Thermally Induced Modifications and Phase Transformations of Red

Assuming that each Mg-calcite single-phase peak is characteristic of a 85 nm grain size and that microstructural defects are equal to the mean value o...
21 downloads 5 Views 4MB Size
Article pubs.acs.org/crystal

Thermally Induced Modifications and Phase Transformations of Red Coral Mg-Calcite Skeletons from Infrared Spectroscopy and High Resolution Synchrotron Powder Diffraction Analyses Nicole Floquet,*,† Daniel Vielzeuf,† Daniel Ferry,† Angèle Ricolleau,† Vasile Heresanu,† Jonathan Perrin,† Didier Laporte,‡ and Andy N. Fitch§ †

CINaM UMR7325, Aix-Marseille Université, CNRS, Campus de Luminy, Case 913, 13288 Marseille, France Laboratoire Magmas et Volcans, Université Blaise Pascal-CNRS-IRD, OPGC, 5 rue Kessler, 63038 Clermont Ferrand, France § European Synchrotron Radiation Facility, BP 220, 38043 Grenoble Cedex, France Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291



S Supporting Information *

ABSTRACT: Thermal modifications and decomposition of synthetic Mg-calcite and biogenic red coral skeletons are studied using thermogravimetry coupled with differential thermal analysis, infrared spectroscopy, and high temperature synchrotron powder diffraction. Synthetic Mg-calcite with 10 mol % Mg is stable up to 1073 K in selfcontrolled CO2, with no phase transformation, a parabolic thermal expansion of the c cell parameter, and noticeably, a nonmonotonic thermal variation of the a cell parameter with a minimum at 673 K. By contrast, biogenic Mg-calcite in the range of 9−15 mol % Mg in the red coral skeletons undergoes anomalous structural modifications prior to a phase transformation at 823 K. The thermally induced structural modifications occur concomitantly to degradation and release of organic molecules occluded within the structure. In the range 423−623 K, the c parameter dilatation increases, leading to the anisotropic increase of (18 ± 11) % in branches and (43 ± 3) % in sclerites of the cell volume expansion. Conversely, from 623 to 823 K, the c parameter dilatation highly decreases, corresponding to the anisotropic decrease of (36 ± 3) % in branches and (60 ± 19) % in sclerites. The relative unit cell contraction, Δa/a = −1.1 × 10−4 and Δc/c = −4.3 × 10−4, measured at ambient in red coral skeletons annealed at 823 K is attributed to the native trace elements such as sulfate evidenced by FTIR analyses. The thermal decomposition starts in the range 823−923 K. The formation of lower Mg-calcite phases as intermediate phases supports the direct formation of calcite via a gradual release of Mg2+ ions to form calcite and periclase. The produced calcite is 400 ± 40 nm in size and shows a c-elongated unit cell than can be ascribed to lattice defects due to incorporated strontium and inorganic sulfate ions. Differences in decomposition steps and degree of conversion observed between branches and sclerites are responses to microstructural and chemical features (size, shape, hierarchical organization, and main trace elements Na, S, Sr, P, K) that are distinct in branches and sclerites. As important additional findings, the XRD-based calibration curves used to determine the Mg content in Mg-calcites are not applicable to biogenic Mg-calcites but efficient to evidence their anisotropic distortion compared to geologic or synthetic calcites.

1. INTRODUCTION Mg-calcite with Mg contents higher than 10 mol % is thermodynamically metastable under ambient conditions.1 Single-phase crystals can be synthesized under high temperature and high pressure.2,3 Despites this fact, Mg-calcite represents the second most important mineral in marine environments. It occurs in marine skeletal hard parts and cements.4−7 Mg-calcite produced by marine organisms is a biomineral, i.e., a mineral formed by living organisms. Environmental variables such as temperature, light, and seawater−carbonate saturation state influence biomineralization and affect the calcite Mg content. Thus, geological, geochemical, and environmental issues are closely linked to physical chemistry of precipitation, dissolution, and ionic diffusion of Mg-calcite compounds. Most biominerals are © 2015 American Chemical Society

nanocomposites in which mineral and organic components are inextricably associated down to the nanolevel. Most importantly, their properties differ from their inorganic or synthetic counterparts.7−11 Toward high Mg-calcite biominerals, echinoid skeletal elements such as seastar ossicles and especially sea urchin spines and teeth have attracted attention due to their availability and preservation as fossils.9,12−16 The apparently single-crystal character of the sea urchin spines and teeth led to numerous studies on their structure, crystalline properties, and composition. Conversely, few studies9 focused on their thermodynamic Received: March 2, 2015 Revised: May 16, 2015 Published: June 16, 2015 3690

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

Crystal Growth & Design

Article

is now well established, even if some particular features are not still well understood (ref 43 and references herein).44,45 Calcite is stable at room pressure up to 873 K, whereupon it breaks down to lime (CaO) and CO2 gas. In a flowing air atmosphere, the thermal decomposition of calcite into lime is complete at T = ∼ 1023 K, it is homogeneous, topotatic in nature, and does not depend on CO2 partial pressure and calcite crystal size.44 Under CO2 overpressure, calcite decomposition is quasi avoided. By heating a calcite sample loaded in a sealed quartz capillary (CO2 pressure increases to 4 atm at 1300 K), calcite structure shows an anisotropic dependence of cell parameters as a function of T and undergoes a reversible phase transition from the ordered R3̅c structure to the disordered phase R3̅m at temperature Tc = ∼1240 K. The transition involves orientational disorder of the CO3 anion groups.46−50 The CO3 groups are planar and along layers perpendicular to the c axis of the hexagonal cell. In R3̅c ordered calcite, successive layers of CO3 groups have apexes that point in opposite directions.50 In R3m ̅ disordered calcite, a CO3 group has an equal probability of pointing in either direction,47 and an umbrella inversion of the CO3 group has occurred.50 The a cell parameter shows a thermal contraction at low T, followed by a plateau at higher T, then a steeper contraction toward Tc = 1240 K, where the CO3 groups disorder rapidly. The c cell parameter shows a large anomalous expansion increasing toward Tc, so that V cell volume increases. The unit cell expansion is the effect of the strong anisotropy expansion of the CaO6 octahedra and the opening of the structure by tilting of octahedral elements.51 The thermal behavior and decomposition of dolomite CaMg(CO3)2 has been extensively studied but is still a subject of debate due to the diversity of experimental and analytical conditions. In air, dolomite thermal decomposition starts at ∼773 K and proceeds at a relatively slow rate up to 1023 K, forming Mg-calcite (Ca1−xMgxCO3), calcite (CaCO3), periclase (MgO), lime (CaO), and CO2 gas. The formation of lime (CaO) starts at T ∼ 973 K.52 Then, a faster, nearly constant decomposition rate is observed up to full decomposition at ∼1173 K in lime (CaO), periclase (MgO), and CO2 gas. After Mg diffusion out of Mg-calcite, the resulting calcite decomposes into CaO plus CO2 gas at T > 1023 K (ref 53 and references therein). In CO2 atmosphere, the dolomite decomposition occurs via an identical two-stage mechanism, the first step starting at higher T than in air. Samtani et al.54 reported periclase (MgO) formation at T = 880 K. By heating up to 875 K a dolomite sample in a sealed quartz capillary, the dolomite structure shows a uniform expansion of a and c cell parameters with increasing temperature, although not linearly.55 Thus, the thermal behavior of dolomite differs markedly from calcite in which the a parameter contracts with increasing T.56,57 Dolomite, like calcite, can be described as a corner-linked structure of filled octahedra and nearly planar CO3 groups. The lower symmetry of dolomite (space group R3̅, compared with R3̅c for calcite) results from alternating Ca and Mg layers perpendicular to the c axis and the slight rotation of CO3 groups which move the oxygen atom off the 2-fold axis that exists in calcite. Dolomite contains both Ca and Mg octahedra, which as a first approximation are similar to those of calcite and magnesite. The octahedra expansion is the major factor responsible for unit-cell thermal expansion. The expansion of the CO3 groups is relatively unimportant. Only a few studies have been devoted to synthetic magnesian calcite Ca1−xMgxCO3 because high crystallinity samples in the composition range x = 0−30 mol % MgCO3 are difficult to

properties such as thermal stability and transformation process by heating. In contrast, biogenic calcite and aragonite were more widely investigated. Their unusual thermal properties, anisotropic lattice distortions, and phase transition temperature are generally ascribed to intracrystalline organic matter.17−28 No such studies dealt with Mg-calcite biominerals, which may be due to marked internal heterogeneities of Mg and organic matrix in a single specimen and at all scales.16,29,30 Here the Mg-calcite thermal properties in red coral skeletons are investigated. Red coral, Corallium rubrum (Linnaeus 1758) (Cnidaria, Anthozoa, Octocorallia), possesses two different biomineral structures with different size, shape, and hierarchical organization:31−35 an axial skeleton and sclerites.36 Both structures are made of Mg-rich calcite (9−15 mol % MgCO3) that contain trace elements (mainly Na, S, Sr, P, K) and organic matter (OM) (1.2−1.7 wt %).35,37,38 The axial skeleton displays an arborescent arrangement of branches (cm size in diameter). A section normal to the axial skeleton branch shows a central crossed shaped medullar zone surrounded by an annular zone composed of concentric crenulated rings.31,36,39−42 Crenulated rings (100−200 μm thick) correspond to annual growth rings and are made of the stacking of micrometer layers with tortuous interfaces. Each layer is made of micrometer fibers. Fibers are superstructures made of submicrometer crystalline units. The annular zone is structurally arranged in radial herringbone strips (150 μm wide and few mm long) orthogonal to the concentric growth rings of the branch.31,32 Sclerites are small (50−90 μm in size) and specifically shaped grains found in the living tissues surrounding the axial skeleton (106 sclerites mg−1 tissue).36,40−42 Their chemical composition is slightly different from the axial skeleton. They display elaborate morphologies with tubercles in opposite tri- or tetrahedral arrangements consistent with the rhombohedral symmetry of calcite.33,34 Like the axial skeleton, sclerites are made of microlayers of well separated Mg-calcite mesocrystals of submicrometer size (80 ± 30 nm). In contrast to the axial skeleton, most sclerites display a single or twin structural arrangement with crystallites oriented along tubercle axes with only a low degree of nonrandom misorientation between them.33,34 Even though the pathways leading to crystallization of metastable high Mg-calcite remain unclear, the key to the biomineralization is the inclusion of trace elements and organic matter (OM) in Mg-calcite by living organisms. In this article, we present ex and in situ studies of the red coral skeleton transformation by heating up to 1073K. The aim is to better understand the structural differences between biominerals and their inorganic counterparts based on differences in thermal properties. A purely inorganic Mg-calcite containing 10 mol % MgCO3 has been synthesized at high temperature and high pressure to serve as reference. The structural transformations of red coral biogenic Mg-calcites, natural minerals (calcite, dolomite, and magnesite) and synthetic Mg-calcite are studied by high resolution X-ray diffraction on a synchrotron beamline and by Fourier transform infrared spectroscopy.

