Shaw meteorite: water-poor and water-rich melt inclusions in olivine and enstatite

Short Communication


During the petrological examination of a double polished thick section of the light-colored lithology of the Shaw meteorite, small melt inclusions, typically with diameters between 10 and 35 μm, were discovered in olivine and orthopyroxene crystals. These inclusions represent two different types of melt inclusions, one low in water (0.57% (g/g)) and the other with extremely high concentrations (15% (g/g)). Extremely water-rich inclusions were also found in olivine, which showed bulk water concentration is between 12 and 13% (g/g). Given the high D2O-concentration of (556 ± 47) ppm and a D/H ratio of (337 ± 32 × 10−6) in the glass phase of the water-rich melt inclusions terrestrial contamination can be excluded. Given the bulk water concentration of the Shaw meteorite is about 1%, typical of L-group chondrites, we propose that the high water concentrations found in some melt inclusions was injected into the melt from a melted projectile during an impact event. As the projectile a Cl-type chondrite with high water concentration is highly probable.


Shaw meteorite Melt inclusion H2O and D2O concentration Impact event 


For the study of melt inclusions we used a 300 μm double polished thick section of the light-colored lithology of the Shaw meteorite (Fig. 1), which is, according to Fredriksson and Mason (1967) an unusual meteorite with chondritic composition, but without the typical chondritic structure. The Shaw meteorite has a very unusual texture with three distinct lithological types, with light, dark, and intermediate grey colored patches, all intermixed in an irregular swirled pattern (Taylor et al. 1979). Remnants of chondrules (Fig. 2) are visible under the microscope particularly in the black regions of the sample. According to Taylor et al. (1979) the Shaw meteorite is an L-group chondrite. The main components are olivine (Fa23) and orthopyroxene (Fs20) with minor amounts of diopside, plagioclase, troilite, and Ni-iron, the last opaque phases are enriched in the dark-colored lithology. An important observation is that we found the water-rich melt inclusions in enstatites inside a chondrule with an unusual internal structure: the colorless entstatite is intergrown with forsterite. The mineralogical composition and structure, as well as the chemical composition of the Shaw meteorite are given by Fredriksson and Mason (1967) and Taylor et al. (1979).
Fig. 1

The studied sample of the Shaw meteorite: a piece of the lighter-colored lithology of the Shaw meteorite with only a small amount of remnant chondrules (about 0.7 to 0.9 mm in diameter) in the darker regions

Fig. 2

Porphyritic chondrule with dark rim in the Shaw meteorite. The white hexagons mark the border of the chondrule, which has a dimension of about 800 × 950 μm. The enstatite (En) crystals with water-rich melt inclusions inside the chondrule are marked by white arrows


Given the rarity of the material and small sample size we exclusively used non-destructive method to preserve the rare material for later, more sophisticated methods in more specialized laboratories.


The intense microscopic study was performed with a polarized light microscope Jenalab Pol for reflected and transmitted light equipped with a ProgRes C10 camera, and a stereomicroscope ZEISS “Stemi 305 trino” with the same camera, or with an Olympus OM-D E-M10 Mark II camera.


Raman spectra were recorded with a Jobin-Yvon LabRam HR800 spectrometer (grating 1800 g/mm), equipped with an Olympus optical microscope and a long-working-distance LMPlanFI 100×/0.80 objective. The 514 and 488 nm excitation of a coherent Ar+ laser, Model Innova 70C, with a power of 45 mW on the sample, at a resolution ≤0.6 cm−1 were used. The spectra were collected at a constant laboratory temperature (20 °C) with a Peltier-cooled CCD detector and the position of the Raman bands were controlled and eventually corrected using the principal plasma lines in the Argon laser. For the water determination in the inclusion glass we used the method presented by Thomas et al. (2009). As a standard we used a synthetic glass with 5.21% (g/g) water, cross-checked with 30 different glasses of well-characterized water concentration (Thomas and Davidson 2007). As standards for the determination of D2O concentration we used rhyolitic glasses doped with up to 5.5% (g/g) D2O (see Leschik et al. 2004; Thomas et al. 2006 and Thomas and Davidson 2007), as well as pure deuterium oxide (heavy water) for NMR spectroscopy (Merck, Darmstadt/Germany) in mixture with pure water. The D2O-Raman stretching bands used are positioned at 2382, 2504, and 2620 cm−1, the weak bending vibration of molecular water at 1203 cm−1 could not be used for quantification because of the low intensity. Note, however, the position of the v1, v3, and 2v2 modes in alkali silicate glasses shift to values of up to 60 cm−1 higher.


