Clay fractions of clastic argillaceous rocks are generally mixtures of diagenetic and detrital mineral components of substantially different ages. Study of the timing of diagenetic-illite formation has not reached its full potential in that it is difficult to separate detrital illite and mica from diagenetic illite and mixed-layered illite-smectite (Ilt-Sme) in the clay fractions of clastic argillaceous rocks. Establishing the timing of illitization in such rocks has long required extensive geochronometric measurements, usually done via the K–Ar method, and cautious interpretation of the resulting age values when physical methods, such as size separation by centrifuging, were not sufficient to separate detrital and diagenetic components (Pevear, Reference Pevear1999). Separation of one of these components from the other by chemical means could allow more accurate timing of diagenesis, but attempts to use intercalation via long-chain octadecylammonium ions for this purpose were not very favourable (Sears et al., Reference Sears, Hesse, Vali, Elliott, Aronson and Martin1998; Chaudhuri et al., Reference Chaudhuri, Środoń and Clauer1999; Yang et al. Reference Yang, Ji, Che and Zhou2002).
Exchange of Na+ for structural interlayer K of micas and the ‘Grundite’ illite was studied extensively at Iowa State University in the 1960s and early 1970s using sodium tetraphenylborate (NaBPh4; Scott et al., Reference Scott, Hunziker and Hanway1960; Scott, Reference Scott1976). Extractant mixtures of NaCl (≥1.0 mol L–1) as the primary source of Na+ and a smaller concentration of NaBPh4 (0.2 or 0.3 mol L–1) were employed. The high concentration of aqueous Na+ and an extremely low concentration of aqueous K+ maintained by precipitation of KBPh4 favoured exchange of aqueous Na+ for structural interlayer K that would not otherwise have been susceptible to exchange. This K extraction method, or some variant thereof (hereinafter, the tetraphenylborate (TPB) method or treatment), has since been used extensively to study soil K that is not readily exchangeable but can become available to plants (Bilias & Barbayiannis, Reference Bilias and Barbayiannis2017 and references therein).
The discovery by Scott & Smith (Reference Scott and Smith1966) that virtually all of the K of muscovite but not all of the K of an illitic underclay from Grundy County, Illinois (‘Grundite’; Grim & Bradley, Reference Grim and Bradley1939), could be extracted via the TPB method foretold that the method might be used to fully extract K from either the authigenic/diagenetic or the allogenic (or inherited) micaceous clay of argillaceous rock. If so, and if the associated radiogenic Ar were released along with the K, the effect of one of the two components on K–Ar age values would be eliminated. Scott (Reference Scott1968, fig. 7) showed that the percentage of structural K that cannot be extracted from micaceous minerals via the TPB method increases with decreasing particle size, and, of the materials he studied, this was most pronounced for fine illite (<0.08 µm). This result provided the basis for the hypothesis examined in the present study: that the TPB method can be used to extract, from clay fractions of clastic argillaceous rocks, nearly all of the interlayer K and radiogenic Ar of detrital muscovite while leaving some fraction of the diagenetic illite’s K and the associated radiogenic 40Ar. K–Ar analysis after such treatment then would yield an age value for the diagenetic illite of an original mixture of detrital and diagenetic components, provided feldspar K is negligible. In the present exploratory study, this hypothesis was tested on three Pennsylvanian-aged underclays, interpreted to be fossilized soils, or palaeosols, from the Illinois Basin. The measurement of the K–Ar age of diagenetic illite helps define whether illite was pedogenically formed in these palaeosols and as such represents a potential quantitative terrestrial climate proxy.
