On local structural changes in lizardite-1T: {Si4+/Al3+}, {Si4+/Fe3+}, [Mg2+/Al3+], [Mg2+/Fe3+] substitutions

Geometrical changes induced by cation substitutions {Si⁴⁺/Al³⁺}[Mg²⁺/Al³⁺], {2Si⁴⁺/2Al³⁺} [2Mg²⁺/2Al³⁺], {Si⁴⁺/Fe³⁺} [Mg²⁺/Al³⁺] or [Mg²⁺/Fe³⁺], where {} and [] indicate tetrahedral and octahedral sheet in lizardite 1T, are studied by ab-initio quantum chemistry calculations. The majority of the models are based on the chemical compositions reported for various lizardite polytypes with the amount of Al in the tetrahedral sheets reported to vary from 3.5% to 8% in the 1T and 2H ₁, up to ~30% in the 2H ₂ polytype. Si⁴⁺ by Fe³⁺ substitution in the tetrahedral sheet with an Al³⁺ (Fe³⁺) in the role of a charge compensating cation in the octahedral sheet is also examined. The cation substitutions result in the geometrical changes in the tetrahedral sheets, while the octahedral sheets remain almost untouched. Substituted tetrahedra are tilted and their basal oxygens pushed down from the plane of basal oxygens. Ditrigonal deformation of tetrahedral sheets depends on the substituting cation and the degree of substitution.     

Sapphirine II

The crystal structure of aP2₁/a polymorph of sapphirine (a=11.286(3),b=14.438(2),c=9.957(2) Å, β=125.4(2) °) of composition [Mg_3.7Fe _0.1 ²⁺ Al_4.1- Fe _0.1 ³⁺ ]^IV[Si_1.8Al_4.2]^IVO₂₀ was refined using structure factors determined by both neutron and x-ray diffraction methods to conventionalR factors of 0.067 and 0.031. respectively, forF _obs>2σ. The results of the two refinements agree reasonably well, but a half-normal probability plot (Abrahams, 1974) comparing the two data sets indicates that the pooled standard deviations of the atomic coordinates have been underestimated by a factor of two. The structure of sapphirine, solved initially by Moore (1969), consists of cubic closest packed oxygens with octahedral and predominantly tetrahedral layers alternately stacked along [100]. The layer in which 70% of the octahedral sites are occupied has an Mg-Al distribution characterized by Mg-rich octahedra sharing edges mainly with Al-rich octahedra. Mean octahedral bond lengths correlate well with Al occupancy determined by neutron site refinement if the relative number of shared octahedral edges is taken into account (see Table 1). The predominantly tetrahedral layer has 10% of the octahedral sites occupied by Al and 30% of the tetrahedral sites occupied by Al-Si in the ratio 2.33∶1. There are single chains of Al-Si tetrahedra parallel toz with corner-sharing wing tetrahedra (T5 andT6) on either side in the (100) plane. The meanT-O distance is highly correlated with Al occupancy, X_Al, as determined from the neutron site refinement: $$\langle T - O\rangle = 1.656 + 0.105X_{Al} (r^2 = 0.995).$$ Details of the neutron refinement are summarized below.     

Effect of early nucleation conditions on synthetic muscovite polytypism as seen by high resolution transmission electron microscopy