During the study of glassy melt inclusions in olivine and orthopyroxene from the Shaw meteorite we generally found two different types of melt inclusions in both minerals, with typical diameters between 10 and 20 μm, rare examples have diameters up to 35 μm. The first and most common inclusion type contains only a very small shrinkage bubble which take up a volume of about 5% (vol/vol) - see (Fig. 3a) in addition to a clear colorless silicate glass (herein type-1 inclusions). The opaque daughter phase is mostly magnetite. The second and rarer type is characterized by a significantly larger “bubble” (herein type-2 inclusions) with a bubble volume of about 30% (vol/vol) - see Fig. 3b. These large bubbles always show two distinct phases, liquid and vapor, with a distinct boundary between them. This type of inclusion contains two different daughter mineral phases in the glass - one opaque, mostly magnetite, and a smaller generally colorless crystal. Raman analysis suggests the colorless phase is apatite indicated by the strong band at 960 cm−1. Exact determination of this phase is tentative because according to the RRUFF data base (Lafuente et al. 2015) the match for the mineral is only 92% and other diagnostic bands coincide with olivine bands.
Fig. 3

Melt inclusions in olivine from the Shaw meteorite: a typical melt inclusion (type-1) in olivine, containing a nearly colorless silicate glass, a small shrinkage bubble, and a small opaque iron mineral. b Type-2 melt inclusion in olivine, characterized by a H2O-containing glass G and a fluid sub-phase composed from liquid water L and a vapor bubble V. This inclusion contains an opaque iron-bearing phase (upper right), mostly magnetite, and a small apatite-like phosphate mineral (upper left)

Water determination with Raman spectroscopy

The bulk water content of the clear inclusion glass in type-1 melt inclusions (Fig. 3a), which represent the predominant inclusion type, determined by Raman-spectroscopy is about 0.57 ± 0.08% (g/g), n = 10 different inclusions. To a first approximation this value corresponds well to the 0.46% (g/g) H2O, typical for L-type chondrites (Table 1.1. in Sears 2004).

The water concentration of glass in type 2 melt inclusions in olivine (Fig. 3b) is significantly higher. From ten different melt inclusions we have obtained as mean (2.3 ± 0.7% (g/g)), a relatively high value for a “bone-dry” glass (see Roedder 1978).

Figure 4 shows the broad Raman band of the melt inclusion glass G (Fig. 3b). Significantly, the large bubble shows an internal meniscus, demonstrating the presence of a liquid water phase. In the high-frequency range, the unpolarized Raman spectrum shows the typical broad asymmetric band for molecular water and hydroxyl groups. After deconvolution, the main components are at 3260, 3483 and 3581 cm−1, corresponding clearly to molecular water and hydroxyl groups, respectively.
Fig. 4

Unpolarized Raman spectrum in the high-frequency region of the type-2 melt inclusions glass (G, Fig. 3), showing three deconvoluted components at 3260, 3483 and 3581 cm−1, corresponding to molecular water and hydroxyl groups

In the large melt inclusion in enstatite (Fig. 5), the presence of molecular water is demonstrated visually by Brownian motion of a vapor bubble moving inside the liquid water sub-phase, especially with weak heating. The water concentration of the glass is 3.3% (g/g) and estimating the liquid volume, and adding this to the water dissolved in the glass gives a bulk concentration of about 15% (g/g) H2O. Similarly, some melt inclusions in olivine have a bulk water concentration between 12 and 13% (g/g). Such high water concentration is not typical for L-group chondrites, however it is typical for wet carbonaceous Cl chondrites with 16.9% (g/g) H2O or more (Tables 1.1 and 2.1in Sears 2004).
Fig. 5