Materials and methods
Materials
Portions of palaeosol samples collected from drill cores of the Pennsylvanian Shelburn and Bond formations of the Illinois Basin (McIntosh et al., Reference McIntosh, Tabor and Rosenau2021, Reference McIntosh, Elliott, Wampler and Tabor2023) were used in the present study. The cores were the Lone Star Cement Company #TH-1 (LSC), the Illinois State Geological Survey #1 City of Charleston (CHA) and the American Coal Company Borehole 7510-20 (HAM), previously presented by Rosenau et al. (Reference Rosenau, Tabor, Elrick and Nelson2013a, Reference Rosenau, Tabor, Elrick and Nelson2013b) and McIntosh et al. (Reference McIntosh, Tabor and Rosenau2021, Reference McIntosh, Elliott, Wampler and Tabor2023, Reference McIntosh, Tabor and Montañez2025). Whole pieces from three palaeosol samples (LSC-16N, CHA-83 and HAM-18; McIntosh et al., Reference McIntosh, Elliott, Wampler and Tabor2023) were chosen to provide inter-sample differences in mineralogy and the K–Ar age values of their clay fractions. The palaeosols are classified as calcic vertisol (LSC-16N), gleyed vertisol (CHA-83) and gleyed protosol (HAM-18) and were buried to different maximum depths ranging from ∼1 km in the north (LSC-16N) to ∼3 km in the south of the basin (HAM-18; McIntosh et al., Reference McIntosh, Elliott, Wampler and Tabor2023). Their clay mineralogy included Ilt-Sme with R0, R1 and R3 stacking orders, mica (muscovite), kaolinite, quartz, chlorite and rare feldspar (Rosenau et al., Reference Rosenau, Tabor, Elrick and Nelson2013a; McIntosh et al., Reference McIntosh, Tabor and Rosenau2021, Reference McIntosh, Elliott, Wampler and Tabor2023). The K–Ar age values of the clay fractions (<2 μm) ranged from ∼270 Ma for LSC-16N to ∼310 Ma for HAM-18, while those of the finest clay subfractions (<0.1 µm) ranged from ∼220 Ma for LSC-16N to ∼265 Ma for HAM-18 (McIntosh et al., Reference McIntosh, Elliott, Wampler and Tabor2023).
The core pieces chosen for the present study (∼20 g each) were gently disaggregated and then treated chemically with buffered sodium acetate–acetic acid (NaOAc–HOAc; pH 5) and sodium bicarbonate–sodium citrate–sodium dithionite (CBD) to remove carbonate and iron oxide cements, respectively (Jackson, Reference Jackson and Jackson1979). The coarse clay (2–5 µm) was separated by timed settling. Medium- and fine-clay fractions (1–2 and <1 µm, respectively) were separated by centrifugation at 1300 rpm for ∼4 min with a Heraeus floor centrifuge. 1 M NaCl and CaCl2 solutions were added to suspensions to aid flocculation. CaCl2 was used to flocculate the finest clay fractions where it could not be settled in larger volumes using NaCl alone. Settled clay fractions were washed repeatedly in deionized water to remove salts. After portions of suspended clay were transferred by pipette to glass slides to prepare orientated mounts for X-ray diffraction (XRD) analysis, the remainder of each clay fraction was dried in air at 80°C.
In addition to the Illinois Basin palaeosol clay fractions, the <2 µm fraction of an Ordovician K-bentonite from Shelbysville, Tennessee, was included in the study to establish how the TPB method affects monomineralic, diagenetic Ilt-Sme (R1 ordered, 80% illite layers; Elliott & Aronson, Reference Elliott and Aronson1987). Fe oxides and carbonate were removed from this K-bentonite prior to size separation (Jackson, Reference Jackson and Jackson1979). This served as a control sample that has no detrital mica detectable by XRD. There was no need to separate this clay to submicron fractions.
Sodium tetraphenylborate (American Chemical Society (ACS) reagent grade, Spectrum Chemical Mfg. Corp.) and certified ACS grades of NaCl, NH4Cl and acetone (Fisher Chemical) were used for TPB treatments. Ultrapure deionized water (18.2 MΩ-cm) was used to prepare all solutions and to rinse treatment products.