Three isobaric (P _{H 2O}=1 kb) and isothermal (T _N=355, 502 and 630° C) nucleation experiments of OH-muscovite have been carefully studied by powder X-ray diffraction and high resolution transmission electron microscopy. The intimate stacking sequences of layers within individual spiral-free crystallites are determined by lattice (or structure)-imaging techniques. At T _N=355° C, disordered stacking involving local 1M and 2M ₁ order, largely dominates. A few nuclei with 1M structure coexist with disordered ones. At T _N=502° C, most of micas are of more or less heavily faulted 1M structure but coexist with heavily faulted 2M ₁ nuclei. At T _N=630° C, almost all the population is of 2M ₁ structure with rare stacking faults often located in the middle of the sequence. The nature of intrinsic and extrinsic stacking faults within 1M and 2M ₁ matrices is described. Decreasing temperature and/or increasing supersaturation are found to promote disorder in muscovite. The role of the style of packing of the two early first condensed layers (and related specific distortions of the single layer structure) on the selection of the basic structure developed through further 2D-nucleation, as suggested by Takeda and Ross (1975), is shown to be counterbalanced by environmental parameters like T and supersaturation. These high resolution transmission electron microscopy (HRTEM) observations on screw dislocation-free crystallites give experimental support to the basic assumptions of a recent faulted matrix model which has been proposed to explain the origin of complex polytypes of mica according to a subsequent spiral growth mechanism.     

Cation partitioning versus temperature in (Mg0.70Fe0.23)Al1.97O4 synthetic spinel by in situ neutron powder diffraction

Neutron powder diffraction experiments in the temperature range 300–1770 K were performed at BENSC, Berlin, Germany, on synthetic (Mg_0.70Fe_0.23) Al_1.97O₄. The cation partitioning over the crystallographic tetrahedral and octahedral sites was determined as a function of temperature through joint Rietveld refinements and advanced minimization techniques. The thermal expansion coefficients of the lattice parameter and inter-atomic bond lengths were also obtained from the full-profile structure refinements. The behaviour of the polyhedral bond-lengths, especially the T−O distances, and of the cell constant upon heating, clearly indicate that the interdiffusion of tetrahedral and octahedral Mg/Al cations starts at about 950 K. This result is straightforwardly supported by the direct analysis of the neutron site scattering factors: Fe always retains tetrahedral coordination at all temperatures, and the cation rearrangement is entirely due to Mg and Al diffusion. Received: 18 November 1997 / Revised, accepted: 23 August 1998     

Temperature dependence of the cation distribution in nickel aluminate (NiAl2O4) spinel: a powder XRD study

NiAl₂O₄ is a largely inverse spinel, which in detail shows increasing randomisation with temperature of Ni and Al between the octahedral and tetrahedral cation sites of the spinel structure. We have used powder XRD to determine this cation distribution in various samples of NiAl₂O₄ quenched after annealing between 700 and 1400° C. The inversion parameter (x) can be measured with a precision of ± 0.004 (one standard deviation), and a comparison of different methods of synthesis, X-ray diffraction and refinement techniques, suggests a probable accuracy of better than 0.01. The results are supported by some preliminary single crystal refinements on flux-grown samples.Below 800° C the rate of cation ordering becomes very slow, and, despite reaching an apparently steady state, it is doubtful if our samples attained complete internal equilibrium. Above 1250° C the cation redistribution becomes so fast that the quenching method becomes unreliable. Between 800 and 1250° C inclusive, the degree of inversion changes smoothly from 0.87 at 800° C to 0.79 at 1250° C, and is accompanied by linear changes in u, the oxygen parameter, from 0.2555 to 0.2563 (±0.0002), and a₀, the lattice parameter, from 8.0462 to 8.0522 Å (±0.0002 Å).     

Systematics of the spinel structure type

Systematic trends in the geometry of 149 oxide and 80 sulfide binary and ternary spinels have been examined from the standpoint of ionic radius and electronegativity. The mean ionic radii of the octahedral and tetrahedral cations, taken together, account for 96.9 and 90.5% of the variation in the unit cell parameter, a, of the oxides and sulfides, respectively, with the octahedral cation exerting by far the dominant influence in sulfides. The mean electronegativity of the octahedral cation exerts an additional, but small, influence on the cell edge of the sulfides. The equation a=(8/3√d)d _tet+(8/3)d _oct, where d _tet and d _oct are the tetrahedral and octahedral bond lengths obained from the sum of the ionic radii, accounts for 96.7 and 83.2% of the variation in a in the oxides and sulfides, respectively, again testifying to the applicability of the hard-sphere ionic model in the case of the spinel structure. Comparison of observed and calculated u values for 94 spinels indicates that up to 40% of the experimentally measured anion coordinates may be significantly in error. In addition to these compounds, u values are given for 52 spinels for which no data have previously been determined. Diagrams are presented for the rapid interpretation of the internal consistency of published data and the prediction of the structural parameters of hypothetical or partially studied spinels.     