2. INFLUENCE OF TEMPERATURE ON THE STABILITY OF CALCITE, DOLOMITE, AND MAGNESIAN CALCITE: AN OVERVIEW OF PREVIOUS DATA The thermal behavior and decomposition of pure calcite as a function of temperature and partial pressure of carbon dioxide 3691

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Crystal Growth & Design

Article

Table 1. Nature, Name, Origin and Chemical Composition of the Studied Samples nature name origin chemical compositiona

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

a

calcite Iceland spar CaCO3

nonbiogenic crystals Mg10calcite Mg5.5calciteS0.1 synthetic synthetic Mg-calcite 9.7 mol % MgCO3

Mg-calcite 5.5 Mg mol % 980 ppm of S

Corallium rubrum skeletons branches and sclerites Triperie Cassis Creus Medes Marseille Coast, Cassis Coast, Cap de Creus, Medes Islands, France France Spain Spain Branches: 12.0 ± 1.0 mol % MgCO3, 4200 ± 500 ppm of Na, 3100 ± 400 ppm of S, 2600 ± 250 ppm of Sr Sclerites: 12.9 ± 1.0 mol % MgCO3, 3500 ± 400 ppm of S

From EMP and IC-PMS analysis.35,42

synthesize. At temperatures below 773 K, high Mg-calcite is expected to transform into dolomite and lower Mg-calcite at a very slow rate.58,59 The partial transformation of echinoid Mgcalcite into dolomite occurred at 573 K due to the presence of a small amount of water.9 In an air atmosphere furnace, full decomposition of Mg-calcite into calcite and periclase has been reported for Mg-calcitic shells and algae: for 6 and 21 mol % of MgCO3, it occurred at 1171 and 1160 K, respectively.60 Thus, the Mg−Ca substitution reduces the decomposition temperature of Mg-calcite. Mg-calcite has the R3c̅ calcite structure. Random substitution of magnesium for calcium results in a decrease of interatomic distances (Mg−O bond length in magnesite is 2.10 Å and Ca−O bond distance is 2.36 Å in calcite), in positional disorder and tilting of the CO3 groups out of the planes parallel to the a axes into the direction of the c axis to provide shorter Mg−O bonds. The c/a cell parameter ratios of synthetic Mg-calcite describe a smooth curve as a function of Mg concentration, the largest deviation occurs at about 25 mol % MgCO3, corresponding to the highest positional disorder of the CO3 groups.10,61−65

MgCO3 content of 5.5 ± 0.1 mol % and a sulfur content of 930 ± 35 ppm for the sulfated synthetic calcite. These numbers are different from the targeted values and indicate that significant amounts of Mg and S entered the melt phase. Nevertheless, enough Mg and S are present in the calcite to be used as a sulfated Mg-calcite standard. No sulfur-bearing mineral other than Mg-calcite was detected by X-ray diffraction (XRD). 3.2. Biomineral Sample Preparation. Colonies of C. rubrum were collected from the rocky coast near Marseille (France), Corsica (France), Cap de Creus (Spain), and the Medes islands (Spain). Soft tissues were removed from the skeletons by immersion in sodium hypochlorite for a few hours, and the skeletons were rinsed in distilled water. Red coral branches (5−8 mm in diameter) were subsampled transversely into pieces, each containing a complete circular section of the branch (referred to as bulk). The 150−200 mg bulk samples were cleaned in distilled water then ethanol and dried in a flow bench. Portions of skeletons and sclerites were finely ground into powder in an agate mortar for in situ thermal treatments. Studied samples are listed in Table 1.

3. SAMPLE PREPARATION 3.1. Synthetic Mg-Calcite. To obtain pure inorganic standards, Mg-calcite with and without sulfur at high pressure and temperature were synthesized. For the 10 mol % MgCO3 Mg-calcite, a starting material composed of a mixture of pure synthetic calcite (Merck supra pure) and natural magnesite from Brumado (Bahia, Brazil) was used. Inductive coupled plasma−atomic emission spectrometry (ICP−AES) analyses of this magnesite indicate trace amounts of Fe (0.15 wt % Fe203) and Ca (0.22 wt % CaO). A calcite−magnesite mixture 9:1 mol % was carefully mixed and ground in a McCrone micronizer agate mortar under ethanol. About 60 mg of the mixture was put in a 7 mm high, 5 mm outer diameter gold capsule, and then the capsule was welded shut. The gold capsule was placed in a salt−glass assembly, pressurized, and heated in a piston− cylinder apparatus at 1000 °C, 1 GPa for 4 days. Run products were checked by scanning electron microscopy (SEM) and showed 100−200 μm nicely shaped euhedral crystals of Mgcalcite. Electron microprobe (EMP) analyses indicate a MgCO3 content of 9.7 ± 0.3 for the synthetic crystals. A similar procedure was used to synthesize a sulfated Mgcalcite. For this starting material, synthetic Alfa Aesar calcite was used and mixed with the Brumado magnesite. The same calcite/magnesite ratio 9:1 mol % was used, and sulfur was added as anhydrous magnesium sulfate (Alfa Aesar). The experiment was carried out in a piston−cylinder apparatus at 1000 °C, 1 GPa, for 4 days. For this run, EMP shows the presence of a small proportion of peripheral quenched phase, indicating that the mixture partially melted during the experiment. A large block of crystalline product was separated from the quenched phase and studied. EMP analyses indicate a

4. EXPERIMENTAL PROCEDURE 4.1. Heat Treatment. Heat treatment of bulk sections and sclerites was performed in an electric furnace under CO2 atmosphere at a flow rate of 750 mL/min. Different samples were kept for 2 h at temperatures of 423, 473, 673, 873, and 1023 K, then rapidly cooled to room temperature in the flow of CO2 prior to removal. Under these conditions, the full decomposition of Mg-calcite into lime is hindered or avoided. Heated bulk samples remain solid and preserve most of the original macroscopic features of the red coral skeleton. The most obvious difference between as-prepared and heated bulk samples and sclerites is their color. After heat treatment, the original red color changes into a range of colors: samples heated at 423 and 473 K are pale pink to brown colored, then they become pale-gray at 673 K, dark-gray at 873 K, and white−gray at 1023 K (Figure 1). 4.2. Thermal Analyses. Thermogravimetry (TG) and differential thermal analysis (DTA) were carried out on a Setaram SENSYS evo TG-DSC 153−1103 K thermal analyzer (LICB, Dijon, France). The powdered sample was placed in a platinum crucible (0.15 cm3) under CO2 atmosphere at a flow rate of 20 mL/min. The rate of heating was fixed at 3 °C/min. Constant weights of sample (∼50 mg) were used to avoid the effect of variation in sample weight on peak shape and temperature. Alpha alumina was used as a reference material. 4.3. Fourier Transform Infrared Attenuated Total Reflection (ATR/FTIR) Spectroscopy. ATR/FTIR spectroscopic studies were carried out at 298 K on a Bruker Vertex 70 spectrometer using a horizontal ATR trough mounted in the sample compartment and equipped with a DLaTGS detector continuously purged with gaseous nitrogen. A MIRacle ATR unit from Pike equipped with a single reflection Ge crystal was used. Absorbance spectra were acquired from powdered samples in the 650−4500 cm−1 range (2 cm−1 spectral resolution), using 800 scans acquired in 20 min. The spectra were apodized applying the Blackman−Harris 3-term function, and the baseline corrected by third-order polynomial and normalized using the Opus 6.1 software supplied by Bruker Optik GmbH, Ettlingen, Germany. 3692

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

Crystal Growth & Design

Article

function in the refined model.67 Subsequently, red coral pattern refinements reached markedly low GoF parameters using the numerical profile function extracted from experimental Bragg (104) peak shape. Assuming that each Mg-calcite single-phase peak is characteristic of a 85 nm grain size and that microstructural defects are equal to the mean value of the synthetic Mg10calcite sample, the Mg content phase distribution within the red coral sample can be extracted directly from the experimental Bragg (104) peak shape. Its maximum and standard deviation correspond to the refined cell parameters and the integral breath square root (√β). 4.4.2. Size-Microstrain Line Broadening Analysis Procedure. Heat treatment of red coral samples is expected to induce lattice parameter distortions but also change the average grain size and crystal lattice microstrain. For such microstructure analyses, the common practice is to estimate coherent domain size ⟨L⟩ and average microstrain values ⟨ε⟩ from the refined profile width parameters. The size effect is given by the Scherrer68 formula: ⟨L⟩ = λ /βcos(θ) where λ, β, and θ are the wavelength, integral breath, and Bragg angle, respectively. The microstrain effect is the variance of the lattice spacing (Δd/d) and is defined by Stokes and Wilson69 with an angular dependence of the form: Δd/d = β/tan(θ). A complete study of line-broadening analysis and Rietveld refinement has been presented by Delhez et al.70 and comprehensively demonstrated by Balzar et al.71 Here we used the common approximation implemented in the Fullprof program,64 where the broadened profile is analyzed by a pseudo-Voigt function (pV). The method is fully described in Supporting Information.

Figure 1. Optical images of red coral branches untreated (a) and after annealing under CO2 atmosphere for 2 h at 473 K (b), 673 K (c), 873 K (d), and 1023K (e). 4.4. High-Resolution Synchrotron X-ray Powder Diffraction (SXRPD) Studies. High-resolution X-ray powder diffraction measurements were performed at the ID31 beamline of the European Synchrotron Radiation Facility (ESRF, Grenoble, France) equipped with a double-crystal monochromator Si (111) and crystal analyzer optical elements in the incident and diffracted beams, respectively. A description of the diffraction instrument is given by Fitch.62 The setup allows high quality powder diffraction patterns with high signal/noise ratio, combined with narrow peaks, accurate positions and intensities. The instrumental contribution to peak width (full width at halfmaximum, fwhm) is not exceeding 0.003° 2θ. The typical resolution is Δd/d ∼10−4. Incident beam size on the sample is typically 1.5 mm × 1.5 mm. The selected wavelength λ = 0.47682(8) Å (26 keV) was calibrated with Si standard NIST 640c (certified cell parameter a = 5.4311946(92) Å). The 1 mm diameter quartz capillaries were filled with the sample powder, then sealed, mounted horizontally, and spun at 1000 rpm during data collection to improve particle statistics. Data were collected in the 1° < 2θ < 45° range in continuous motion 5°/ min and rebinned in 2θ step of 0.003°. Data acquisition time was 30 min for each sample at room temperature (RT) to ensure good counting statistics. For in situ thermal treatment, the capillaries were heated with a hot air blower mounted vertically, perpendicular to the capillary. Temperature was controlled by a thermocouple located in the hot air stream. The measurements (three recording patterns of 10 min at each T) were performed at RT, from 323 to 1023 K in steps of 50 K (heating rate of 10 K/min) and again at room temperature after cooling. A wavelength οf 0.39992(8) Å (31 keV) was used. 4.4.1. SXRPD Pattern Analysis Procedure. Diffraction patterns were analyzed with the Rietveld technique63 using the Fullprof program.64 The calcite and Mg-calcite structure parameters (space group R3̅c, hexagonal setting) given by Althoff65 and Malsen et al.66 were used as initial parameters in the refined model. Atomic positions and displacement parameters were preserved. Site occupancy of the Ca and Mg cations was constrained to keep the stoichiometry of the Mgcalcite. Pseudo-Voigt and Thompson−Cox−Hastings pseudo-Voigt profile functions were selected for reference calcite and synthetic Mgcalcite pattern refinement. By contrast, any standard analytical profile function yields reasonable goodness-of-fit (GoF) parameters for pattern refinement of red coral skeletons. As quoted in the introduction, red coral samples are microcrystalline mixtures of multiple Mg-calcite phases. The grain size distribution determined by atomic force microscopy (AFM) and SEM is unimodal (mean size 85 nm; standard deviation 30 nm, range 35−155 nm), and the Mg content determined by EMP is an almost log-normal distribution (mean Mg content 12.0 ± 1.0 mol % (standard deviation); range 9− 15 mol % Mg content).31−35 As a result, a red coral diffraction pattern is a superposition of a continuous series of single Mg-calcite phase patterns. The profile fitting issue is solved by selecting a numerical expression of the experimental diffraction peak profile as a profile

5. RESULTS AND DISCUSSION 5.1. Thermal Analyses. Figure 2 shows TG and DSC curves of a powdered bulk sample. The TG curve consists of

Figure 2. Simultaneous thermogravimetry (red line) and differential scanning calorimetry (blue line) of red coral skeleton. The weight loss can be assigned to three successive steps: loss of water, exothermic decomposition of organic matter, and endothermic decomposition of Mg-calcite into MgO and CO2.

