This is a (100) projection of the crystal structure of clinochlore, a Mg-rich triclinic chlorite of the common polytype IIb4. We have just refined this structure in space group C-1 and been able to associate the Moessbauer peaks with ferrous iron in M1 (lt blue) and M2 (blue-green) in the talc-like layer and with ferric iron in M4 (dark gray) in the bruceite-like layer. Al and Si are disordered over the tetrahedral sites. Note the hydrogen bonds from the hydroxyls of the bruceite-like layer to the oxygens of the tetrahedral layer.
The following manuscript is has just appeared in Clays and Clay Minerals.45, p 544-550 (1997).
1Department of Geological Sciences, University of Colorado, Boulder, CO 80309-0250 2Department of Geology and Astronomy, West Chester University, West Chester, PA 19383 3U. S. Geological Survey, Water Resources Division, 3215 Marine Street, Boulder, CO 80303 4U. S. Geological Survey, Reston, VA 22092
Keywords: Clinochlore, Chlorite, M”ssbauer, Crystal Structure, Cation Ordering
Chlorite is a common to abundant mineral in low- to intermediate-grade metamorphosed mafic and ultramafic rocks. It is particularly abundant in altered sea floor basalts and may provide a vehicle by which water is subducted into the upper mantle. Inasmuch as water is a major cause of partial melting, the thermodynamic stability of chlorite is of significance to models of subduction, dewatering, and magma genesis. Chlorites are also significant components of metamorphic terranes that are impacted by acidified atmospheric deposition in the eastern U.S. and Canada, and thus their weathering characteristics (May et al, in preparation) influence water chemistry and watershed response to acidic inputs. The thermodynamic stability of chlorite is known to be affected by various degrees of ordering of the octahedral and tetrahedral cations, so that rates of weathering for chlorites are affected by cation occupancies of the various structural sites.
In the present study, we seek to characterize the composition and crystal structure of a natural clinochlore using X-ray diffraction, electron microprobe, wet chemistry, and M”ssbauer spectroscopy, with a goal of assessing octahedral and tetrahedral cation ordering. X-ray diffraction data are used to clarify the interpretation of M”ssbauer results, and site assignments from the two independent methods are compared. For this study, a large single crystal sample approximately 4 by 8 cm in size from West Chester PA (USA) was obtained from the Smithsonian Institution. The specimen, NMNH #R4513, was originally described by Clark and Schneider (1890), and was also used in our study of chlorite weathering (May et al, in prep.).
The structural chemistry of the chlorite group has been reviewed by Bailey (1988). The chlorite structure consists of a talc-like 2:1 layer with an octahedral sheet sandwiched between two tetrahedral sheets (Fig. 1). The 2:1 layers alternate with brucite-like interlayers. The talc-like layer has a chemical formula (Mg,Fe)3Al1+xSi3-xO10(OH)2 with a net charge of -(1+x) per formula unit. The brucite-like interlayer has a formula of Mg2-x(Al,Fe,Cr)1+x(OH)6 with a net charge of +(1+x). Although other stacking polytypes are possible, polytype IIb is the most common and exists with monoclinic (C2/m) and triclinic modifications (C). The site nomenclature is the same for both modifications. The tetrahedral sheets are each composed of equal numbers of two distinct tetrahedra, T1 and T2. The octahedral sheet in the 2:1 layer comprises two distinct octahedra, M1 and M2, with M1 at the origin, trans-bonded to OH, and M2 in a general position cis-bonded to OH. Similarly, the brucite-like layer comprises two distinct octahedra, M3 and M4, with M3 in a general position and M4 on an inversion, so that there are half as many M4 sites per layer.
