Showing papers in "American Mineralogist in 1991"

Alpha-decay event damage in zircon

Abstract: Based on density measurements, X-ray diffraction analysis, and high-resolution trans-mission electron microscopy of a suite of natural zircon samples from Sri Lanka (0.06 xl0r5 to 6.3 x l0rs a-decay events/mg), three stages of damage accumulation may bedelineated. Stage I ( 8 x 10'5 a-decayevents/mg) consists of material that is entirely aperiodic as far can be determined byX-ray or electron diffraction. There was no evidence for the formation of ZtO, or SiO, asfinal products during the last stage of metamictization. Based on rnodeled density changes,aperiodic regions continue to experience a change in structure as they are redamaged.During stage II of the process, the modeled density of aperiodic regions changes from 4.5g/crn3 lo 4.1 g/cm3. Fission fragment damage does not contribute to the process of meta-mictization. The amorphization process is consistent with a model for the multiple overlapof displacement cascades, suggesting amorphization occurs as a result of defect accumu-lation rather than directly within a single displacement cascade.Comparison of results for natural zircon with those for Pu-doped zircon showed thatdose-rate variations (even as great as a factor of 108) had no substantial effect on thedamage accumulation process. Unit-cell parameters increased and density decreased morefor the Pu-doped zircon than for natural zircon in early stages of damage accumulation(<3 x 10,5 a-decay events,/mg), suggesting that annealing of point defects in the earlystages of the damage accumulation process occurs in natural zircon under ambient con-ditions. This accounts for the distinct sigmoidal shape of the damage cnryes for naturalzircon and the apparent incubation period before onset ofamorphization.IxrnonucrroN ppm U and2to 2000 ppm Th (Ahrens etal.,1967;Gorz,Zircon (ZrSiOo, I4r/amd, Z : 4) is a common acces- 1974), it is one of the most important phases used insory mineral in crustal igneous, metamorphic, and sedi- U-Th-Pb dating techniques (Faure, 1977;JiLger and Hun-mentary rocks, as well as in lunar materials, meteorites, z1ke4 1979). Of particular interest are processes by whichand tektites. Because zircon typically contains 5 to 4000 the U-Th-Pb systematics are disturbed (e.g., Silver and0003-004x/9 1/09 l 0- l 5 I 0$02.00 l 5 r0

Pressure-induced silicon coordination and tetrahedral structural changes in alkali oxide-silica melts up to 12 GPa: NMR, Raman, and infrared spectroscopy

Abstract: The 29 Si and 23 Na NMR, Raman, and infrared spectra of glasses quenched from liquids at pressures up to 12 GPa for three alkali silicate compositions (Na 2 Si 2 O 5 , Na 2 Si 4 O 9 , K 2 Si 4 O 9 ) and silica (SiO 2 ) reveal systematic changes in the melt structure with pressure. The most novel change is the occurrence of [5] Si and [6] Si species at high pressures identified by peaks in the 29 Si MAS NMR spectra of the alkali silicate glasses. -from Authors

Plagioclase-melt equilibria in hydrous systems.

Abstract: Although plagioclase is the most common minera! in crustal rocks, plagioclase-melt equilibrium relationships in hydrous silicate systems are poorly understood. This study is based on analyses of 54 new plagioclase-melt pairs produced in HrO-saturated experiments on natural andesite and basalt in the pressure range of 1-4 kbar.We combined these data with Hro-saturated data from the literature, and we evaluated the quality of available activity-composition models for silicate melts by calculating equilibrium constants for the exchange of albite and anorthite components between plagioclase and melt as functions of reciprocal temperature. Plagioclase-melt equilibrium relationships are shown to be strongly dependent on both the temperature and the HrO content of the melt. Because the HrO content of natural melts is generally not known, we present two expressions that relate the temperature of the system to the HrO content ofthe melt, one for albite exchange and the other for anorthite exchange. Successful application ofthese relationships to natural rocks requires both a good estimate of preeruptive temperature and knowledge of the equilibrium plagioclase and melt compositions. Under optimal conditions, the uncertainties in estimated HrO contents are 0.54 and 0.33 wto/o for the albite and anorthite exchange equations, respectively.