three weight loss steps in the temperature ranges 300−390, 390−820, and 820−930 K, respectively. The first two low slopes can be assigned to the loss of adsorbed water (∼0.4 wt %) and to the exothermic degradation of organic matter (∼1.9 wt %). In the second step, the substep at 490 K is identified by Raman spectrometry as the degradation of the red pigments. The third step is steeper and can be attributed to the partial decomposition of Mg-calcite into MgO and CO2. This decomposition is correlated with a sharp endothermic peak in the DSC curve at 904 K. The calculated weight loss in this region is 5.1 wt % and is characterized by a slope change at 880 K. If we assume that all CO2 released during this step originates from MgCO3 decomposition, the 5.1 wt % weight loss corresponds to 11.6 mol % of MgCO3 in the biogenic Mgcalcite. This amount is consistent with chemical analyses presented earlier (Figure 3 in ref 35). This result suggests that 3693

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Crystal Growth & Design

Article

The observed absorption frequencies are presented in Table 2. The infrared spectrum of calcite is characterized by absorption bands assigned to the fundamental vibrational modes of the carbonate ion group in calcite:72−77 the doubly degenerated out-of-plane bending (ν4) at 713 cm−1, the asymmetric stretching (ν2) at 872 cm−1, and the doubly degenerated in-plane-bending (ν3) modes at 1398 and 1405 cm−1. Besides the fundamental modes, the combination (ν1 + ν4) and (ν1 + ν3) modes are observed at 1796 and 2513 cm−1, respectively. The satellite band of the ν2 band due to the isotopic substitution of 12C by 13C is well observed at 848 cm−1 (natural abundance = 1.1% 13C). The 24 cm −1 shift corresponding to frequency ratio ν2(13CO32−)/ν2(12CO32−) equal to 0.97 is in agreement with the experimental and theoretical values determined for rhombohedral carbonates (0.969).78,79 The weak band at 1090 cm−1 is assigned to the symmetric stretching (ν1) mode of the carbonate ion group even though IR forbidden. The splitting of the doubly degenerate ν4 and ν3 vibration modes as well as the infrared activation of ν1 are connected to changes in the electronic structure and distortions to the equilibrium positions of the CO32− dipolar anion. Actually, the small splitting of 7 cm−1 for ν3 and the very low intensity for ν1 make difficult the assignment of a special significance to this relaxation in selection rules. The frequencies of the vibrational bands are in good agreement with literature values (Table 2 in ref 80). The infrared spectrum of synthetic Mg10calcite exhibits a broad ν4 band at 716 cm−1 resolved into two overlapping bands at 711 and 717 cm−1 and a ν3 band at 1404 cm−1 with a shoulder at 1450 cm−1. On the other hand, the small splitting of ν3, 7 cm−1, band in calcite is no longer observed. The ν2 and ν1 bands are located at 873 and 1087 cm−1, respectively. The splitting of the doubly degenerate ν4 and ν3 vibration modes is assigned to the structural distortion of CO32− from regular planar symmetry due to the Ca2+ substitution by Mg2+. The shift toward higher wavenumbers results from higher cation Mg2+ polarizing power in comparison to that of Ca2+ in pure calcite.10,79,81,82 The infrared spectrum of synthetic Mg5.5calciteS0.1 exhibits a broad ν4 band at 714 cm−1 resolved into two bands at 711 and 715 cm−1 and a ν3 band at 1400 cm−1 with asymmetric shape. No marked shoulder or splitting of ν3 band is observed. The ν2 and ν1 bands at 871 and 1087 cm−1 are similar to those of calcite. The weak and very wide band with two well-resolved peaks at 1143 and 1167 cm−1 cannot be attributed to calcite or Mg-calcite. This additional band can be likely assigned to the ν3 vibration of SO42− anions. Indeed, the sulfate group has four fundamental vibrational modes. In a solid-state sulfate, internal vibrational features generally appear at ∼1050−1250 (ν3),

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

the entire magnesite content of the solid-solution is decarbonated at 880 K. 5.2. FTIR Analyses. Figure 3 displays the FT-IR spectra of reference calcite (Iceland spar), synthetic Mg-calcite with 9.7 ±

Figure 3. ATR/FT-IR spectra of Iceland spar (black line 1), Mg10calcite (blue line 2), Mg5.5calciteS0.1 (green line 3) and red coral branches Triperie (red line 4) in the wavenumber ranges: (a) 600−2600 cm−1, (b) 680−750 cm−1, (c) 820−920 cm−1, (d) 1050− 1310 cm−1.

0.3 mol % MgCO3 (Mg10calcite), synthetic Mg-calcite with 5.5 ± 1.0 mol % MgCO3 and 980 ppm sulfur (Mg5.5calciteS0.1), and red coral Triperie. The FTIR spectra of powdered bulk samples (Triperie) and sclerites show similar peak values and shapes. Thus, the sclerite spectrum is not shown and the Triperie sample is taken as representative example.

Table 2. Assignment of the Vibrational Modes of Carbonate Ion and Sulfate Ion for Calcite, Synthetic Mg-Calcite, and Red Coral Triperie vibrational modes and frequencies (cm−1) carbonate ion a

calcite Iceland spar Mg10calcite Mg5.5calciteS0.1 Triperie a

sulfate ion

ν4

ν2(13C)

ν2(12C)

ν1

ν3

ν1 + ν4

ν1 + ν3

ν3

712 713 716 714 711−717

848 848

876 872 873 871 873

1090 1087 1087 1088

1435 1398−1405 1404 1400 1405

1812 1796 1799 1800 1799

2545 2513 2523 2519 2523

1143−1167 1145−1166

sample

From Huang and Kerr.77 3694

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

Crystal Growth & Design

Article

∼1000 (ν1), ∼500−700 (ν4), and ∼400−500 (ν2) cm−1. The ν3 band is dominant. The ν2 band is known to be significantly weaker than the ν1 mode and commonly not observable in the sulfate infrared spectra.83 The ν3 band is a triply degenerate asymmetric stretching band, found at 1121−1201 cm−1 for calcium sulfate and shifted to higher frequency (larger wavenumber) for magnesium sulfate.84−87 The large splitting of the ν3 bands is related to a large distortion of the sulfate tetrahedron. The presence of ν3 band is consistent with trace amounts of sulfur present in the sample as a sulfate group. The ν3 vibrational features are linked to cationic environments of sulfate sites. The infrared spectrum of Triperie (red coral skeleton) exhibits a broad ν4 band at 714 cm−1 resolved into two bands at 711 and 715 cm−1 and a ν3 band at 1405 cm−1 with asymmetric shape, a shoulder at 1450 cm−1, and no splitting as previously observed in calcite. The ν2 and ν1 bands at 873 and 1088 cm−1 are similar to those of calcite. The wide splitting of ν4 band could be connected to the large histogram of Mg distribution (12% ± 3%), and to a lesser extent to trace amount of Na, Sr, and K in the sample.35 The weak and broad band with two well-resolved peaks at 1145 and 1166 cm−1 is observed, together with a shoulder at 1248 cm−1. This band is similar to the one observed in Mg5.5calciteS0.1. This observation suggests an assignment to the ν3 vibration of SO42− as a result of the incorporation of sulfur as SO42− anion in the Mg-calcite structure in the red coral skeleton. This assignment is likely but not exclusive. Indeed, a band in this 1100−1250 cm−1 region could be also assigned to the inorganic/organic component of the red coral. Indeed, it coincides with strong infrared absorptions of most organic molecules (C−N stretching of proteins (glycosaminoglycans),88,89 C−C stretching of red pigments (trans-canthaxanthin carotenoid and trans-polyacetylene molecules),90−95 S−O bending of sulfate sugars (chondroitin),96,97 and S−O bending of sulfate sugars (chondroitin).96,97 Such components could be present in the organic matter occluded in the Mg-calcite crystals of the red coral skeleton.35 Figure 4 displays the FT-IR spectra of the red coral branches Triperie untreated and annealed for 2 h under CO2 atmosphere at 423, 473, 673, 873, and 1023 K. The observed absorption frequencies are listed in Table 3. At 423 and 473 K, temperatures at which only loss of water occurs, the two FTIR spectra are similar; the four fundamental bands and their combination assigned to Mg-calcite together with the additional bands in the 1100−1200 cm−1 region are unchanged. At 673 K, the degradation of organic matrix is almost complete. The three ν2, ν1, and ν3 bands assigned to CO32− anion at 873, 1088, and 1404 cm−1 and the combination bands are practically unchanged in shape and position. Only two bands are affected. The fundamental ν4 band assigned to CO32− anion is slightly narrowed, resolved in two peaks at 712 and 716 cm−1. The tighter splitting of the ν4 band indicates a reduction of coordination effect of the cations and less distortion of site symmetry. By this slight change, the carbonate vibration part of the spectrum becomes comparable to what is observed in Mg10calcite (Figure 3b). The major change occurs in the 1100−1200 cm−1 region attributed to the vibration bands of structurally substituted sulfate and carbon−carbon bond of occluded organic molecules. This large band flattens, broadens toward lower wavenumbers, and is resolved in at least three overlapping bands (Figure 4d). At this stage, the origin of the change is ascribed to the degradation of active organic

Figure 4. ATR/FT-IR spectra of red coral branches Triperie untreated (RT, red line) and after annealing under CO2 atmosphere for 2 h at 423 K (blue line), 473 K (green line), 673 K (pink line), 873 K (purple line), and 1023 K (black line), respectively, in the wavenumber ranges: (a) 600−2600 cm−1, (b) 680−750 cm−1, (c) 820−920 cm−1, and (d) 1050−1310 cm−1.

molecules such as chonchroitin98 and red pigments.99 Comparable evolution of the sulfate ν3 band is observed between mono- and bicationic sulfate minerals.85,100,101 In addition, the 1248 cm−1 shoulder is observed in Mg2Ca(SO4)3 sulfate.85 This band observed in previous studies of Corallium species remained unassigned.82,91,94 In the absence of OM destroyed at this T, this large band is assigned to triply degenerate ν3 band (plus extra satellite peaks) of the SO42− occupying various cationic environment in Mg-calcite. At 873 K, the endothermic decomposition of Mg-calcite into MgO and CO2 begins. The carbonate vibration part of the spectrum is almost comparable to the one observed in the Iceland spar. The main changes correspond to the sharpening of the ν4 band and the disappearance of the ν3 band shoulder. The sulfate vibration part in the 1100−1250 cm−1 range is markedly affected; the large sulfate ν3 band is a large bump shifted to lower wavenumbers. The shoulder previously observed at 1248 cm−1 is absent. At 1023 K, the decomposition of Mg-calcite into calcite, MgO, and CO2 is complete. In the carbonate vibration part of the spectrum, the 7 cm−1 splitting of the ν3 band is marked. Both position and shape of the carbonate bands are consistent with Mg-free calcite (Iceland spar). In the sulfate vibration part, the large smooth ν3 band changes to an irregular bump resolved in three peaks at 1104, 1124, and 1160 cm−1. This modification reflects the environmental change of sulfate sites 3695

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Crystal Growth & Design

Article

Table 3. Assignment of the Vibrational Modes of Carbonate and Sulfate Ions for Red Coral Branches Triperie Untreated (RT) and after Annealing under CO2 Atmosphere for 2 h at 423, 473, 673, 873, and 1023 K vibrational modes and frequencies (cm−1) carbonate ion ν4

ν2

ν1

ν3

ν1 + ν4

ν1 + ν3

ν3

RT (untreated)

