Synchrotron powder diffraction data for some smectite clay mineral standards

Abstract

Synchrotron powder diffraction data is presented for a series of relatively phase-pure smectite clay mineral standards obtained from the Clay Minerals Society. Rietveld refinement using a model for turbostratic disorder was performed to estimate the lattice parameters and mineral impurities in the smectite standards. Bragg reflection lists and raw data have been provided for inclusion in the Powder Diffraction File.

I. INTRODUCTION

Clay minerals are notable for both their importance to a broad array of industries (Harvey and Murray, 1997) and the significant challenges they present for correct identification, characterization, and quantification (Moore and Reynolds, 1989). Clays are layered phyllosilicates that are difficult to analyze due to their presence in complex heterogeneous mixtures, extensive faulting and disorder, layer swelling and contraction with water content, and a propensity for preferred orientation. Smectite clays in particular are challenging to physically model due to the presence of turbostratic disorder, where adjacent layers are randomly translated and rotated with respect to one another (Ufer et al., 2004; Gilg et al., 2020), giving rise to patterns with anisotropic broadening and asymmetric peak shapes. These features make smectite powder diffraction patterns difficult to adequately describe with a traditional d-spacing–intensity (d-I) list.

Synchrotron data collection with a capillary in transmission, particularly using a large two-dimensional (2D) detector, can provide some advantages for examination of mineralogical and clay samples. Capillary mounting with rotation during data collection and azimuthal integration of full 2D data sets are excellent strategies for minimizing both preferred orientation (clays) and graininess (quartz and other coarse minerals), making it easier to obtain accurate intensity estimates from challenging, heterogeneous samples.

The Powder Diffraction File (PDF; Gates-Rector and Blanton, 2019) began including raw diffraction data with experimental entries, when available, in PDF-4+ in 2008. The inclusion of raw data with PDF entries significantly improves the description of disordered, poorly crystalline, and amorphous materials like clays and polymers, facilitating better identification. Raw data deposited in the PDF has even been used to solve the crystal structures of pharmaceuticals that were previously unpublished (Reid et al., 2016; Reid and Kaduk, 2021). This work used synchrotron powder diffraction data to examine four relatively phase-pure smectite clay mineral standards obtained from the Clay Minerals Society (CMS) to provide high-quality powder patterns for the PDF. The patterns were examined with Rietveld refinement to determine their lattice parameters, estimate the crystalline phase compositions, and to aid indexing of the Bragg peaks.

II. EXPERIMENT

The smectite clay mineral standards examined in this work were obtained from the CMS and are summarized in Table I. Nominal chemical formulas obtained from previous studies (Van Olphen and Fripiat, 1979; Keeling et al., 2000) were assumed for the clays. Samples were ground with an agate mortar and pestle, and specimens were mounted in 0.3 mm ID polyimide capillaries for data collection.

TABLE I. The smectite clay mineral standards examined in this work.

Powder X-ray diffraction (PXRD) patterns were collected using the Canadian Macromolecular Crystallography Facility bend magnet beamline 08B1-1 (Fodje et al., 2014) at the Canadian Light Source (CLS). 08B1-1 is a bending magnet beamline with a Si (111) double crystal monochromator. 2D diffraction patterns were collected on a Pilatus3 S 6M detector with an active area of 423.6 mm × 434.6 mm. The patterns were collected at an energy of 18 keV (λ = ~0.6888 Å) and a sample-detector distance of ~350 mm.

The 2D PXRD patterns were calibrated and integrated using the GSASII software package (Toby and Von Dreele, 2013). The sample-detector distance, detector centering, tilt, and wavelength were calibrated using a series of patterns obtained from a lanthanum hexaboride (LaB₆) standard reference material (NIST SRM 660a LaB₆) and the calibration parameters were applied to all patterns. After calibration, the 2D patterns were integrated to obtain standard point detector powder diffraction patterns. Three to six short exposure 2D patterns (1 or 2 min) were obtained for each clay sample, to ensure that no hydration changes were taking place during data collection. Search/match phase identification was performed on the smectite clay patterns using the Powder Diffraction File, PDF-4+ (Gates-Rector and Blanton, 2019) to determine the mineral impurities.

Rietveld refinements were performed on the patterns using the BGMN software program (Bergmann et al., 1998) via the Profex graphical user interface (Döbelin and Kleeberg, 2015). BGMN uses a fundamental parameters approach for peak shape definition which was primarily designed for laboratory diffractometers. Establishment of appropriate instrument parameters for the synchrotron beamline was an iterative, empirical exercise. Optical elements of a previous synchrotron instrument parameter file (Döbelin, 2019) were modified in small increments and compiled at each step, to iteratively obtain a satisfactory starting description of the instrument parameters using a diffraction pattern from NIST SRM 660a LaB₆.