Single crystal Raman spectra of forsterite, fayalite, and monticellite

Abstract: Polarized single crystal Raman spectra of the fundamental modes offorsterite (Mg2Si04), olivine (M&.88,Feo.12),fayalite (Fe2Si04), and monticellite (CaMgSi04) are presented. Seven low energy modes «450 cm-I) in forsterite differ from those of all previous studies. All modes predicted by symmetry for forsterite and monticellite were observed; 34 out of 36 modes were observed for fayalite. Assignment of the modes was determined by systematics in frequency changes resulting from cation substitution. Although the lattice modes are generally mixed, likely mode assignments are made using the following observations. The lowest energy modes are assigned to Si04 translations and appear to be mixed with the cation translations. The lattice modes between 330 and 435 cm-I in forsterite that changed the least with composition were assigned as Si04 rotations. The modes from 300 to 390 cm-I that showed the greatest variation with composition or showed signs of two-mode behavior in the olivine Mg88were assigned as M2 translations. The Si04 internal stretching and bending modes were assigned to the highest frequencies; they vary little from forsterite to monticellite to fayalite and appear to depend more on cation mass than volume. These assignments are consistent with previous single-crystal infrared studies of forsterite and fayalite and with mode Griineisen parameters 'Yimeasured vs. pressure in that the highest 'Yivalues should be associated with the M2 cations and the lowest with the Si04 internal modes. INTRODUCTION important modes are those below 450 cm-I, as the estiThe vibrational modes of olivines, important geophysmat~s of heat capacity and entropy vs. tem~~rature at ical and cosmochemical phases, have been studied extenambIent temperatures and below are very sensItIve to the sively by Raman spectroscopy (Iishi, 1978; Servoin and number of these mo~es and their distribution. These are, Piriou, 1973; Piriou and McMillan, 1983; Stidham et aI., h,owever, the most dIfficult to measure because of exten1976; Hohler and Funck, 1973; Chopelas, 1990), infrared ~IV~over~apping: breadth ofp~aks a~~ t~o-mode behavspectroscopy (Hofmeister, 1987. Tarte 1963. Paqueslor In sohd solutIons, and low IntensItIes In Raman specLedent and Tarte, 1973; Kovach 'et aI., 1'975; Hofmeister tra because ~f the low polarizabilities of the octahedral et aI., 1989; plus some references given above for Raman structural unIts. spectroscopy), inelastic neutron scattering (Rao et aI., T~e polarized ,Raman spectrum of forsterite has been 1988), and lattice dynamical calculations (Price et aI., prevl0.u~ly descnbe~ (~ohler and Funck, 1973; Servoin 1987a, 1987b). Study of these properties yields insight and Plnou, 1973; IIS~I, 1978), bu~ none, of the results into the interatomic forces within the crystal structure agree for the ~r~qUencies o,f the low IntensIty modes. The and a basis for understanding the effect of cation substionly other ohvlne for whIch a (nearly) complete singletutions on the thermodynamic properties. For example, crystal, Rama~ spectrum has been des~ribed is tephroi,te the heat capacity and entropy can be estimated to better (Mn2SI04) (StIdham et aI., 1976); partIal data are avallthan 50/0over a moderate temperature range (, , 500 K, a~l~ for Ca-beari~g phases (e.g., ~ohler and Funck, 1973; Kieffer, 1979) from the mode frequency information ob~lnou and McMIllan, 1983). ThIS study was prompted, tained by infrared and Raman spectroscopy. The accuIn part, by the general lack of Raman data in the low racy in estimating these quantities can increase to better frequency range for other olivines and, in part, by disthan 0.5% over a larger temperature range (> 1000 K) if crepancies with the previously reported data for forsterite all of the infrared and Raman modes are enumerated and (Chopelas, 1990). assigned to atomic motions. The detailed mode assignIn this study, single-crystal polarized Raman spectra of ments provide an accurate method for estimating the freforsterite, olivine (Mgo.88,Feo.12)2Si04,monticellite, and quencies of inactive or unobserved modes, thereby yieldfayalite are presented. Mode assignments to molecular ing more precise information on the frequency distribution motions are made by comparison of analogue modes in or density of states (see Hofmeister, 1987; Hofmeister these olivines and previous results of tephroite (Stidham and Chopelas, 1991a; Chopelas, 1990, 1991). The most et aI., 1976) and 'Y-Ca2Si04 (Piriou and McMillan, 1983). 0003-004X/91/0708-1101$02.0