711 717 711 717 711 717 717 713 713

873

1088

1405

1799

2523

1145−1166

873

1088

1405

1799

2523

1145−1166

873

1088

1405

1799

2523

1145−1166

873 873 872

1088 1091 1091

1404 1400−1405 1398−1405

1799 1796 1796

2523 2513 2513

1124−1142−1165 1124−1144−1162 1104−1124−1160

423 473 673 873 1023

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

sulfate ion

annealing temperature

Ca−O and Mg−O bond lengths (>10%), significant cation and carbonate group positional disorder, octahedral distortion, and carbonate group tilting have been identified in the Mg-calcite structure.10,56,65,102 Mg10calcite pattern features could also reflect some locally restricted ordering in the distribution of Mg and Ca over the octahedral sites in the calcite structure. Various correlation diagrams have been proposed to determine Mg contents from lattice parameters.11,25,26,61,103 Negative deviations from linearity were measured from high pressure and temperature synthetic samples61 within a large Mg content range. By using the cell parameters (a and c), c/a ratio, volume (V), and (104) interplanar distance (d(104)) calibration curves provided by Bischoff et al.,61 the Mg content determined in the synthetic Mg10calcite are 9.71, 9.93, 9.89, 9.73, and 9.85%, respectively. These different values and the mean value 9.82 ± 0.10% are in good agreement with the EMP chemical composition of 9.7 ± 0.3 mol % MgCO3. The best degree of agreement from the highest to the lowest is successively observed from a, V, d(104), c/a, and c calibration curves. As a result, the cell parameter a is the best reliable parameter to estimate the Mg content from the Bischoff’s calibration curve; otherwise, note that the largest Mg content difference (0.23%) is related to a misfit of the c parameter Δc = 0.0043 Å, i.e., a relative misfit Δc/c = 2.55 × 10−4. 5.3.1.2. Mg Distribution in Red Coral Skeletons. Compared to synthetic Mg10calcite pattern, the Triperie red coral sample yields an unsymmetrical broadening peak pattern slightly shifted toward lower 2θ positions. As described in section 4.2, the red coral patterns are refined as a single Mg-calcite phase pattern using a numerical profile function generated from the (104) peak shape. Goodness-of-fit parameters are low values, peak positions are well fitted, and differences in intensity are observed for some reflections. Mg content was calculated from the lattice parameters according to the different calibrations of Bischoff et al.61 The Mg content determined from a, c, c/a, and V parameters and some characteristics of the Mg distribution determined from (104) profile for all samples are reported in Table S1 in Supporting Information. On the other hand, the Mg distribution curve was deduced from the (104) peak numerical profile deconvoluted from grain size and microstructural defects response. Table 4 summarizes the refined structural parameters. The hexagonal cells of all coral samples differ significantly from those of synthetic samples with similar Mg content, Mg10calcite, and those published by Bischoff et al.61 The hexagonal cell deformation is shown by plotting c versus a parameters (see Figure 6a) and emphasized by plotting c/a

in relation to the loss of Mg to form MgO. The shoulder at 1248 cm−1 is still absent, confirming its link with Mg and not with biomolecules entrapped in calcite. 5.3. Synchrotron X-Ray Powder Diffraction (SXRPD) Analyses. 5.3.1. Structural Parameters and Magnesium Quantification in “As Prepared” Branches and Sclerites. Powdered red coral axial skeleton and sclerites display analogous SXRPD patterns with broad and dissymmetric peaks at positions corresponding to Mg-calcite diffraction pattern.34 For clarity, sclerites diffractograms are not reported in the following figures. Figure 5 displays the SXRPD patterns

Figure 5. SXRPD patterns for calcite Iceland spar (black), synthetic Mg-calcite Mg10calcite (blue), and red coral Triperie (red) (wavelength of 0.39992(8) Å (31 keV)): (a) Bragg peak indexation in the 2θ range from 7.2 to 12.5 (deg); (b) Bragg (104) peak in the 2θ range from 7.48 to 7.78 (deg).

collected at room temperature for the Triperie sample, Iceland spar, and the synthetic Mg-calcite (Mg10calcite). Refined structural parameters are reported in Table 4. Refined unit cell parameters of the reference calcite agree well with those of the literature.51,61 5.3.1.1. Mg Content Determination in Synthetic Mg10calcite: Accuracy of the Calibration Curves. The synthetic Mg10calcite pattern shows well-shaped nonsplit diffraction peaks that allow excellent refinement in terms of single-phase, MgxCa1−xCO3 solid solution. Compared to Iceland spar, the goodness of fit parameters are slightly higher and the peak shape is broadened (fwhm (104) = 0.0069° (2θ) compared to 0.0050° (2θ) for the Iceland spar). These differences arise from structural defects associated with the substitution of Ca by Mg. Indeed, given the large difference in 3696

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Crystal Growth & Design

Article

Table 4. Refined Unit Cell Parameters, Microstructural Parameters (fwhm Broadening, Size ⟨L⟩V and Microstrain ⟨ε⟩), and Conventional Rietveld R-Factors for Untreated Samples: Calcite (Iceland Spar), Synthetic Mg-calcite (Mg10calcite), Red Coral Branches (Cassis, Triperie, Creus, and Medes), and Red Coral Sclerites (Cassis, Medes) (Wavelength of 0.47682(8) Å (26 keV)) unit cell parameters samples

calcite

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

a (Å) 4.9906240(9) c (Å) 17.065273(4) fwhm (104) 2θ (deg) 0.0046 541 ⟨L⟩V (nm) ⟨ε⟩ 10−3 0.197 Conventional Rietveld R-Factors 11.10 Rp % 15.30 Rwp % 12.61 R exp % GoF 1.47

branches

sclerites

Mg10calcite

Cassis

Triperie

Creus

Medes

Cassis

Medes

4.9472632(4) 16.864304(3) 0.0069 397 0.582

4.94812(5) 16.89286(6) 0.036

4.94826(6) 16.89294(2) 0.032

4.94705(1) 16.88401(2) 0.030

4.94662(4) 16.87694(5) 0.032

4.94486(9) 16.86584(3) 0.039

4.94194(1) 16.84792(7) 0.048

10.80 14.30 4.67 9.39

13.20 16.40 4.85 11.4

12.00 15.30 4.84 9.93

10.20 13.40 4.58 8.60

10.70 13.90 4.61 9.16

10.90 14.30 4.85 8.69

12.30 16.70 9.39 3.17

samples than the standard EMP multipoint analyses (see details in the Supporting Information). Indeed, the Mg distribution curves plotted in Figure 7 show that (1) mean values of the Mg

Figure 7. Mg content distribution curves: (a) for red coral branches Cassis (purple), Triperie (red), Creus (green), and Medes (blue); (b) for red coral sclerites Cassis (purple) and Medes (blue).

distribution of branches of different colonies are comparable and differ slightly from the sclerites that are Mg enriched by about 1.3 mol %, and (2) standard deviation of the Mg distribution curve are larger for the sclerites than for the skeletons. 5.3.2. Structural and Microstructural Parameters in Branches and Sclerites Annealed in Controlled CO2 pressure. 5.3.2.1. Thermally Induced Modifications of Biogenic Magnesium Calcite. Figure 8 displays the SXRPD patterns collected at room temperature of the red coral Triperie samples annealed at various increasing temperatures, the reference calcite (Iceland spar), and the synthetic Mg-calcite (Mg10calcite) are shown for comparison. As a reminder, each annealed sample of coral branches and sclerites are distinct but obtained from the same coral branch and colony, respectively. Heating the red coral samples results in SXRPD pattern modifications indicative of changes of structural and microstructural parameters, or Mg distributions, or both. Table 5 reports refined structural and microstructural parameters. Up to 673 K, temperature at which the degradation of organic matrix is complete, SXRPD patterns do not change significantly. Only slight shift of the peak positions and increase of the peak fwhm are observed (Figure 9). These modifications may come from a change of lattice parameters, a decrease of average grain size, an increase of crystal lattice microstrain (Table 5), and variation of the initial Mg distributions (sampling effect). Calculated lattice parameters (a and c) and

Figure 6. Hexagonal a and c unit-cell parameters for untreated red coral branches (Cassis, Triperie, Creus, and Medes) (red squares), sclerites (Cassis, Medes) (green circles), and synthetic Mg-calcite (Mg10calcite) (blue plus), second-order regression curve (black line) after Bischoff et al. (1983)61 for synthetic Mg-calcite samples (gray double cross): (a) unit cell c parameter versus a parameter; (b) cell edge ratio c/a versus a parameter.

ratio versus a parameter (Figure 6b). The values in coral samples are out of the trend line determined by Bischoff et al.61 Each offset reflects an elongation of c versus a parameter in the hexagonal cell. In other words, for a given a parameter, the c parameter is significantly elongated in coral branches (Δc/c = 7 × 10−4 to 1.22 × 10−3) and a bit less in sclerites (Δc/c = 3.2 × 10−4 and 5.4 × 10−4) and thus significantly above Δc/c = 2.55 × 10−4 measured for the synthetic Mg10calcite. As a result of this large crystalline anisotropy, the use of the cell parameter calibration curves for Mg content in coral samples is not applicable (see Table S1 in the Supporting Information). This conclusion confirms the statement of Bischoff et al. (1983) that existing X-ray calibration curves based on synthetic phases can lead to errors of over 5 mol % in the Mg content of biogenic Mg-calcites. Nevertheless, the use of the (104) profile is still of interest and more efficient to determine the standard deviation of Mg content in coral 3697

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

Crystal Growth & Design

Article

calcite with 1.82 ± 1.05 mol % Mg and 10.3(1) mol % periclase corresponding to a fraction decomposed of MgCO3 (i.e., a degree of conversion α) equal to 95.1%. On the other hand, sclerites transformed into a mixture of 88.9(4) mol % Mgcalcite with 2.08 ± 2.6 mol % Mg and 11.1(7) mol % periclase, i.e., a degree of conversion α = 85.7% (Supporting Information Table S2 and Figures 9 and 10). Typically, branches decomposed into calcite, low Mg-calcites, and periclase. In contrast, no pure calcite is observed in sclerites composed of a large series of low Mg-calcites asymmetrically distributed toward high Mg-calcite and periclase. No lime was found. The decomposition into calcite and periclase is incomplete after 2 h, indicating a low rate of decomposition of the coral samples in CO2 partial pressure. All samples are partially and differentially decomposed via intermediate low Mg-calcite phases, and sclerites are slightly less decomposed than branches according to the periclase amount. As already noted, the isothermal decomposition process results in the formation of low Mg-calcites in all samples. This observation excludes the direct formation of calcite and supports a step-by-step process via metastable solid-solution precursors according to the reaction:

Figure 8. SXRPD patterns at RT in the 2θ range from 8.95 to 9.25 (deg) showing the (104) peak for calcite Iceland spar (black crosses), synthetic Mg-calcite Mg10calcite (blue plus), and red coral branches Triperie untreated (red squares) and after annealing under CO2 atmosphere for 2 h at 423 K (blue circles), 473 K (green triangles), and 673 K (purple lozenges) (wavelength of 0.47682(8) Å (26 keV)).

distortions (Δa/a and Δc/c) are variable for the branches but decrease with increasing temperatures for the sclerites. fwhm increase of the heated sclerites can be ascribed to the microstrain value changes from ⟨ε⟩ = (0.58 ± 0.01)103 at 273 K to (1.24 ± 0.01)103 at 673 K. The hexagonal cell deformation can be put forward by plotting c/a ratio versus a or c parameter (Figure 10): from RT to 673 K, annealed sclerites values continuously decrease toward the trend line of the synthetic Mg-calcites; in contrast, annealed branches values are scattered around the initial value at RT. The anisotropic lattice distortion decrease observed in the sclerites could be readily associated with the intracrystalline organic matter degradation completed at 673 K. Such an effect was also expected in branches. It was not observed certainly because each untreated sample had different initial Mg content (due to the large Mg distribution present in the same branch) that actually biases the cell parameter modifications induced by annealing. 5.3.2.2. Thermal Decomposition of Biogenic Magnesium Calcite. In our experiments, thermal decomposition of the coral branches and sclerites starts at 873 K. At this T, SXRPD patterns change significantly (Figure 11). The dissymmetric and large peaks characteristic of the Mg-calcite distribution in the coral samples disappear. Instead, SXRPD patterns show peaks with a dissymmetry toward high 2θ diffraction angles, consistent with the presence of calcite and low Mg-calcite distribution and small and broad peaks attributed to periclase (MgO). Thus, after 2 h at 873 K, “Triperie” sample tranforms into a mixture of 60.5(0) mol % calcite and 29.2(5) mol % Mg-

Ca(1 − x)MgxCO3(s) → Ca(1 − x − y)Mg(x − y)CO3(s) + y MgO(s) + yCO2 (g)