Influence of stacking faults on the spiral growth of polytype structures in mica

A systematic theoretical deduction of polytype structures of mica that can result by the spiral growth mechanism operating in faulted 1M, 2M ₁ and 3T basic matrices is reported. As a prerequisite, all possible intrinsic and extrinsic stacking fault configurations in each of the basic matrices have been worked out and their stacking fault energy (SFE) estimated. The deduction of polytype structures on the basis of the “faulted-matrix model” takes into account (i) the introduction of each of the low energy fault configurations in the exposed ledge of the screw dislocations, (ii) the change in the layer-position of the fault within the exposed ledge and (iii) the variation of the strength of the generating screw dislocation. At each stage, the spirally-grown polytypes are deduced for each basic structure. The most probable structures are predicted on the basis of the lowest SFE for the same strength of the screw dislocation and are then compared with the polytype structures reported in the literature. It was found that the faulted matrix model accounts successfully for the origin of all the polytype structures in mica. Furthermore, it may provide a basis for limiting the number of trial structures for determining the structures of long period polytypes.     

Determination of the mechanism of cation ordering in magnesioferrite (MgFe2O4) from the time- and temperature-dependence of magnetic susceptibility

The kinetics of non-convergent cation ordering in MgFe₂O₄ have been studied by measuring the Curie temperature (T _c) of synthetic samples as a function of isothermal annealing time. The starting material was a synthetic sample of near-stoichiometric MgFe₂O₄, synthesised from the oxides in air and quenched from 900 °C in water. Ordering experiments were performed using small chips of this material and annealing them at temperatures between 450 °C and 600 °C. The chips were periodically removed from the furnace, and their Curie temperatures were determined from measurements of alternating-field magnetic susceptibility (χ) as a function of temperature (T) to 400 °C. The Curie temperature of MgFe₂O₄ is very sensitive to the intracrystalline distribution of Fe³⁺ and Mg cations between tetrahedral and octahedral sites of the spinel crystal structure, and hence provides a very sensitive probe of the cation ordering process. The χ-T curve for the starting material displays a single sharp magnetic transition at a temperature of 303 °C. During isothermal annealing, the χ-T curve develops two distinct magnetic transitions; the first at a temperature corresponding to T _c for the disordered starting material and the second at a higher temperature corresponding to T _c for the equilibrium ordered phase. The size of the magnetic signal from the ordered phase increases smoothly as a function of time, until equilibrium is approached and the shape of the χ-T curve corresponds to a single sharp magnetic transition for the homogeneous ordered phase. These observations demonstrate that cation ordering in MgFe₂O₄ proceeds via a heterogeneous mechanism, involving the nucleation and growth of fine-scale domains of the ordered phase within a matrix of disordered material. Disordering experiments were performed by taking material equilibrated at 558 °C and annealing it at 695 °C. The mechanism of isothermal disordering is shown to involve nucleation and growth of disordered domains within an ordered matrix, combined with continuous disordering of the ordered matrix. This mixed mechanism of disordering may provide an explanation for the difference between the rates of ordering and disordering observed in MgFe₂O₄ using X-ray diffraction. The origin of the heterogeneous ordering/disordering mechanism is discussed in terms of the Ginzburg-Landau rate law. It is argued that heterogeneous mechanisms are likely to occur in kinetic experiments performed far from equilibrium, whereas a homogeneous mechanism may operate under slow equilibrium cooling. The implications of these observations for geospeedometry are discussed. Received: 12 May 1998 / Accepted: 25 June 1998     

Influence of structural factors on 29Si and 27Al NMR chemical shifts of phyllosilicates 2:1

