Complex F,Cl Apatite Solid Solution Investigated Using Multinuclear

Subscriber access provided by READING UNIV

Article

Complex F,Cl Apatite Solid Solution Investigated Using Multinuclear Solid-State NMR Methods John S. Vaughn, Donald H. Lindsley, Hanna Nekvasil, John M. Hughes, and Brian L. Phillips J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.7b09912 • Publication Date (Web): 08 Dec 2017 Downloaded from http://pubs.acs.org on December 13, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry C is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Complex F,Cl Apatite Solid Solution Investigated Using Multinuclear Solid-State NMR Methods JOHN S. VAUGHN†*, DONALD H. LINDSLEY‡, HANNA NEKVASIL‡, JOHN M. HUGHES§, BRIAN L. PHILLIPS‡ † Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, CA 94550 ‡ Department of Geosciences, Stony Brook University, Stony Brook, NY 11794 § Department of Geology, University of Vermont, Burlington, VT 05405 Corresponding Author Email: [email protected] ABSTRACT: A series of synthetic F,Cl apatites with low hydroxyl content was investigated using 31

P, 19F, and 35Cl solid-state magic angle spinning (MAS) NMR spectroscopic methods.

31

P single-pulse

(SP) NMR spectra show that for each composition the phosphate 31P chemical shift depends on occupancy of the nearest anion site for each composition. Overall the average chemical shift deviates from linear across the F/Cl series. For intermediate compositions 19F SP NMR spectra reveal complex, broadened spectral profiles that are not correlated to composition.

19

F{35Cl} TRAPDOR results indicate that the

spectral profiles do not reflect occupancy of F and Cl in adjacent column anion sites. We propose that the complex 19F lineshapes are principally due to broad distributions of 19F-Ca distances arising from displacement of F along the anion channel. Approximate fluorine atomic positions are estimated based on 19

F chemical shifts of alkaline earth fluoride salts, and the results are in good accord with those proposed for

intermediate composition F,Cl apatites from XRD data.

INTRODUCTION: Apatite (Ca5(PO4)3X, where X = F, OH, or Cl) is an important mineral for scientists in fields ranging from Earth and space sciences to biomaterials and medicine. Much research is focused on chemical composition and crystal structure, as these molecular-scale properties are related to physical properties such as mechanical strength or biocompatibility.1 In the geological realm apatite

1 ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 28

composition is used to provide insight into the source of the hosting magma and its volatile content, which dictate both magma evolution and explosivity.2 Changes in apatite composition or structure often have significant impacts on physical or chemical properties, spawning a great deal of interest in this field. The most remarkable feature regarding the apatite crystal structure is its ability to accommodate cations and anions of broadly different sizes. Consequently apatite materials have been utilized for a variety of applications; such as nuclear waste storage materials3 and as electrolytes in solid oxide fuel cells.4 The molecular-scale mechanisms by which substituents of varying size and charge are accommodated in the apatite structure are poorly understood, but are of fundamental importance to geoscientists and materials scientists. These mechanisms can provide clues into the geological conditions of crystallization, and suggest pathways for the preparation of novel materials with desirable properties. The current work focuses specifically on structural adaptations related to anion substitutions among fluorine, chlorine, and hydroxyl groups.

The monovalent anions in naturally occurring apatite (F-, OH-, and Cl-) occupy a 1-dimensional channel parallel to the crystallographic c-axis.5 Solid solution among end-members is controlled by the sequence and atomic positions of anions in this channel. The ‘walls’ of the channel are composed of Ca2 triads that define a plane orthogonal to the c-axis, positioned at z = 0.25 and 0.75. The positions of column anions are often described relative to these triads. The fluoride ion in fluorapatite is centered within the Ca2 triads consistent with crystallographic mirror planes observed parallel to the a-b plane. Hydroxyl groups and chloride ions are too large to occupy the fluorine position in fluorapatite due to steric constraints from the triad of Ca2 atoms, and are displaced in end-member compositions 0.35Å and 1.3Å above or below the Ca2 triads for hydroxylapatite and chlorapatite, respectively.5 Average hexagonal symmetry can be maintained for end-member hydroxlapatite and chlorapatite by a balanced distribution of anion occupancies above and below the Ca2 triad.6 As a consequence of greatly varying end-member column anion positions, combining end-member structures to create single-phase binary (F,OH; F,Cl; OH,Cl) and ternary (F,OH,Cl) apatite compositions results in unacceptably small interatomic anion distances. Therefore, preparation of


Page 3 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


intermediate binary and ternary apatite compositions requires a structural adaptation of some kind not reflected in end-member structures (e.g. a reduction in symmetry from hexagonal to monoclinic, thereby lifting restraints on possible column anion positions, or occupancy of new anion positions that preserve hexagonal symmetry).

Previously we have described a series of apatite compositions close to the F-Cl binary, for which dilute hydroxyl groups were shown to play an important role in occupying positions at the interface between fluorine and chlorine atoms.7 However, the concentration of hydroxyl groups was low (< 2 mol percent), so only a minor fraction of fluorine/chlorine arrangements could feature hydroxyl groups at the interface. The present work focuses on the more abundant F-Cl configuration, because the disparate end-member anion positions in fluorapatite and chlorapatite would seem to require significant structural re-arrangement to accommodate anions of such different size, which could more easily permit observation of the corresponding adaptation mechanism. We apply 19F, 31P and 35Cl NMR methods to this same set of nearbinary F,Cl samples examined in previously,7 in order to explore the arrangement of major column anions. The 31P NMR spectra appear sensitive only to the identity of the nearest anion column occupant to the phosphate group, but not to the details of the F/Cl arrangement.

19

F NMR spectra reveal complex and non-

linear changes in fluorine environments with composition. Complex 19F line shapes are observed for intermediate compositions that likely reflect a distribution of F-Ca2 distances in response to intimate mixing of F and Cl in the anion channels, indicating that the fluorine atoms are disordered in the anion channel.