The hydrous components in garnets; grossular-hydrogrossular

Abstract: Grossular garnets from 33 localities were examined for indications of OH- or H_2O in their infrared spectrum. All contained OH-. Classical hydrogrossular with more than 5 wt% H_2O displays systematics pectroscopic behavior consistent with the hydrogarnet substitution consisting of two absorption bands at 3598 and 3662 cm^(-1). These spectroscopic characteristics were generally not observed in other glossular samples. Instead, 20 distinct absorption bands have been identified in the spectra of common grossular, occurring in groups of four to ten bands. Both the number and intensities of these bands show a large variation, which does not correspond with the garnet's composition. Seven classes of spectra were identified in the OH region based on the position of the most intense absorption band. The spectroscopic data suggest that in addition to the tetrahedral site, OH groups exist in multiple other environments. The OH content of grossular garnets can be obtained from infrared spectra using the equation H_2O wt% = 0.0000786 x integrated absorbance per cm in the OH region near 3600 cm^(-1). The OH content of macroscopic grossular crystals (expressed as weight percent H_2O) ranged from 12.8% doown to less than 0.005%. Macroscopic grossular typically contains less than 0.3 wt% H_2O. Grossular from rodingites and low-temperature alteration vugs contained much more than that from skarn or contact metamorphic environments.

Static lattice energy minimization and lattice dynamics calculations on aluminosilicate minerals

Abstract: Lattice energy minimization and lattice dynamics calculations for the minerals andalusite, sillimanite, kyanite, diopside, cordierite, gehlenite, leucite, orthozoisite, grossular, and pyrope are compared with experimental data and previous calculations. The potential models used in this study included bond-bending interactions, short-range Born-Mayer forces, effective dispersive interactions, long-range Coulomb interactions, and harmonic core-shell interactions for the 0 ions.. Parameters for the potential models were generally taken from the literature, but the core-shell force constant was modified to give better agreement with experimental data for refractive indices. It was necessary to include bondbending interactions for AI-O polyhedra with coordination numbers even larger than four. A method for describing effective potentials with Al-Si disorder and solid-solution is presented. Modified Morse and Buckingham potentials were used to model O-H bonds. Relaxed energy-minimum structures were calculated, allowing cell parameters to change and treating atomic cores and shells as independent entities within the adiabatic approximation. Calculated phonon frequencies for the relaxed structures were used to construct thermodynamic functions. Elastic and dielectric constants were also calculated. Comparisons between calculated structures and other properties with experimental data have shown that the model is genuinely transferable and gives reasonable predictions of crystallographic, physical, and thermodynamic properties. Detailed analysis gives a measure of the reliability of the model.

The metamictization of zircon: Radiation dose-dependent structural characteristics

Abstract: A suite of gem zircon samples from Sri Lanka has been studied using infrared (IR) spectroscopy, X-ray diffraction, and chemical analysis. The degree of metamictization of the zircon, as indicated by unit-cell parameters, increases systematically with U-Th content up to the point of total metamictization. The appearance of IR spectra also varies as a function of metamictization; band widths increase and intensities decrease with increasing U-Th contents. Persistence of bands related to Si-O bonds and disappearance of bands related to Zr-O bonds indicates that the structure of metamict zircon consists of distorted and disoriented isolated silica tetrahedra with few if any undisplaced Zr cations. All spectroscopic indicators of crystalline order show that total metamictization is reached at an accumulated radiation dosage of ~4.5 x 10^(15) alpha decay events per mg. Hydrous components enter the structure only after total metamictization, but the amounts are not correlated with U-Th content. In all cases OH was the only hydrous species detected.

