设。但是，没有矿物的弹性特性是各向同性的，这可能会导致计算 P 和 T 的错误。这也意味着当宿主处于室温时测量单个夹杂物压力仅提供对截留条件的一个约束。因此，人们只能计算 PT 空间中的一条线（截留 isomeke），它代表包裹体的可能截留条件（Rosenfeld 和 Chase，1961 年；Angel 等人，2014b）。在这里，我们描述了如何利用矿物包裹体的各向异性来确定岩石最后记录的独特 PT 条件。这种方法背后的基本思想是各向异性夹杂物将沿不同的结晶方向表现出不同的应力和应变。通过测量单个夹杂物的这些应变，我们获得了两个或三个独立的数据，其中，结合包裹体矿物和主体的晶胞参数的已知 PT 变化，能够确定包裹体的 P 和 T 或弹性平衡。来自 MIR 管（雅库特）的榴辉岩 Xenolith Mir 管（雅库特金伯利岩，俄罗斯）是一种相对年轻的金伯利岩（360 Ma），其喷发温度约为 1000 °C，上升速度非常快或停留时间很短（<0.1 my） (Korsakov et al., 2009; Zhukov and Korsakov, 2015)。Mir 管状金伯利岩携带由绿辉石、石榴石和金红石制成的榴辉岩捕虏体（Taylor 等人，2003 年；Tomilenko 等人，2005 年；Zhukov 和 Korsakov，2015 年）。这些榴辉岩捕虏体的主要和微量元素整体成分表明它们可能来自俯冲的洋壳（Shimizu 和 Sobolev，1995年；泰勒等人，2003 年）。所研究的捕虏体含有粗粒均质石榴石，其中含有相对较大的原生石英 *电子邮件：matteo.alvaro@unipv.it 2019 年第 XX 月在线发布 下载自 https://pubs.geoscienceworld.org/gsa/geology/article-pdf/4858978 /g46617.pdf 来宾于 2019 年 10 月 31 日 2 www.gsapubs.org | 卷XX | 编号 XX | 地质 | 美国地质学会包裹体（图 1）。先前的模型表明，在从地幔向上运输过程中，石榴石的塑性流动不会导致任何包裹体压力的显着重置（Zhong 等，2018）。石英包裹体的估计残余压力为 1.0-1.2 GPa，明显高于 0.1-0，证实了这一点。6 GPa 报告了来自柯石英级和金刚石级超高压 (UHP) 岩石的石榴石中的其他石英包裹体（参见 Korsakov 等，2009）。包裹体的较大尺寸和高压以及石榴石主体的均质成分，再加上快速的金伯利岩上升可能不足以重置造岩矿物，使这些榴辉岩成为测试和证明其潜力的理想案例各向异性弹性地球气压测量法。夹杂物应变的确定 四个选定夹杂物周围的双折射光晕（见图 1）表明夹杂物中存在显着的残余应力（Korsakov 等人，2007 年、2009 年；Howell 等人，2010 年；Cammenosi 等人， 2018 年；)。拉曼谱带的上移（图 2A；更多细节，参见 GSA 数据存储库 1) 在夹杂物的不同位置确认夹杂物处于显着的残余应变下。应变的定量值已通过使用石英的 Grüneisen 张量模式从 128、206、464 和 696 cm-1 的拉曼谱带的波数位移确定（Murri 等人，2018 年，2019 年；Angel 等人，2019 年）。 , 2019; Bonazzi 等人, 2019)。以这种方式从同一石榴石中的几个内含物确定的应变在估计的标准偏差内是相同的，并且与单晶 X 射线衍射的测量结果一致（表 1）。通过 X 射线衍射和微拉曼光谱 (MRS) 确定的残余应变不能直接用于反计算截留条件下的残余压力，因为它们是两个过程的产物：主体和夹杂物在 P 和 T 上的弹性特性的对比导致夹杂物表现出超压，以及由这种超压驱动的系统的相互弹性松弛（Angel 等，2014b）。在使用主矿物和包裹体矿物的状态方程计算包封条件之前，必须对残余应变进行弹性松弛校正（Angel 等，2017b）。弹性各向同性系统中球形包裹体弹性松弛的校正仅取决于主体和包裹体的弹性特性（Zhang，1998；Angel 等，2017b）。对于多面或复杂形状的包裹体，以及所有弹性各向异性包裹体，最终应力状态取决于系统的几何形状和包裹体-主体对的弹性各向异性（Eshelby，1957；朱可夫和科萨科夫，2015 年；Campomenosi 等人，2018 年；Mazzucchelli 等人，2018 年、2019 年）。因此，对于所有这些情况，必须使用夹杂物的确切形状及其完全弹性特性以及主体的特性，以数值方式计算弹性松弛（Mazzucchelli 等人，2018 年、2019 年）。我们通过瑞士光源 (SLS; Paul Scherrer Institut) 的 TOMCAT（断层显微和相干放射学实验）光束线的 X 射线显微断层扫描测量确定了整个样品的三维（3D）模型（见图 1E） （Stampanoni 等人，2006 年）。从这些图像中，我们创建了一个网格化 3-D 模型，以执行有限元 (FE) 分析，从而使我们能够计算发生的弹性松弛量（Mazzucchelli 等人，2018 年）。由于石英的各向异性，不同方向的弹性松弛量不同。只有在对弹性松弛进行校正后，我们才能证明测量的石英包裹体受到各向同性应变（表 1；图 2B），这是由于立方主体矿物尺寸随 P 和 T 的变化而导致的。校正后的应变在估计的不确定性范围内，所有四个包裹体中的 1 个是相同的，这与它们在相同条件下被捕获并经历了相同的捕获后历史一致。1GSA 数据存储库项目 2020006，计算的理论背景，可在 http://www.geosociety.org/datarepository/2020/ 在线获取，或应editing@geosociety.org 的要求提供。AB 只有在对弹性松弛进行校正后，我们才能证明测量的石英包裹体受到各向同性应变（表 1；图 2B），这是由于立方主体矿物尺寸随 P 和 T 的变化而导致的。校正后的应变在估计的不确定性范围内，所有四个包裹体中的 1 个是相同的，这与它们在相同条件下被捕获并经历了相同的捕获后历史一致。1GSA 数据存储库项目 2020006，计算的理论背景，可在 http://www.geosociety.org/datarepository/2020/ 在线获取，或应editing@geosociety.org 的要求提供。AB 只有在对弹性松弛进行校正后，我们才能证明测量的石英包裹体受到各向同性应变（表 1；图 2B），这是由于立方主体矿物尺寸随 P 和 T 的变化而导致的。校正后的应变在估计的不确定性范围内，所有四个包裹体中的 1 个是相同的，这与它们在相同条件下被捕获并经历了相同的捕获后历史一致。1GSA 数据存储库项目 2020006，计算的理论背景，可在 http://www.geosociety.org/datarepository/2020/ 在线获取，或通过editing@geosociety.org 请求。AB 所有四种包裹体的校正应变在估计的不确定性内是相同的，与它们在相同条件下被捕获并经历了相同的捕获后历史一致。1GSA 数据存储库项目 2020006，计算的理论背景，可在 http://www.geosociety.org/datarepository/2020/ 在线获取，或应editing@geosociety.org 的要求提供。AB 所有四种包裹体的校正应变在估计的不确定性内是相同的，与它们在相同条件下被捕获并经历了相同的捕获后历史一致。1GSA 数据存储库项目 2020006，计算的理论背景，可在 http://www.geosociety.org/datarepository/2020/ 在线获取，或应editing@geosociety.org 的要求提供。AB