X-ray diffraction studies (Rule and Bailey 1989, Zheng and Bailey 1989, Joswig et al 1980, 1989) have indicated little if any ordering of Al and Si in the two tetrahedral sites, the sites being nearly identical in volume and average T-O distance. MAS NMR studies of both 27Al and 29Si however (Welch et al 1995), have indicated considerable short-range order and Al-O-Al avoidance in the tetrahedral layer, even in synthesized samples, although significant ordering based on the crystallographic distinction between the tetrahedral sites has not been documented. However, X-ray diffraction studies have indicated considerable ordering of octahedral cations. The two octahedra in the brucite-like layer differ considerably in volume, distortion, and mean cation-oxygen distance. Trivalent cations (Al, Cr, and Fe3+) concentrate in the brucite-like layer, creating a net positive charge to balance the net negative charge on the talc-like layer. Further, the trivalent cations tend to concentrate into the smaller M4 site. If the cations are disordered over the two sites, the symmetry can be monoclinic, whereas ordering of trivalent cations into M4 causes a difference (D) in the mean cation oxygen distances of the two sites that strongly correlates with the deviation of a from 90§. Evidence for ordering of Fe and Mg in the octahedral sheet of the talc-like layer is less clear.
M”ssbauer studies have done little to clarify the site assignments of Fe due to confusion in the literature regarding the appropriate number and assignment of doublets in the spectra. This is due largely to the fact that Fe atoms can occupy any of the octahedrally-coordinated sites in the structure, leading to a potential for four distinct doublets corresponding to each oxidation state of Fe in each spectrum. Most M”ssbauer studies report only one or two Fe2+ doublets (Taylor et al 1968, Coey 1975, Ericsson et al 1977, Goodman and Bain 1978, Blaauw et al 1980, Borggaard et al 1982, Kodama et al 1982, Ballet et al 1985, Raclavsk and Raclavsky 1988, Aramu et al 1989, Christofides et al 1994) that are generally assigned to the M1 and M2 sites, but are recognized to be potentially indistinguish- able from M3 and M4 doublets. In many of these studies (e.g., Christofides et al 1994), the authors were primarily interested in Fe3+/Fe2+ ratios, and so did not pursue more elaborate fitting schemes to determine Fe occupancy of the brucite layer sites. A recent study by Gregori and Mercader (1994) fit four doublets to the M”ssbauer spectra of some Argentine chlorites; however, their site assignments suggested Fe2+ in M1, M2, and tetrahedral coordination, with Fe3+ assigned to the brucite-like layer. Four recent studies have modeled both brucite layer and octahedral site occupancy of Fe2+ in the M”ssbauer spectra of chlorite (Townsend et al, 1986, DeGrave et al, 1987, de Parseval et al, 1991, and Pal et al, 1992). Their collective assignments of the M”ssbauer doublets in chlorite spectra will be used in this paper and are given in Table 1. Note that the parameters for Fe2+ cis-M2 and Fe2+ "hydroxide" are very similar, with the quadrupole splitting of the Fe2+ "hydroxide" doublet being only slightly higher than that for Fe2+ cis-M2; this may well have contributed to past confusion in the literature. Only when both doublets are fit to the same spectrum can their distinct contributions be recognized. This makes it almost impossible (without independent data from other techniques) to determine accurate Fe2+ site occupancies in samples where Fe2+ is ordered in either Fe2+ cis-M2 or Fe2+ "hydroxide", but not in both.
The sample was analyzed on the electron microprobe at the University of Colorado. An average of ten spot analyses gave an iron content of 2.81 weight percent with all iron as FeO. Fe contents were variable, ranging from 2.6 to 3.2 wt% FeO. Small (typically less than 10mm in size) opaque inclusions are also visible in the optical microscope. These are sporadically distributed with some areas of the sample free of the inclusions. The average chemical analysis is given in Table 2, with ferrous-ferric ratios determined from M”ssbauer spectroscopy and by wet chemical analysis. H2O content was determined by difference of the oxide total from 100%. The calculated H content of the cell was then 8.2 atoms per formula unit which is indistinguishable from the ideal composition of 8.0. The total of other cations in the formula unit is 9.911, which indicates the possible presence of cation vacancy in the structure.
