American Mineralogist, Volume 93, pages 1865–1873, 2008
Common gem opal: An investigation of micro- to nano-structure EloïsE Gaillou,1,2,* EmmanuEl Fritsch,1 BErtha aGuilar-rEyEs,3 BEnjamin rondEau,1 jEFFrEy Post,2 alain BarrEau,1 and mikhail ostroumov3 Université de Nantes, Nantes Atlantique Universités, CNRS, Institut des Matériaux Jean Rouxel (I.M.N.), UMR 6502, 2 rue de la Houssinière, B.P. 32229, Nantes, F-44000 France 2 Department of Mineral Sciences, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20064, U.S.A. 3 Universidad de Michoacán de San Nicolas de Hidalgo, Ciudad universitaria, Fransisco J. Mujica S/N, Apartado postal 52B, C.P. 58000 Morelia, Michoacán, Mexico 1
aBstract The microstructure of nearly 200 common gem opal-A and opal-CT samples from worldwide localities was investigated using scanning electron microscopy (SEM). These opals do not show play-of-color, but are valued in the gem market for their intrinsic body color. Common opal-AG and opal-CT are primarily built from nanograins that average ~25 nm in diameter. Only opal-AN has a texture similar to that of glass. In opal-AG, nanograins arrange into spheres that have successive concentric layers, or in some cases, radial structures. Common opal does not diffract light because its spheres exhibit a range of sizes, are imperfectly shaped, are too large or too small, or are not well ordered. Opal-AG spheres are typically cemented by non-ordered nanograins, which likely result from late stage fluid deposition. In opal-CT, nanograins have different degrees of ordering, ranging from none (aggregation of individual nanograins), to an intermediate stage in which they form tablets or platelets, to the formation of lepispheres. When the structure is built of lepispheres, they are generally cemented by non-ordered nanograins. The degree of nanograin ordering may depend on the growth or deposition rate imposed by the properties of the gel from which opal settles, presumably, fast for non-ordered nanograin structures in opal-CT to slow for the concentric arrangement of nanograins in the spheres of opal-AG. Keywords: Opal-A, opal-CT, common opal, structure, SEM, nanograin
introduction Opals are natural hydrous silica with either amorphous (opal-A) or disordered cristobalite/tridymite structures (opalCT). Gem opal is best known for the highly prized variety showing diffraction of visible light, called play-of-color opal. Yet, the most widespread gem varieties, so-called common opals, do not show play-of-color but are valued in the gem trade for their attractive body colors. The only detailed studies of common opals reported to date are for Australian potch opals (Bayliss and Males 1965; Barnes et al. 1992) and biogenic opals (e.g., Kastner et al. 1977; Botz and Bohrmann 1991; Graetsch 1994; Elsass et al. 2000). The picture emerging from the previous studies is that opal-A consists of regular spheres and opal-CT of spherical aggregates of plate-like cristobalite crystallites, called lepispheres. In this study, we characterized a large number of common gem opals from a wide variety of geologic settings and localities to provide a more complete understanding of the structure of these materials and to determine if they are consistent with this model. We do not consider here biogenic opals, which are not used as gems.
BackGround There are three recognized opal varieties: opal-A (amorphous); opal-CT (cristobalite-tridymite, which consists of disordered α-cristobalite with tridymitic stacking), and opal-C (which * E-mail:
[email protected] 0003-004X/08/1112–1865$05.00/DOI: 10.2138/am.2008.2518
the proportion of cristobalite much greater than that of tridymite; we did not encounter this type in our study). The basic types are typically identified using X-ray diffraction (XRD) (e.g., Jones and Segnit 1971; Elzea and Rice 1996), but opal-A can also be distinguished from opal-C and -CT on the basis of Raman scattering spectroscopy (Ostrooumov et al. 1999). Langer and Flörke (1974) subdivided opal-A into two groups on the basis of features observed in small angle X-ray and neutron-scattering experiments: (1) opal-AN (network), or “hyalite,” which shows only diffuse scattering of X-rays or neutrons at small angles, suggesting that it has a glass-like structure; and (2) opal-AG (gel), which is the most widespread variety. In small-angle X-ray or neutron patterns, it exhibits obvious intensity maxima superimposed upon the diffuse scattering, indicating a structure consisting of packed silica spheres. The term opal-A typically is synonymous with opal-AG. Structure of opal-AG The first scanning electron microscopy (SEM) study of the structure of opal was published by Jones et al. (1964) for an Australian play-of-color opal-AG. They demonstrated that it is constructed of a near-perfect 3-D stacking of monodisperse silica spheres, which diffracts visible light if the spheres have diameters ranging from ~150 to 300 nm (Sanders 1964). Common opal-AG, named potch opal by Australian miners, appears to be made of the same type of spheres (Darragh and Gaskin 1966; Rau and Amaral 1969; Sanders and Darragh 1971). The absence of visible
1865
1866
GAILLOU ET AL.: STRUCTURE OF COMMON GEM OPAL