Such a reaction mechanism referred to later as “partial decomposition” is comparable to what was first described by Hashimoto et al.59 and then other workers55,104−106 for natural dolomite decomposition in CO2 atmosphere. The differences observed in decomposition steps and degree of conversion are likely a response of the microstructural and chemical features (size, shape, hierarchical organization, and trace element amount) that are distinct in branches and sclerites. At 1023 K, branches are fully transformed into calcite and periclase, without any trace of lime. The amount of periclase produced (10.9 (3) Mg mol %) is in agreement with the initial content of Mg in branches (Supporting Information Table S2). In contrast, sclerites transformed into calcite, periclase, and low Mg-calcite (1.74 mol % MgCO3). The amount of low Mgcalcite and periclase corresponds to 0.39 Mg mol % and 12.66 (6) Mg mol %, respectively, and the total of 13.05 Mg mol % is in good agreement with the mean values calculated from IDICPMS measurements in sclerites (Vielzeuf, unpublished results). The produced calcite is characterized by almost symmetric peaks with a Lorentzian broadening corresponding to average sizes in the range 432−372 nm. All peaks are slightly shifted from the Iceland spar. Typically, the intense (104) peak

Table 5. Refined Unit Cell Parameters, Parameter Distortions, Microstructural Parameters (fwhm Broadening, Average Size ⟨L⟩V, and Microstrain ⟨ε⟩), and Mg Distributions for Untreated and Annealed Samples for 2 h under CO2 Atmosphere: Branches Triperie and Sclerites Cassis (Wavelength of 0.47682(8) Å (26 keV)) structural parameters red coral sample

annealing temperature (K)

a (Å)

branches Triperie

293 423 473 673 293 473 673

4.948265(2) 4.946727(9) 4.949024(2) 4.948889(2) 4.9446(8) 4.9432(2) 4.9425(6)

sclerites Cassis

104 Δ(a)/a

c (Å) 16.892942(2) 16.885780(4) 16.898081(2) 16.890602(2) 16.8658(4) 16.8548(9) 16.8457(6)

3.11 1.54 1.26 2.95 4.29 3698

104 Δ(c)/c −4.24 3.04 −1.38 −6.49 −11.92

fwhm (104) 2θ (deg)

⟨L⟩V (nm)

⟨ε⟩ 10−3

0.038(8) 0.042(8) 0.041(8) 0.040(9) 0.047(8) 0.052(6) 0.052(4)

85 60 60 61 85 85 23

0.58 0.69 0.61 1.12 0.58 1.22 1.24

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Crystal Growth & Design

Article

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

Figure 9. (a) a unit cell parameter. (b) c unit cell parameter. (c) fwhm for the (104) peak as a function of temperature of the annealed samples, red coral branches Triperie (red squares), sclerites Cassis (green circles). Data for untreated calcite Iceland spar (black cross) and synthetic Mg-calcite (Mg10calcite) (blue plus) are reported for comparison (wavelength of 0.47682(8) Å (26 keV)).

Figure 10. (a) Cell edge ratio c/a versus a parameter. (b) Cell edge ratio c/a versus c parameter for red coral branches Triperie (red squares) and sclerites Cassis (green circles) untreated (plain marks) and annealed under CO2 atmosphere for 2 h at 423, 473, 673, 873, and 1023 K (empty symbols). Data for calcite (Iceland spar) (black cross), synthetic Mg-calcite (Mg10calcite) (blue plus), and second order regression curve (black line) of Bischoff et al.61 for synthetic Mg-calcite samples (gray double cross) are also reported.

discrepancy argues for crystallite growth by a sintering effect during the experiments. 5.3.3. Structural and Microstructural Parameters in Branches and Sclerites Annealed “in Situ” under SelfControlled CO2 Pressure. SXRPD analysis realized on “ex situ” annealed samples provides significant information on the thermal decomposition steps. However, “ex situ” results are biased by sampling effect on the thermally induced modifications and by sintering effect on the thermal decomposition products. Conversely, SXRPD analysis of “in situ” annealed coral samples can bring relevant information on the thermal expansion of the biocrystalline phases and any thermally induced modification of the structural and microstructural parameters and phase transformation, because sampling effects are avoided and data are collected during the course of the decomposition reaction. Furthermore, “in situ” experimental data should provide new insights on the formation mechanism of the low Mg-calcite transient phase observed in our “ex situ” experiments, knowing that it is a debated subject in the case of half decomposition steps of dolomite as underlined by Rodriguez-Navarro et al.53 and Galai et al.106 5.3.3.1. Thermal Expansion of Iceland Spar and Synthetic Mg-Calcite (Mg10calcite). Thermal measurements were performed on both calcite (Iceland spar) and synthetic Mgcalcite (Mg10calcite) samples in the step-mode heating procedure described in section 4.4 (Supporting Information Table S3). Iceland spar was heat treated up to 1073 K. Under these experimental conditions, neither phase decomposition

is shifted to lower 2θ positions; at higher angles, the (300) peak is practically not shifted and the (0 0 12) peak is shifted to lower 2θ (Figure 11c). Thus, compared to Iceland spar, the unit cell of the produced calcite is anisotropically deformed, with Δa/a = −3.735 × 10−4 to −5.779 × 10−4 and Δc/c = 6.579 × 10−4 to 8.289 × 10−4, which corresponds to a smaller volume with ΔV/V from −3.265 × 10−4 to −8.941 × 10−5 depending of the coral sample. The shrinkage of the cell parameters may indicate that the produced calcite contains some Mg. However, only the a parameter is smaller and thus the discrepancy of the cell parameters can be more likely attributed to trace elements (mainly Na, S, Sr, P, K) initially present in the coral, in particular strontium (2600 ppm) and sulfate ions evidenced by FTIR spectra up to 1023 K and confirmed by EMP measurement. The periclase phase is characterized by small and broad peaks with marked dissymmetry toward low 2θ diffraction angles. The average sizes determined from peak broadening analysis range from 68 to 101 nm. The unit cell parameter of the produced periclase, a = 4.21402(5) Å, increases by Δa/a = 7.17 × 10−4 in comparison with a standard MgO (aMgO = 4.211 Å).107 This high value of oxide unit cell is confirmed by the dissymmetric broadening of the peaks that leads to a parameter values ranging from 4.2119 and 4.2193 Å. This lattice expansion is consistent with a periclase containing between 0.15% and 1.5% atoms of Ca. The average sizes of the produced calcite (ca. 432−372 nm) and periclase (ca. 68−101 nm) are large compared to those reported in XRD analysis of the thermal decomposition of dolomite.53,108 This size 3699

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Crystal Growth & Design

Article

acalcite = 4.99138 − 2.9924 × 10−5t + 1.65109 × 10−8t 2 ; αa = −3.267 × 10−6 K−1 ccalcite = 17.0533 + 4.5205 × 10−4t + 1.0152 × 10−7t 2 ; αc = 3.140 × 10−5 K−1 Vcalcite = 367.944 + 5.3366 × 10−3t + 4.566 × 10−6t 2 ; αV = 2.473 × 10−5 K−1

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

(1)

where t is the temperature (K) and α is the mean linear thermal coefficient between 298 and 1073 K.51 The variation of fwhm of the Bragg peaks as a function of T is small. fwhm of the most intense (104) peak plotted versus temperature in Figure12d presents a mean width of (0.00460942 ± 0.00025718)°(2θ), with minimal values around 673 K. Like calcite, the synthetic Mg10calcite structure is stable under heat treatment and no phase decomposition is observed up to 1073 K. Noticeably, the a cell parameter variation with temperature is parabolic with a minimum at 673 K, while the c parameter exhibits a thermal expansion described by the following thermal coefficients (eq 2): aMg10calcite = 4.94748 − 1.34795 × 10−5t + 1.69651 × 10−8t 2 ; αa = 1.044 × 10−7 K−1 cMg10calcite = 16.8525 + 4.47497 × 10−4t + 6.18743 × 10−8t 2 ; αc = 2.956 × 10−5 K−1 VMg10calcite = 357.245 + 7.51616 × 10−3t + 3.78647 × 10−6t 2 ; αV = 2.977 × 10−5 K−1 (2)

where t is the temperature (K) and α is the mean linear thermal coefficient between 298 and 1073 K. The thermal variation of the fwhm is inverse to the one observed for calcite; it corresponds to (0.00950839 ± 0.00035955)° (2θ) for the most intense (104) peak with maximal values around 673 K as shown in Figure 12d. To summarize, no phase transformation is observed in Mg10calcite in the range 273−1073 K, but its cell parameters change. The negative thermal expansion of a parameter becomes positive at ∼673 K. The thermal behavior of Mg10calcite differs from dolomite whose decomposition starts around 773−823 K up to half decomposition fully achieved at 1033 K. 5.3.3.2. Thermally Induced Modifications of Biogenic Magnesium Calcite. Contrary to the ex situ SXRDP analysis that definitely shows phase decomposition products for all coral samples annealed at 873 K for 2 h, the in situ SXRDP patterns indicate no phase decomposition product at temperatures of 773 and 823 K for branches and 873 K for sclerites. Compared to ex situ experiments, in situ decomposition proceeds at lower rate and the kinetic discrepancies in the decomposition of branches and sclerites are confirmed. The variations of the cell parameters and (104) fwhm are reported in Supporting Information Table S4 and plotted as a function of temperature in Figure 13. The a and c parameters and the (104) fwhm present nonuniform temperature dependence compared to the monotonic thermal variation of calcite and Mg10calcite. While a and c parameters highly fluctuate with temperature, the ratio c/a, the volume V, and the (104) fwhm

Figure 11. SXRPD patterns at RT for calcite Iceland spar (black crosses), synthetic Mg-calcite Mg10calcite (blue plus signs), and red coral branches Triperie annealed under CO2 atmosphere for 2 h at 823 K (yellow circles) and 1023 K (purple squares) in the 2θ range: (a) 8.95−9.25° showing the (104) peak; (b) 12.9−13.2° showing the (200) peak of periclase and the (202) peak of calcite; (c) 18.95− 19.45° showing the (300) and (0012) peaks of calcite (wavelength at 0.47682(8) Å (26 keV)).

nor order−disorder phase transition is observed. Thermal variations of the cell parameters measured are fully consistent with previous data47,51 (and references therein). The unit cell shows a thermal contraction and expansion for the a and c parameters, respectively. The temperature dependence of both parameters presented in Figure 12, is fitted by a polynomial expression, yielding the following thermal coefficients (eq 1): 3700

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

Crystal Growth & Design

Article

Figure 12. Thermal variation of: (a) a unit cell parameter; (b) c unit cell parameter; (c) c/a cell edge ratio; (d) fwhm for the (104) peak as a function of temperature for Iceland spar (black crosses) and synthetic Mg10calcite (blue plus) from SXRPD patterns recorded in situ at different temperatures under self-controlled CO2 (wavelength of 0.39992(8)Å (31 keV)).

Figure 13. Thermal variation for the red coral samples, relative to the thermal variation of Mg10calcite (eq 2) plotted in Figure 12, of: (a) a unit cell parameter, (b) c unit cell parameter, (c) c/a cell edge ratio, (d) fwhm for the (104) peak. Red coral branches (plain line, triangles) Triperie (green), Cassis (red), and Medes (blue) and for sclerites (dotted line, circles) Cassis (red) and Medes (blue) from SXRPD patterns recorded at different temperatures up to 873 K under self-controlled CO2. Thermal variation deviation for synthetic Mg10calcite (blue plus) is reported for comparison (wavelength of 0.39992(8)Å (31 keV)).

% in sclerites of the cell volume expansion. Conversely, from 623 K and below the phase decomposition temperature, the c parameter dilatation highly decreases, leading to the anisotropic decrease of (36 ± 3) % in branches and (60 ± 19) % in sclerites of the cell volume expansion. Similar thermal tensilelike and compressive-like effects on the crystalline structure

describe smoother curves with marked slope changes at around 423 and 623 K. The successive change of slope is particularly well marked for sclerites. Up to 423 K, no significant modification is observed. In the range 423−623 K, the dilatation of the c parameter increases. This leads to the anisotropic increase of (18 ± 11) % in branches and (43 ± 3) 3701

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Crystal Growth & Design

Article

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

Figure 14. SXRPD patterns recorded at 973 K under self-controlled CO2 for calcite Iceland spar (black crosses), synthetic Mg-calcite Mg10calcite (blue plus) and red coral branches Cassis (red triangles) and Medes (blue triangles) in the 2θ range: (a) 7.44−7.64° showing the (104) peak; (b) 10.7−11.2° showing the (200) peak for periclase and the (202) peak for calcite and Mg10calcite (wavelength of 0.39992(8) Å (31 keV)).

