Professional Documents
Culture Documents
Inclusions Are Grain of Truth
Inclusions Are Grain of Truth
9, 1671–1700
doi: 10.1093/petrology/egy075
ABSTRACT
Inclusions in minerals, whether fluids, melts or crystalline phases, are small pieces of the large-
scale puzzle of Nature, time-consuming to investigate and often of difficult interpretation. Yet they
are windows into the past of their host mineral. Mineral inclusions provide the opportunity to un-
ravel the genesis of their host, and the increasingly refined understanding of their elastic behaviour
provides the basis for alternative, equilibrium-independent geobarometry. Fluid and melt inclu-
sions reveal information about material transfer in the Earth system, from shallow mineralization
to mantle re-fertilization via subduction. The study of inclusions is thus one of the most intriguing
and fertile branches of micropetrology. In this contribution, we focus on two recent developments:
the use of elasticity models to extract the formation conditions of the host crystal, and the discov-
ery and investigation of melt inclusions in metamorphic rocks. We also discuss how to evaluate the
information provided by inclusions, given that they are no longer at the pressure and temperature
conditions of entrapment. We discuss how to understand and quantify the changes undergone dur-
ing cooling and depressurization, and how metastability-related phenomena in inclusions, such as
crystallization of rare polymorphs and preservation of the original content of volatiles in fluid and
melt inclusions, provide direct evidence that inclusions represent closed systems. The field of
study of inclusions in minerals still has a largely untapped potential. The most fruitful avenues for
future research will emerge from continuous technological innovation in analytical and imaging
techniques, the application of experimental petrology, and the development and application of
new theoretical models for coupled mineral behaviour under changing P–T conditions.
C The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
V 1671
1672 Journal of Petrology, 2018, Vol. 59, No. 9
of diamonds [first discovered by Sobolev & Shatsky what inclusions tell us should we believe? In a nutshell,
(1990)] and coesite [first reported by Chopin (1984)] to what extent are inclusions really ‘grains of truth’? In
enclosed in metamorphic garnet are often the only dir- this contribution, we address these questions in the
ect mineralogical evidence of an ultra-deep meta- light of recent developments and results. We first re-
morphic history (e.g. Majka et al., 2014; Klonowska view the thermodynamics of inclusion systems to
et al., 2017). The discovery of these ultrahigh-pressure emphasize the common principles of behaviour of MI,
(UHP) phases in crustal rocks provides an excellent ex- FI and solid inclusions, and show where they differ fun-
ample of how the study of mineral inclusions may damentally from the thermodynamics of an open sys-
cause paradigm shifts; for example, leading to the the- tem. Then we show how these concepts can be applied
ory that continental crust can be subducted to depths of to the understanding of single-phase mineral inclusions
more than 100 km in collisional orogens (Spear et al., as a basis for interpreting the more complex situations
2016). However, as our understanding of Nature con- found in melt inclusions, to provide insights into the
tinuously evolves, the equation diamond/coesite ¼ geological history of the rocks in which they are found.
Elastic processes are, by definition, reversible, and For the opposite case of diamond inclusions inside
calculations of elastic stresses can be made over any garnets trapped at P and T within the stability field of
convenient path in P–T space, between entrapment con- diamond, Kinc > Khost, and therefore the sense of the
ditions and any subsequent condition of interest. pressure difference between the inclusion and the ex-
Therefore, the thermodynamic concept of an isomeke ternal pressure is reversed (Fig. 1b). Decompression
(Rosenfeld & Chase, 1961; Adams et al., 1975) can pro- from the same entrapment isomeke at room tempera-
vide a useful basis for understanding the thermody- ture then leads to dPinc > dPext. Therefore, when the
namics of inclusions. The use of the isomeke in garnet host is at room conditions (Pext 0) thermody-
calculations is only a convenience; it does not imply namics predicts that the diamond inclusion should be
that the host–inclusion system in a rock actually fol- under negative pressure (Fig. 1b). In the absence of
lowed the isomeke on exhumation. bonding between the garnet and diamond surfaces, this
Garnet–diamond host–inclusion systems can serve means that void space would be developed at the dia-
to illustrate the thermodynamic principles. From the mond–garnet interface, the diamond would ‘rattle’ in
Phase equilibria in inclusions strength of the host mineral: the result is the failure of
Equilibrium thermodynamics as used in petrology the host mineral, with more or less extensive decrepita-
assumes free boundary conditions, with the rock vol- tion and material loss (Ferrero et al., 2012; Bartoli et al.,
ume free to expand at constant P and T. Even if it is not 2013a; Cesare et al., 2015). Therefore, to obtain bulk
clear that this is generally a valid boundary condition chemical analyses of partially crystallized MI, they must
for rocks (e.g. Mancktelow, 1995; Vrijmoed et al., 2009; be re-melted under a confining pressure to prevent de-
Hacker & Gerya, 2013; Moulas et al., 2013; Tajcmanová crepitation. The pressure required to obtain re-homoge-
et al., 2014, 2015), it is certainly not valid for an inclusion nization of a MI without decrepitation places a lower
because it is enclosed by the host mineral and is there- bound on the original entrapment pressure (Ferrero
fore not free to expand. As a consequence of thermal et al., 2018b).