Potassium extraction via the TPB method
Interlayer K was extracted at room temperature over varying time periods from 250 mg portions of each clay fraction (2–5, 1–2 and <1 µm) of each palaeosol and of the K-bentonite clay by exchange with Na+ in NaCl–NaBPh4 solution. Each clay portion was suspended in an amber-coloured 50 mL polypropylene centrifuge tube in 10 mL of a mixed solution of NaCl (1.0 mol L–1), NaBPh4 (0.2 mol L–1) and disodium ethylenediaminetetraacetic acid (Na-EDTA; 0.01 mol L–1). The suspensions were slowly rotated end over end (5 revolutions per min; Fisherbrand™ tube rotator) for either 1 h, 1 day, 22 days or 27 days. The longer 27 day period was used for the fine <1 µm fraction of each palaeosol after treatment with TPB and the TPB-treated fractions of the K-bentonite.
At the end of the preset period, the reaction in each suspension was quenched with 20 mL of a solution of NH4Cl (0.5 mol L–1) in a 60:40 acetone: water mixture, and the then larger suspension was manually shaken for 1 min. The quenching step saturates the available cation sites with NH4+ and Na+. This exchange was shown to prevent the reuptake of soluble K into the mica interlayers (Hanway, Reference Hanway1954; Reed, Reference Reed1963). The K-TBP precipitate remaining would be more soluble in acetone and thus be concentrated in acetone during centrifugation. The suspensions were centrifuged at 5600 rpm for 10 min (Thermo Fisher Scientific Sorvall ST 8FR centrifuge). The supernatant solutions were decanted transferred to 40 mL fluorinated ethylene propylene (FEP) centrifuge tubes. The treated clays were resuspended, each in 20 mL of the quenching solution, with the aid of a vortex mixer. After an additional 20 mL of quenching solution (NH4Cl (0.5 mol L–1) in a 60:40 acetone: water mixture) had been added, each suspension was manually shaken and centrifuged as before. After decantation, the process of resuspension in quenching solution, centrifugation and decanting was repeated once more for three successive washes with quenching solution. The treated clays were then rinsed three times with ultrapure deionized water, with phase separation by centrifugation as before, and then suspended in ∼5 mL of a 50:50 methanol: water mixture. This 50:50 methanol:water suspension was centrifuged, and, after decanting, the treated clay was resuspended in ultrapure deionized water. Repetitive washing with the quenching solution composed of Na-EDTA, NH4Cl and NaCl further prevented any uptake of K by mica and solubilized any remaining K-TPB precipitate. The final wash with deionized water removed excess quenching solution and produced a clay fraction free from loosely adsorbed ions for subsequent K–Ar analyses. A portion of each suspension was used for analysis. The remainders were dried in an oven at 80°C, and the solids were held for K–Ar analyses.
X-ray diffractometry
Suspensions of treated and untreated clay fractions were pipetted onto petrographic slides and dried at room temperature and humidity to prepare orientated clays for XRD. One of three orientated clays prepared from each untreated clay fraction was saturated with ethylene glycol, and another was heated at 575°C. These preparations were scanned on a PANalytical X’Pert Pro XRD device with a 3071/00 flat sample stage with Ni-filtered Cu-Kα radiation generated at 45 kV and 40 mA, using a fixed anti-scatter slit at 1/32°, at a rate of ∼1 min per °2θ. The d values were assigned to the observed diffraction peaks using the HighScore software package.
K–Ar geochronology
A single test portion of each sample was used for both the Ar isotopic analysis and K determination to obtain K–Ar age values of the TPB-treated clay fractions and the untreated clay fractions (Stephens et al., Reference Stephens, Anderson, Gullet-Young, Wampler and Elliott2007; McIntosh et al., Reference McIntosh, Elliott, Wampler and Tabor2023). The one-weigh method was used to decrease analytical error for the K–Ar determination. Approximately 20 mg of each air-dried clay fraction was weighed into a copper-foil capsule (Denver Instrument M-220 balance), which was then closed by folding but not sealed. The remainder of the procedure was similar in process to that used by McIntosh et al. (Reference McIntosh, Elliott, Wampler and Tabor2023), in which the test portion in each capsule was heated to ∼1000°C for 15 min to extract Ar. The Ar was diluted with a known volume of 38Ar, purified and isotopically analysed in static mode with a modified MS-10 mass spectrometer (Associated Electrical Industries, Inc.). Ion beam currents for 36Ar, 38Ar and 40Ar were measured over nine scans to reduce the effect of random error caused by the absence of a working emission controller for the ion source. Ion beam current ratios were converted to molar isotope ratios (40Ar/38Ar and 40Ar/36Ar) using mass discrimination corrections determined by repeated isotopic analyses of a reference argon of known isotopic composition.