MATERIALS AND METHODS Apatite Synthesis Preparation of the present samples has been described previously,7 but is summarized here for completeness. Near-binary low-OH fluor-chlorapatite was prepared from mixtures of tricalcium-phosphate, calcium fluoride (99.99%, Alfa Aesar), and calcium chloride (99.99%, Alfa Aesar) at high temperature. Tricalcium phosphate (TCP) was synthesized by high-temperature (1250 °C) reaction of calcium carbonate



Page 4 of 28

(99.5%, Alfa Aesar) and ammonium phosphate monobasic (98+%, Sigma-Aldrich). All reflections detected by powder X-ray diffraction (PXRD) could be indexed to apatite, indicating absence of significant fractions of crystalline impurity phases. Anhydrous calcium chloride and calcium fluoride were purchased in sealed ampules and opened immediately before use. Mixed powders of reagents, in proportion to the desired target composition, were ground together and loaded into a Pt capsule. The Pt capsule was then inserted into a silica glass tube, and a capillary drawn using an oxygen torch. The sample was subsequently dried under vacuum at 800°C for 30 min, after which the capillary was severed by melting, leaving the Pt capsule sealed under vacuum. The evacuated silica tube + Pt capsule/sample assembly was then heated to 1200°C and reacted for 18 days, yielding transparent hexagonal crystals 10 µm in diameter. Samples are designated FxCly, where x and y represent the nominal mol% of fluor- and chlorapatite component, respectively. Endmember compositions are designated F100 and Cl100 for fluorapatite and chlorapatite, respectively. Solid State NMR Solid-state magic angle spinning (MAS) NMR spectra were obtained on a Varian Infinityplus 500 MHz (11.7 T) spectrometer operating at 202.318 MHz for 31P, 470.179 MHz for 19F, and 48.98 MHz for 35

Cl using standard single-pulse (SP) techniques.

31

P and 35Cl spectra were collected at a spinning rate of 10

kHz, and 19F spectra at 15 kHz. A Varian-Chemagnetics T3-type HX probe configured for 4mm rotors was used for all single-pulse experiments. The rotor assembly consisted of ZrO2 sleeves, and vespel tips and spacers. The 31P pulse width (π/2) was 5 µs, and 128 transients were collected at a relaxation delay of 30s. At longer relaxation delays signal intensity did not increase further, so at this delay the spectra are fully relaxed.

31

P chemical shifts were measured relative to the phosphate resonance in stoichiometric

hydroxylapatite set to δP = 2.65 ppm. The 19F pulse width was 5µs (π/2), and the relaxation delay was 600s for full relaxation, determined by acquisition of fluorapatite spectra with relaxation delays between 5 and 800s. After 600s signal intensity did not increase further, so these spectra are considered fully relaxed under these conditions. The number of transients varied between 4 and 148 owing to the varied concentration of fluorine.

19

F chemical shifts were measured relative to CFCl3 (l) set to δF = 0 ppm. The solution (non-


Page 5 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


selective) 35Cl π/2 pulse width was set to 4µs using signal from a 1M aqueous solution of NaCl. For solids acquisition 2µs pulses were used (selective π/2).8 The relaxation delay was 2s, and the number of transients varied between 2400 and 170,000 to obtain suitable signal-to-noise due to the varied concentration of chlorine. Chemical shifts were measured relative to NaCl (aq) set to δCl = 0 ppm. No signal was observed in background spectra for 35Cl, 19F, and 31P collected using an empty 4mm rotor.

19

F{35Cl} TRAPDOR9 data

were collected for the F50Cl50 composition at a spinning rate of 20 kHz, using a probe outfitted for 3.2 mm (O.D.) rotors. The 19F pulse width (π/2) was 5 µs, and the 35Cl B1 field was 42 kHz. The interpulse delay (τ) was 750 µs, the relaxation delay was 120s, and 345 scans were collected. RESULTS 31

P NMR The 31P single-pulse (SP) MAS NMR spectra of the F,Cl apatites investigated in this work are

shown in Figure 1, and the results of least-squares fitting of the spectra to a sum of Gaussian curves are shown in Table 1. All spectra could be adequately fit using two peaks, however the position, width, and relative contribution of the component peaks vary considerably between differing compositions. Despite the complex, asymmetric nature of the observed line shapes, a few general trends are apparent. The line shapes of those samples at or near end-member composition (Cl100, F100, Cl90F10, and F10Cl90) feature prominent, relatively narrow (0.4-0.7 ppm FWHM) signals with shoulders of greater width and lesser intensity. For the end-members, the broadened shoulder occurs at higher chemical shift relative to the prominent narrow signal. The origin of these shoulders in F100 and Cl100 is uncertain, however it is possible that they arise from a minor impurity phase. For those binary compositions nearest the endmembers (F10Cl90 and F90Cl10), the shoulders appear on opposing sides of the primary resonance; for F10Cl90 the shoulder appears at lower chemical shift to the main signal, and for F90Cl10 the shoulder is observed at higher chemical shift relative to the main signal.

Previous solid state 31P NMR analysis of apatite has shown that the phosphate 31P chemical shifts of fluorapatite and hydroxylapatite are similar,10,11 but 31P chemical shifts of fluorapatite and chlorapatite differ



Page 6 of 28

sufficiently that the peaks can be readily resolved.12 O’Donnell et al.12 demonstrate that phosphate tetrahedra associated with fluorine give a higher chemical shift relative to those associated with chlorine. Similar results are observed in the present end-member 31P spectra, although the values differ somewhat (2.7 vs. 1.7 ppm in the present work for Cl100, 3.3 vs. 2.5 ppm in the present work for F100). Based on this information we assign the shoulder in F90Cl10 at lower chemical shift to phosphate groups adjacent to chlorine, and assign the shoulder at higher chemical shift in F10Cl90 to phosphate groups adjacent to fluorine. These assignments are relatively straightforward because phosphate groups in apatite are closely associated with only a single column anion position. For example, the shortest column anion-phosphorus distance in fluorapatite is 3.6Å and the next shortest is 4.98Å,5 probably too far to significantly affect phosphate chemical shifts. As a result only two principal chemical environments can be resolved, phosphate proximal to fluorine, or phosphate proximal to chlorine. The relative signal intensities of the two peaks for F90Cl10 and F10Cl90 are in reasonable agreement with expected 1:9 or 9:1 ratio based on nominal F:Cl stoichiometry considering the uncertain compositions and incomplete resolution. A similar agreement is observed for F67Cl33 and F33Cl67; 79 and 33 percent of the 31P intensity is attributed to the peak at greater chemical shift (P near F) for F67Cl33 and F33Cl67, respectively. The resolution of two signals appears much poorer in F33Cl67, so fitted intensity values for this spectrum carry more uncertainty.