equilibration on the crystal length-scale being fast com- If there were no elastic relaxation, the stable phase
pared with most geological timescales, the thermo- assemblage in an inclusion would be the one with the
dynamic equilibrium behaviour of the inclusion is lowest Helmholtz free energy, in contrast to the assem-
governed by the external T and the volume imposed by blage with the lowest Gibbs free energy in the rock ma-
the host crystal, together with the mutual elastic relax- trix. With relaxation, neither free energy is an
ation. Thus, Pinc becomes a dependent variable for the appropriate predictor of the equilibrium phase assem-
thermodynamics of an inclusion. blage in an inclusion. The equilibrium thermodynamic
A simple example is provided by melting. Melting in consequences of inclusion entrapment therefore often
a rock matrix as a result of an increase in T leads to an appear to represent metastable equilibrium when
increase in volume and a decrease in density at con- viewed from the perspective of the rock as a whole. The
stant P. In contrast, melting of a constrained system simplest example is provided by polymorphic transi-
such as that of a crystallized melt inclusion does not tions in silica. Mineral inclusions such as quartz have
change the density of the inclusion itself. Its chemistry thermal expansion coefficients that are similar to those
and its volume are kept approximately constant by en- of the host garnets, so the isomeke slopes tend to be
capsulation, with small changes arising only from ther- shallow in P–T space. Consequently, the pressures in
mal expansion of the host and changes in the quartz inclusions originally entrapped in garnet cores
relaxation. Therefore melting of an inclusion leads to a during the early stages of subduction lag so far behind
pressure increase in the inclusion (see also Vrijmoed the external P that they often do not pass into the stabil-
et al., 2009). This is clearly visible when crystallized for- ity field of coesite, even when the garnet is at peak P
mer melt inclusions are experimentally heated to melt- [for details of the calculation see Angel et al. (2015b)].
ing T without applying any confining pressure; for The apparent metastability of quartz in garnet at pres-
example, in a classic heating stage used for magmatic sures in the coesite field is thus just a consequence of
MI studies (Audetat & Lowenstern, 2014). As the crystal- the contrast of elastic properties of the host and inclu-
line phases in the inclusion start to melt, the internal sion and the constrained state of the inclusion, and it ac-
P of the inclusion rises until it very often overcomes the tually represents equilibrium thermodynamics.
Journal of Petrology, 2018, Vol. 59, No. 9 1675
MINERAL INCLUSIONS: INSIGHTS INTO THE such as geothermobarometry and phase equilibria
GENESIS OF THE HOST modelling.
The ‘bread and butter’ of metamorphic petrology has
long been the detailed investigation of inclusion-rich Isotropic elastic geobarometry
porphyroblasts, the most exploited—and often the Although the simple identification and composition of
most aesthetically appealing—including spectacular inclusions trapped during growth of the host phase
garnet porphyroblasts in metasediments (Yardley et al., have long been used to infer P–T points along meta-
1991; Passchier & Trouw, 2005). Inclusions of sillimanite morphic paths, the advent of accurate P–V–T equations
and andalusite in garnets are commonly preserved for of state allows the analysis of the elastic evolution of a
the same reason that quartz is preserved in garnets that single-crystal inclusion to be used to constrain the con-
reach pressures in the coesite stability field; the inclu- ditions of entrapment, and thus provide an alternative
sions are softer than the garnet host, and thus along the method of barometry (e.g. Ashley et al., 2016; Kouketsu
prograde path above the entrapment isomeke Pinc et al., 2016). This has been variously termed ‘piezobar-
remains significantly below the external pressure. ometry’ (Adams et al., 1975), ‘host–inclusion thermo-
These mineral inclusions therefore provide direct con- barometry’ (e.g. Zhang, 1998), ‘QuiG thermobarometry’
straints for the P–T conditions of host growth, and thus when applied to quartz in garnet, and ‘Thermoba-
a first-order estimate of the P–T trajectory, especially Raman-try’ when specifically involving inclusion meas-
when coupled with data on chemical zoning of the host urement by Raman spectroscopy (Kohn, 2014).