After the argon had been extracted and analysed for all test portions in a set, the capsules were removed from the extraction line and digested in fluorocarbon containers with a mixture of HF, HNO3 and HClO4. After evaporation of most of the remaining acid, a Cs-bearing (0.01 mol kg–1) test solution was prepared from each residue for potassium measurement by atomic absorption spectrophotometry (Perkin Elmer 3110). The potassium reference solutions were prepared from standard KCl (National Institute of Standards and Technology SRM 999). K–Ar age values were calculated (Supplementary File 1), using the 40K decay constants and the K and Ar isotopic abundances of Steiger & Jäger (Reference Steiger and Jäger1977).
Estimates of the effect of random analytical error on the age values were calculated as 2σ error values based on the precision of K determination and Ar isotopic analyses observed in repeated analyses of other materials following the same procedures in this laboratory from the past few years. There were no repeated analyses of the materials in the present study. To obtain a 2σ error value, a standard deviation for the ratio of radiogenic Ar to K was calculated by addition in quadrature of the estimated precision of each of four contributing quantities (Dalrymple & Lanphere, Reference Dalrymple and Lanphere1969, eq. 7-1). The result was then multiplied by 2 and by a factor that accounts for the non-linearity of the K–Ar age equation (Baksi, Reference Baksi1982).
Results
Mineralogy of the materials before TPB treatment
Ilt-Sme, muscovite (herein, discrete 10.0 Å minerals) and kaolinite were identified by XRD (Table 1 & Fig. S1a–c in Supplementary File 2) in all of the untreated palaeosol clay fractions. Only a trace of mica was seen in the LSC-16N fine clay. Chlorite was found only in the HAM-18 clay fractions. Quartz and feldspar were identified in all of the coarser clay fractions, with distinctly more present in the 2–5 µm fractions than in the 1–2 µm fractions of HAM-18 and LSC-16N; these minerals were not found in any fine clay fraction except for a trace of quartz in the CHA-83 fine fraction. The observed Ilt-Sme ordering was R0 for LSC-16N, R1 for CHA-83 and R3 for HAM-18. These results are generally consistent with what McIntosh et al. (Reference McIntosh, Elliott, Wampler and Tabor2023) found for other portions of the same core samples at the same depths, but there was no feldspar in the LSC-16N clay studied earlier.
Minerals identified by XRD in Illinois Basin palaeosol clay fractions and the Shelbysville K-bentonite clay fraction (KB-5) before and after TPB treatment.
Table 1 Long description
The table lists minerals detected by XRD in clay-size fractions from three Illinois Basin palaeosol samples (HAM-18, LSC-16N, CHA-83) and one K-bentonite sample (KB-5), split by particle-size range and TPB exposure time (none, one hour, one day, about three to four weeks). For untreated palaeosol clays, illite–smectite is reported with Reichweite ordering that differs by sample: HAM-18 is R3, LSC-16N is R0, and CHA-83 is R1 across size fractions. After TPB treatment, the illite–smectite column is no longer populated and an interstratified complex is identified instead for all treated palaeosol fractions and for KB-5. Kaolinite is consistently present in all palaeosol fractions before and after treatment, and chlorite is present in HAM-18 but not reported for LSC-16N or CHA-83. Quartz and feldspar are common in the coarser palaeosol fractions but diminish with decreasing size, often dropping to trace or absent in the less than one micrometre fractions. In HAM-18, mica-like ten angstrom phases persist through treatment, while in LSC-16N the mica-like phase is strong in coarser fractions but only trace in the finest untreated fraction. KB-5 shows no minerals listed when untreated, but after TPB only the interstratified complex is marked, with other phases not reported. Interpretation should note that treated portions were assessed from air-dried patterns only, so mineral identifications after TPB are not directly comparable in method detail to untreated identifications.