The widths of all peaks increase as the F:Cl ratio approaches 1:1, suggesting a distribution of 31P environments due to structural disorder, possibly reflecting variations in occupancy of column anion sites beyond those closest to P. Least-squares fitting of the F50Cl50 spectrum using two peaks resulted in two broad resonances, with the peak at higher chemical shift 4 times more intense than the signal at lower chemical shift. However the reported relative intensities of these signals may be inaccurate, owing to the poor resolution, uncertain peak positions, and assumption of symmetrical Gaussian peak shapes. 19

F NMR The 19F single-pulse (SP) NMR spectra of the F,Cl apatites investigated in this work are shown in

Figure 2, and the results of least-squares fitting of the spectra to a sum of Gaussian curves are shown in


Page 7 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 2. The first moment of the 19F (SP) NMR spectra are shown in Figure 3. The 19F NMR spectrum of F100 features a single narrow peak, indicative of a well-defined fluorine chemical environment as expected for single-phase fluorapatite. The 19F NMR spectrum of F90Cl10 shows a narrow prominent resonance, and trailing intensity at higher chemical shift with at least two additional peaks. Figure 2b more clearly shows the signals from F90Cl10 at expanded horizontal scale. The signals at higher chemical shift are not believed to arise from 19F environments associated with OH impurities because of the low concentration measured previously7 (less than two mol%) compared to the much larger intensity of the peaks at -96 to -101 ppm (Table 2).

The 19F NMR spectra of F67Cl33, F50Cl50, F33Cl67, and F10Cl90 exhibit broadened spectral profiles relative to those of F90Cl10 and F100. The center bands of these spectra were fit with two symmetrical signals: a relatively narrow signal at lower chemical shift and a broadened shoulder at higher chemical shift, and in the case of F10Cl90 a third peak at -93.2 ppm. The broad shoulder in these spectra is most prominent for F50Cl50. The 1st spectral moments were determined by intensity-weighted averaging of the chemical shifts at each data point ( = ∑ / ∑ ), were Yi is the intensity. These values are shown in the right-most column of Table 2, and increase as the F:Cl ratio approaches 1:1. The end-member or near-end-member compositions, F100, F90Cl10, and F10Cl90 have similar first spectral moments, approximately -102 ppm. 19

F{35Cl} TRAPDOR NMR Spectroscopy The results of the 19F{35Cl} TRAPDOR NMR experiment for F50Cl50 are shown in Figure 4. The

TRAPDOR9 experiment is a double-resonance pulse sequence that detects dipolar coupling between spin-½ and quadrupolar nuclei, such as 19F and 35Cl in the present case. The upper spectrum (Figure 4a) is a spin echo spectrum, showing signal from all 19F environments with some attenuation due to T2 relaxation. The spin echo and single-pulse spectra appear very similar, reflecting the absence of significant differential T2 relaxation effects for this sample under these conditions. The lower spectrum (Figure 4b) shows the 19

F{35Cl} TRAPDOR difference spectrum, obtained by subtracting from the spin-echo spectrum another that



Page 8 of 28

was collected under the same conditions, but with 35Cl irradiation that reduces intensity from 19F near 35Cl. This difference spectrum shows only those 19F environments that are coupled to 35Cl. The TRAPDOR difference spectrum has been scaled vertically to facilitate comparison to the upper spin echo spectrum. The ratio between the intensity of the broad shoulder and the narrower signal at lower chemical shift remains unchanged between the spin echo spectrum and the TRAPDOR difference, which indicates that the resolved spectral features are unrelated to the occupancy of adjacent column positions. This result is in good agreement with a report by McCubbin et al.13 for a F/Cl apatite that contained significant OH component. From simulations of the TRAPDOR effect these authors show that the difference spectrum arises dominantly from F having Cl in adjacent column sites.

35

Cl SP MAS NMR Spectroscopy The 35Cl single-pulse (SP) NMR spectra are shown in Figure 5. Except for that of F90Cl10, these

spectra can broadly be described as featuring a sharp left edge with trailing intensity to lower chemical shift. Peak shapes of this type have been ascribed to a distribution of quadrupolar coupling constants (CQ), arising from a distribution of electronic environments of varying asymmetry but with a relatively small range of chemical shifts. The 35Cl spectrum of F90Cl10 features prominent edges and horns, suggesting a significant fraction of the Cl occurs in a well-ordered environment, in addition to a broad tail indicative of Cl in more disordered regions. Estimates for the distribution of quadrupolar coupling constants (CQ) and chemical shifts were obtained by fitting spectra to lineshapes calculated using a method described in Coster et al.14 which assumes a random distribution of electric field gradients (EFGs). Using this method the three principal components of the EFG are perturbed by the imposition of a Gaussian distribution, and from this perturbed EFG, new quadrupolar coupling constants and asymmetry parameters (η) are calculated. MAS spectra calculated from these spectral parameters are then weighted and summed to generate the lineshape. Average CQ values and associated width of the CQ distribution (σ) that best match the spectral profiles using this process are shown in Table 3. Since data are available for only one magnetic field these values can be