The Fe2+/Fe3+ ratio was determined by wet chemical methods developed for clay mineral studies in the laboratory of Joseph Stucki at the University of Illinois. His colorimetric method (Stucki 1981, Stucki and Anderson 1981, Komadel and Stucki 1988) includes an organic chromophore during the sample digestion which complexes the iron as it is released into solution, thereby protecting against aerial oxidation. This method has a greater specificity for iron than do oxidimetric methods. In the analyses, 1,10-phenanthroline (phen) is used, which forms a red complex with Fe2+, and a colorless complex with Fe3+. Fe2+ content is then determined colorimetrically using a spectrophotometer. Several aspects of this procedure are significant. First, Stucki's group has done an extensive analysis of the sources of variability in the method. Second, they have demonstrated that the instability of the Fe3+-phen complex, which was problematic in earlier colorimetric analyses, is due to photoreduction and can be eliminated by excluding light. Third, the photoreduction of the Fe3+-phen complex to the Fe2+-phen complex can be carried out quantitatively. Samples are digested while excluding light, Fe2+ is measured colorimetrically, the sample is then exposed to light to reduce all Fe3+, and finally Fe2+ is measured again, this time to obtain total iron. Thus, both Fe2+ and total iron are obtained from a single dissolution. This method gave a total iron content of 2.59 wt% as FeO with 57.4 % of this as Fe2+, and 42.6% as Fe3+. Errors are estimated at less than + 5%. The total iron content obtained by this method is thus slightly lower than the average of those obtained by electron microprobe.
Initial orientation was determined by means of a rotation photograph from which 15 peaks were indexed using an automatic indexing routine. After application of an appropriate transformation matrix a unit cell consistent with C1 or C (triclinic) was obtained. The cell was refined from centered angles of 24 strong reflections with 9§ < 2q < 35§ centered in both positive and negative 2q positions. The average peak widths were about 1.5§ 2q. The refined cell parameters are a = 5.3262(6)Ź; b = 9.2263(13)Ź; c = 14.334(3)Ź; a = 90.559(17)§; b = 97.467(16)§; and g = 89.979(9)§. The basal spacing computed from these values is d001 = 14.212(4)Ź. The cell parameters are consistent with polytype IIb-4 (triclinic).
Using scattering factors for fully ionized cations and O-1 in the refinement program RFINE (Finger and Prince 1975), the Fe and Mg occupancies of the octahedral sites were refined so that total Fe occupancy was constrained to that of the microprobe chemical analysis (Table 2). The occupancy of M3 refined to 0.996 + 0.004 Mg and M4 to 0.918 + 0.005 Al (+Mg) and 0.082 Fe. The occupancy of M1 refined to 0.040 + 0.005 Fe and M2 to 0.048 Fe, with all sites constrained to full total occupancy. This model gave an anomalously small equivalent isotropic atomic displacement parameter 0.45 for M4. Also, the total amount of Al and Mg are somewhat larger than indicated by microprobe chemical analysis. The chemical analysis indicates a slight deficiency in octahedral cations. The X-ray structure refinement method can only determine the number of electrons at a given site, so that if there is the possibility of octahedral vacancy, the method cannot give a unique result. The X-ray results thus indicate that the number of electrons in the octahedral sites decreases as M4 > M2 ~ M1 > M3. If we arbitrarily assign all trivalent cations (except tetrahedral Al), plus the trace Ti, to M4, the smallest of the octahedral sites, we are left with a total occupancy at this site of 0.95. If we incorporate this 5% vacancy at M4 into the model and refine occupancies with constraints otherwise as above, we obtain a model that is consistent with the M”ssbauer spectroscopy and wet chemical analyses, and within the range of Fe contents observed by electron microprobe. This model gives Fe occupancies of 0.034 + 0.006 for M1, 0.040 + 0.004 for M2, 0.004 + 0.004 for M3, and 0.110 + 0.006 (= 0.102Fe3+ + 0.004Cr + 0.004 Ti) for M4. The remaining light cation (Al or Mg) at M4 agrees with the amount of octahedral Al (Table 2). Also, the equivalent isotropic atomic displacement parameter (B = 0.77) is consistent with those of the other M sites. The amount of Fe at this site is thus consistent with total Fe3+ determined by both wet chemical analysis and M”ssbauer spectroscopy. This gives the preferred formula of (Mg0.966Fe0.034)M1 (Mg0.962Fe0.038)M22 O10(OH)6 (Mg0.996Fe0.004) M32 (Al0.841 Fe3+0.102 Cr0.004 Ti0.004)M4(OH)6 which contains slightly more total Mg atoms (4.88) than indicated by electron microprobe (4.84) but is within the analytical error of the probe.
