X-ray Diffraction and Clay Mineralogy

A crystal is a repeating lattice of atoms, and the spacing between its planes is similar in scale to the wavelength of X-rays (roughly 0.5 to 5 angstroms). When X-rays hit the crystal, waves reflected from successive parallel planes travel different path lengths. If the path-length difference equals a whole number of wavelengths, the waves reinforce each other (constructive interference) and a diffraction peak is observed. If not, they cancel. The condition for a peak is given by Bragg's law: nλ = 2d sin(θ), where n is an integer (usually 1), λ is the X-ray wavelength (fixed for a given source, commonly 1.5406 angstroms for Cu Kα radiation), d is the interplanar spacing, and θ is the glancing angle.

Because each mineral has unique interplanar spacings, it diffracts at unique angles. In a powder diffractometer, the sample is a randomly oriented powder (or an oriented clay mount) and the detector sweeps through angles from about 2 to 70 degrees 2-theta, recording the intensity at each angle. The resulting trace, plotted as intensity vs. 2-theta, is the powder diffractogram. Peak positions identify mineral phases; peak intensities (and ideally peak areas) allow quantification.

Clay minerals are phyllosilicates: sheet-like structures built from layers of silica tetrahedra and aluminium (or magnesium) octahedra, stacked along the c-axis. The repeat distance along this stacking direction, the basal or 001 d-spacing, is the primary diagnostic measurement. Different clay groups have characteristically different layer thicknesses:

• Kaolinite: a 1:1 clay (one tetrahedral sheet bonded to one octahedral sheet). Basal d-spacing is 7.1-7.2 angstroms and does not change with humidity or glycolation. Formed by intense weathering of feldspars under acidic, well-drained conditions. Common in tropical and subtropical soils.
• Illite: a 2:1 clay (two tetrahedral sheets sandwiching one octahedral sheet) with potassium ions fixed between the layers. Basal d-spacing is 10 angstroms, again non-expandable. Formed from muscovite mica by weathering or diagenesis; common in temperate soils and shales.
• Smectite: a 2:1 expandable clay. The interlayer holds water and exchangeable cations; the basal d-spacing swells from about 12-15 angstroms in air-dried conditions to around 17 angstroms after treatment with ethylene glycol. This expansion on glycolation is diagnostic. Common in soils derived from mafic and intermediate rocks and in poorly drained, seasonally wet settings.
• Chlorite: a 2:1+1 structure with a brucite-like interlayer that makes it non-expandable. Basal d-spacing around 14 angstroms, unaffected by glycolation, and stable on heating (distinguishing it from smectite). Common in low-grade metamorphic soils and in mafic rock weathering.
• Mixed-layer clays: most natural soils contain some interstratified clays where two layer types alternate. Illite-smectite mixed-layer clay is very common in diagenetically altered sediments. Its d-spacing and peak shape are intermediate and distinctive: a broad, asymmetric peak between the pure-phase positions.

Two preparation approaches are used for soil XRD. Bulk powder mounts handle the whole soil (everything from sand down to clay, typically sieved to less than 75 micrometres), giving an overview of all crystalline phases: quartz, feldspars, carbonates, iron oxides, and any clay present. Oriented clay mounts are prepared from just the fine clay fraction (less than 2 micrometres), extracted by repeated centrifugation and dispersion in water, and settled as a film on a glass slide or membrane filter.

The oriented mount works because clay platelets are flat; when they settle from suspension onto a flat surface they align with their basal planes horizontal. This means the 001 basal reflections are strongly reinforced while non-basal reflections weaken. The diagnostic clay peaks become intense and easy to measure.

1. Clay separation
   Soil is dispersed in distilled water (sometimes with sodium hexametaphosphate as a dispersant), allowed to settle until only particles smaller than 2 micrometres remain in suspension (Stokes' law gives the settling time), then centrifuged to collect the clay fraction as a paste.
2. Oriented slide preparation
   The clay paste is pipetted onto a glass slide and allowed to air-dry to a uniform film. Three duplicate slides are prepared for the three diagnostic treatments.
3. Air-dried scan
   The first slide is scanned from 2 to 30 degrees 2-theta in the air-dried state. Peak positions identify kaolinite (7.1 A), illite (10 A), smectite (~14 A), and chlorite (~14 A).
4. Ethylene glycol solvation
   The second slide is exposed to ethylene glycol vapour at 60 degrees Celsius for at least 8 hours. Smectite expands to 17 angstroms; kaolinite, illite, and chlorite are unchanged. This shift is the definitive test for expandable 2:1 clays.
5. Heating to 550 degrees Celsius
   The third slide is heated in a muffle furnace at 550 degrees for one hour. Kaolinite's structure collapses (its 7.1 A peak disappears). Smectite collapses to about 10 A. Chlorite retains its 14 A peak (its interlayer is structural, not water). Comparing the heated scan with the air-dried and glycolated scans allows unambiguous identification of all four groups.