Page 9 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


considered only approximate, providing values for comparison between samples. We were not able to obtain a suitable fit in the 35Cl spectrum of F90Cl10, which appears to contain distinct contributions from well-ordered site plus a broad tail arising from a distribution of more distorted environments. The 35Cl isotropic chemical shifts (δiso) were also estimated from the simulations, reported in Table 3. The isotropic chemical shifts demonstrate that as the chlorine content increases, the isotropic chemical shifts decrease, in general agreement with the 19F and 31P NMR results discussed previously. The F90Cl10 and F67Cl33 35Cl spectra feature significant trailing intensity to smaller chemical shift, indicating that the Cl atoms in these samples may experience a greater range of electronic environments than the other samples. The amount of 35Cl data in the literature available for comparison is rather limited, owing to difficulty of such measurements attributable to the low γ and poor sensitivity of the 35Cl nucleus. McCubbin et al.13 described a F,Cl composition in which the target F,Cl composition was F50Cl50, and observed a second-order quadrupolar line shape with an isotropic 35Cl chemical shift of +115 ppm. The 35Cl CQ from that work (1.6 MHz) is comparable to the average CQ reported in this work (1.67 MHz). The CQ reported previously for chlorapatite (0.8 MHz)15, however, is much smaller relative to the CQ for the present chlorapatite sample (1.43 MHz, from Table 3). DISCUSSION Variation in phosphorus environments The 31P results indicate that for the present samples, the NMR chemical shift of apatite phosphate groups is sensitive mainly to the nearest anion column occupant, reflected in the variation of signal intensity from two resolved peaks with F:Cl ratio. However, a non-linear relationship between composition and average

31

P chemical shift is apparent in Figure 6, which shows the first

31

P spectral moment (δP,AVG,

intensity weighted average chemical shift) of the center band region versus the nominal fluorine content in each sample. Measuring the 1st

31

P spectral moment eliminates uncertainties associated with fitting the

intensity of two overlapping peaks, such as in peak position and width. Figure 6 reveals a fairly strong variation of δP,AVG with composition from Cl100 to F50Cl50, but with little apparent change in δP,AVG from F50Cl50 to F100. The deviations from linearity may be due to structural ordering associated with F,Cl solid



Page 10 of 28

solution, or errors in measurement due to very broad peak shapes. The uncertainty in these measurements is rather large, particularly for F50Cl50.

Effect of OH on spectra of F/Cl Apatite Comparison of the present results with previous work shows the importance of excluding water for preparing binary F/Cl apatite. O’Donnell et al.12 report a

31

P chemical shift of 2.7 ppm for chlorapatite,

which differs significantly from that observed here at 1.7 ppm. This chemical shift difference might be attributed in part to inter-laboratory variation in chemical shift referencing; for example O’Donnell et al.12 reported a fluorapatite 31P chemical shift of 3.3 ppm whereas the F100 31P shift in the present work is 2.8 ppm. Both fluorapatite compositions are presumed to be nearly identical, yet the reported

31

P NMR

chemical shifts differ by 0.5 ppm. This presumed difference in chemical shift referencing (i.e. 0.5 ppm) is too small however to account for the difference in the chlorapatite 31P chemical shift, which might reflect differences in hydroxyl component as discussed below.

The samples described by O’Donnell et al.12 and those in the present work were both prepared by high temperature synthesis, the principal difference being the extensive measures taken in the present study to minimize the fraction of hydroxyl component (see Methods). The observation of significant hydroxyl component in F/Cl apatite by McCubbin et al.13 was published just before publication of the work by O’Donnell, so the need for extensive drying and atmospheric control during F/Cl apatite synthesis was not yet widely appreciated. These measures are particularly important for chlorine rich compositions, due to the hygroscopic nature of the starting materials required to prepare chlorapatite, especially calcium chloride. In the absence of specific techniques to exclude water, it is probable that the chlorine-rich compositions described by O’Donnell et al.12 contain a significant amount of hydroxyl component.

The likely presence of OH in the chlorapatite composition described by O’Donnell et al.12 is reflected in the

31

P NMR data.

The nearly identical

31

P chemical shifts of fluorapatite and


Page 11 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


fluorhydroxylapatite reported previously,10,11 suggests that a 31P spectrum of chlorapatite with some fluorine component could be expected to resemble chlorapatite with some hydroxyl component. The presence of either F or OH in a dominantly Cl-rich apatite composition would shift the

31

P peak position to higher

chemical shift. The higher chemical shift for chlorapatite observed in O’Donnell et al.12 compared with the present result is consistent with a higher concentration of hydroxyl component.

Comparison of peak widths provides further evidence for the effects of OH on NMR spectra of Clrich F/Cl-apatite. The 31P lineshapes in the present work are narrow for fluorapatite (FWHM = 0.7 ppm) and become progressively broader for compositions close to F:Cl = 1:1, (FWHM = 1.6 ppm) which was also observed in the 31P NMR work by O’Donnell et al.12 However the line widths observed by O’Donnell et al.12 remain broad (1-1.5 ppm FWHM) for compositions between F:Cl = 1:1 (F50Cl50) and end-member chlorapatite.

In contrast, the present

31

P NMR results show progressively narrower line widths with

increasing Cl-content, reaching 0.4 ppm for end-member chlorapatite. This difference in peak width for Clrich compositions can most likely be attributed to a more ordered distribution of P environments at low hydroxyl content. As the chlorine content increases, so does the susceptibility for incorporation of hydroxyl component during synthesis owing to a larger fraction of hygroscopic CaCl2. Presence of hydroxyl is likely to yield a more complex 31P peak shape owing to contributions from phosphate groups proximal to chlorine atoms and those near hydroxyl groups. This effect is most clearly evident by noting the mirror-image resemblance between the chlorapatite 31P line shape reported in O’Donnell et al.12 and that for the F67Cl33 composition in the present work. The F67Cl33 31P spectrum features a prominent (79% relative intensity) signal at 2.9 ppm, due to phosphate groups proximal to fluorine and a less prominent shoulder at 1.6 ppm, arising from phosphate groups with chlorine as the nearest column anion. The chlorapatite lineshape in O’Donnell et al.12 features a prominent (84% relative intensity) signal at 2.7 ppm, likely representing phosphate groups proximal to a chlorine, plus a less intense shoulder at 4 ppm, probably due to phosphate groups proximal to hydroxyl groups. This interpretation is supported by comparison with the effect of OH on spectra of F/Cl apatite in the study by McCubbin et al.13 From electron probe microanalysis (EPMA)



Page 12 of 28

these authors found that the composition of a F/Cl apatite synthesized at a nominal 1:1 ratio is F50Cl37OH13. The 31P line shape for that sample most closely resembles that of F67Cl33 in the present work. Because the

31

P chemical shift from phosphate proximal to hydroxyl groups is similar to that for

phosphate groups with fluorine as the closest column anion, it follows that the

31

P NMR spectrum of a

ternary composition such as F50Cl37OH13 would resemble F67Cl33.