light diffraction by these common opals is attributed primarily to inhomogeneity in the sphere diameters, which makes regular stacking impossible. Lack of light diffraction may also be due to irregular sphere shapes (Darragh et al. 1966; Sanders and Darragh 1971), spheres that are too large (up to 1000 nm; Cole and Monroe 1967), or spheres having the same refractive index as the cement in-between (Graetsch and Ibel 1997). Cementation and section-orientation effects. The silica spheres in opal-AG are generally cemented by small particles (Sanders and Darragh 1971). The cement is sufficiently strong that a fracture induced to prepare an SEM sample passes through the spheres rather than between them. On a fresh break, it is typically difficult to observe the spherical structure, which is only revealed by etching. It is not straightforward to measure the diameters of the spheres (Sanders and Darragh 1971), as the fracture might pass anywhere through a sphere, not necessarily through the full diameter. Therefore, only the largest measure might serve as an estimate of sphere diameter. In some play-of-color opals from South Australia, Akizuki (1970) observed a ripple-like pattern at low magnification on the SEM (and with an optical microscope). This effect is a manifestation of a fracture that is oriented obliquely to a plane of packed spheres (Gauthier 1986). Internal structure of the spheres. Etched Australian opal spheres commonly show a concentric shell-like structure (Darragh and Gaskin 1966). The spheres range from solid to multishelled, with up to three shells. It has been suggested that the shell-like structure is the result of the deposition around a nucleus (Rau and Amaral 1969). The shells are composed of ~50 nm-sized silica particles (Darragh et al. 1966; Sanders and Darragh 1971), although such particles are not always observed (Dódony and Takacs 1980), and are commonly referred to as primary particles (Darragh et al. 1966; Sanders and Darragh 1971). Similar primary particles were observed in synthetic opal (Sanders and Darragh 1971). Darragh et al. (1966) and Sanders and Darragh (1971) showed that the concentric structure is typical of opals from the Central Australian fields, and stated that “American opals or equivalent opals” (which may mean opal-CT) do not exhibit such a structure. However, Sanders (1976) described a star opal from Idaho in which the spheres have six shells. Dodóny and Takacs (1980), on the other hand, reported that silica spheres in opals from Cervenica, Slovakia showed neither a layered internal structure nor any primary particles. Most observations described above (which were done primarily in the 1960s or the 1970s) were performed on replicas, not directly on opal. This approach might introduce artifacts or might not preserve certain extremely fine details. Relatively few studies have been reported using direct observation on fresh opal surfaces (e.g., Gauthier 1986; Fritsch et al. 1999, 2002, 2004, 2006). Structure of opal-CT Opal-CT is often considered a transitional material between opal-AG and quartz (Flörke et al. 1976), as is the case for biogenic opals (e.g., O’Neil 1987), such as those formed from diatoms (e.g., Clarke 2003). Most previous structural studies of opal-CT were done on marine biogenic opals. Biogenic and sedi-
mentary opal-CT is made of 2–5 µm-sized lepispheres (Flörke et al. 1975) that can be arranged into a diffracting network (Fritsch et al. 2002). In addition to the lepispheric structure, opal-CT can exhibit spherulitic or fibrolitic structures (Heaney et al. 1994 and references therein). Fritsch et al. (1999, 2002, 2004, 2006) showed that several varieties of opal-CT are composed of 10–40 nm particles that can give rise to granular (especially for fire opals) or fibrous (in some common pink opals) structures.
matErials and ExPErimEntal mEthods Materials Samples of common opals were chosen to cover the widest range of gem materials found in the trade. They come from mines in Mexico (77 samples), Ethiopia (18), Australia (18), Honduras (13), France (8), Turkey (8), Slovakia (7), Brazil (6), Mali (6), Tanzania (6), Venezuela (5), Kazakhstan (3), Madagascar (3), Peru (3), the U.S.A. (3), Austria (2), Serbia (1), and Tchequia (1). Their body colors span almost the entire visible spectrum (blue, green, yellow, orange, red, pink, white, brown, gray, and black), and range from transparent to opaque. As far as was possible for each locality and type of material, a majority of samples deemed typical were analyzed along with some unusual ones. Opal-CT samples come from all the countries cited above, and common opal-AG samples are from Australia, Honduras, Mexico, Slovakia, France, the U.S.A., Austria, and Brazil. Samples came from volcanic as well as sedimentary environments. A complete list of sample descriptions is in Appendix 1. Part of our study involved a contract with ECOS to study Mexican opals, which explains the large number of samples from that country. Opals were collected in the field during four trips to the Mexican high plateaus. Two samples from Australia (no. 733 and 824) with slight play-of-color were selected to show the subtle differences at the microscopic scale between common and play-of-color opals.
