M^ord to theW^ise

JOHN RAKOVANDepartment of Geology

Miami UniversityOxford. Ohio 45056

rakovajf@muohio.edu

Figure I. Rutile epitaxic on hematite, Novo Horizonte, BahiaBrazil. The orientation relationship hetween the rutile andhematite is the same as that shown in figure 2. HoustonMuseum of Natural Science specimen, Julian Gray photo.

Many mineral specimens—and by definition almostall rocks—are intergrowths of crystals. Crystalsin an intergrowth arc usually randomly oriented,

but those exceptions, where we find nonrandom or crystal-lographically controlled intergrowths, are fascinating andoften quite beautiful in their symmetry. One of several typesof nonrandom (or regular) cry.stal intergrowths is epitaxy.

Generally speaking, epitaxy is the oriented overgrowthof one mineral or crystalline material on another. In thiscase, specific crystallographic directions of the substrateand overgrowth are aligned. This occurs when certain struc-tural planes in the overgrowth and the substrate have similarspacing.s between atoms. In such instances the directionsof similar atomic spacing in the two crystals prefer to bealigned rather than randomly oriented. There are manyinstances of epitaxy in the mineral kingdom. A few com-mon examples are pyrite on marcasite, rutile on hematite(figs. 1, 2), and chalcopyrite on sphalerite (figs. 4, 5). It isnot difficult to imagine why two isomorphic minerals (samestructure but different chemistry) will have similar atomicspacings and will form epitaxic relationships. Examples ofthis include albite epitaxic on microcline and pyromorphiteepitaxic on apatite. In some cases the two minerals involved

Dr. John Rakovan, an executive editor o/Rocks & Minerals,is a professor of mineralogy and geochemistry at MiamiUniversity in Oxford, Ohio.

Figure 2. Schematic of rutile expitaxic on hematite formed dueto structural meshing hetween the (100) planes of rutile and the(001) planes of hematite. In this case, the f-axis of every rutilecrystal is parallel to one of the three equivalent rt-axes of thehematite, and the hematite c-axis is parallel to the rutile n-axis.Modified from Herting-Agthe (2006).

in an epitaxic relationship are polymorphs (i.e., they havethe same composition but different atomic arrangements,such as pyrite and marcasite or sphalerite and wurtzite). Yet,in many cases both the chemistry and .structure of the twominerals are different; however, along a certain few direc-tions in the two crystals there is similar spacing betweentheir constituent atoms.

Recognition of epitaxy between two crystals can be easyif the substrate and overgrowth have well-developed mor-phologies from which the orientation of the crystallographicaxes are readily apparent. Thus, if you can see that the axesof one crystal are parallel to those of the other, you knowthat an oriented relationship exists between the two. Inmany mineral samples, you find numerous small crystalsof the overgrowth on a single larger substrate crystal (figs.1, 4). Often the crystals in the overgrowth are physicallyseparated from one another. If you rotate the sample to aposition where light reflects directly off a face on one of theovergrowth crystals and into your eye {specular rejlection),you will notice that many or all of the other crystals are giv-ing specular reflection at the same time (fig. 4). For this tooccur, all of the overgrowth crystals must have a similar ori-entation. If they are physically separated from one anotheron the substrate surface, their mutual orientation is often,but not always, due to an epitaxic relationship between theovergrown crystals and the substrate.

Epitaxy, in the strictest sense, refers to the oriented rela-tionship between two different minerals or crystalline mate-

Volume 61, )uty/August 2006 317

Solution

Homogeneous nucleus

Heterogeneous nucleus

MarcasJte

Figure 3. Schematic of a pyrite nucleus freely suspended insolution {homogeneous nucleus) and a pyrite nucleus thathas formed on the surface of a pre-existing marcasite crystal(heterogeneous nucleus).

rials, such as the examples given above. However, it hasbecome commonplace in the physics and materials sciencecommunities to use the term when referring to overgrowthsof the same material as the substrate. This is especially truewhen describing an overgrowth and substrate that havesome property that differs dramatically, such as their electri-cal conductivity or electronic structure (e.g., the band gap insemiconductors). A property such as electrical conductivitycao change dramatically if a crystal is doped (partially sub-stituted) with an impurity. For example, pure diamond is aneiectricai insulator; however, when a small number of boronatoms substitute for carbon in the structure it becomes asemiconductor. Interestingly, this change in electronic struc-ture, is what gives boron-doped diamonds their blue color.To distinguish between the two, homoepitaxy is sometimesused to describe overgrowths of the same material andheteroepitaxy to describe an overgrowth on a different mate-rial. The International Mineralogical Society has sanctionednomenclature describing mineral intergrowths, includingepitaxy (Bailey 1977)"; it restricts the use of epitaxy to caseswhere the two minerals involved are different species.