(Thompson et al., 1977; Tracy et al., 1976; Krogh, 1982). Inclusion geobarometry does not require thermo-
The same approach has been more recently applied to dynamic equilibrium between the minerals to have
quartz and coesite inclusions in garnets in UHP meta- existed at the time of inclusion entrapment. In contrast,
morphic rocks (e.g. Parkinson, 2000). Another example because elastic equilibration is by definition instantan-
was recently provided by Barich et al. (2014), who inves- eous, elastic geobarometry should record the pressure
tigated rocks from the anatectic sequence of Jubrique at the moment of entrapment. A comparison of entrap-
(Betic Cordillera, southern Spain). Here, MI occur along- ment conditions inferred by elastic geobarometry with
side mineral inclusions of kyanite in garnet cores and growth conditions inferred from the composition of the
with sillimanite in garnet rims (Fig. 3). This assemblage host offers a unique opportunity to determine the de-
of inclusions immediately constrains partial melting to gree of over-stepping during metamorphic reactions
have occurred during garnet growth, which itself (e.g. Spear, 2014; Wolfe & Spear, 2018).
occurred first at high pressure in the stability field of The method, as currently applied, consists of meas-
kyanite, and was then followed by a second growth epi- uring the remnant pressure Pinc in the inclusion, while
sode at supra-solidus conditions in the sillimanite sta- the host is at room conditions, by Raman spectroscopy
bility field (Barich et al., 2014). This example shows how or in situ X-ray diffraction (e.g. Nestola et al., 2011). The
the identification of three different types of inclusions in difference between Pinc and room pressure, and the
different microstructural positions allows a direct quali- elastic properties of the two phases, is used to calculate
tative description of metamorphic history, without the DPrelax, and thus the Pthermo (Angel et al., 2017). From
employment of more sophisticated petrological tools Pthermo the pressure of the entrapment isomeke at room
Journal of Petrology, 2018, Vol. 59, No. 9 1677
temperature can be calculated. Then the equations of quartz inclusions in a single host (e.g. Kouketsu et al.,
state of the host and inclusion can be used to calculate 2016; Viete et al., 2018). Finite-element modelling shows
the entrapment isomeke, which represents all P–T that some of these problems arise from the inclusions
points at which the inclusion may have been trapped being non-spherical. When the axial ratios of the inclu-
(e.g. Fig. 1a and b). It should be noted that, because sions significantly exceed 2:1:1, the elastic relaxation is
only one quantity, the Pinc, is measured on the inclu- changed, leading to a Pinc that differs from that
sion, a unique entrapment P and T cannot be deter- expected for a spherical inclusion entrapped under the
mined from this isotropic analysis. An independent same conditions (Mazzucchelli et al., 2018). Once the
determination of either the P or the T of entrapment effects of shape and other geometric effects, such as
(e.g. from classical barometry using phase equilibria) proximity to external surfaces and other inclusions,
then allows the entrapment conditions to be uniquely have been allowed for, experimentally synthesized
determined from the measurement of Pinc. inclusions measured by Raman spectroscopy exhibit
Elastic geobarometry, as currently applied to single- slightly higher Pinc than expected from a spherical iso-
the inclusion was elastically re-equilibrated (e.g. by ‘nanogranitoids’ (Cesare et al., 2015). Regardless of
plastic flow in the garnet host) at high temperatures and their crystallization state, in the absence of cracks con-
lower pressures in the stability field of graphite (Fig. 1c), necting the inclusion with the rock matrix they can be
consistent also with the observed partial inversion of regarded as preserved and untouched volumes of ori-
diamond to graphite in some inclusions. In this case, ginal melt (Ferrero et al., 2016b). The majority of nano-
the elastic analysis of the diamond-in-garnet inclusions granitoids with sizes 10 mm in garnet from high-grade
using the isomeke concept provides qualitative inde- terranes show fully developed negative crystal shapes;
pendent evidence that the Variscan granulites experi- that is, they mimic the shape of the host garnet (Fig. 4b
enced pressure reduction as a result of partial and c). Although negative crystal is the most common
exhumation at high temperatures in agreement with the shape, tubular shapes have been also observed in
P–T path inferred from conventional petrology (Haifler nanogranitoids hosted in garnet. For example, Darling
& Kotková, 2016). et al. (1997) reported tubular polycrystalline inclusions
interpreted as former MI in garnet megacrysts from the
does not contain any preserved nanogranitoids, as dem- & Zhou, 2017); however, the distinction between the
onstrated by Bartoli (2018), who used nanogranitoid two can be achieved by the identification of primary MI
compositions from rocks with similar protoliths and (Cesare et al., 2015).
melted under similar P, T and H2O conditions (Bartoli The most common structural evidence for partial
et al., 2016a). These results provide a strong argument in melting is the presence of well-defined leucosome
support of the use of anatectic MI compositions as a vi- domains in the rock, whereas at the microscale pseudo-
able approach to model the processes of partial melting morphs after melt-filled pores (e.g. Holness & Sawyer,
and melt extraction. 2008), euhedral crystals and/or crystal overgrowths
Melt inclusions are also useful in other ways; for ex- formed during melting (peritectic phases) or crystallized
ample, they allow the identification of the true genesis from the melt (e.g. feldspar; Vernon, 2011) are often
of garnets. In medium- to high-grade metamorphic observed. Such evidence can, however, be obliterated
crustal rocks, garnet may form on the prograde path as a result of melt loss, deformation and extensive
well before partial melting conditions are reached (e.g. chemical re-equilibration during retrogression
Spear, 1993). Thus, the distinction between sub-solidus (Marchildon & Brown, 2001; White et al., 2011). In the
and supra-solidus garnets may be problematic (e.g. Xia absence of such evidence, MI in refractory phases with
1680 Journal of Petrology, 2018, Vol. 59, No. 9
Fig. 6. Melt re-integration procedure applied to partially melted granulitic enclave from El Hoyazo, Neogene Volcanic Province [fig-
ures provided by O. Bartoli and modified after Bartoli (2017)]. In (a), the initial protolith has been back-calculated using both the
composition of anatectic MI and that of the glass still preserved in the enclave’s matrix; the calculated protolith shows a melt prod-
uctivity of 50 wt% at 850 C, consistent with petrographic constraints (see discussion by Bartoli, 2017).