Alpha-decay damage in titanite

Abstract: Titanite can incorporate minor amounts of radioactive impurity components (particularly U and Th) that affect the crystal structure by ɑ- and β-decay events; in particular, ɑ-decay gives rise to significant knock-on structural damage by causing atomic displacements. We have examined a series of chemically and structurally well-characterized titanite samples by a variety of techniques to follow the progress of the metamictization process at low radiation doses. Ten titanite samples were characterized by electron microprobe analysis, powder X-ray diffraction, and powder infrared (IR) spectroscopy. The X-ray diffraction patterns vary from sharp and well resolved to almost totally degraded, reflecting the various degrees of ɑ-decay damage. The powder IR spectra show a similar variation, and the order of increasing pattern degradation is almost the same as that for the X-ray diffraction patterns, indicating that both features reflect the same physical property of the material. However, the annealing behavior is different: powder X-ray diffraction patterns become sharp and well resolved; powder IR spectra sharpen slightly, but do not recover to anywhere near the same extent. Four representative samples were selected for further work. The crystal structures were refined using MoKɑ X-ray single-crystal diffraction data. The crystals were then annealed at 1090 °C under Ar and the intensity data were again measured. For small degrees of ɑ-decay damage, the structure seems to be completely restored on annealing; this is not the case for titanite with the largest amount of ɑ-decay damage. Polarized single-crystal IR spectra of undamaged titanite show a single sharp (OH) stretching band at ~3490 cm^(-1) with a little fine structure reflecting local cation disorder around the OH. With increasing ɑ-decay damage, the sharpness of the absorption band decreases and a wide wing appears on the low-energy side of the sharp (OH) band. Mossbauer spectroscopy shows only Fe^(3+) to be present in undamaged titanite; as ɑ-decay damage increases, the amount of Fe^(2+) increases, suggesting that radioactive decay causes reduction as well as atomic displacement. Fe^(2+) is easily oxidized on heating. The ^(29)Si MAS-NMR peak width is strongly correlated with increasing radiation damage and increasing Fe content, and no signal was observed from the most damaged titanite. With increasing ɑ-decay damage, single-crystal electron diffraction patterns develop diffuse halos indicative of a mean atomic spacing of 3.6 A. In bright field, undamaged material shows continuous lattice fringes. Small amounts of damage are characterized by mottled diffraction contrast superimposed on largely continuous lattice fringes. The most damaged titanite shows mottled diffraction contrast with coexisting crystalline and aperiodic domains produced by overlapping ɑ-recoil tracks, corresponding to damage on the order of 30-50% of that required to render the structure fully aperiodic. Ti XANES spectra show intensification of the principal pre-edge feature (ls → 3d transition) with increasing damage, indicative of increasing local asymmetry around the Ti position. Loss of resolution in the EXAFS spectra also indicates increasing disorder around Ti with increasing damage. There is no sign of any ^([4])Ti in even the most damaged samples, although it was detected in a glass of titanite composition. The metamictization process begins by the formation of isolated ɑ-recoil and ɑ-particle tracks. With increasing dose, the ɑ-tracks begin to overlap, producing aperiodic domains; in the most damaged titanite examined, there were approximately equal amounts of coexisting crystalline and aperiodic material. At this stage, the crystalline domains still retain their original orientation, except where affected by low-temperature annealing. As undamaged titanite does not (usually) contain significant Fe^(2+), it seems that the damage process is accompanied by reduction of Fe^(3+) → Fe^(2+), which resides in the aperiodic domains. These domains incorporate much more hydrogen (as OH) than is contained in crystalline titanite, presumably a result of postdamage diffusion of H into the structure. All information is consistent with the Ewing model for metamict materials, an aperiodic random network structure; HRTEM images show patterns of random contrast consistent with the random network model, with no evidence to support any microcrystalline model of the metamict state. High-temperature annealing only partly restores the structure, the apparent degree of recovery being dependent on the coherence length of the experimental technique used to characterize the material. The degree of recovery is also dependent on the amount and pattern of damage. We suggest that the original structure is recovered when the ratio of surface area to volume for the damaged material is high, and the interface can exert a strong memory effect on the amorphous material; when large equant aperiodic domains form, they are annealed to a more defect-free and relatively stable aperiodic network structure.