The ²⁹Si and ²⁷Al nuclear magnetic resonance (NMR) analysis of synthetic trioctahedral phyllosilicates 2:1, with tetrahedral ratios Al _T/(Si + Al _T) ranging from 0 to 0.5, has shown that the ditrigonal distortion of tetrahedral rings (angle ) is the main factor controlling chemical shift values of tetrahedral components in both signals. The increase of ditrigonal rotation angle shifts these components towards more positive values. For each sample, the composition of tetrahedral and octahedral sheets determine the value of , and from this parameter, the mean tetrahedral Tot angle and the chemical shift values of components are deduced. For a given environment, variations on ditrigonal angle are responsibles for the observed evolution of chemical shift values with bulk composition. The comparative analysis of micas and saponite samples has demonstrated that the location of compensating charge (interlayer and octahedral sheet) does not affect chemical shift values.     

IR characteristics of hydrotalcite-like compounds

The infrared spectra of well characterized hydrotalcite-like compounds with M ²⁺/M ³⁺= 2/1 and M ²⁺/M ³⁺=3/1 chemical ratios indicate that octahedral cation ordering is only present for the 2/1 composition. For M ²⁺/M ³⁺ ratios greater than 2/1, the octahedral sheets are disordered although local cation order can be present through cation segregation. The order-disorder characteristics of these compounds can be explained by the total cation charge around any hydroxyl which cannot exceed seven units nor be lower than six. In addition, the infrared spectra suggest that the interlayer anions are symmetrically perturbed, the electrostatic interaction being greater for the 2/1 compositions.     

The crystal structure of disordered (Zr,Ti)O2 solid solution including srilankite: evolution towards tetragonal ZrO2 with increasing Zr

Crystal structure data are presented for seven synthetic samples of disordered zirconium-titanate solid solution (Zr,Ti)O₂, ranging in composition from x _Ti=0.43 to 0.67, thus covering compounds such as ZrTiO₄, Zr₅Ti₇O₂₄, and ZrTi₂O₆ (srilankite). The compounds, synthesized at high temperatures and various pressures in their respective stability fields, are well crystallized and of homogeneous composition. The resulting structure data are less scattered compared to previous studies that were based on compounds synthesized metastably at low temperatures and room pressure. The compounds have the structure of scrutinyite (α-PbO₂) with space group Pbcn, Z=4, unit cell parameters a=4.8495(3) Å, b=5.4635(3) Å, c=5.0462(3) Å at x _Ti=0.425 to a=4.7112(2) Å, b=5.4944(1) Å, c=4.9962(1) Å at x _Ti=0.666. The first structure refinement of pure, synthetic srilankite is presented, which is in good agreement with that of the natural counterpart. Structural trends observed in disordered zirconium-titanate solid solution along the binary join ZrO₂–TiO₂ are relatively smooth and continuous, except for rapid lengthening of an unshared octahedral edge which is anomalously short in scrutinyite-structure TiO₂. The shortness of this edge may explain the observed instability of this structure with the relatively small Ti as the dominant cation. With increasing Zr content, the average cation position moves off-centre inside the octahedron, away from two shared edges, which permits the 12 closest cation–cation distances in the structure to become more equal. The shortening of the b dimension with increasing amount of the larger cation Zr decreases the distance between octahedral Zr and two additional oxygens in an adjacent chain of edge-sharing octahedra, implying that the Zr environment is evolving towards eightfold coordination. If the two additional oxygens are considered as part of the Zr coordination polyhedron, the bonding topology of tetragonal zirconia is obtained. The compositional evolution of the cell parameters, Zr atomic coordinates and Zr coordination environment is consistent with the idea that the structure is evolving towards that of tetragonal ZrO₂. Group-theoretical relationships between scrutinyite, tetragonal zirconia, baddeleyite and fluorite structures show that the sequence of structures fluorite > tetragonal zirconia > scrutinyite > baddeleyite are all related by potentially diffusionless phase transitions driven by wavelike displacements of the oxygen substructure. The scrutinyite and tetragonal structures can act outside their stability fields as “transition states” between the structures on either side.     