Rietveld refinements of the smectite clay data sets employed a turbostratic disorder model for smectite (Ufer et al., 2004, 2008) using tetrahedral-octahedral-tetrahedral (TOT) layer coordinates determined by electron diffraction (Tsipursky and Drits, 1984). Modifications were made to the starting octahedral and tetrahedral site occupations (based on the chemical formulas in Table I) for individual samples. Minor deviations were allowed in the occupations during refinement using hard limits on refined occupancies and constraints on total site occupation for individual sites.

III. RESULTS AND DISCUSSION

The powder patterns for the smectite clay minerals are illustrated with phase identification and indexing results in Figure 1 (nontronite samples NAu-1 and NAu-2) and Figure 2 (montmorillonite samples SAz-2 and STx-1b). While NAu-1 appeared phase-pure, minor impurities were observed in NAu-2 (quartz) and SAz-2 (quartz and feldspar). In addition to quartz and feldspar, STx-1b contained a large, broad peak attributable to opal or cristobalite. This was consistent with previous baseline powder diffraction studies of this clay, where the peak was attributed to opal-CT (Chipera and Bish, 2001).

Figure 1. Plots of the raw data and Bragg reflections for the observed phases in the nontronite samples, NAu-1 (top) and NAu-2 (bottom).

Figure 2. Plots of the raw data and Bragg reflections for the observed phases in the montmorillonite samples, SAz-2 (top) and STx-1b (bottom).

Rietveld refinements are illustrated for the four data sets in Figure 3 (nontronite samples NAu-1 and NAu-2) and Figure 4 (montmorillonite samples SAz-2 and STx-1b), and the lattice parameters, refinement indices, and mineral impurity contents are summarized in Table II. The refinement for the phase-pure NAu-1 yielded a β-angle very close to 100.4°. It was generally found that refining the β-angle resulted in minimal improvement to the refinements, but increased the estimated standard deviations in the lattice parameters. The β-angle was subsequently fixed at 100.4° for all four refinements. The a and b lattice parameters are larger in the nontronites than the montmorillonites, consistent with higher iron content in the octahedral sheets. The cis-trans site occupancies in the octahedral sheets refined to completely cis-vacant (NAu-1, NAu-2, and STx-1b) or predominantly cis-vacant (trans occupancy of 0.956(9) for SAz-2) for all refinements, consistent with most smectite clay minerals (Tsipursky and Drits, 1984; Gilg et al., 2020).

Figure 3. Plots of the final Rietveld refinements obtained for the nontronite samples, NAu-1 (top) and NAu-2 (bottom).

Figure 4. Plots of the final Rietveld refinements obtained for the montmorillonite samples, SAz-2 (top) and STx-1b (bottom).

TABLE II. The Rietveld-refined unit cell parameters for the smectite clays examined in this work, obtained using Profex/BGMN.

Three of the Rietveld refinements were excellent in terms of refinement indices and visually reasonable fits, but the refinement for NAu-2 was clearly inferior to the other three. The smectite model was unable to fully account for some of the features observed between ~10 and 18° in the experimental data set (Figure 3, bottom). The refinement was still suitable for estimating the lattice parameters and the quartz impurity content (<1 wt.%), which were the primary purposes in this work.

For the Rietveld refinement with the STx-1b data set, the opal/cristobalite phase was modeled using cristobalite with significant peak broadening. While this may not be a completely physically correct representation of the opal-CT phase, broadened cristobalite provided a reasonable fit yielding a phase composition that was consistent with previously reported results (Chipera and Bish, 2001).

Unit cells, Bragg reflection lists, and raw data for all four clays were provided in CIF format for inclusion in the PDF database. Three of the Bragg reflection lists (NAu-1, SAz-2, and STx-1b) were indexed consistent with turbostratic disorder considered as completely uncorrelated stacking (Ufer et al., 2004), where all reflections belong to either {hk0} (h or k can be zero) or {00l} sets. NAu-2 required deviation from completely uncorrelated stacking to account for a small number of Bragg reflections.

IV. DEPOSITED DATA

Crystallographic Information Framework (CIF) files containing unit cells, d-spacing–intensity lists and raw data were deposited with ICDD. The data can be requested at [email protected].