Figure 15. SXRPD patterns recorded in situ at 973 K for calcite Iceland spar (black crosses). Synthetic Mg-calcite Mg10calcite (blue plus) and red coral branches Triperie (green triangles) and sclerites Cassis (red circles) in the 2θ range: (a) 7.44−7.64° showing the (104) peak; (b) 10.7−11.2° showing the (200) peak for periclase and the (202) peak for calcite and Mg10calcite (wavelength of 0.39992(8) Å (31 keV)).

were previously observed for biogenic aragonite18,19 and biogenic calcite25 and generally linked to the incorporated organic molecules. These unusual lattice distortions were formulated by these authors as stress components imposed by organic matter on the mineral. In the coral samples, these structural changes occur within a range of temperature corresponding to the degradation and the release out of the organic matter trapped within the crystallites. From thermal analysis (section 5.1), the full range 423−823 K corresponds to the second step of the TG and DSC curves (Figure 2) assigned to the exothermic degradation of the organic matter. The degradation of the red pigments is confirmed by Raman spectroscopy on the coral branch samples annealed at 673 K (unpublished data). In the range 623−823 K, the release of the organic sulfur is monitored by the change of the additional mode assigned to the ν3 vibration of SO42− anion in the infrared spectrum. Furthermore, the decrease of the cell parameters toward synthetic Mg-calcite observed in annealed coral samples is confirmed: Δa/a = −1.1 × 10−4 and Δc/c = −4.3 × 10−4 were calculated for the “Medes” sclerites measured after heat treatment at 873 K (data reported in Figure 13). Thus, as assumed by previous authors, the large increase and decrease of the cell volume expansion could be caused by the internal pressure induced by the thermal degradation of OM

starting at 423 K98 and their release starting at 623 K. The stress components resulting from the lattice distortions could not be evaluated because elastic constants for single-crystal Mg10calcite are not known. These data confirm the complexity of biominerals and the implication of OM in their physical properties. 5.3.3.3. Thermal Decomposition of Biogenic Magnesium Calcite. Thermal decomposition is not observed in calcite and synthetic Mg10calcite up to 1073 K “in situ” experiments. In contrast, the thermal decomposition of coral samples occurs as soon as 823 K. SXRDP patterns show the first decomposition products within the 823−923 K interval. Periclase, depleted Mg-calcites, and Mg rich-calcites are identified in all samples, but pure calcite is present in “Medes” and “Cassis” branch samples only. At 973 K, more periclase, increasingly depleted Mg-calcite, are formed with either more calcite in “Medes” and “Cassis” branch samples (Figure 14) or more Mg rich-calcite in “Triperie” branch and “Cassis” sclerite samples (Figure 15). After cooling at ambient T, the produced periclase and calcite in “Medes” and “Cassis” branch samples are similar to those produced at 1023 K in “ex situ” experiments. The periclase and calcite crystals have an average size of about 60 and 300 nm, respectively. The unit cell parameter of periclase, a = 4.2165(7) Å, is longer with a Δa/a of 1.19 × 10−3 than the one of 3702

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Crystal Growth & Design

Article

ences (size, shape, hierarchical organization, and main trace elements Na, S, Sr, P, K).

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

standard MgO (aMgO = 4.211 Å). The unit cell parameters of produced calcite are anisotropically deformed, with Δa/a = −8.00 × 10−4 and Δc/c = 2.11 × 10−4, compared to the reference calcite (Iceland spar). At 1023 K, the half decomposition into calcite and periclase is still incomplete after 30 mn long heat treatment. SXRPD pattern of the “Triperie” sample heat treated during 30 mn at 1023 K and cooled to ambient (Figure 16) still shows peaks

6. CONCLUSIONS As discussed in the Introduction, the challenge of this structural study was to deal with the large magnesium calcite distribution specific to red coral skeletons compared to other studied calcite biominerals. This work led to firm and conclusive results based on the large previous knowledge obtained on the red coral in various areas, but also owing to the concomitant data obtained on Iceland spar and synthetic magnesium calcite with 10 Mg mol %. The SXRPD study on our synthesized Mg-calcite standard pointed out the problem of quantification of the Mg content in magnesium calcites by Rietveld-based XRD analysis. Although the lattice parameter a is the best parameter to determine the average composition of biogenic Mg-calcites (Titschack et al.11), this quantification is still unreliable compared to chemical analysis obtained by EMP. As a corollary, these calibration curves are useful to evidence the anisotropic distortion of biogenic materials compared to geologic or synthetic ones and should be definitively used for that purpose. However, SXRPD analysis remains a good tool to determine the histogram of Mg distribution in biogenic materials from diffraction peak profiles. Accurate structural parameters obtained from high-resolution synchrotron radiation diffraction analysis in pure calcite, synthetic Mg-calcite, and different red coral skeletons pointed out: (1) the general phenomenon of anisotropic unit cell distortions of biogenic materials and (2) its link to intracrystalline organic molecules occluded within individual crystallites as shown by TG/DSC analyses and FTIR/Raman spectroscopies. These structural properties unique to biogenic materials should have important implications on mechanical, optical, and other physical characteristics of biogenic crystals (dissolution, precipitation, ionic diffusion). The thermal stability of the red coral skeletons is different from calcite and specifically from synthetic magnesium calcite at 10 Mg mol % that is found stable up to 1073 K. Thus, the red coral skeletons made of Mg-calcite, with 10−15 mol % Mg content, including trace elements (Na, S, Sr, P, K), are not stable above 823 K in self-controlled CO2 (“in situ” experiments). Their thermal decomposition mechanism is comparable to the one observed in dolomite (50 Mg mol %). Furthermore, the presence of low Mg-calcite phases as intermediate phases supports the formation of calcite via a magnesium calcite and gradual release of Mg2+ ions. Besides, the produced calcite maintains a unit cell distortion along the c parameter probably linked to crystallinity defects associated with remaining trace elements. For instance, sulfate ions are still present within the calcite crystals after annealing according to the infrared spectroscopy results. Tetrahedral sulfate ions

Figure 16. SXRPD patterns recorded at RT in the 2θ range 7.48−7.68 (deg) showing the (104) peak for untreated calcite Iceland spar (black crosses) and red coral branches Triperie (green triangles) after annealing under self-controlled CO2 at 1023 K (wavelength of 0.39992(8)Å (31 keV)).

with dissymmetry toward high 2θ diffraction angles, which is consistent with a range of compositions from Mg-rich calcite to low Mg-calcite. A small broad peak is attributed to the periclase MgO (Table 6). Thus, after 30 mn at 1023 K, the “Triperie” sample is made of 55.33 mol % Mg-calcite with ∼1.2 mol % MgCO3, 35.45 mol % Mg-calcite with ∼3.00 mol % MgCO3, and 9.22 mol % periclase corresponding to a degree of conversion α equal to 84.2%. The amount of Mg rich calcite, low Mg-calcite, and periclase corresponds to 0.66, 1.06, and 9.22 Mg mol %, respectively. The total amount of 10.94 Mg mol % is in good agreement with the mean value of coral branches calculated from EMP measurements. The results of this “in situ” study support the conclusion of the “ex situ” study on the thermal decomposition of coral samples (section 5.3.2.1): (1) the thermal decomposition starts in the range 823−923 K; (2) low Mg-calcite phases are intermediate phases of the decomposition; (3) the unit cell of produced calcite is c-elongated; (4) the lattice of the produced periclase is extended; (5) differences in decomposition steps and degree of conversion are observed between branches and sclerites, as responses to microstructural and chemical differ-

Table 6. Composition, Structural Parameters (Refined Unit Cell Parameters, Parameter Distortions), and Microstructural Parameters (fwhm Broadening, size ⟨L⟩V of the Thermal Decomposition Products of Branches Triperie after Annealing at 1023 K (Wavelength at 0.47682(8) Å (26 keV)) chemical composition

a

formula

Mg mol %

MgxCa(1−x)CO3 MgxCa(1−x)CO3 MgO

1.2 3.0 100

structural parameters a

quantity mol %

a (Å)

55.33 35.45 9.22

4.9846(9) 4.9761(6) 4.2145(2)

10 Δa/a 4

c (Å)

V

17.0628(8) 17.0194(3)

367.16(3) 364.97(5) 74.588(4)

8.31

ΔV/V

fwhm (104) °2θ

⟨L⟩V (nm)

2.493 × 10−3

0.020(9) 0.030(7) 0.037(5)

328 220 101

The Mg mol % content for Mg-calcite compositions is calculated from the a parameter of Bischoff et al.’s calibration curve.61 3703

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Crystal Growth & Design

Article

(8) Rosenberg, G. D., The “Vital Effect” on Skeletal Trace Element Content as Exemplified by Magnesium. In Skeletal Biomineralization: Patterns, Processes and Evolutionary Trends; Carter, J. G., Ed.; American Geophysical Union: Washington, DC, 2013. (9) Dickson, J. A. D. Transformation of echinoid Mg-calcite skeletons by heating. Geochim. Cosmochim. Acta 2001, 65, 443−454. (10) Bischoff, W. D.; Sharma, S. K.; Mackenzie, F. T. Carbonate disorder in synthetic and biogenic magnesian calcitesa Raman spectral study. J. Geol. 1985, 70, 581−589. (11) Titschack, J.; Goetz-Neunhoeffer, F.; Neubauer, J. Magnesium quantification in calcites [(Ca,Mg)CO3] by Rietveld-based XRD analysis: revisiting a well-established method. Am. Mineral. 2011, 96, 1028−1038. (12) Turner, J. V.; Anderson, T. F.; Sandberg, P. A.; Goldstein, S. J. Isotopic, chemical and textural relations during the experimental alteration of biogenic high-magnesian calcite. Geochim. Cosmochim. Acta 1986, 50, 495−506. (13) Gaffey, S. J. H2O and OH in Echinoid Calcite A Spectroscopic Study. Am. Mineral. 1995, 80, 947−959. (14) Long, X.; Ma, Y.; Qi, L. Biogenic and synthetic high magnesium calcitea review. J. Struct. Biol. 2014, 185, 1−14. (15) Weber, J. N. Temperature Dependence of Magnesium in Echinoid and Asteroid Skeletal Calcite: A Reinterpretation of its Significance. J. Geol. 1973, 81, 543−556. (16) Moureaux, C.; Pérez-Huerta, A.; Compère, P.; Zhu, W.; Leloup, T.; Cusack, M.; Dubois, P. Structure, composition and mechanical relations to function in sea urchin spine. J. Struct. Biol. 2010, 170, 41− 49. (17) Berman, A.; Hanson, J.; Leiserowitz, L.; Koetzle, T. F.; Weiner, S.; Addadi, L. Biological control of crystal texturea widespread strategy for adapting crystal properties to function. Science 1993, 259, 776−779. (18) Pokroy, B.; Fitch, A. N.; Lee, P. L.; Quintana, J. P.; Caspi, E. N.; Zolotoyabko, E. Anisotropic lattice distortions in the mollusk-made aragonite: a widespread phenomenon. J. Struct. Biol. 2006, 153, 145− 150. (19) Wardecki, D.; Przenioslo, R.; Brunelli, M. Internal pressure in annealed biogenic aragonite. CrystEngComm 2008, 10, 1450−1453. (20) Stolarski, J.; Mazur, M. Nanostructure of biogenic versus abiogenic calcium carbonate crystals. Acta Palaeontol. 2005, 50, 847− 865. (21) Stolarski, J.; Przenioslo, R.; Mazur, M.; Brunelli, M. Highresolution synchrotron radiation studies on natural and thermally annealed scleractinian coral biominerals. J. Appl. Crystallogr. 2007, 40, 2−9. (22) Lee, S. W.; Kim, Y. M.; Kim, R. H.; Choi, C. S. Nanostructured biogenic calcite: a thermal and chemical approach to folia in oyster shell. Micron 2008, 39, 380−386. (23) Okumura, T.; Suzuki, M.; Nagasawa, H.; Kogure, T. Microstructural Variation of Biogenic Calcite with Intracrystalline Organic Macromolecules. Cryst. Growth Des. 2012, 12, 224−230. (24) Gilow, C.; Zolotoyabko, E.; Paris, O.; Fratzl, P.; Aichmayer, B. Nanostructure of Biogenic Calcite Crystals: A View by Small-Angle XRay Scattering. Cryst. Growth Des. 2011, 11, 2054−2058. (25) Pokroy, B.; Fitch, A. N.; Marin, F.; Kapon, M.; Adir, N.; Zolotoyabko, E. Anisotropic lattice distortions in biogenic calcite induced by intracrystalline organic molecules. J. Struct. Biol. 2006, 155, 96−103. (26) Zolotoyabko, E.; Caspi, E. N.; Fieramosca, J. S.; Von Dreele, R. B.; Marin, F.; Mor, G.; Addadi, L.; Weiner, S.; Politi, Y. Differences between Bond Lengths in Biogenic and Geological Calcite. Cryst. Growth Des. 2010, 10, 1207−1214. (27) Zolotoyabko, E.; Pokroy, B. Biomineralization of calcium carbonate: structural aspects. CrystEngComm 2007, 9, 1156−1161. (28) Frølich, S.; Sørensen, H. O.; Hakim, S. S.; Marin, F.; Stipp, S. L. S.; Birkedal, H. Smaller Calcite Lattice Deformation Caused by Occluded Organic Material in Coccoliths than in Mollusk Shell. Cryst. Growth Des. 2015, 15, 2761−2767.