Journal of Petrology, 2018, Vol. 59, No. 9 1681
In some cases, OH-bearing phases are absent and Frezzotti & Ferrando, 2015), and this is expected to have
the only remaining evidence for the former presence of caused fluid and/or material loss as well as chemical re-
volatiles is provided by rounded voids preserved in the equilibration with the matrix, thus preventing complete
inclusions (Gao et al., 2017). As for lower-P nanograni- re-homogenization on re-heating (Ferrero et al., 2012).
toids, re-homogenization under pressure in the piston Besides the experimental approach, the most direct
cylinder is the most direct way to prove that these HP/ evidence that HP/UHP inclusions originally contained
UHP inclusions were melt droplets at the moment of melt is provided by the presence of glassy inclusions
entrapment [see, e.g. Ferrero et al. (2015) and Stepanov (Borghini et al., 2017) and/or residual glass, as is the
et al. (2016)]. However, this approach requires that the case for near-UHP nanogranitoids from the Bohemian
inclusions are preserved; that is, they did not undergo Massif (Ferrero et al., 2015, 2016b), in which glass was
decrepitation and material loss. Unfortunately, exten- identified via Raman mapping (Fig. 9). Previously, re-
sive decrepitation cracks around HP/UHP multiphase sidual glass has been identified using transmitted elec-
inclusions are common (see, e.g. Gao et al., 2013; tron microscopy of multiphase inclusions in
Journal of Petrology, 2018, Vol. 59, No. 9 1683
microdiamond from the Kokchetav Massif (Hwang independence of the inclusions from the surrounding
et al., 2004, 2006) and in the Erzgebirge (Hwang et al., environment does not make them intrinsically metasta-
2006), and interpreted as evidence that deep melts play ble; as we have shown for mineral inclusions, many
a role in the formation of microdiamonds in subducted examples that appear to be metastable are simply
crust. exhibiting the stable behavior of an enclosed system
under its own peculiar stress state. On the other hand,
examples of real metastable behaviour in fluid inclu-
UNDERSTANDING THE MESSAGE OF MI sions have been extensively reported; for example, the
Metastable phenomena in small containers? failure of vapour bubbles and/or ice to nucleate on cool-
While inclusions remain isolated, their behaviour is ing (Roedder, 1967; Bakker & Baumgartner, 2012; Qiu
expected to be independent from what happens in the et al., 2016) or the growth of graphite in COH fluid inclu-
surrounding environment. This condition, coupled with sions, apparently triggered by the Raman laser beam
the small size of the inclusions, may favor the formation (Van den Kerkhof et al., 1991), as well as evidence of
and preservation of phases well outside their stability delayed crystallization of daughter minerals in igneous
field. It is important to note, however, that the MI (Roedder, 1971).
1684 Journal of Petrology, 2018, Vol. 59, No. 9
2012). However, it is difficult to imagine how the natural protolith; kokchetavite has been reported in nanograni-
process of exhumation of deeply subducted rocks can toids from metagranitoids (Ferrero et al., 2016b; Figs 8a
cause instantaneous cooling of the host rocks. and 10a), and mafic (Borghini et al., 2017) and ultramaf-
In summary, the available data are consistent with ic rocks (Ferrero et al., 2018a).
the hypothesis that these polymorphs form as result of
rapid crystallization of a melt, which is not caused by SiO2 polymorphs
rapid cooling but by the peculiar undercooled and Cristobalite is another polymorph recognized in nano-
supersaturated conditions achieved on cooling by a granitoids (Fig. 10a) as well as in mineral inclusions in
melt confined in a small cavity, perhaps further metamorphic rocks. In metamorphic rocks cristobalite
enhanced by the availability of the host mineral surface was first identified within polycrystalline inclusions,
to act as nucleation sites. Metastable growth as a result which are possibly former melt or solute-rich fluid, in
of the small size of the inclusions was also recently pro- garnets of the Gore Mountains in the Adirondacks, USA
posed to explain the unexpected coexistence of corun- (Darling et al., 1997), and, a few years later, in ompha-
dum and quartz in inclusions from the felsic gneisses of cites of the Kokchetav Massif (Hwang et al., 2004),
the Athabasca Granulite Terrane reported by Tacchetto where it often coexists with kokchetavite. More recently,
et al. (2018). Other suggestions [e.g. that the presence cristobalite was recognized in nanogranitoids from fels-
of kokchetavite is an indicator of a metasedimentary ic granulites from both the Bohemian Massif (Ferrero
protolith (Mikhno et al., 2017)] are not consistent with et al., 2016b) and the Acadian orogen in Connecticut
the observation that the crystallization products of MI [Ferrero et al. (2017); see also Axler & Ague (2015) for
depend exclusively on the chemistry of the trapped P–T formation conditions]. Cristobalite was reported as
melt and are not an inherited feature of the original mineral inclusions in medium-P garnets by Ashley et al.