The 19F MAS NMR spectra obtained by O’Donnell et al.12 also differ significantly from the present results, although there are differences in acquisition conditions. For example, O’Donnell et al.12 used a lower field strength NMR spectrometer (4.7T vs. 11.7T in the present work), and a lower MAS rate (10kHz vs. 15kHz in the present work), however these differences are not expected to significantly affect the centerband profile. As with the 31P NMR spectra, the presence of hydroxyl component likely contributes to the differences between the two present

19

19

F data sets. Although the chemical shifts of the narrow signal in the

F NMR spectra are in good correspondence with that reported in O’Donnell et al.12 (shown in

Table 4) the lineshapes in that previous study lack the prominent shoulder toward higher chemical shift observed here. The spectral profiles instead generally have a broad, symmetric appearance, becoming progressively broader as the F:Cl ratio is decreased. Hydroxyl groups have been demonstrated to facilitate average hexagonal anion arrangements between F and Cl, so it is likely that the hydroxyl component believed to be present in those compositions in O’Donnell et al.12 reduces the strain in the anion column, resulting in more symmetric 19F lineshapes. For a nominally F50Cl50 apatite, McCubbin et al.13 observed a 19F lineshape similar to that in the present work, except that the intensity of the broad shoulder at higher δ is smaller compared to that observed here. McCubbin et al.13 noted that this sample contains 13 mol% hydroxyl component. Hydroxyl component has been shown to facilitate F/Cl distributions that feature fluorine in its end-member position, and preserve average hexagonal symmetry.7,16 The reduced intensity of the peak at higher chemical shift and more symmetrical 19F lineshape relative to the present work most likely reflects a larger fraction of 19F nuclei in the end-member fluorine position.


Page 13 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


These results further emphasize that minimizing hydroxyl component is important in investigating F,Cl apatite, because hydroxyl component forms readily and can impact the resulting structure in meaningful ways. Variation in fluorine environments The 19F NMR line shapes for F67Cl33, F50Cl50, F33Cl67, and F10Cl90 appear markedly different from those of F90Cl10 and F100; the width of the narrower, prominent feature is much larger in the former set of spectra, and a broad shoulder is observed at higher chemical shift. The lineshapes for these intermediate and Cl-rich compositions do not change markedly as a function of F:Cl ratio, suggesting that the 19F spectra do not relate simply to the nearest anion column occupant or average F/Cl ratio. Figure 7 demonstrates that the first 19F spectral moment exhibits non-linear variation with composition, further indicating that line shapes are not directly correlated with anion column chemistry. The 19F{35Cl} TRAPDOR data support this interpretation as well; neither of the spectral features, the narrower prominent signal nor the shoulder at higher chemical shift are selectively attenuated upon 35Cl irradiation, indicating that the 19F spectral features do not relate to differences in F-Cl distance or occupancy of adjacent column positions.

Lacking evidence for an effect from local column anion chemistry, the most plausible explanation for the present complex 19F NMR line shapes is a variation in bond distance between fluorine and the coordinating Ca2+ ions in the Ca2 triad. For example, Kiczenski and Stebbins17 noted that the 19F chemical shift generally increases with the radius of the adjoining cation. For constant metal oxidation state and structure type it is apparent that the relationship between 19F chemical shift and cation radius can be extended to metal-fluoride (M-F) interatomic distance. This effect is shown for a series of halite and fluorite-type metal fluorides in Figure 8 and Table 5. LiF has been excluded because chemical shift trends reverse with further decrease in radius ratio, as reported elsewhere.18 There are deviations from the purely linear trends between 19F chemical shift and M-F distance, particularly for the alkali halite-type fluorides, however these data indicate generally that increased M-F interatomic distance results in a higher chemical



Page 14 of 28

shift. Therefore, it is likely that the broad shoulder observed at higher chemical shift in the present 19F NMR result arises from F atoms with longer Ca2-F distances than the Ca2-F distance in end-member fluorapatite.

From a structural standpoint increasing the Ca2-F distance requires either dilation of the anion channel through displacement of the Ca2 triad in the a-b directions, or location of fluorine atoms away from the Ca2 triad along c. Displacement of the Ca2 triad is unlikely, as it would result in non-linear changes in the a lattice parameter, that were not observed for this series by refinement of powder X-ray diffraction data.7 The existence of additional atomic positions in the anion channel, on the other hand, has been observed in apatite with intermediate compositions along both the F,Cl and F,OH binaries.19,20 Such additional atomic positions could be occupied without large deviations from linear variation in lattice parameters. We assert that the 19F NMR line shapes in the present work most likely reflect variation in fluorine atomic position with respect to the Ca2 triad, yielding distributions of Ca-F bond distance. The narrower resonance at lower δ is assigned to fluorine near the end-member fluorine atomic position, within the Ca2 triad or slightly displaced above or below it (giving rise to the increase in peak width) which would yield the shortest Ca-F distance. The broad shoulder at higher δ is assigned to fluorine at a range of distances from the Ca2 triad, where the left-most edge of the shoulder represents fluorine atoms at the largest distance from Ca2. An estimate for this range in Ca2-F distance was obtained from 19F chemical shifts using the correlation in Figure 8 for the alkaline earth fluorides as the closer analogy in terms of charge and coordination number, although admittedly imperfect considering structural differences. The parameters used in this calculation are listed in Table 6. The column second to the left, labeled “Peak to Peak” lists the distance between the peak position of the narrow resonance at lower chemical shift and the peak position of the broad shoulder at higher chemical shift. The position of the peak in the narrow feature is assumed to represent F environments in the fluorapatite position, and the difference between this position and the first moment is assumed to reflect the average displacement of fluorine away from the Ca2 triad. Measuring the position of the broad shoulder relative to the narrow signal at lower δ compensates for the deshielding effect