Finally, H atoms were put in the model and refined using isotropic displacement factors. The positions of H atoms in the brucite-like layer were refined without constraint, however the H1 position drifted to a position that gave an unreasonably long O-H distance and large displacement parameter. The z coordinate of this H was constrained to that of a significant maximum in the difference Fourier while x, y and B were refined without constraint. The final weighted R (Rw) for the refinement reduced to 0.059 for reflections greater than 3s, and 0.064 for all reflections. Final position and temperature factors are given in Table 4. Selected nearest neighbor oxygen distances and coordination polyhedron volumes and distortions are given in Table 5.
Similarly, nearly identical values of polyhedral volume and mean M-O distance for the M1 and M2 octahedra of the talc-like layer indicate similar Fe and Mg occupancies of these sites. The X-ray refinement using scattering factors from ionized cations (Cromer and Mann 1968) indicates approximately 3.4% occupancy of M1 and 3.8% occupancy of M2 by iron, for a total of 0.110 cations of Fe per formula unit in these two sites. The scattering intensities and volumes of these two sites are nearly identical indicating only a slight tendency for iron ordering between these two sites. Both M”ssbauer spectroscopy and X-ray diffraction thus indicate a slight preference of Fe2+ for M2. Also, M1 and M2 are nearly identical in volume, distortion, and mean nearest neighbor distances. They differ in that M1 is on an inversion center with OH in trans positions, whereas M2 is in a general position with OH in cis positions.
Based on the observed occupancies and multiplicities of the two sites, we assign the M”ssbauer peak with (=1.16 mm/s and (=2.365 mm/s to Fe2+ in M1, and that with (=1.134 mm/s and (=2.681 to Fe2+ in M2. The observed peak areas then give occupancies of 3.2 % Fe2+ in M1 and 4.3% Fe2+ in M2 (Table 3), in very close agreement with those determined by X-ray. The latter assignment can be made with confidence because corroborating XRD results argue against its being interpreted as Fe2+ in M3 or M4. We then assign the remaining M”ssbauer doublet with (=0.328 mm/s and (=0.80 mm/s to Fe3+ at M4. This gives an Fe3+ occupancy of 10.2% in M4 in very close agreement with the 10.1 % Fe occupancy determined by X-ray diffraction. This X-ray result is, of course, dependent on our assignment of 5% vacancy to this site. This assignment is not entirely arbitrary. Without the vacancy, the Fe occupancy of M4 refined to 8.2%, but this model gave a somewhat anomalous atomic displacement parameter. We are able to eliminate this anomaly and obtain essentially complete agreement between chemical analysis, M”ssbauer spectroscopy, and X-ray diffraction by assignment of 5% vacancy to this site.
The anisotropic atomic displacement factors observed in this X-ray refinement are somewhat anomalous in being strongly prolate parallel to c, particularly for M1, M2, M3, M4, T1, T2, O1, O2, and OH1, but much less strongly for the remaining O and OH positions. This has been observed by previous authors (Zheng and Bailey 1989) and ascribed to possible error in absorption correction. This is a possible explanation, but we tried several different arbitrary crystal shapes for the absorption correction and were unable to eliminate the effect.