Transmission electron microscope study of dislocations in orthopyroxene (Mg,Fe)2Si2O6

The orthopyroxene crystal structure can be viewed as the stacking of alternating tetrahedral and octahedral layers parallel to the (100) plane. Easy glide occurs in the (100) plane at the level of the octahedral layer to prevent breakage of the strong Si-O bonds. Dislocations with c and b Burgers vectors have been activated in (100) by room temperature indentation in an orthoenstatite gem quality single crystal. Investigations in transmission electron microscopy show that the b dislocations (b?9 Å) are not dissociated while the c's (c=5.24 Å) are dissociated into four partials. This result is interpreted by considering the oxygen sublattice as a distorted FCC one. The four c partials are thus Shockley partials bounding three stacking faults. For the two outer ones, synchroshear of the cations is necessary to keep unchanged their sixfold coordination; the oxygen sublattice is locally transformed into a HCP lattice. This accounts for the observed low splitting (?100 Å) of these faults as compared to the median one (?500 Å) which does not affect the oxygen sublattice and does not require cation synchroshear. In a Fe rich orthopyroxene (eulite), semi coherent exsolution lamellae have been studied. Either only c edge dislocations or both b and c edge dislocations occur in the phase boundaries depending upon the thickness of the lamellae. Only the c dislocations are dissociated. From the observed spacing between these mismatch dislocations a crude estimate of the exsolution temperature is proposed T _ex ? 700° C.     

Pressure dependence of structural parameters of paragonite

The compressibility and structure of a 2M₁ paragonite with composition [Na_0.88K_0.10Ca_0.01Ba_0.01] [Al_1.97Ti_0.007Fe_0.01Mn_0.002Mg_0.006]Si_3.01Al_0.99O₁₀OH₂ were determined at pressures between 1 bar and 41 kbar, by single crystal X-ray diffraction using a Merrill-Bassett diamond anvil cell. Compressibility turned out to be largely anisotropic, linear compressibility coefficients parallel to the unit cell edges being β_a=3.5(1)·10^?4, β_b=3.6(1)·10^?4, β_c=8.3(3)·10^?4 kbar^?1 (β_a:β_b·β_c=1:1028:2.371). The isothermal bulk modulus, calculated as the reciprocal of the mean compressibility of the cell volume, was 650(20) kbar. The main features of the deformation mechanism resulting from structural refinements at pressures of 0.5, 25.4, 40.5 kbar were: –?variation in sheet thickness, showing that compression of the c parameter was mainly due to the interlayer thickness reduction from 3.07 Å at 0.5 kbar to 2.81 Å at 40.5 kbar; –?the compressibility of octahedra was greater than that of tetrahedra, the dimensional misfit between tetrahedral and octahedral sheets increased with P, so that tetrahedral rotation angel α increased from 15° at 0.5 kbar to 21.6° at 40.5 kbar; –?the basal surface corrugation (Δz) of the tetrahedral layer, due to the different dimensions of M1 and M2 octahedra and to the octahedral distortion, decreased with P (Δz=0.19 and 0.12 Å at 0.5 and 40.5 kbar respectively). Comparison of the new data on paragonite with those of a K-muscovite and a Na-rich muscovite (Comodi and Zanazzi 1995) revealed a clear trend toward decreasing of compressibility when Na substitutes for K atoms in the interlayer sites.     

Effect of the Si/Al ordering on structural parameters and the energetic stabilization of vermiculites – a theoretical study

The effect of the Si/Al distribution in the tetrahedral sheets of the vermiculite mineral has been investigated employing density functional theory. The structures of six models for vermiculite with the structural formula (Mg₄)(Mg₁₂)(Si₈Al₈)O₄₀(OH)₈·24(H₂O) per unit cell were fully optimized. The models differ by the T···Mg²⁺···T coordination of the interlayer Mg²⁺ cations by two central cations from the adjacent tetrahedral sheets of the 2:1 vermiculite layers (T,T=Si,Al). We observed the formation of very strong hydrogen bonds between water molecules solvating the interlayer Mg²⁺ cations and the surface basal oxygen atoms of the 2:1 layers. The directionality of hydrogen bonds is the major factor determining the layer stacking in the vermiculite structure. Results showed that the most stable model is that where only silicon atoms in the tetrahedral sheets coordinate all interlayer Mg²⁺ cations.     