Early XRD quantification used the intensity of a single diagnostic peak relative to an internal standard. This is fast but problematic when peaks from different minerals overlap (quartz and feldspar, or smectite and chlorite) or when crystallite sizes and disorder vary between samples. Rietveld refinement addresses these limitations by fitting the entire diffraction pattern, not just selected peaks.

The method requires a crystal structure model for each phase in the mixture. Software such as TOPAS, FullProf, or MAUD reads the measured diffractogram and simultaneously adjusts the scale factor (proportional to weight fraction), unit cell parameters, peak-shape parameters, and background for every phase. The scale factors at convergence give the weight percentages of each mineral. The method is validated using certified mixtures and by checking that quantified phases sum to 100%.

For forensic soil comparison, Rietveld-derived mineral abundances are more defensible than qualitative descriptions because they carry quantitative uncertainty estimates and are reproducible across laboratories using the same crystal structure models. A report stating '38 +/- 3 wt% kaolinite, 22 +/- 2 wt% illite, 15 +/- 2 wt% smectite' is far more actionable for source comparison than 'predominantly kaolinite with some illite.'

Clay mineral assemblages reflect the parent rock, the weathering regime, and the geological age of the soil. Kaolinite-dominant soils form by intense weathering in tropical, acidic, well-drained settings (laterites of West Africa, south and south-east Asia, Australia). Illite-dominant soils come from mechanical weathering of micas in cooler, drier settings or from burial diagenesis of shales. Smectite forms from mafic rocks (basalts, dolerites) under poorly drained, seasonally wet conditions, or from silica-rich volcanic ash. Mixed-layer illite-smectite is a hallmark of diagenetic conversion of smectite to illite in buried sediments and is common in Mesozoic and Palaeozoic shales.

These associations are broadly applicable and stable enough for forensic comparison. A soil dominated by kaolinite cannot have come from a recently basalt-weathered setting; a smectite-rich soil is not from a well-leached laterite. When combined with XRF elemental data and particle size, clay mineralogy provides a third independent axis of discrimination. Two soils from different geological settings that happen to share the same colour and major-element chemistry are very unlikely to also share the same clay assemblage.

Forensic XRD is not limited to soil. Building and construction environments produce mineral dusts with characteristic phase assemblages that can be matched to specific sites or processes. The most commonly encountered phases are:

• Portland cement and its hydration products: fresh cement contains tricalcium silicate (alite, Ca3SiO5), dicalcium silicate (belite, Ca2SiO4), tricalcium aluminate, and calcium aluminoferrite. Hydrated cement shows portlandite (calcium hydroxide, d = 4.9 A and 2.63 A), calcite from carbonation, and amorphous calcium silicate hydrate (CSH). The presence of unhydrated alite indicates fresh or recently mixed cement.
• Gypsum: calcium sulphate dihydrate (CaSO4.2H2O). A distinctive set of peaks at d = 7.6, 4.3, and 3.8 angstroms. Used in plasterboard, plaster coatings, and as a cement retarder. Hemihydrate (CaSO4.0.5H2O, plaster of Paris) has a different pattern (d = 6.0 A main peak).
• Calcium carbonate: calcite (the stable polymorph, d = 3.035 A strong peak) and aragonite (d = 3.396 A main peak). Calcite is ubiquitous in limestone building stone, mortar, concrete aggregate, and chalk dust. Aragonite is rarer, found in some marine building materials and shells.
• Quartz and other silica phases: quartz (d = 3.34 A, the most intense peak) is present in sand and concrete aggregate. Cristobalite and tridymite appear in some ceramic and industrial dusts. Silica phase identification is relevant when evaluating silicosis risk but also provides a background mineral in construction dust comparisons.

No single analytical technique individualises a soil source; the strength of a forensic soil comparison comes from combining independent lines of evidence. XRF gives the elemental chemistry (what atoms are present and in what amounts). ICP-MS adds trace-element detail and REE patterns. XRD supplies the mineralogy (how those atoms are arranged, which minerals are present). Colour measurement (Munsell chart or reflectance spectrometry) adds a physical property. Particle-size distribution adds another.

A questioned soil that matches a reference in elemental chemistry but differs in clay mineralogy is more likely from a different source than a soil that matches on both. Conversely, two soils from the same geological formation may share clay assemblage and elemental composition but differ in particle-size distribution if one has been mechanically disturbed. Reporting all available evidence, not just the technique that gives the most striking match, is best practice in forensic soil analysis.