The natural clinochlore sample studied here thus shows a very high degree of cation order in the brucite layer as evidenced by the difference between the mean M4-OH (1.922(4)Ź) and M3-OH (2.083(4)Ź) distances, D, which is equal to 0.157Ź. Phillips et al (1980) noted a correlation between the a cell parameter and cation ordering in the brucite layer which was further discussed by Joswig and Feuss (1989). The results of our analysis show the largest alpha value (90.56§) and the largest D yet reported for IIb-4 chlorites for which there are high precision single crystal structure refinements. These parameters are plotted in Fig. 3. which shows the strong correlation and is consistent with the high degree of trivalent cation ordering in the current sample. The only sample that falls slightly off of the smooth trend is one of the high Cr samples described by Philips et al (1980). The strong correlation indicates that the degree of cation order can be inferred from cell parameters which are readily refined to high precision from powder data by Rietveld methods (Walker and Bish, 1992).
Bailey, S.W. 1988. Chlorites: structures and crystal chemistry. Rev. Mineralogy 19, 347-398.
Ballet, O., J.M.D. Coey and K.J. Burke. 1985. Magnetic properties of sheet silicates; 2:1:1 layer minerals. Phys. Chem. Minerals, 12, 370-378.
Blaauw, C., G. Stroink, and W. Leiper. 1980. M”ssbauer analysis of talc and chlorite. Journal de Physique, Colloque C1, 41, C1-411-C1-412.
Borggaard, O.K., H.B. Lindgreen, and S. Morup. 1982. Oxidation and reduction of structural iron in chlorite at 480 řC. Clays & Clay Minerals 30, 353-364.
Christofides, G., G. Thimiatis, A. Koroneos, S. Sklavounos, and G. Eleftheriadis. 1994. Mineralogy and chemistry of Cr-chlorites associated with chromites from Vavdos and Vasilika Ophiolite Complexes (Chalkidiki, Macedonia, N. Greece). Chemie der Erde, Geochemistry, 54, 151-166.
Clark, F.W. and E.A. Schneider. 1890. Experiments upon the constitution of the natural silicates. Am. Jour. Sci. 40, 406-415.
Coey, J.M.D. 1975. The clay minerals: Use of the M”ssbauer effect to characterize them and study their transformation. In: Hrynkiewicz, A.Z., Sawicki, J.A. (Eds.), Proceedings of the International Conference on M”ssbauer spectroscopy, vol. 2, Akademia Gorniczo-Hutnicza. Krakowie, 333-354.
Cromer, D.T, and J. Mann. 1968. X-ray scattering factors computed from numerical Hartree-Fock wave functions. Acta Cryst. A24, 321-324.
DeGrave, E., J. Vandenbruwaene, and M. Van Bockstael. 1987. 57Fe M”ssbauer spectroscopic analysis of chlorite. Phys. Chem. Minerals, 15, 173-180.
deParseval, P., L. Fournes, J.-P. Fortune, B. Moine, J. and Ferret. 1991. Distribution du fer dans les chlorites par spectrom‚trie M”ssbauer (57Fe): Fe3+ dans les chlorites du gisement de talc-chlorite de Trimouns (Pyr‚n‚es, France). Acad‚mie des Sciences de Paris, 312(II), 1321-1326.
Dyar, M.D. 1984. Precision and interlaboratory reproducibility of measurements of the M”ssbauer effect in minerals. Amer. Mineral. 69, 1127-1144.
Dyar, M.D. 1990. M”ssbauer spectra of biotite from metapelites. Amer. Mineral. 75, 656-666.
Ericsson, T., R. Waeppling, and K. Punkivi. 1977. M”ssbauer spectroscopy applied to clay and related minerals. Geologiska F”reningens i Stockholm F”rhandlinger 99, 229-244.
Finger, L.W. and E. Prince. 1975. A system of FORTRAN IV computer programs for crystal structure computations. National Bureau of Standards Technical Note 854, 128pp.
Goodman, B.A., and D.C. Bain. 1978. M”ssbauer spectra of chlorites and their decomposition products. Proceedings of the International Clay Conference, 1978 (Oxford), 65-74. Elsevier, Amsterdam.