1686 Journal of Petrology, 2018, Vol. 59, No. 9
(2015), who proposed it to be the result of post- near-UHP nanogranitoids of the Bohemian Massif, as
entrapment transformation under-pressurization of for- they show no evidence of H2O loss (Ferrero et al., 2015).
mer quartz inclusions. Interestingly, this finding comes In this example, cristobalite coexists with residual glass,
from the Adirondacks Highlands, where cristobalite was indicating direct crystallization from the melt at T much
first reported in garnet by Darling et al. (1997). lower than 875 C (Ferrero et al., 2016b), also favoured
The most intriguing feature shared by all of these by the similarities between the structure of cristobalite
case studies is that none of the host rocks experienced and the melt (e.g. Huang & Kieffer, 2004). In summary,
the extreme conditions at which cristobalite is thermo- the available experimental evidence and observations
dynamically stable (i.e. T > 1470 C and P < 02 GPa; on natural inclusions support the idea that the presence
Heaney, 1994), thus suggesting its crystallization as a of metastable polymorphs of SiO2 in inclusions is the
metastable phase. Cristobalite was previously observed result of rapid melt crystallization in a confined volume.
to crystallize in the tridymite field (Schairer & Bowen,
1956), and thus Darling et al. (1997) proposed that OH
loss from the polycrystalline inclusions may have low- Thermodynamics of MI
ered the Pinc of the inclusion, still at high-T conditions, It is important to consider the thermodynamics of MI to
thus bringing the inclusions into the stability field of investigate whether the preservation of glass or appar-
tridymite, at T 900 C. This mechanism cannot, how- ently metastable polymorphs could actually represent
ever, be invoked for the nucleation of cristobalite in the stable behaviour when viewed from the perspective of
Journal of Petrology, 2018, Vol. 59, No. 9 1687
several case studies of fluid inclusions in HP/UHP meta- spectrometry (nanoSIMS) measurement (Ferrero et al.,
morphic rocks show evidence of H2O loss during ex- unpublished data) after re-homogenization. These inclu-
humation and depressurization, commonly visible as sions in fact contain 62 wt % of H2O (measured via
decrepitation cracks (e.g. Frezzotti & Ferrando, 2015). In NanoSIMS; Fig. 14b) after re-homogenization, an
some cases, such cracks are partially healed into trails amount consistent with that observed in experimental
of sub-micrometric fluid inclusions, and only transmis- melts (Hermann & Spandler, 2008) formed under simi-
sion electron microscopy (TEM) observation can prove lar P–T conditions [900 C and 27 GPa; see also discus-
their existence (Vityk et al., 2000; Viti & Frezzotti, 2001; sion by Ferrero et al. (2015)]. No H2O transfer seems to
Ferrero et al., 2011). H2O loss can also occur via diffu- have occurred during cooling and exhumation from
sion of hydrous components from the inclusion to the these tiny volumes of melt, and no cracks are visible
matrix through the host (Frezzotti et al., 2012; Frezzotti connecting the inclusions with the host matrix (Ferrero
& Ferrando, 2015), a phenomenon also supported by et al., 2016b). That re-homogenized nanogranitoids
experiments on garnet (Wang et al., 1996). maintain their pristine H2O content was already proven
Recent studies support the idea that inclusions in in low-P inclusions after a detailed NanoSIMS investiga-
UHP rocks may instead preserve an H2O content con- tion by Bartoli et al. (2014).