‘Interstratified complex’ denotes mixed-layered material with artificially altered interlayers. ‘Mca’ denotes undifferentiated 10.0 Å phases. ‘Kln’ denotes kaolinite. ‘Chl’ denotes chlorite. ‘Qz’ denotes quartz. ‘Fsp’ denotes feldspar.
‘R’ denotes Reichweite notation and ‘tr.’ denotes trace. R values were determined from XRD traces of ethylene glycol-solvated preparations.
Identification of minerals in untreated clays was based on XRD traces of air-dried, ethylene glycol-solvated and heated preparations. Only air-dried patterns were obtained for the TPB-treated portions.
Peak intensities indicate abundant mica in the coarse and medium clay fractions of HAM-18, consistent with the relatively large amount of muscovite found in a clay fraction of HAM-18 (McIntosh et al., Reference McIntosh, Elliott, Wampler and Tabor2023). The relative intensities of mica and Ilt-Sme peaks indicate that: (1) mica is distinctly less abundant relative to the Ilt-Sme in the fine clay fractions of all three palaeosols than in the coarser fractions; and (2) unlike the HAM-18 coarse clay, the LSC-16N and CHA-83 coarse clay fractions do not have appreciably more mica than their respective medium clay fractions.
The XRD traces of the Shelbysville K-bentonite clay before TPB treatment (Fig. S1d) are consistent with the original finding that it is monomineralic and consists of illite-rich R1-ordered Ilt-Sme (Elliott & Aronson, Reference Elliott and Aronson1987). The main feature of the XRD trace for the K-bentonite’s air-dried clay is a broad Ilt-Sme peak centred at 10.8 Å.
Mineralogy of TPB-treated clays
The XRD traces for the air-dried K-bentonite clay after each TPB treatment (Fig. S2) show a large, broad peak at almost the same position as the main Ilt-Sme peak of the air-dried untreated K-bentonite clay, but their intensities are considerably greater. Small peaks are similarly close to the original small peak at 3.25 Å, but the intensity differences are much less pronounced.
On the XRD traces for air-dried TPB-treated palaeosol clay fractions (Figs 1, S3 & S4), 10.0 Å mica peaks are seen only as shoulders on the broad main peaks of interstratified Ilt-Sme material. These shoulders are most prominent on the XRD traces for 1 h treatment of the coarse clay fractions. They are also easily seen, but have reduced intensity, on the patterns for 1 day treatment of the HAM-18 and CHA-83 coarse clay fractions. They are not present on any XRD trace for 22 or 27 day treatments.
XRD traces of the air-dried 2–5 µm fractions of HAM-18, LSC-16N and CHA-83 before treatment and after TPB treatment for 1 h, 1 day and 22 days.
Figure 1 Long description
The image contains three X-ray diffraction plots for air-dried 2–5 micrometer fractions of HAM-18, LSC-16N and CHA-83. Each plot compares untreated samples with those treated with TPB for 1 hour, 1 day and 22 days. The x-axis is labeled ′2 Theta′ in degrees, ranging from 0 to 30. The y-axis is labeled ′Counts,′ ranging from 0 to 6000. For HAM-18, peaks are observed at 10.0 Å, 15.4 Å and 3.34 Å, with intensity changes after treatment. In LSC-16N, peaks at 12.5 Å and 3.34 Å are prominent, with noticeable shifts in intensity and position after treatment. CHA-83 shows peaks at 7.15 Å and 3.34 Å, with variations in peak intensity and position across treatments. The plots illustrate how TPB treatment affects the diffraction patterns, indicating changes in the clay mineral structure over time.
The XRD trace for each of the nine TPB-treated air-dried palaeosol clay fractions has a relatively large and broad peak to the high-d side of 10.0 Å. Their highest points are variable in position, but all are in the range 10.3–12.3 Å. For the LSC-16N fractions, these peaks are very broad, as were the corresponding peaks for this R0 Ilt-Sme before treatment, but the peaks for the treated clay are centred at distinctly smaller d values than are those for the original clay.