replacing planar carbonate ions may cause a distortion of the calcite unit cell along the c axis. This study leaves many questions open; nevertheless, it contributes to understanding the mechanisms of thermal decomposition of carbonates and the structure of carbonate skeletal structures of the living organisms.



ASSOCIATED CONTENT

S Supporting Information *

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

Size-microstrain line broadening analysis procedure. Mg distribution in red coral skeletons. Tables of chemical composition, structural, and microstructural parameters are reported for coral samples ex situ and in situ annealed and for in situ annealed reference samples (Iceland spar and synthetic Mg10calcite). The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/ acs.cgd.5b00291.



AUTHOR INFORMATION

Corresponding Author

*Phone: 33662922855. Fax: 33491829197. E-mail: floquet@ cinam.univ-mrs.fr. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work has been supported by the Centre National de la Recherche Scientifique (CNRS)−Institut National des Sciences de l’Univers (INSU) through grant INTERRVIE 2009 and 2013, by the Agence Nationale pour la Recherche through ANR CoRo 2011−2015, and by the Centre Interdisciplinaire de Nanoscience de Marseille (CINaM) through internal grants. This work is also part of the European Union COST action TD0903. We thank G. Weber and C. Paulin (Laboratoire Interdisciplinaire Carnot de Bourgogne (LICB), Dijon, France) for the implementation of the TG-DSC thermal analyses and J. L. Devidal ((LMV), Clermont Ferrand, France) for his assistance with EMP analyses. This is contribution ANR CoRo no. 05. We acknowledge the European Synchrotron Radiation Facility for provision of synchrotron radiation facilities, and we thank Yves Watier for assistance in using beamline ID31.



REFERENCES

(1) Chave, K. E.; Deffeyes, K. S.; Weyl, P. K.; Garrels, R. M.; Thompson, M. E. Observations on the Solubility of Skeletal Carbonates in Aqueous Solutions. Science 1962, 137, 33−34. (2) Goldsmith, J. R.; Graf, D. L.; Heard, H. C. Lattice constants of the calcium−magnesium carbonates. Am. Mineral. 1961, 46, 453−457. (3) Érenburg, B. G. Artificial mixed carbonates in the CaCO3− MgCO3 series. J. Struct. Chem. 1961, 2, 167−171. (4) Morse, J. W.; Mackenzie, F. T. Geochemistry of Sedimentary Carbonates; Elsevier: Amsterdam, 1990; p 707. (5) Chave, K. E. Aspects of the biogeochemistry of magnesium. 1. Calcareous marine organisms. J. Geol. 1954, 62, 266−283. (6) Goldsmith, J. R.; Graf, D. L.; Joensuu, O. I. The occurence of magnesian calcites in nature. Geochim. Cosmochim. Acta 1955, 7, 212− 230. (7) Mackenzie, F. T.; Bischoff, W. D.; Bishop, F. C.; Loijens, M.; Schoonmaker, J.; Wollast, R., Magnesian calcites: Low-temperature occurrence, solubility and solid-solution behavior. In Carbonates: Mineralogy and Chemistry; Reviews in Mineralogy; Reeder, R.J., Ed.; Mineralogical Society of America: Chantilly, Virginia, 1983; Vol. 11, pp 97−144. 3704

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

Crystal Growth & Design

Article

(29) Magdans, U.; Gies, H. Single crystal structure analysis of sea urchin spine calcites: systematic investigations of the Ca/Mg distribution as a function of habitat of the sea urchin and the sample location in the spine. Eur. J. Mineral. 2004, 16, 261−268. (30) Rocha, J. H. G.; Lemos, A. F.; Agathopoulos, S.; Kannan, S.; Valerio, P.; Ferreira, J. M. F. Hydrothermal growth of hydroxyapatite scaffolds from aragonitic cuttlefish bones. J. Biomed. Mater. Res., Part A 2006, 77A, 160−168. (31) Vielzeuf, D.; Garrabou, J.; Baronnet, A.; Grauby, O.; Marschal, C. Nano to macroscale biomineral architecture of red coral (Corallium rubrum). Am. Mineral. 2008, 93, 1799−1815. (32) Vielzeuf, D.; Floquet, N.; Chatain, D.; Bonnete, F.; Ferry, D.; Garrabou, J.; Stolper, E. M. Multilevel modular mesocrystalline organization in red coral. Am. Mineral. 2010, 95, 242−248. (33) Floquet, N.; Vielzeuf, D. Mesoscale twinning and crystallographic registers in biominerals. Am. Mineral. 2011, 96, 1228−1237. (34) Floquet, N.; Vielzeuf, D. Ordered Misorientations and Preferential Directions of Growth in Mesocrystalline Red Coral Sclerites. Cryst. Growth Des. 2012, 12, 4805−4820. (35) Vielzeuf, D.; Garrabou, J.; Gagnon, A.; Ricolleau, A.; Adkins, J.; Gunther, D.; Hametner, K.; Devidal, J. L.; Reusser, E.; Perrin, J.; Floquet, N. Distribution of sulphur and magnesium in the red coral. Chem. Geol. 2013, 355, 13−27. (36) Grillo, M. C.; Goldberg, W. M.; Allemand, D. Skeleton and Sclerite Formation in the Precious Red Coral Corallium rubrum. Mar. Biol. 1993, 117, 119−128. (37) Bramanti, L.; Movilla, J.; Guron, M.; Calvo, E.; Gori, A.; Dominguez-Carrio, C.; Grinyo, J.; Lopez-Sanz, A.; Martinez-Quintana, A.; Pelejero, C.; Ziveri, P.; Rossi, S. Detrimental effects of ocean acidification on the economically important Mediterranean red coral (Corallium rubrum). Glob. Change Biol. 2013, 19, 1897−1908. (38) Debreuil, J.; Tambutté, E.; Zoccola, D.; Deleury, E.; Guigonis, J.M.; Samson, M.; Allemand, D.; Tambutté, S. Molecular Cloning and Characterization of First Organic Matrix Protein from Sclerites of Red Coral, Corallium rubrum. J. Biol. Chem. 2012, 287, 19367−19376. (39) Marschal, C.; Garrabou, J.; Harmelin, J. G.; Pichon, M. A new method for measuring growth and age in the precious red coral Corallium rubrum (L.). Coral Reefs 2004, 23, 423−432. (40) Lacaze-Duthiers, H. Histoire Naturelle du Corail: Organisation− Reproduction−Pêche en Algérie−Industrie et Commerce; J. B. Baillière: Paris, 1864; p 371. (41) Weinberg, S. Revision of the common Octocorallia of the Mediterranean circalittoral I. Gorgonacea. Ser. Miscellaneous Publ. 1976, 24, 63−104. (42) Perrin, J.; Vielzeuf, D.; Ricolleau, A.; Dallaporta, H.; Valton, S.; Floquet, N. Block-by-block and layer-by-layer growth modes in coral skeletons. Am. Mineral. 2015, 100, 681−695. (43) L’vov, B. V. Mechanism and kinetics of thermal decomposition of carbonates. Thermochim. Acta 2002, 386, 1−16. (44) Rodriguez-Navarro, C.; Ruiz-Agudo, E.; Luque, A.; RodriguezNavarro, A. B.; Ortega-Huertas, M. Thermal decomposition of calcite: mechanisms of formation and textural evolution of CaO nanocrystals. Am. Mineral. 2009, 94, 578−593. (45) Matvienko, A. A.; Chizhik, S. A.; Sidel’nikov, A. A. A new kinetic model of calcite thermal decomposition. Dokl. Phys. Chem. 2013, 451, 184−186. (46) Dove, M. T.; Swainson, I. P.; Powell, B. M.; Tennant, D. C. Neutron powder diffraction study of the orientational order−disorder phase transition in calcite, CaCO3. Phys. Chem. Miner. 2005, 32, 493− 503. (47) Antao, S. M.; Hassan, I.; Mulder, W. H.; Lee, P. L.; Toby, B. H. In situ study of the R-3c → R-3m orientational disorder in calcite. Phys. Chem. Miner. 2009, 36, 159−169. (48) Ballirano, P. Laboratory parallel-beam transmission X-ray powder diffraction investigation of the thermal behavior of calcite: comparison with X-ray single-crystal and synchrotron powder diffraction data. Period. Mineral. 2011, 80, 123−134.