sistent with their formation conditions. For instance, Further support for this idea is provided by the fact
Malaspina et al. (2017) recently reported large amounts that we can completely re-homogenize these inclusions
of preserved H2O detected via Fourier transform infra- at the formation conditions of the garnet (Ferrero et al.,
red (FTIR) mapping (Fig. 14) in multiphase inclusions 2015). Any H2O loss would move the liquidus to higher
from UHP garnet orthopyroxenites of the Maowu temperatures (Johannes & Holtz, 1996) and thus cause
Ultramafic Complex, China. These rocks equilibrated at the nanogranitoids to melt at higher temperatures, in-
4 GPa and 750–800 C (see also Malaspina et al., 2006) consistent with the results of independent geobaromet-
and the inclusions, hosted in garnet, are the result of ric and thermodynamic calculations (Cesare et al.,
the crystallization of a supercritical aqueous fluid. The 2015). Hþ loss through the garnet would be as easily de-
presence of H2O suggests that these inclusions still be- tectable in the experimental products of re-
have as a locally closed system (Malaspina et al., 2017). homogenization experiments of nanogranitoids. Such
The same also appears to be true for melt droplets ori- loss would cause the persistence of a shrinkage bubble
ginally trapped near the HP–UHP transition: nanograni- on re-homogenization owing to the loss of material,
toids from the Polish Sudetes still contain their original and thus incomplete re-homogenization. Moreover, Hþ
H2O content, as verified via Raman investigation loss would cause excess oxygen to be present in the
(Ferrero et al., 2015) and nanoscale secondary ion mass MI, and as this tiny volume of melt is generally enclosed
1690 Journal of Petrology, 2018, Vol. 59, No. 9
by an Fe-rich host (i.e. the garnet), an iron oxide would microstructures in alkali feldspars whose orientation
form under re-heating, as reported by Danyushevsky and scale can be used as a geospeedometer (Parsons
et al. (2002) for magmatic MI hosted in Fe-bearing oliv- et al., 2015). Other types of CORs might be expected to
ine. Further details about the problems of H2O loss or form when inclusions are trapped by, rather than
gain in nanogranitoids have been given in the discus- exsolved from, the host. If the inclusions are remnant
sion by Bartoli et al. (2014). In summary, despite the fragments of the protolith, one might expect a com-
high mobility of H2O in metamorphic environments, pletely random orientation of the inclusions relative to
inclusions of partial melt as well as (some) fluid inclu- the host minerals. On the other hand, if the inclusion is
sions seem to preserve their original H2O content, even a product of the host-forming reaction it may well have
when these inclusions formed under extreme condi- nucleated and grown on the existing host crystal, in
tions in deeply subducted rocks. which case some non-random COR may be expected. In
the latter scenario, if the nucleation and subsequent
growth is epitaxial, a specific single COR will be exhib-
Angel, R. J., Mazzucchelli, M. L., Alvaro, M. & Nestola, F. (2017). shoulder (Adam’s Diggings) of Rio Grande Rift: evidences
EosFit-Pinc: a simple GUI for host–inclusion elastic thermo- for metasomatism. European Mineralogical Congress,
barometry. American Mineralogist 102, 1957–1960. Rimini.
Angel, R. J., Alvaro, M. & Nestola, F. (2018). 40 years of mineral Bodnar, R. J. (1989). Synthetic fluid inclusions: a novel tech-
elasticity: a critical review and a new parameterisation of nique for experimental water-rock interaction Studies. In:
Equations of State for mantle olivines and diamond inclu- Miles, D. L. (ed.) Water–Rock Interaction. Rotterdam:
sions. Physics and Chemistry of Minerals 45, 95–113. Balkema.
Arlt, T. & Angel, R. J. (2000). Pressure buffering in a diamond Bodnar, R. J. (2003). Reequilibration of fluid inclusions. In:
anvil cell. Mineralogical Magazine 64, 241–245. Samson, I., Anderson, A. & Marshall, D. (eds) Fluid
Ashley, K., Steele-MacInnis, M., Bodnar, R. J. & Darling, R. S. Inclusions: Analysis and Interpretation. Mineralogical
(2016). Quartz-in-garnet inclusion barometry under fire: Association of Canada, Short Course Series 32, 213–230.
reducing uncertainty from model estimates. Geology 44, Bodnar, R. J., Lecumberri-Sanchez, P., Moncada, D. & Steele-
699–702. MacInnis, M. (2014). Fluid inclusions in hydrothermal ore
Audetat, A. & Lowenstern, J. B. (2014). Melt inclusions. In: deposits. In: Holland, H. D. & Turekian, K. K. (eds) Treatise
Holland, H. D. & Turekian, K. K. (eds) Treatise on on Geochemistry, 2nd edn. Oxford: Elsevier, pp. 119–142.
Cesare, B., Ferrero, S., Salvioli-Mariani, E., Pedron, D. & garnet from metapelitic enclaves, El Hoyazo, Spain. Lithos
Cavallo, A. (2009a). Nanogranite and glassy inclusions: the 124, 117–131.
anatectic melt in migmatites and granulites. Geology 37, Ferrero, S., Bartoli, O., Cesare, B., Salvioli-Mariani, E., Acosta-
627–630. Vigil, A., Cavallo, A., Groppo, C. & Battiston, S. (2012).
Cesare, B., Rubatto, D. & Gómez-Pugnaire, M. T. (2009b). Do ex- Microstructures of melt inclusions in anatectic metasedi-
trusion ages reflect magma generation processes at depth? mentary rocks. Journal of Metamorphic Geology 30,
An example from SE Spain. Contributions to Mineralogy 303–322.
and Petrology 157, 267–279. Ferrero, S., Braga, R., Berkesi, M., Cesare, B. & Laridhi Ouazaa,
Cesare, B., Acosta-Vigil, A., Ferrero, S. & Bartoli, O. (2011). Melt N. (2014). Production of metaluminous melt during
inclusions in migmatites and granulites. Journal of the fluid-present anatexis: an example from the Maghrebian
Virtual Explorer 38, paper 2. basement, La Galite Archipelago, central Mediterranean.