The main peaks for the air-dried HAM-18 and CHA-83 TPB-treated clay fractions are distinctly narrower and centred at smaller d values than those for LSC-16N. For HAM-18, these peaks exhibit a distinct trend of increasing peak intensity with increasing treatment duration if intensities are normalized by the observed intensity of the 7.2 Å peak for kaolinite. As for the treated K-bentonite clay, these peaks are narrow, and most are distinctly taller than the corresponding Ilt-Sme peak of the untreated clay.
K contents
The K contents (as K) of the untreated palaeosol clay fractions ranged from ∼2% to 4% by mass (Table 2), with the values for LSC-16N fractions all being <3% and those for HAM-18 all being >3%, which is in general agreement with the values found by McIntosh et al. (Reference McIntosh, Elliott, Wampler and Tabor2023) for similar materials. The K content determined for the Shelbysville K-bentonite clay was 5.10% (6.15% K2O), which is in good agreement with the published value of 6.34% K2O (Elliott & Aronson, Reference Elliott and Aronson1987).
K–Ar data for Illinois Basin palaeosol clay fractions and the Shelbysville K-bentonite clay fraction before and after TPB treatment.
Table 2 Long description
The table reports potassium content, radiogenic argon measures, and resulting K–Ar ages for clay size fractions from three palaeosol samples (HAM−18, LSC−16N, CHA−83) and one K-bentonite sample (KB−5), before and after TPB treatment for 1 hour, 1 day, and about 22 to 27 days. For untreated palaeosol fractions, potassium is about 2.08 to 4.20 percent and radiogenic argon is about 88.7 to 95.1 percent, with ages roughly 260 to 316 million years depending on size fraction. With TPB treatment, potassium generally drops (often to about 0.61 to 2.90 percent) and radiogenic argon percent also declines (down to about 38.6 to 79.7 percent), with the lowest radiogenic argon percent in CHA−83, 2 to 5 micrometre, 22 days. Despite these compositional shifts, most ages remain in a similar range, commonly about 245 to 325 million years, though some long-treatment results are lower, such as HAM−18 and CHA−83 near 245 million years for 22-day treatments and LSC−16N reaching about 237 million years after 27 days in the less than 1 micrometre fraction. Across size fractions, the coarser 2 to 5 micrometre fractions often yield older ages than finer fractions in untreated and short-treatment conditions, for example HAM−18 and CHA−83. The KB−5 bentonite (less than 2 micrometres) shows potassium decreasing from 5.10 percent untreated to about 1.30 percent after 27 days, while ages remain near the mid 250s to low 270s million years. Interpretation should note that treatment reduces radiogenic argon percentages and potassium, which can increase uncertainty and may affect comparability across treatment durations.
The K content of the Shelbysville K-bentonite clay portion that was TPB treated for 1 h was only 2.31%, less than half that of the untreated clay (Fig. 2 & Table 2). The K contents were even smaller after longer treatments, at 1.52% and 1.30% for the portions treated for 1 day and for 27 days, respectively. For each palaeosol clay fraction, K contents of the treated portions became progressively smaller as treatment duration increased; the differences were quite small for the HAM-18 fine clay, but in most cases larger differences produce a pronounced step-down pattern whereby the K content of the longest-treated portion is well below that of both the untreated clay and the 1 h-treated portion. For each of the three palaeosol samples, the smallest relative difference in K content between the untreated clay and the nearly 1 month-treated portion was for the finest clay, and the largest difference in K was for the coarsest clay.
K–Ar age values and K contents of Illinois Basin palaeosol clay fractions (HAM-18, LSC-16N and CHA-83) and the Shelbysville K-bentonite (KB-5) clay fraction before and after TPB treatment. Gray = untreated clay fractions; colours = clay fractions TPB treated for 1 h (yellow), 1 day (green) and 22 or 27 days (blue); horizontal bars = K contents (% K by mass).