Experimental methods Samples were identified as opals based on their gemological properties, including a specific gravity between 1.97 and 2.2, and index of refraction between 1.42 and 1.46. They were further classified as opal-AG or -CT based on the position of the main broad Si-O-Si band in Raman scattering spectra, 423 ± 17 cm–1 for opal-AG and 335 ± 11 cm–1 for opal-CT (Ostrooumov et al. 1999; Rondeau et al. 2004). A refractometer was used to measure the index of refraction with an optical contact liquid having n = 1.79. The specific gravity was measured by the hydrostatic method. The variety of opal (-A or -CT) was determined with a Bruker RFS 100 Fourier transform Raman spectrometer using operating conditions described by Smallwood et al. (1997) and Ostrooumov et al. (1999); 1000 scans (Appendix 11) were accumulated at a power of 350 mW and a resolution of 4 cm–1, using a 1064 nm Nd YAG laser for excitation, which eliminates most luminescence. High-resolution images were obtained with a JEOL 6400 SEM equipped with a field-emission electron gun, using a current of 7 kV and 6 × 10–11 A. Two types of samples were imaged for each opal: freshly fractured, and fractured followed by etching in 10 vol% HF for 30 s (standard etch to reveal opal microstructures). Samples were coated with ~5 nm of a gold-palladium alloy.
rEsults Structure of common opal-AN Opal-AN is rarely used as a gem. The results from the two samples that we analyzed (from Bohemia, Czech Republic, and Mexico) are consistent with those of Langer and Flörke (1974). Opal-AN is typically botryoidal and colorless. The fracture surDeposit item AM-08-053, Appendix 1. Deposit items are available two ways: For a paper copy contact the Business Office of the Mineralogical Society of America (see inside front cover of recent issue) for price information. For an electronic copy visit the MSA web site at http://www.minsocam.org, go to the American Mineralogist Contents, find the table of contents for the specific volume/issue wanted, and then click on the deposit link there. 1
GAILLOU ET AL.: STRUCTURE OF COMMON GEM OPAL
1867
faces appear smooth in SEM images, without obvious micro- to nanostructural features either on freshly broken surfaces or after HF etch, and are similar to those of Libyan Desert glass (Fig. 1). Structure of common opal-AG
FiGurE 1. The smooth texture of the freshly broken surface in (a) opal-AN (“hyalite”) from Tchequia (Bohemia, no. 942) compared to that of (b) Lybian desert glass (both images 30 000×).
Arrangements of spheres not giving rise to play-of-color. We observed several conditions for which opal-AG does not exhibit play-of-color. Different size spheres. Our study determined that 48% of the common (no play-of-color) opal-AG specimens consist of silica spheres that are not of uniform diameter. This finding is consistent with results from previous studies concluding that polydisperse spheres are the primary cause of lack of play-ofcolor in opals (Darragh and Gaskin 1966; Rau and Amaral 1969; Sanders and Darragh 1971). We observed that a variation of only 5% in sphere diameters is sufficient to preclude diffraction. In one Mexican common opal sample (Fig. 2a), diameters range
F i G u r E 2. SEM images (30 000×) of common opal-A samples, after HF etching (except e and f). (a) Opaque orange opal-AG from Mexico (Mina Iris, Queretaro, no. 759). The silica spheres vary in diameter from 2 to ~250 nm. Both small and large spheres show concentric structures, and the larger the sphere, the more numerous the layers. (b) Opaque orange opalAG from France (SaintNectaire, no. 950) with spheres ~6.5 to 7.5 µm in diameter. Note the overall botryoidal appearance. (c) Gray opal-AG from Australia (Lightning Ridge mine, New South Wales, no. 235C) showing spheres that are not spherical, but commonly elongated, leading to an imperfect packing, which cannot diffract light. (d) Transparent opal-AG from Honduras (no. 684C) consisting of well-ordered spheres too large (~650 nm) to diffract light. Concentric layering is seen in the center of spheres and a radial structure in the rims. (e) Fire opal-AG from Slovakia (Dubník, no. 637) consisting of spheres that are not ordered and are too small (~80 nm in diameter) to diffract light. (f) White opal-AG from Honduras (no. 671) with spheres ~280 nm in diameter (adequate for a red play-of-color), but do not show a regular arrangement.
1868
GAILLOU ET AL.: STRUCTURE OF COMMON GEM OPAL
from ~250 to 2000 nm. Similar, large variations were observed for ~23% of the samples, especially for opals from Honduras, France, and Slovakia. Interestingly, our results revealed that contrary to conclusions from the earlier studies, which were based on a limited number of samples, most Australian common opals have monodisperse spheres. In the rare cases where the silica spheres in Australian common opals are polydisperse, the variation in their diameters is only 5–15%. Yellow to orange opaque opals from Austria (called forcherite) and France (Saint-Nectaire) consist of exceptionally large spheres with diameters ranging from ~2000 to 8000 nm, which are the largest we observed. The spheres commonly coalesce to build botryoidal-like structures (Fig. 2b). Imperfectly shaped spheres. Our study indicates that common opal-AG with imperfectly shaped spheres is more abundant than originally believed (Sanders and Darragh 1971). We observed this type of irregularity in ~25% of our samples, but only in those from Coober Pedy, South Australia and the Lightning Ridge area, New South Wales, Australia (Fig. 2c). Spheres in these opals are commonly elongated, making orderly stacking impossible. The fact that they never appear broken implies that
they were soft when they were stretched, or that they originally formed in that shape. Spheres that are too large or too small. We found only one common opal-AG, from Honduras (sample no. 684), in which spheres are in a well-packed arrangement but have diameters too large (~650 nm) to diffract light (Fig. 2d). In theory (Sanders 1964), spheres that are too small (