Cesare, B., Acosta-Vigil, A., Bartoli, O. & Ferrero, S. (2015). Journal of Metamorphic Geology 32, 209–225.
What can we learn from melt inclusions in migmatites and Ferrero, S., Wunder, B., Walczak, K., O’Brien, P. J. & Ziemann,
granulites? Lithos 239, 186–216. M. A. (2015). Preserved near ultrahigh-pressure melt from
Cheskonov, B. V., Lotova, E. V., Pavlyuchenko, V. S., Usova, L. continental crust subducted to mantle depths. Geology 43,
Garcia-Arias, M. & Stevens, G. (2017). Phase equilibrium mod- multi-technique approach. European Journal of Mineralogy
elling of granite magma petrogenesis: A. An evaluation of 24, 563–573.
the magma compositions produced by crystal entrainment Huang, L. & Kieffer, J. (2004). Amorphous–amorphous transi-
in the source. Lithos 277, 131–153. tions in silica glass. I. Reversible transitions and thermome-
Gardien, V., Thompson, A. B., Grujic, D. & Ulmer, P. (1995). chanical anomalies. Physical Review B 69, 1–11.
Experimental melting of biotite þ plagioclase þ quartz 6 Huddlestone-Holmes, C. R. & Ketcham, R. A. (2005). Getting the
muscovite assemblages and implications for crustal melt- inside story: using computed X-ray tomography to study in-
ing. Journal of Geophysical Research: Solid Earth 100, clusion trails in garnet porphyroblasts. American
15581–15591. Mineralogist 90, ea1–e17.
Griffiths, T. A., Habler, G. & Abart, R. (2016). Crystallographic Huddlestone-Holmes, C. R. & Ketcham, R. A. (2010). An X ray
orientation relationships in host–inclusion systems: new computed tomography study of inclusion trail orientations
insights from large EBSD data sets. American Mineralogist in multiple porphyroblasts from a single sample.
101, 690–705. Tectonophysics 480, 305–320.
Griffiths, T. A., Habler, G., Schantl, P. & Abart, R. (2017). Hwang, S.-L., Shen, P., Chu, H. T., Yui, T. F., Liou, J. G., Sobolev,
Crystallographic orientation relationships (CORs) between N. V., Zhang, R. Y., Shatsky, V. S. & Zayachkovsky, A. A.
Krogh, E. J. (1982). Metamorphic evolution of Norwegian Mikhno, A., Musiyachenko, K. A., Shchepetova, O. V., Korsakov,
country-rock eclogites, as deduced from mineral inclusions A. V. & Rashchenko, S. V. (2017). CO2-bearing fluid inclu-
and compositional zoning in garnets. Lithos 15, 305–321. sions associated with diamonds in zircon from the UHP
Kueter, N., Soesilo, J., Fedortchouk, Y., Nestola, F., Belluco, L., Kokchetav gneisses. Journal of Raman Spectroscopy 48,
Troch, J., Wälle, M., Guillong, M., Von Quadt, A. & Driesner, 1566–1573.
T. (2016). Tracing the depositional history of Kalimantan dia- Milani, S., Nestola, F., Alvaro, M., Pasqual, D., Mazzucchelli, M.
monds by zircon provenance and diamond morphology L., Domeneghetti, M. C. & Geiger, C. (2015).
studies. Lithos 265, 159–176. Diamond–garnet geobarometry: the role of garnet com-
Laurie, A. & Stevens, G. (2012). Water-present eclogite melting pressibility and expansivity. Lithos 227, 140–147.
to produce Earth’s early felsic crust. Chemical Geology Milani, S., Nestola, F., Angel, R. J., Nimis, P. & Harris, J. (2016).
314–317, 83–95. Crystallographic orientations of olivine inclusions in dia-
Lecumberri-Sanchez, P., Steele-MacInnis, M. & Bodnar, R. J. monds. Lithos 265, 312–316.
(2015). Synthetic fluid inclusions XIX. Experimental deter- Moore, S. J. & Carlson, W. D. (2015). Evaluation of a combined
mination of the vapor-saturated liquidus of the system HRXCT/EBSD method for detecting epitaxial nucleation of
H2O–NaCl–FeCl2. Geochimica et Cosmochimica Acta 148, garnet porphyroblasts. European Journal of Mineralogy 27,
O’Brien, P. J. (2001). Subduction followed by collision: Alpine Roedder, E. (1967). Metastable superheated ice in liquid-water inclu-
and Himalayan examples. Physics of the Earth and sions under high negative pressure. Science 155, 1413–1417.
Planetary Interiors 127, 277–291. Roedder, E. (1971). Metastability in Fluid Inclusions. Society of
Palke, A. C., Renfro, N. D. & Berg, R. B. (2017). Melt inclusions in Mining Geology Japan 3, 327–334.
alluvial sapphires from Montana, USA: formation of sap- Roedder, E. (1984). Fluid Inclusions. Mineralogical Society of
phires as a restitic component of lower crustal melting? America, Reviews in Mineralogy 12.