The article on sakura ishi in this issue (pages 284^292)describes the formation mechanism of these unusual micapseudomorphs after cordierite-indialite intergrowths. Theyoriginate as multiple epitaxic overgrowths of cordierite onthe prism faces of indialite. The cause of this and otherexamples of epitaxy is related to the energy involved in theformation of a crystal nucleus, or embryonic crystal, andhow nuclei form. It turns out that there is an energy bar-rier, an impediment, to the formation of crystal nuclei fromany medium, be it a solution, melt, or vapor. This barrier isthe result of certain interactions between the surface of the

Figure 4. Chalcopyrite epitaxic on sphalerite, Ballard mine,Treece, Kansas. The orientation relationship between the twominerals is the same as that shown in figure 5. Author'sspecimen and photo.

nucleus and the surrounding growth medium. The conse-quence of these interactions is that it takes the addition ofenergy for a crystal nucleus to form.

One way to lower the energy barrier associated with nucle-ation is to decrease the amount of surface area of the crystalnucleus in contact with the growth medium. This occurs ifthe nucleus forms on the surface of a pre-existing solid sub-strate. The substrate can be any type of solid, including glass,organic materials, or crystalline solids, in such a case, partof the nucleus is in contact with the substrate; thus, the sur-face area in contact with the growth medium is diminishedcompared to a freely suspended nucleus of equal size (fig. 3).This type of nucleation is referred to as heterogeneous nucle-ation., whereas the formation of a freely suspended nucleusis referred to as homogenous nucleation.

The interaction between a heterogeneous nucleus andits substrate also creates an energy barrier to its forma-tion. However, this is less than the barrier resulting frominteraction with the growth medium; thus, heterogeneousnucleation requires the addition of less energy than homo-geneous nucleation. Epitaxy occurs because of a special kindof heterogeneous nucleation in which the nucleus takes aspecific (nonrandom) orientation as it forms on a crystallinesubstrate. This occurs because in certain situations there isa smaller barrier to formation for an oriented overgrowthcompared to one of random orientation. These special situ-ations arise when one or more of the structural dimensions

•Although some use epitactic as the adjectival form of epitaxy, the editorsof Acta CrystaUographica and the Journal of Applied Crystallography prererepitaxic for etymological reasons (Schneider i%3J, The use of epitaxial,however, is not acceptable.

318 ROCKS & MINERALS

Cu Fe

chalcopyritesphalerite

Zn

in the surface of the substrate is very close to (usually within10 percent) one or more of the structural dimensions in theovergrowth (Sunagawa 2005). In such a case, the directionsof similar structural dimension in the two phases want to bealigned. Hence, there is a control on the orientation of theovergrowth as it forms. For example, in the case of rutile andhematite (figs. 1, 2), there are directions of similar spacingbetween the atoms in the (100) plane of rutik and the (001)plane of hematite. If rutile heterogeneously nucleates onhematite, these directions prefer to line up with one another.This results in the a and c erystallographic axes of the rutileovergrowth being parallel to the c and one of the three aerystallographic axes of hematite, respectively.

Another common epitaxy found in nature is chalcopyrite(CuFeS,) oriented on sphalerite (ZnS) (fig. 4). Classic exam-ples inciude those from the Mississippi Valley-type depositsof the Tri-State district in the United States (Rakovan 2006).It is instructive to look at the atomic arrangement of theinterface between two minerals in an epitaxic relationshipto see how they mesh. Figure 5 is a ball-stick model of theinterface between the chalcopyrite and sphalerite. Spheresrepresent atoms, and sticks show the bonds between them.The three-dimensional structure of these two minerals isvery similar. In sphalerite each zinc atom is bonded to foursulfur atoms in a tetrahedral arrangement. Likewise, thecopper and iron atoms In chalcopyrite are in tetrahedralcoordination, and the linkages between tetrahedra are thesame in the two structures. The atomic planes that definethe interface shown in figures 4 and 5 are the (111) ofsphalerite and the (112) of chalcopyrite. If chaicopyrite het-erogeneously nucleates on a (111) face of sphalerite, all ofthe copper and iron atoms at the interface can retain theirtuU tetrahedral coordination by bonding to sulfur atoms onthe sphalerite surface. This is preferable because it decreasesstrain at the interface, but it is only possible if the chalcopy-rite grows in a specific orientation relative to the sphalerite.This orientation is one in which the similar atomic spacingsof the two minerals are aligned.