Figure 2 Long description
Four separate grouped bar charts titled HAM-18, LSC-16N, CHA-83 and KB-5 less than 2 micrometer. HAM-18 chart: Left vertical axis label Age value (Myr) with range 0 to 350. Right vertical axis label K content (percent K by mass) with range 0 to 7. Horizontal axis categories: less than 1 micrometer, 1 dash 2 micrometer, 2 dash 5 micrometer. Legend entries: Untreated, 1 hour, 1 day, 1 month, K content. Less than 1 micrometer: Age value bars at approximately 280 (Untreated), 270 (1 hour), 270 (1 day), 250 (1 month). K content bar at approximately 2. 1 dash 2 micrometer: Age value bars at approximately 300 (Untreated), 290 (1 hour), 290 (1 day), 240 (1 month). K content bar at approximately 3. 2 dash 5 micrometer: Age value bars at approximately 330 (Untreated), 320 (1 hour), 320 (1 day), 240 (1 month). K content bar at approximately 4. LSC-16N chart: Left vertical axis label Age value (Myr) with range 0 to 350. Right vertical axis label K content (percent K by mass) with range 0 to 7. Horizontal axis categories: less than 1 micrometer, 1 dash 2 micrometer, 2 dash 5 micrometer. Legend entries: Untreated, 1 hour, 1 day, 1 month, K content. Less than 1 micrometer: Age value bars at approximately 280 (Untreated), 270 (1 hour), 260 (1 day), 230 (1 month). K content bar at approximately 1.5. 1 dash 2 micrometer: Age value bars at approximately 280 (Untreated), 270 (1 hour), 260 (1 day), 250 (1 month). K content bar at approximately 1.5. 2 dash 5 micrometer: Age value bars at approximately 300 (Untreated), 290 (1 hour), 280 (1 day), 260 (1 month). K content bar at approximately 2. CHA-83 chart: Left vertical axis label Age value (Myr) with range 0 to 350. Right vertical axis label K content (percent K by mass) with range 0 to 7. Horizontal axis categories: less than 1 micrometer, 1 dash 2 micrometer, 2 dash 5 micrometer. Legend entries: Untreated, 1 hour, 1 day, 1 month, K content. Less than 1 micrometer: Age value bars at approximately 280 (Untreated), 290 (1 hour), 260 (1 day), 250 (1 month). K content bar at approximately 2. 1 dash 2 micrometer: Age value bars at approximately 300 (Untreated), 290 (1 hour), 270 (1 day), 250 (1 month). K content bar at approximately 2. 2 dash 5 micrometer: Age value bars at approximately 330 (Untreated), 320 (1 hour), 300 (1 day), 240 (1 month). K content bar at approximately 1. KB-5 less than 2 micrometer chart: Left vertical axis label Age value (Myr) with range 0 to 350. Right vertical axis label K content (percent K by mass) with range 0 to 7. Horizontal axis shows four treatment categories: Untreated, 1 hour, 1 day, 1 month. Legend entries: Untreated, 1 hour, 1 day, 1 month, K content. Age value bars: approximately 270 (Untreated), 260 (1 hour), 250 (1 day), 260 (1 month). K content bars: approximately 5.5 (Untreated), 2.5 (1 hour), 2.0 (1 day), 2.0 (1 month).
Special note needs to be taken regarding the K content of the 1 h-treated fine clay and coarse clay of LSC-16N, for which K contents were larger than and approximately the same as, respectively, that of the untreated clay. The radiogenic Ar contents of each of these fractions (Table 2) were larger after the 1 h treatment than before, and by large factors (20% and 10% for the fine and coarse clay fractions, respectively). That independently measured K and radiogenic Ar both were anomalously large indicates that the problem probably was not due to mistakes in their measurement. Instead, the explanation for the anomalous increases is probably that the mass of each treated portion was substantially less after the 1 h treatment than before. Some decrease in mass of the clay may be attributed to changes in interlayer materials during the K extraction and during the subsequent processes of quenching with NH4Cl in acetone–water solution, rinsing and air drying, but such decreases were probably not more than ∼5% (Supplementary File 2). The masses of other treated portions also were probably less after treatment than before. Because two fractions of LSC-16N appear to have experienced mass decreases of much more than 5%, an additional cause of mass decrease during the first hour of TPB treatment remains to be found for at least these two fractions.