The Crystal Structure of Lisiguangite

The crystal structure of lisiguangite,CuPtBiS3,from Yanshan mountains,Chengde Prefecture,Hebei Province,China has been determined by single crystal X-ray diffraction.It belongs to orthorhombic space group P2_12_12_1 with a = 7.7372(15) A,b = 12.844(3) A,c = 4.9062(10) A,V =487.57(17) A~3,Z = 4.The final full-matric least-square refinement on F2 converged with Rl = 0.0495 and wR2 = 0.0992 for 704 observed reflections[I≥2σ(I)].Lisiguangite is the isomorph of known CuNiSbS_3 and CuNiBiS_3· Pt~(2+) and Bi~(3+) have the distorted octahedral coordination enviroments composed of two metal and four S and Cu~(+2) has a distorted tetrahedral coordination environment with four S atoms.Each S atom is surrounded by four metals to give a tetrahedral environment.The crystal structure is a complex 3 dimensional network.     

Microtextures and crystal chemistry in P21/c pigeonites

Summary ?A single-crystal X-ray investigation was performed on crystals of P2₁/c natural pigeonite with varying Ca and Fe^* ( = Fe²⁺ + Mn²⁺) contents, in order to verify the effect of microtextural disorder on structure refinements and to constrain the crystal chemistry of pigeonite. Antiphase domains and exsolution lamellae affect differently the refinement results. In a crystal free of exsolution the structure obtained after refinement with all reflections is an average of that of the antiphase domains and of their boundaries, whereas in an exsolved crystal it represents only the structure of the prevailing pigeonite lamellae. The refinement using only h + k odd reflections seems to give the structure of the Ca-free pigeonite characteristic of the antiphase domains rather than that of Ca-rich domain walls. The ratio of the scale factors in refinements with all reflections and with only h + k odd reflections allows the ratios of the exsolved augite and pigeonite phases to be estimated. The crystal chemistry of the investigated samples follows the trends outlined by data on Ca-free and Fe-free synthetic samples. In particular, it is shown that Ca and Fe^* substitution for Mg induce similar changes in the average structure, i.e. both induce an expansion in the M1 polyhedron and decrease the difference between the M2–O3 distances. Received October 18, 2001; revised version accepted February 15, 2002     

Cation distributions in olivines and orthopyroxenes an interlaboratory study

The determination of the (Fe, Mg)-cation distribution in important rock forming minerals such as olivines and orthopyroxenes from single crystal X-ray diffraction data is well established. The findings obtained from good quality data are, however, subject to the structure model, refinement strategy, data weighting etc. Moreover, they do not necessarily agree with those of comparable Mössbauer studies. In order to assess the reliability of the X-ray results, i.e. precision and accuracy of the site occupancies and ensuing cation distribution coefficients K _D , an interlaboratory study was undertaken. Two olivine and two orthopyroxene data sets, one experimental and three synthetic, were dispatched to interested applicants for structure refinements according to their own choice. Approximately correct structure parameters were given, but no information about the chemical composition, i.e. the Fe-content, was communicated. The total of 25 refinements for each data set, carried out with different programs, valence models, and weighting schemes, yielded some considerable differences between the results, which are summarized and discussed in terms of precision and accuracy. The uncertainties of the refined site occupancies, following from the qualities of the fits, are only realistic, when they are dominated by the statistical errors in the data. Errors in the independently refined Fe-concentrations on the M1-and M2-sites are not fully transferred to the thermodynamically relevant cation distribution coefficient. Therefore, correlations between the site occupancies, clearly recognizable from the compilations of all respective results, can be combined with an independently obtained chemical constraint (microprobe analysis) in order to yield reliable K _D -values with realistic uncertainties. In summary, all results indicate the need for multiple refinement techniques, if reliable results on the absolute scale are desired.     