Page 15 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


of Cl atoms and other structural parameters on 19F chemical shifts. The values labeled “peak to left edge” correspond to the chemical shift difference between the left-most edge of the broad shoulder and the center of the narrower peak. We propose that this difference reflects the maximum displacement of fluorine away from center of the Ca2 triad. The values labeled “average Ca2-F” and “max Ca2-F” represent the average and maximum Ca2-F distance obtained from the peak to peak and peak to left edge values, respectively, via the relation: (δ19F shift (ppm) ÷ 19F/dAEF) + dFAp = Ca2-F (Å) Where 19F/dAEF is the slope (296.4 ppm/Å) of the 19F chemical shift: metal-fluorine distance correlation for fluorite-type alkaline earth fluorides Figure 8, and dFAp is the Ca2-F distance for fluorine in its end-member position in fluorapatite (2.3109Å).5 From these estimated Ca2-F distances, the displacements of F along the c-axis from the Ca2 triad centroid can be determined from simple geometrical considerations: Vertical F displacement relative to Ca2 triad (Å) = (FCl ) − (F ) These estimated F displacements can be compared with F positions determined for the F50Cl50 composition from ScXRD refinement in a previous contribution.21 In that work, a partially occupied fluorine position was located displaced 0.57Å away from the center of the Ca2 triad. The average F positions in the second to right-most column in Table 6 should correspond to those values reported by XRD. However, comparison of these values reveals that the F displacements approximated from this analysis of the 19F NMR data are significantly less than those reported via ScXRD. The maximum F displacements shown in the right most column of Table 6 exhibit a better correspondence, although the maximum represents only a small fraction of the total fluorine population. The trend of the slope from CN=6 (halite) to CN=4 (fluorite) suggests that we are probably underestimating the slope of the line for apatite (CN=3). It should be noted, however that the value reported from XRD and those from 19F NMR are of the same order of magnitude, suggesting that our interpretation of the broad shoulder at higher chemical shifts as representing fluorine displacements away from the Ca2 triad is reasonable.

The 19F spectrum of F10Cl90 features a peak at -93.2 ppm, at a higher chemical shift than the 19F



Page 16 of 28

shift in fluorapatite and the shoulder for other F,Cl compositions. The position of this peak suggests the existence of a favored F position at a Ca2-F distance greater than that corresponding to the broad shoulder. From the above calculations this signal has a Ca2-F distance of 2.4Å, and is displaced approximately 0.46Å from the centroid of the Ca2 triad. This displacement is closer to the F position described by Hughes et al.,21 however ScXRD data for F10Cl90 are not available due to prohibitively small crystal size. The welldefined nature of this peak suggests it arises from F located at a preferred F position that is not currently described. 35

Cl and 19F Spectra of F90Cl10 The 19F NMR spectra of all intermediate compositions show similarly broadened spectral profiles,

with the exception of F90Cl10. The F90Cl10 19F spectrum more closely resembles that of F100 apatite, featuring a narrow prominent signal at -103.8 ppm, but with trailing intensity towards higher chemical shift containing distinct features. The 35Cl spectrum of F90Cl10 is somewhat anomalous as well in that it exhibits sharp horns at near +100 and +20 ppm, superimposed on a broad tail extending out to approximately -150 ppm, indicative of both ordered and disordered Cl environments. The disordered anion position regime proposed for the other intermediate F,Cl compositions is probably not present in this composition. Most of the fluorine atoms appear to be located near the end-member positions within the Ca2 triad, giving rise to the narrow 19F feature. Additionally, a subset of the chlorine atoms likely occupy a welldefined atomic position, which gives rise to the greater detail in the 35Cl NMR spectrum.

CONCLUSIONS & FUTURE WORK By careful synthesis we were able to prepare F,Cl apatites with exceptionally low hydroxyl component. Comparison of solid-state NMR spectra of the present samples to those samples believed to have greater concentrations of hydroxyl component has shown that the presence of hydroxyl groups affects both NMR chemical shifts and overall line shapes in significant ways. This results emphasizes not only the importance of the effect of hydroxyl groups in fluorine/chlorine arrangements, but also the necessity of minimizing hydroxyl content in investigating the F,Cl apatite binary.


Page 17 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The chemical shifts of all nuclei used in this study were observed to move to lower chemical shift as the chlorine content is increased, reflecting the importance of anion column chemistry on the lattice as a whole. However, significant deviations from linear mixing were observed in all cases, suggesting a complex mechanism of solid solution along this binary. The 19F NMR spectra are most telling, and through careful analysis of these spectra we suggest that these spectral profiles indicate a distribution of F positions in terms of displacements along c from the centroid of the Ca2 triad. This distribution is complex and quasicontinuous, but with distinct features representing more likely positions. These positions are not related to local occupancy of adjacent anion positions, but must represent long-range response to the steric requirements of accommodating both F and Cl in the anion channel. These results provide evidence for intimate mixing of F and Cl in the anion channel. Further analysis is warranted to determine definitively the variation of fluorine positions with respect to the Ca2 triad. Several apatite grains may be of sufficient size for single-crystal X-Ray diffraction at a synchrotron source, which would provide more precise information regarding the average fluorine position in the anion channel, and a more suitable comparison with the present 19F results. For example, ScXRD has revealed the presence of incommensurate scattering in a comparable F,Cl apatite composition13. It is possible that the inferred F displacements in the channel anions in the present samples also relate to an incommensurate modulation of the atomic position.

TABLES

F100 F90Cl10 F67Cl33 F50Cl50

Chemical Shift (ppm)

FWHM (ppm)

Relative Intensity

3.4(3) 2.5(2) 2.7(2) 1.1(3) 2.9(6) 1.6(2) 2.7(2) 1.6(2)

0.9(1) 0.7(2) 0.7(3) 0.8(1) 1.3(2) 1.3(4) 1.6(4) 1.2(3)

0.13(3) 0.87(3) 0.96(2) 0.04(1) 0.79(6) 0.21(7) 0.80(6) 0.20(4)

1st Spectral Moment (ppm) 2.7 2.7 2.6 2.6



F33Cl67 F10Cl90 Cl100

2.6(1) 2.2(2) 2.8(1) 1.8(3) 1.9(2) 1.70(4)

2.0(2) 1.8(1) 0.9(3) 0.7(3) 1.4(3) 0.4(2)

Page 18 of 28

0.33(12) 0.67(10) 0.21(11) 0.79(11) 0.29(20) 0.71(20)

2.4 2.0 1.8

Table 1: 31P NMR Parameters. Full peak width at half maximum (FWHM) shown in the second column. 1st spectral moment obtained by intensity weighted average. Uncertainty obtained by fitting spectra thrice.