In extreme cases the Influence ofthe crystal substrate canbe even more profound than just causing an overgrowth togrow in a certain orientation. The interaction of adsorbedatoms with a surface can cause the overgrowth to assume anatomic structure that would normally not form under the

Figure 5. Structure diagram of theinterface (dashed hlack line) betweenchalcopyrite and sphalerite in an epitaxicrelationship controlled hy meshing ofatomic spacings hetween the (111) planeof sphalerite and the (112) plane ofchalcopyrite. By taking the orientationshown, all of the copper and iron atomsof the chalcopyrite can retain iheir let-rahedral coordination hy forming bonds(dashed gray lines) with S atoms on thesurface of the sphalerite. In this view ofthe structures, only three out of everyfour honds to each Zn, Cu, and Fe atomcan he seen.

figure 6. Chalcopyrite epitaxic on sphalerite, Dal'negorsk,Primorskiy Kray, Russia. Specimen is 7 x 6.1 x 4 cm. CarnegieMuseum of Natural History specimen (#CM27709), DehraWilson photo.

Figure 7. Galena epitaxic on pyrrhotite, Primorskiy Kray, Rus-sia. Specimen is 5.1 x 3.2 x 0.9 cm. Carnegie Museum of NaturalHistory specimen (#CM27654), Debra Wilson photo.

conditions existent during its growth (Bloss 1994). In suchcases another polymorphic structure is more stable underthe conditions in which the overgrowth is formed. Thus, the

Volume 8 1 , July/August 2006 319

structure that does form because of the epitaxic control issaid to be metastable.

Mimicking what is found naturally has led to technologi-cal advances, and epitaxy is the foundation of the design andmanufacture of many materials and devices used in industrytoday, such as semiconductors and photonic devices. In fact,epitaxy is the only affordable method for the high-qualityproduction of many semiconductor materials, including thetechnologically important materials gallium arsenide andindium phosphide that are used extensively in telecommu-nications devices. The growth of doped (containing addedimpurities) silicon on the surface of pure silicon crystals isan example of homoepitaxy that leads to the formation of asemiconducting material commonly used in power devicesfor a wide range of products, from pacemakers to vendingmachine controllers. As mentioned above, the interaction ofadsorbed atoms with a surface can result in an overgrowthassuming a metastable atomic structure. This has greatpotential in different technological applications. The forma-tion of a metastable polymorph of a particular material byepitaxic growth may be advantageous because that poly-morph cannot be synthesized homogeneously. For example,epitaxic stabilization allows synthesis of ferromagnetic, fer-roelectric BiMnO, films that cannot be nucleated homog-enously at ambient pressures (Santos et al. 2002; Ohshima etal. 2000). Indeed, without an understanding of epitaxy and

the ability to create it in advanced materials, the world welive in would be very different today.

ACKNOWLEDGMENTSI thank John laszczak, Pete Richards, and Bob Feldman for their

thorough reviews of this column. I also thank Jiilian Gray and DebniWilson for the photographs and Bill Diimeron. Richard Dietrich,Tony Kampf, Mark Mauthner, Peter Modreski, Bob Ramik, ArtSmith, and John Wliite for their helpful discussions on the topicand its presentation to a broad audience.

REFERENCESBailey, S. W. 1977. Report ofthe LM.A.-LU.Cr. Joint Committee on

Nomenclature. American Mineralogist 62:411-15.Bloss, F. D. 1994. Crystallography and crystal chemistry. Washington,

D.C; Mineralogical Society of America.Her ting-Agthe, S. 2006. http://www. mineral ogische-sammlungen.

de/strohstern-engl.htm.Ohshima, E., Y. Saya, M. Nantoh, and M. Kawai. 2000. Synthesis

andmagneticproperty of theperovskiteBi|^Sr^MnO, thin film.Solid State Communications 116 (2): 73-76.

Rakovati, ]. 2006. Word to the Wise—Mississippi Valley-typedeposits. Rocks & Minerals 81:69-71.

Santos, A. M., S. Parashar, A. R. Raju, Y. S. Zhao, A. K. Cheetham,and C. N. R. Rao. 2002. Evidence for the likely occurrence ofmagnetoferroelectricity in the simple perovskite. BiMnOj. SolidState Communications 122:49-52.

Schneider, H. G. 1963. Bemerkung zur Terminologie des BegriffesEpitaxie. Affrt CrystaUographica 16:1261-62.

Sunagawa, L 2005. Crystals: Growth morphology ami perfection.Cambridge, NY; Cambridge University Press. Q

OcTOBER131415

GREATER DETROITGEM, MINERAL, FO//IL

& JEWELRY/HOWPresented by the Michigan Mineralogical Society

www.mtchmin.org

Dazzling Exhibits from theTop Museums of the US and CANADA

Over 50 Choice Dealers

FeaturingSome of the

Best Copper CrystalsTo come out

Of Michigan'sCopper Country

"MARVELOUS MICHIGAN"

Macomb Community College Expo Center, bIdg. P,South Campus, 12 Mile & Hayes, Warren, Ml

320 ROCKS & MINERALS