Melt inclusions in enstatite from the Shaw meteorite (a) Water-rich melt inclusion (MI) in nearly colorless enstatite (En) about 80 μm deep from the sample surface - an overview. Note the internal structure of enstatite, composed from forsterite remnants. b Details of the melt inclusion: G - silicate glass, L - water-rich liquid, V - vapor bubble. The inset in (b) shows the liquid phase with bubble taken a short time later, demonstrating that the vapor bubble is moving inside the liquid phase, thus proving that phase marked L is not a second glass

That the high water concentration of this inclusion type is not a result by contamination with terrestrial water can be demonstrated by the unusual high D2O concentration determined by Raman spectroscopy. For the fluid phase we have obtained a clear and relatively strong Raman band between 2300 cm−1 and 2700 cm−1 corresponding to the O-D stretching. Exact quantification was not possible because the Raman band for water is suppressed by a strong luminescence in the high-frequency region between 3200 cm−1 and 3700 cm−1 corresponding to the O-H stretching. For estimation of D2O in the melt inclusion glass and in the reference glass (Leschik et al. 2004) we used the integral intensity between 2400 and 2600 cm−1 (see Thomas et al. 2006). The D2O values for the inclusion glass is only an estimation, because in the low concentration range the discrimination between D2O and HDO is difficult (see Rank et al. 1934 and Leschik et al. 2004). Figure 6 shows a set of Raman spectra of pure D2O, mixtures with H2O, and of the inclusion glass of the melt inclusion (Fig. 3b).
Fig. 6

Raman spectra of water (dark spectrum) with 12.5% D2O, heavy water (pale spectrum). The insert b) shows the Raman spectrum of the melt inclusion glass of type-2 inclusions (black) with small, but, significant amounts of D2O

For five inclusion glasses (Table 1) we have determined 556 ± 47 ppm D2O, which corresponds to a D/H ratio (337 ± 32) * 10−6, a value which is significantly higher than the D/H ratio of Cl chondrites (see discussion by Robert 2003). The D/H ratio of the bulk Earth is 149 * 10−6 (Robert 2003; Hallis 2017) and the mean of the Cl chondrites is 173 ± 18 (n = 18). Our relative high value may be explained by enrichment of D2O in the glass phase by diffusion, because the light water diffuses faster than the heavy water. According to Gaetani et al. (2012) such diffusion can be identified by an elevated deuterium-to-hydrogen ratio within the inclusions. In melt inclusions from terrestrial rocks we have never found detectable amounts of D2O. Furthermore, the formation of a glass with about 15% bulk water is also exceptional for melts in terrestrial olivine or enstatite.
Table 1

D2O-concentration in the studied melt inclusion glasses and the resulting D/H ratio

Melt inclusion

Integral intensity

D2O (ppm)

D/H (× 10−6)






















556 ± 47

337 ± 32


5626.9 ± 293

2080 ± 108


For the estimation of the D/H ratio we used as the water concentration 3.3% (g/g) for the glass. If we take the determined mean, corresponding to 2.3% (g/g) H2O, the D/H ratio increases to a value of 483 × 10−6

Generally, relative to type-1 melt inclusions, type-2 inclusions are very heterogeneous with respect to phase composition, and especially regarding the water concentrations, which vary from 5.3 to 16.4% (g/g).


The type-1 and type-2 melt inclusions found in olivine and orthopyroxene clearly implies the crystallization of both minerals from melts (see to this also the discussion in Roedder 1981), however with a very different origin. The clear evidence of melt inclusions of different composition confirms the results of Taylor et al. (1979) that the Shaw meteorite has a complicated igneous history rather a metamorphic origin. Metamorphic processes normally would have eliminated or at least partially leveled compositional and textural differences. The observed very large differences in water concentration would not have survived.

The two different melt inclusion types in olivine and enstatite with strong contrasting water concentrations and high D/H ratio requires an explanation. L-group chondrites such as the Shaw meteorite are normally water-poor, typical values are 0.46% (g/g). On the other hand, Cl-type chondrites contain between 2.9 and 15.75% (g/g) H2O (Robert 2003) or up to 16.9 according to Sears 2004.