Grant, C.A. 1995. Sources of experimental and analytical error in measurements of the M”ssbauer effect in amphibole. Ph.D. Thesis, University of Oregon, Eugene, OR, USA.
Gregori, D.A., and R.C. Mercader. 1994. M”ssbauer study of some Argentinean chlorites. Hyperfine Interactions 83, 495-498.
Joswig, W., and H. Feuss, R. Rothbauer, Y. Takeuchi, and S.A. Mason. 1980. A neutron diffraction study of a one-layer triclinic chlorite (penninite). Amer. Mineral. 65, 349-352.
Joswig, W., and H. Feuss. 1989. Refinement of a one-layer triclinic chlorite. Clays & Clay Minerals 38, 216-218.
Joswig, , W., and H. Feuss, and S.A. Mason. 1989. Neutron diffraction study of a one-layer monoclinic chlorite. Clays & Clay Minerals 37, 511-514.
Kodama, H., G. Longworth, and M.G. Townsend. 1982. A M”ssbauer investigation of some chlorites and their oxidation products. Can. Mineral. 20, 585-592.
Komadel, P., and J.W. Stucki. 1988. The quantitative assay of minerals for Fe2+ and Fe3+ using 1,10-phenanthroline: III. A rapid photochemical method. Clays and Clay Minerals 36, 379-381.
Long, G.J., T.E. Cranshaw, and G. Longworth. 1984. The ideal M”ssbauer absorber thickness. M”ssbauer Effect Reference and Data Journal 6, 42-49.
May, H. M., J.G. Acker, J.R. Smyth, O.P. Bricker, and M.D. Dyar. 1996. Aqueous dissolution of clinochlore at 25C. Clays & Clay Minerals, (in preparation).
Pal, T., D. Dipankar, and S. Mitra. 1992. 57Fe M”ssbauer investigation of naturally oxidized chlorite. Hyperfine Interactions 73, 313-321.
Phillips, T.L., J.K Loveless, and S.W. Bailey. 1980. Cr3+ coordination in chlorites: a structural study of ten chromian chlorites. Amer. Mineral. 65, 112-122.
Raclavsk , H., and K. Raclavsky. 1988. Chlorites of the Desn Unit, Hruby Jesenˇk Mountains. Tenth Conference on Clay Mineralogy and Petrology, Ostraca, 113-118.
Rule, A.C., and S.W. Bailey. 1989. Refinement of the crystal structure of a monoclinic ferroan clinochlore. Clays & Clay Minerals 35, 129-138.
Stucki, J.W. 1981. The quantitative assay of minerals for Fe2+ and Fe3+ using 1,10-phenanthroline: II. A photochemical method. Soil Science Society of America Journal 45, 638-641.
Stucki, J.W., and W.L. Anderson. 1981. The quantitative assay of minerals for Fe2+ and Fe3+ using 1,10-phenanthroline: I. Sources of variability. Soil Science Society of America Journal, 45, 633-637.
Taylor, G.L., A.P. Ruotsala, and R.O. Keeling, Jr. 1968. Analysis of iron in layer silicates by M”ssbauer spectroscopy. Clays & Clay Minerals 16, 381-391.
Townsend, M.G., G. Longworth, and H. Kodama. 1986. Magnetic interaction at low temperature in chlorite and its products of oxidation: A M”ssbauer investigation. Canadian Mineralogist 24, 105-115.
Walker, J.R., and D.L. Bish. 1992. Application of Rietveld refinement techniques to a disordered IIb Mg-chamosite. Clays & Clay Minerals 40, 319-322.
Welch, M.D., J. Barras, and J. Klinowski. 1995. A multinuclear NMR study of clinochlore. Amer. Mineral. 80, 441-447.
Zheng, H. and S.W. Bailey. 1989. Structures of intergrown triclinic and monoclinic IIb chlorites from Kenya. Clays & Clay Minerals 37, 308-318.