Chemical Shift (ppm)

1st Spectral F moment (ppm)

FWHM (ppm)

Relative Intensity

1.5(1)

1

-102(1)

2.7(7) 3.2(4) 1.2(3) 1.7(1)

0.03(3) 0.14(4) 0.11(4) 0.72(5)

-101.4(8)

11.1(1) 4.8(4)

0.49(1) 0.50(1)

-100.3(5)

13.5(2) 5.0(4)

0.65(1) 0.35(1)

-98.3(5)

13.3(2) 4.92(4)

0.61(1) 0.39(1)

-100(3)

6.5(2) 8.4(5) 4.1(4)

0.20(1) 0.34(2) 0.46(1)

-102.0(4)

F100 -102.6(2) F90Cl10 -96.3(8) -99.3(9) -101.3(2) -103.8(1) F67Cl33 -97.3(1) -103.0(1) F50Cl50 -96.2(1) -103.0(4) F33Cl67 -97.2(4) -104.1(2) F10Cl90 -93.2(1) -101.2(1) -106.5(2)

19

Table 2: Results of fitting 19F MAS NMR Spectra. Full peak width at half maximum (FWHM) shown in the second column. 1st spectral moment obtained by intensity weighted average.

Composition

χ2

F90Cl10 F67Cl33 F50Cl50 F33Cl67 F10Cl90

δiso (ppm) 135 120 119 110 107

Avg. CQ (MHz) N/A 2.0 1.4 1.7 1.3

σ (MHz) N/A 1.1 0.8 0.9 0.7

117.3 36.24 32.05 78.6 49.89

Cl100

106

1.7

0.7

21.88

Table 3: Simulated 35Cl NMR parameters, showing the isotropic chemical shift (δiso) average quadrupolar coupling constant (CQ), and width of CQ distribution (σ). Chi squared values (χ2) from spectral fitting shown in the right-most column.

Nominal Composition

δF Present work (ppm)

F100

-102.6

δF O’Donnell et al. (ppm) -102.8


Page 19 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


F90Cl10 F50Cl50 F10Cl90

-103.8 -103.0 -106.5

-102.8 -103.4 -105.4

Table 4: 19F Chemical shifts of the narrow component in the present work compared to a previous work 19

Structure

dM-F(Å)

CaF2 SrF2 BaF2 NaF KF RbF

2.365 22 2.516 23 2.685 23 2.385 24 2.680 25 3.172 25

Fδ (ppm) -109 -84 -15 -225 -132 -90

Table 5: M-F interatomic distances and 19F chemical shifts of halite and fluorite-type metal fluorides. from Kiczenski and Stebbins17 and refs therein.

Sample

Peak to Peak (ppm)

Peak to left edge (ppm)

Average Ca2-F (Å)

Max Ca2-F (Å)

F67Cl33 F50Cl50 F33Cl67 F10Cl190

5.7 6.8 7.0 5.4

16.0 18.5 18.1 16.5

2.3 2.3 2.3 2.3

2.4 2.4 2.4 2.4

19

Average F displacement relative to Ca2 triad 0.30 0.33 0.33 0.29

F chemical shifts obtained

Furthest F displacement relative to Ca2 triad 0.50 0.54 0.54 0.51

Table 6: Approximation of Ca2-F distances 19F NMR results. Peak to peak and peak to left edge correspond to the difference in ppm between the center of the two spectral components and the center of the narrow component to the left edge of the broad shoulder, respectively. Average and max Ca2-F represent the Ca-F distance derived from the peak to peak and peak to left edge measurements, respectively. Average and furthest F displacement relative to Ca2 triad represent the displacement of F relative to the F position in fluorapatite derived from the average and max Ca-F distances, respectively

ACKNOWLEDGEMENTS This work was supported by the grant NSF EAR 1249696 to J. Hughes, H.Nekvasil and B. Phillips, which is gratefully acknowledged. Prepared by LLNL under Contract DE-AC52-07NA27344.

REFERENCES

1.

LeGeros RZ. Calcium Phosphate-Based Osteoinductive Materials. Chemical Reviews. 2008, 108, 4742-4753.

2.

Piccoli PM, Candela PA. Apatite in Igneous Systems. Reviews in Mineralogy and



Page 20 of 28

Geochemistry. 2002, 48, 255-292. 3.

Ewing RC, Wang L. Phosphates as Nuclear Waste Forms. Reviews in Mineralogy and Geochemistry. 2002, 48, 673-699.

4.

Kendrick E, Islam MS, Slater PR. Developing Apatites for Solid Oxide Fuel Cells: Insight into Structural, Transport and Doping Properties. J Mater Chem. 2007, 17, 3104-3111.

5.

Hughes JM, Rakovan J. The Crystal Structure of Apatite, Ca5(PO4)3(F,OH,Cl). Reviews in Mineralogy and Geochemistry. 2002, 48, 1-12.

6.

Hughes JM, Cameron M, Crowley KD. Structural Variations in Natural F. OH, and Cl Apatites. American Mineralogist. 1989, 74, 870-876.

7.

Vaughn JS, Woerner WR, Lindsley DH. Hydrogen Environments In Low-OH, F, Cl Apatites Revealed By Double Resonance Solid-State NMR. The Journal of Physical Chemistry C. 2015, 119, 28605-28613.

8.

Bryce DL, Bernard GM, Myrlene G, Lumsden MD, Eichele K, Wasylishen RE. Practical Aspects of Modern Routine Solid-State Multinuclear Magnetic Resonance Spectroscopy: One-Dimensional Experiments. Canadian Journal of Analytical Sciences and Spectroscopy. 2001, 46, 46-82.