(49) Thompson, S. P.; Parker, J. E.; Tang, C. C. Thermal breakdown of calcium carbonate and constraints on its use as a biomarker. Icarus 2014, 229, 1−10. (50) Ishizawa, N.; Setoguchi, H.; Yanagisawa, K. Structural evolution of calcite at high temperatures: phase V unveiled. Sci. Rep 2013, 3, 1− 11. (51) Markgraf, S. A.; Reeder, R. J. High temperature structure refinements of calcite and magnesite. Am. Mineral. 1985, 70, 590−600. (52) Sasaki, K.; Qiu, X.; Hosomomi, Y.; Moriyama, S.; Hirajima, T. Effect of natural dolomite calcination temperature on sorption of borate onto calcined products. Microporous Mesoporous Mater. 2013, 171, 1−8. (53) Rodriguez-Navarro, C.; Kudlacz, K.; Ruiz-Agudo, E. The mechanism of thermal decomposition of dolomite: new insights from 2D-XRD and TEM analyses. Am. Mineral. 2012, 97, 38−51. (54) Samtani, M.; Dollimore, D.; Alexander, K. S. Comparison of dolomite decomposition kinetics with related carbonates and the effect of procedural variables on its kinetic parameters. Thermochim. Acta 2002, 392, 135−145. (55) De Aza, A. H.; Rodríguez, M. A.; Rodríguez, J. L.; De Aza, S.; Pena, P.; Convert, P.; Hansen, T.; Turrillas, X. Decomposition of Dolomite Monitored by Neutron Thermodiffractometry. J. Am. Ceram. Soc. 2002, 85, 881−888. (56) Reeder, R. J.; Markgraf, S. A. High temperature crystal chemistry of dolomite. Am. Mineral. 1986, 71, 795−804. (57) Antao, S. M.; Mulder, W. H.; Hassan, I.; Crichton, W. A.; Parise, J. B. Cation disorder in dolomite, CaMg(CO3)(2), and its influence on the aragonite plus magnesite ↔ dolomite reaction boundary. Am. Mineral. 2004, 89, 1142−1147. (58) Busenberg, E.; Plummer, L. N. Thermodynamics of magnesian calcite solid solutions at 25 degrees C and 1 atm total pressure. Geochim. Cosmochim. Acta 1989, 53, 1189−1208. (59) Bertram, M. A.; Mackenzie, F. T.; Bishop, F. C.; Bischoff, W. D. Influence of temperature on the stability of magnesian calcite. Am. Mineral. 1991, 76, 1889−1896. (60) Smykatz-Kloss, W. Differential thermal analysis of Mg-bearing carbonates and sheet silicates. J. Therm. Anal. Calorim. 2002, 69, 85− 92. (61) Bischoff, W. D.; Bishop, F. C.; Mackenzie, F. T. Biogenically produced magnesian calcite inhomogeneities in chemical and physical propertiescomparison with synthetic phases. Am. Mineral. 1983, 68, 1183−1188. (62) Fitch, A. N. The high resolution powder diffraction beam line at ESRF. J. Res. Natl. Inst. Stand. Technol. 2004, 109, 133−142. (63) Rietveld, H. M. A Profile Refinement Method for Nuclear and Magnetic Structures. J. Appl. Crystallogr. 1969, 2, 65. (64) Rodriguez-Carvajal, J. FullProf: A Program for Rietveld Refinement and Profile Matching Analysis of Complex Powder Diffraction Patterns (ILL). Physica B 1993, 192, 55. (65) Althoff, P. L. Structural refinements of dolomite and a magnesian calcite and implications for dolomite formation in marine environment. Am. Mineral. 1977, 62, 772−783. (66) Maslen, E. N.; Streltsov, V. A.; Streltsova, N. R.; Ishizawa, N. Electron density and optical anisotropy in rhombohedral carbonates. III. Synchrotron X-ray studies of CaCO3, MgCO3 and MnCO3. Acta Crystallogr., Sect. B: Struct. Sci. 1995, 51, 929−939. (67) Hepp, A.; Baerlocher, C. Learned peak shape functions for powder diffraction data. Aust. J. Phys. 1988, 41, 229−236. (68) Scherrer, P. Bestimmung der Grösse und der inneren Struktur von Kolloidteilchen mittels Rö ntgenstrahlen. Nachr. Ges. Wiss. Göttingen 1918, 26, 98−100. (69) Stokes, A. R.; Wilson, A. J. C. A method of calculating the integral breadths of Debye−Scherrer lines. Proc. Cambridge Philos. Soc. 1942, 38, 313−322. (70) Delhez, R.; Keijser, T. H.; Langford, J. I.; Louer, D.; Mittemeijer, E. J.; Sonneveld, E. J., In The Rietveld Method (; International Union of Crystallography Monographs on Crystal no. 55; Young, R. A., Ed.; Oxford University Press: London, 1993. 3705

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706

Downloaded by UNIV OF NEBRASKA-LINCOLN on August 30, 2015 | http://pubs.acs.org Publication Date (Web): July 2, 2015 | doi: 10.1021/acs.cgd.5b00291

Crystal Growth & Design

Article

(71) Balzar, D.; Audebrand, N.; Daymond, M. R.; Fitch, A.; Hewat, A.; Langford, J. I.; Le Bail, A.; Louer, D.; Masson, O.; McCowan, C. N.; Popa, N. C.; Stephens, P. W.; Toby, B. H. Size-strain linebroadening analysis of the ceria round-robin sample. J. Appl. Crystallogr. 2004, 37, 911−924. (72) Adler, H. H.; Kerr, P. F. Infrared absorption frequency trends for anhydrous normal carbonates. Am. Mineral. 1963, 48, 124−137. (73) Adler, H. H.; Kerr, P. F. Infrared spectra, symmetry and structure relations of some carbonate minerals. Am. Mineral. 1963, 48, 839−853. (74) Ross, D. Inorganic Infrared and Raman Spectra; McGraw Hill: London, 1972; p 414. (75) White, W. B. The Infrared Spectra of Minerals;Farmer, V. C.Ed.; : Mineralogical Society: London, 1974; Vol. 4. (76) Frost, R. L.; Dickfos, M. J. Raman and infrared spectroscopic study of the anhydrous carbonate minerals shortite and barytocalcite. Spectroc. Acta Part A 2008, 71, 143−146. (77) Huang, C. K.; Kerr, P. F. Infrared study of carbonate minerals. Am. Mineral. 1960, 45, 311−324. (78) Golyshev, S. I.; Padalko, N. L.; Pechenkin, S. A. Fractionation of stable oxygen and carbon isotopes in carbonate systems. Geochem. Int. 1981, 10, 85−99. (79) Bottcher, M. E.; Gehlken, P. L.; Steele, D. F. Characterization of inorganic and biogenic magnesian calcites by Fourier transform infrared spectroscopy. Solid State Ion. 1997, 101, 1379−1385. (80) Andersen, F. A.; Brecevic, L. Infrared spectra of amorphous and crystalline calcium carbonate. Acta Chem. Scand. 1991, 45, 1018−1024. (81) Borzecka-Prokop, B.; Weselucha-Birczynska, A.; Koszowska, E. MicroRaman, PXRD, EDS and microscopic investigation of magnesium calcite biomineral phases. The case of sea urchin biominerals. J. Mol. Struct. 2007, 828, 80−90. (82) Dauphin, Y. Infrared spectra and elemental composition in recent biogenic calcites: relationships between the nu(4) band wavelength and Sr and Mg concentrations. Appl. Spectrosc. 1999, 53, 184−190. (83) Hezel, A.; Ross, S. D. Forbidden transitions in infrared spectra of tetrahedral anions 0.3. Spectra structure correlations in perchlorates sulfates and phosphates of formula MXO4. Spectrochim. Acta 1966, 22, 1949−1961. (84) Lane, M. D. Mid-infrared emission spectroscopy of sulfate and sulfate-bearing minerals. Am. Mineral. 2007, 92, 1−18. (85) Smith, D. H.; Seshadri, K. S. Infrared spectra of Mg2Ca(SO4)(3), MgSO4, hexagonal CaSO4, and orthorhombic CaSO4. Spectrochim. Acta Part A 1999, 55 (4), 795−805. (86) Bishop, J. L.; Perry, K. A.; Darby Dyar, M.; Bristow, T. F.; Blake, D. F.; Brown, A. J.; Peel, S. E. Coordinated spectral and XRD analyses of magnesite−nontronite−forsterite mixtures and implications for carbonates on Mars. J. Geophys. Res.: Planets 2013, 118, 635−650. (87) Böke, H.; Akkurt, S.; Ö zdemir, S.; Göktürk, E. H.; Caner Saltik, E. N. Quantification of CaCO3-CaSO3·0.5H2O−CaSO4·2H2O mixtures by FTIR analysis and its ANN model. Mater. Lett. 2004, 58, 723−726. (88) Rahman, M. A.; Oomori, T. Aspartic Acid-Rich Proteins in Insoluble Organic Matrix Play a Key Role in the Growth of Calcitic Sclerites in Alcyonarian Coral. Chin. J. Biotechnol. 2008, 24, 2127− 2128. (89) Rahman, M. A.; Oomori, T.; Worheide, G. Calcite Formation in Soft Coral Sclerites Is Determined by a Single Reactive Extracellular Protein. J. Biol. Chem. 2011, 286, 31638−31649. (90) Merlin, J. C. Resonance Raman spectroscopy of carotenoids and carotenoid containing systems. Pure Appl. Chem. 1985, 57, 785−792. (91) Karampelas, S.; Fritsch, E.; Sklavounos, S.; Soldatos, T. In Polyacetylenic Pigments Found in Pearls and Corals, 30th International Gemmological Conference, Moscow, 2007; Institute of Problems of Chemical Physics, Russian Academy of Sciences: Moscow, 2007; pp 49−51. (92) Fritsch, E.; Karampelas, S. Comment on: Determination of canthaxanthin in the red coral (Corallium rubrum) from Marseille by

HPLC combined with UV and MS detection (Cvejic et al. Mar. Biol. 152:855−862, 2007). Mar. Biol. 2008, 154, 929−930. (93) Urmos, J.; Sharma, S. K.; Mackenzie, F. T. Characterization of some biogenic carbonates with Raman spectroscopy. Am. Mineral. 1991, 76, 641−646. (94) Kupka, T.; Lin, H. M.; Stobinski, L.; Chen, C. H.; Liou, W. J.; Wrzalik, R.; Flisak, Z. Experimental and theoretical studies on corals. I. Toward understanding the origin of color in precious red corals from Raman and IR spectroscopies and DFT calculations. J. Raman Spectrosc. 2010, 41, 651−658. (95) Brambilla, L.; Tommasini, M.; Zerbi, G.; Stradi, R. Raman spectroscopy of polyconjugated molecules with electronic and mechanical confinement: the spectrum of Corallium rubrum. J. Raman Spectrosc. 2012, 43, 1449−1458. (96) Dauphin, Y. Mineralizing matrices in the skeletal axes of two Corallium species (Alcyonacea). Comp. Biochem. Physiol., Part A: Mol. Integr. Physiol. 2006, 145, 54−64. (97) Kim, S. B.; Ji, C. I.; Woo, J. W.; Do, J. R.; Cho, S. M.; Lee, Y. B.; Kang, S. N.; Park, J. H. Simplified purification of chondroitin sulphate from scapular cartilage of shortfin mako shark (Isurus oxyrinchus). Int. J. Food Sci. Technol. 2012, 47, 91−99. (98) Rosca, C.; Novac, O.; Lisa, G.; Popa, M. I. Polyelectrolyte complexes of chitosan with dextran sulphate. Synthesis and characterization. Cell Chem. Technol. 2011, 45, 185−189. (99) Crouzet, J.; Kanasawud, P.; Lester, P., [7] Formation of volatile compounds by thermal degradation of carotenoids. In Methods in Enzymology; Academic Press: New York, 1992; Vol. 213, pp 54−62. (100) Myneni, S. C. B.; Traina, S. J.; Waychunas, G. A.; Logan, T. J. Vibrational spectroscopy of functional group chemistry and arsenate coordination in ettringite. Geochim. Cosmochim. Acta 1998, 62, 3499− 3514. (101) Vassallo, A. M.; Finnie, K. S. Infrared emission spectroscopy of some sulfate minerals. Appl. Spectrosc. 1992, 46, 1477−1482. (102) Paquette, J.; Reeder, R. J. Single crystal X-ray structure refinements of 2 biogenic magnesian calcite crystals. Am. Mineral. 1990, 75, 1151−1158. (103) Zhang, F. F.; Xu, H. F.; Konishi, H.; Roden, E. E. A relationship between d(104) value and composition in the calcitedisordered dolomite solid-solution series. Am. Mineral. 2010, 95, 1650−1656. (104) Spinolo, G.; Beruto, D. Thermodynamic analysis of the thermal decomposition of dolomite. J. Chem. Soc., Faraday Trans. 1 1982, 78, 2631−2642. (105) Beruto, D. T.; Vecchiattini, R.; Giordani, M. Solid products and rate-limiting step in the thermal half decomposition of natural dolomite in a CO2 (g) atmosphere. Thermochim. Acta 2003, 405, 183− 194. (106) Galai, H.; Pijolat, M.; Nahdi, K.; Trabelsi-Ayadi, M. Mechanism of growth of MgO and CaCO3 during a dolomite partial decomposition. Solid State Ion. 2007, 178, 1039−1047. (107) Hazen, R. M. Effects of temperature and pressure on cell dimension and X-ray temperature factors of periclase. Am. Mineral. 1976, 61, 266−271. (108) Hashimoto, H.; Komaki, E.; Hayashi, F.; Uematsu, T. Partial decomposition of dolomite in CO2. J. Solid State Chem. 1980, 33, 181−188.

3706

DOI: 10.1021/acs.cgd.5b00291 Cryst. Growth Des. 2015, 15, 3690−3706