Chemical and structural study of 1M- and 2M 1-phlogopites coexisting in the same Kasenyi kamafugitic rock (SW Uganda)

In the present work, crystal chemical variations between 1M and 2M ₁ phlogopites coexisting in the same rock sample from kamafugite of Kasenyi (southwest Uganda, west branch of the East African Rift) were explored by electron probe microanalyses, single crystal X-ray diffraction and M?ssbauer spectroscopy. Chemical analyses revealed close similarity both within and between the two polytypic arrangements as well as high TiO₂ (~4.9?wt%) and Al₂O₃ (~12.9?wt%), and low Cr₂O₃ (~0.8?wt%), F (~0.3?wt%) and BaO (~0.2?wt%) contents. Room temperature ⁵⁷Fe M?ssbauer investigation proved that the studied mica is a tetraferriphlogopite with: ^IVFe³⁺?=?19(1)?%, ^VIFe²⁺?=?58(1)?%, ^VIFe³⁺?=?23(1)?%. Single crystal refinement showed that both polytypes have narrow range of variation in terms of some relevant unit cell parameters and similar values in terms of mean bond lengths, mean atomic numbers and distortion parameters. Similar substitutions were active in the structure of the 1M and 2M ₁ studied phlogopites. However, in 2M ₁ polytypes the oxy-type substitutions were found to occur to a greater extent. Comparison of unit layer of 1M mica (in the 2M ₁ setting) with that of the 2M ₁ ones showed that the 2M ₁ polytypes are affected to different extent by relative shifts of the upper and lower triads of octahedral oxygens along the ±b directions. This effect did not cause any symmetry lowering in the T-O-T layer of the studied samples.     

High-pressure perovskites on the join CaTiO3-FeTiO3

An exploratory high-pressure study of the join CaTiO₃-FeTiO₃ has uncovered two intermediate perovskites with the compositions CaFe₃Ti₄O₁₂ and CaFeTi₂O₆. These perovskites have ordering of Ca²⁺ and Fe²⁺ on the A sites. Both of these perovskites are unusual in that the A sites containing Fe²⁺ are either square planar or tetrahedral, due to the particular tilt geometries of the octahedral frameworks. For CaFe₃Ti₄O₁₂, the structure has been refined from a powder using the Rietveld technique. This compound is a cubic double perovskite (SG Im $\bar 3$ , a = 7.4672 Å), isostructural with NaMn₇O₁₂. Fe²⁺ is in a square-planar A site (similar to Mn³⁺ in NaMn₇O₁₂) with Fe-O = 2.042(3) Å, with distant second neighbors in a rectangle at Fe-O = 2.780(6) Å. Calcium is in a distorted icosahedron with Ca-O =2.635(5) Å. CaFeTi₂O₆ crystallizes in a unique tetragonal double perovskite structure (SG P4₂/nmc, a = 7.5157(2), c = 7.5548(2)), with A-site iron in square-planar (Fe-O = 2.097(2) Å) and tetrahedral (Fe-O = 2.084(2) Å) coordination, again with distant second neighbor oxygens near 2.8 Å. Rietveld refinement was also performed for the previously known perovskite-related form of FeTiO₃ recovered from high pressure (lithium niobate type). This compound is trigonal R3c, with a = 5.1233(1) and c = 13.7602(2). The ordered perovskites appear to be stable at 1215 GPa and CaFe₃Ti₄O₁₂ is found as low as 5 GPa. Thus these perovskites may be important to upper mantle mineralogy, particularly in kimberlites. These compounds are the first known quenchable perovskites with large amounts of A-site ferrous iron, and add greatly to the known occurrences of ferrous iron in perovskites.