Lithos 278–281, 43–53. Rosenfeld, J. L. & Chase, A. B. (1961). Pressure and temperature
Pamukcu, A. S., Gualda, G. A. R. & Rivers, M. L. (2013). of crystallization from elastic effects around solid inclusion
Quantitative 3D petrography using X-ray tomography 4: minerals? American Journal of Science 259, 519–541.
assessing glass inclusion textures with propagation Sawyer, E. W., Cesare, B. & Brown, M. (2011). When the contin-
phase-contrast tomography. Geosphere 9, 1704–1713. ental crust melts. Elements 7, 229–234.
Papa, S., Pennacchioni, G., Angel, R. J. & Faccenda, M. (2018). Scambelluri, M. & Philippot, P. (2001). Deep fluids in subduction
The fate of garnet during (deep-seated) coseismic frictional zones. Lithos 55, 213–227.
heating: the role of thermal shock. Geology 46, 471–474. Schairer, J. F. & Bowen, N. L. (1956). The system
Parisatto, M., Turina, A., Cruciani, G., Mancini, L., Peruzzo, L. & Na2O–Al2O3–SiO2. American Journal of Science 254, 129–196.
and Interactions II. Geological Society of America Special Torquato, S. (2002). Random Heterogeneous Materials:
Papers 523, 31–74. Microstructure and Macroscopic Properties. New York: Springer.
Stachel, T. & Luth, R. W. (2015). Diamond formation—where, Touret, J. L. R. (1981). Fluid inclusions in high grade meta-
when and how? Lithos 220–223, 200–220. morphic rocks. In: Hollister, L. S. & Crawford, M. L. (eds)
Steele-MacInnis, M., Lecumberri-Sanchez, P. & Bodnar, R. J. Short Course in Fluid Inclusions: Applications to Petrology.
(2015). Synthetic fluid inclusions XX. Critical PTx properties Calgary, AB: Mineralogical Association of Canada, pp.
of H2O–FeCl2 fluids. Geochimica et Cosmochimica Acta 148, 182–208.
50–61. Touret, J. L. R. (2009). Mantle to lower-crust fluid/melt transfer
Stepanov, A. S., Hermann, J., Rubatto, D., Korsakov, A. V. & through granulite metamorphism. Russian Geology and
Danyushevsky, L. V. (2016). Melting history of an ultrahigh- Geophysics 50, 1052–1062.
pressure paragneiss revealed by multiphase solid inclu- Tracy, R. J., Robinson, P. & Thompson, A. B. (1976). Garnet
sions in garnet, Kokchetav Massif, Kazakhstan. Journal of composition and zoning in the determination of temperature
Petrology 57, 1531–1554. and pressure of metamorphism, central Massachusetts.
Sterner, S. M. & Bodnar, R. J. (1984). Synthetic fluid inclusions American Mineralogist 61, 762–775.
in natural quartz. I. Compositional types synthesized and Van den Kerkhof, A. M., Touret, J. L. R., Maijer, C. & Jansen, J.
Whitney, D. L. (1991). Calcium depletion halos and Fe–Mn–Mg Yakymchuk, C. & Brown, M. (2014). Consequences of
zoning around faceted plagioclase inclusions in garnet from a open-system melting in tectonics. Journal of the Geological
high-grade pelitic gneiss. American Mineralogist 76, 493–500. Society, London 171, 21–40.
Whitney, D. L. & Evans, B. W. (2010). Abbreviations for names of Yardley, B. W. D. (2009). The role of water in the evolution of the
rock-forming minerals. American Mineralogist 95, 185–187. continental crust. Journal of the Geological Society, London
Whitney, D. L., Cooke, M. L. & Du Frane, S. A. (2000). Modeling 166, 585–600.
of radial microcracks at corners of inclusions in garnet using Yardley, B. W. D., Mackenzie, W. S. & Guilford, C. (1991). Atlas
fracture mechanics. Journal of Geophysical Research: Solid of metamorphic rocks and their textures. Terra Nova 3,
Earth 105, 2843–2853. 217–218.
Wilke, F. D. H., O’Brien, P. J., Schmidt, A. & Ziemann, M. A. Yoshida, K., Hirajima, T., Miyake, A., Tsuchiyama, A., Ohi, S.,
(2015). Subduction, peak and multistage exhumation meta- Nakano, T. & Uesugi, K. (2016). Combined FIB microsam-
morphism: traces from one coesite-bearing eclogite, Tso pling and X-ray microtomography: a powerful tool for the
Morari, western Himalaya. Lithos 231, 77–91. study of tiny fluid inclusions. European Journal of
Williame, C. & Brown, W. L. (1974). A coherent elastic model for Mineralogy 28, 245–256.
the determination of the orientation of exsolution bounda- Zhang, Y. (1998). Mechanical and phase equilibria in