Following Taylor et al. (1979) the simplest explanation is that the Shaw meteorite is a partly-melted shock-breccia and the light-colored portion of the Shaw was totally melted. However, the appearance of two different melt inclusion types with strong contrasting compositions argues against the total melting of the light-colored portions of the Shaw meteorite. We infer that the high water concentrations was injected into the melt from the melted projectile during the impact event. As projectile a Cl chondrite is highly probable. Additionally, the high water content of about 15% (g/g) in a olivine-forming melt requires high temperatures and pressures (≥ 1200 °C, > 10 kb) - compare Sood (1981).



The contribution is dedicated to Klaus Keil in memory of our meeting in Freiberg and Jena. Lhiric Agoyaoy, Anton Beran, and two anonymous reviewers are thanked for their constructive comments which help us to improve the manuscript.


  1. Fredriksson K, Mason B (1967) The Shaw meteorite. Geochim Cosmochim Acta 31:1705–1709CrossRefGoogle Scholar
  2. Gaetani GA, O'Leary JA, Shimizu N, Bucholz CE, Newville M (2012) Rapid reequilibration of H2O and oxygen fugacity in olivine-hosted melt inclusions. Geology 40:915–918CrossRefGoogle Scholar
  3. Hallis LJ (2017) D/H ratios of the inner Solar System. Phil Trans R Soc A 375:20150390CrossRefGoogle Scholar
  4. Lafuente B, Downs RT, Yang H, Stone N (2015) The power of databases: the RRUFF project. In: Armbruster T, Danisi RM (eds) Highlights in mineralogical crystallography. W. De Gruyter, Berlin, pp 1–30Google Scholar
  5. Leschik M, Heide G, Frischat GH, Behrens H, Wiedenbeck M, Wagner N, Heide K, Geißler H, Reinholz U (2004) Determination of H2O and D2O contents in rhyolitic glasses. Phys Chem Glasses 45(4):238–251Google Scholar
  6. Rank DH, Larsen KD, Bordner ER (1934) The Raman spectrum of heavy water vapor. J Chem Phys 2:464–467CrossRefGoogle Scholar
  7. Robert F (2003) The D/H ratio in chondrites. Space Sci Rev 106:87–101CrossRefGoogle Scholar
  8. Roedder E (1978) Silicate liquid immiscibility in magmas and in the system K2O-FeO-Al2O3-SiO2: an example of serendipity. Geochim Cosmochim Acta 42:15973–11617CrossRefGoogle Scholar
  9. Roedder E (1981) Significance of Ca-Al-rich silicate melt inclusions in olivine crystals from the Murchison type II carbonaceous chondrite. Bull Mineral 104:339–353Google Scholar
  10. Sears DWG (2004) The origin of Chondrules and Chondrites. Cambridge Planetary Sciences, 222 pp.
  11. Sood MK (1981) Modern igneous petrology. A Wiley-Interscience Publication, John Wiley & Sons, New York - Chichester - Brisbane - Toronto, p 244Google Scholar
  12. Taylor GJ, Keil K, Berkley JL, Lange DE, Fodor RV, Fruland RM (1979) The Shaw meteorite: history of a chondrite consisting of impact-melted and metamorphic lithologies. Geochim Cosmochim Acta 43:323–337CrossRefGoogle Scholar
  13. Thomas R, Davidson P (2007) Progress in the determination of water in glasses and melt inclusions with Raman spectroscopy: a short review. Acta Petrol Sin 22(12):15–20Google Scholar
  14. Thomas R, Kamenetsky VS, Davidson P (2006) Laser Raman spectroscopic measurements of water in unexposed glass inclusions. Am Mineral 91:467–470CrossRefGoogle Scholar
  15. Thomas S-M, Koch-Müller M, Reichart P, Rhede D, Thomas R, Wirth R, Matsyuk S (2009) IR calibration for water determination in olivine, r-GeO2, and SiO2 polymorphs. Phys Chem Miner 36:489–509CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Austria, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Helmholtz-Centre Potsdam, German Research Centre for Geoscience – GFZ, Section 4.3. Chemistry and Physics of Earth MaterialsPotsdamGermany
  2. 2.CODES, Centre for Ore Deposit and Earth ScienceUniversity of TasmaniaHobartAustralia

Personalised recommendations