9.

Grey CP, Vega AJ. Determination of the Quadrupole Coupling Constant of the Invisible Aluminum Spins in Zeolite HY with 1H/27Al TRAPDOR NMR. Journal of the American Chemical Society. 1995, 117, 8232-8242.

10.

Braun M, Jana C. 19F NMR Spectroscopy of Fluoridated Apatites. Chemical Physics Letters. 1995, 245, 19-22.

11.

Braun M, Hartmann P, Jana C. 19F and 31P NMR Spectroscopy of Calcium Apatites. Journal of Materials Science: Materials in Medicine. 1995, 6, 150-154.

12.

O’Donnell MD, Hill RG, Law RV, Fong S. Raman spectroscopy, 19F and 31P MAS-NMR of a Series of Fluorochloroapatites. Journal of the European Ceramic Society. 2009, 29, 377-384.

13.

McCubbin FM, Mason HE, Park H, Phillips BL, Parise JB, Nekvasil H, Lindsley DH. Synthesis and Characterization of Low-OH Fluor-chlorapatite: A Single-Crystal XRD and NMR Spectroscopic Study. American Mineralogist. 2008, 93, 210-216.

14.

Coster D, Blumenfeld AL, Fripiat JJ. Lewis Acid Sites and Surface Aluminum in Aluminas and Zeolites: A High-Resolution NMR Study. Journal of Physical Chemistry. 1994, 98, 6201-6211.

15.

Bryce DL, Sward GD. Solid-State NMR Spectroscopy of the Quadrupolar Halogens: Chlorine-35/37, Bromine-79/81, and Iodine-127. Magnetic Resonance in Chemistry. 2006, 44, 409-450.


Page 21 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


16.

Hughes JM, Cameron M, Crowley KD. Crystal Structures of Natural Ternary Apatites: Solid Solution in the Ca5(PO4)3X (X= F,OH,Cl) System. American Mineralogist. 1990, 75, 295304.

17.

Kiczenski TJ, Stebbins JF. Fluorine Sites in Calcium and Barium Oxyfluorides: F-19 NMR On Crystalline Model Compounds and Glasses. Journal of Non-Crystalline Solids. 2002, 306, 160-168.

18.

Hayashi S, Hayamizu K. Accurate Determination of NMR Chemical Shifts in Alkali Halides and their Correlation with Structural Factors. Bulletin of the Chemical Society of Japan. 1990, 63, 913-919.

19.

Mackie PE, Young RA. Fluorine-Chlorine Interaction in Fluor-Chlorapatite. Journal of Solid State Chemistry. 1974, 11, 319-329.

20.

Young RA, Lugt W, Elliott JC. Mechanism for Fluorine Inhibition of Diffusion in Hydroxyapatite. Nature. 1969, 223, 729-730.

21.

Hughes JM, Nekvasil H, Ustunisik G, Lindsley DH, Coraor AE, Vaughn JS, Phillips BL, McCubbin FM, Woerner WR. Solid Solution in the Fluorapatite-Chlorapatite Binary System: High-Precision Crystal Structure Refinements of Synthetic F-Cl Apatite. American Mineralogist. 2014, 99, 369-376.

22.

Swanson HE, Tatge E. Standard X-Ray Diffraction Powder Patterns. National Bureau of Standards Circular 539. 1966, 1, 69.

23.

Sobolev BP, Ippolitov EG, Zhigarnovskii BM. Phase Composition of the Systems CaF 2— YF 3, SrF2—YF3, and BaF2—YF3. Izvestiia Akademii Nauk SSSR, Neorganic Materialy. 1965, 1, 362.

24.

Kolditz L, Wilde W, Hilmer W. Phase Determinations by X-Rays on the Therman Dissociation and the Hydrolysis of Aklali Hexafluorogermanates at Higher Temperatures. Zeitschrift für anorganische und allgemeine Chemie. 1984, 512, 48-58.

25.

Davey WP. Precision Measurements of Crystals of the Alkali Halides. Physical Reviews. 1923, 21, 143-161.



Page 22 of 28

FIGURES

Figure 1: 31P SP NMR spectra of F,Cl apatites. 128 scans collected with a pulse delay of 30s


Page 23 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2a: 19F SP MAS/NMR spectra of F,Cl apatites, collected using 5µs π/2 pulses and 600s pulse delays at a 15 kHz spinning rate. Number of acquisitions varied for samples owing to the concentration of 19F; 4 scans for F100 and F90Cl10, and 8, 16, and 32 scans for F67Cl33, F50Cl50, and F33Cl67, respectively. 148 scans collected for both F10Cl90 and the rotor blank. Fig 2b: The center band region of the 19F NMR spectrum of F90Cl10 at expanded scale. Asterisks denote spinning side band positions (SSBs)



Page 24 of 28

Figure 3: Plot 19F chemical shifts vs. fluorine content for broadened F,Cl apatite compositions

Figure 4: 19F{35Cl} TRAPDOR NMR of F50Cl50. Spin echo spectrum shown in the upper spectrum, TRAPDOR difference spectrum shown below, scaled by a factor of 2.27 to facilitate comparison with the spin echo. 345 scans were collected at a spinning rate of 20kHz, using a 120s relaxation delay. The dephasing period was 0.75ms.


Page 25 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5: 35Cl SP experimental and simulated NMR Spectra of F,Cl Apatites. For the experimental spectra, the pulse delay was 2s for all samples, and 170,000, 80,000, 44,000, 5000, 4000, and 3700 scans were collected for F90Cl10, F67Cl33, F50Cl50, F33Cl67, F10Cl90, and Cl100, respectively.



Page 26 of 28

Figure 6: Plot of 1st 31P spectral moment vs. fluorine content. Uncertainties obtained by fitting spectra thrice and calculating the range of spectral moments for each fit.

Figure 7: Plot 19F chemical shifts vs. fluorine content


Page 27 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 8: 19F chemical shifts vs. M-F distance for halite and fluorite type metal fluorides



Page 28 of 28

TOC GRAPHIC:


Complex F,Cl Apatite Solid Solution Investigated Using Multinuclear

Recommend Documents