Accurate determination of the amount of K extracted would require that it be measured directly, not obtained from differences in K content, but this information is not available for the present study. In the absence of such information, the difference in K content between an untreated clay and a treated portion of it may be taken as a semiquantitative indicator of the amount of K extracted, but only if the difference is relatively large. Thus, when differences are large, semiquantitative estimates of the amounts of K and radiogenic Ar extracted are possible, providing an indication of the age of the minerals from which they were extracted.
K–Ar age values
The age value of the untreated K-bentonite clay of 271 ± 4 Myr (Table 2) agrees with the value for this clay determined earlier at 274 ± 13 Ma (Elliott & Aronson, Reference Elliott and Aronson1987). Although the TPB treatments caused large decreases in K content, none of them caused much change in the K–Ar age value. The age values of the treated bentonite clay trend downward from those of the untreated clay by amounts that, for the two longer treatments, are a little more than the sum of the 2σ errors.
One hour of TPB treatment produced little change in age value of any of the HAM-18 fractions, but the age values of most fractions of the two other palaeosols were larger after such treatment by amounts that are small (∼15 Myr) but exceed the sum of the 2σ errors. One day of treatment also produced no large change in age value of any of the palaeosol clay fractions. In contrast to the shorter treatment, however, most changes upon 1 day of treatment appear as decreases relative to that of the untreated clay, with some significant and some not. Nearly 1 month of TPB treatment caused significant decreases in K–Ar age values relative to those of the untreated clay fractions across all fractions of each palaeosol. For LSC-16N, the decreases were relatively small (∼25 Myr) and little, if at all, dependent on grain size. For the two other palaeosols, the amount of decrease increased notably with increasing grain size, and for the two coarse clay factions the decreases were ∼70 Myr. A striking result of the nearly 1 month treatments is that the age values of all three fractions of each of the three palaeosols are grouped around the 1 month grand average of 247 ± 5 (1σ) million years, and tightly so except for the fine-fraction values of HAM-18 and LSC-16N, which differ from the average by a little more than their 2σ errors.
Discussion
A conceptual framework for further discussion
The mineralogical and geochronological results of the present study show clearly that the K and Ar in the detrital mica components of the palaeosol clays were more completely extracted by the TPB method than were the K and Ar in the diagenetic Ilt-Sme. This assertion depends, however, on a conceptual framework built from earlier work on K extraction from mica and illite by the TPB method and from past and present mineralogical and geochronological work on untreated Illinois Basin palaeosol clay fractions and the Shelbysville K-bentonite clay.
The K-bentonite clay, consisting entirely of K-rich, R1-ordered Ilt-Sme, lost a large fraction of its K in 1 h of treatment and somewhat more but not all during longer treatments (Fig. 2). The material remaining after the treatments was still interstratified phyllosilicate, closely resembling the original Ilt-Sme in XRD trace. These results are similar to what Scott & Smith (Reference Scott and Smith1966) found for the clay fraction of Grundite illite. The present results, however, are more definitive in showing the behaviour of Ilt-Sme, because the clay fraction of Grundite had more discrete illite than Ilt-Sme (Gaudette et al., Reference Gaudette, Eades and Grim1964). These findings confirm, as expected from the work of Scott & Smith (Reference Scott and Smith1966) and Scott (Reference Scott1968), that extended (i.e. for weeks or longer) TPB treatment extracts some, but not all, of the interlayer K from diagenetic Ilt-Sme, and that most of that K is extracted within the first hour. The absence of a significant change in the K–Ar age value shows that, to a good approximation, associated radiogenic Ar, but no other radiogenic Ar, was lost as the K was extracted.
The present study’s XRD traces show Ilt-Sme and 10 Å material in all of the untreated palaeosol clay fractions, but only a trace of the latter in the fine clay of LSC-16N. The amounts of 10 Å material increased with increasing grain size (Fig. S1a–c), as is typical of mudstones (Pevear, Reference Pevear1999). Given existing evidence that the 10 Å materials include detrital components older than 300 Ma and that the diagenetic components are considerably younger than that (McIntosh et al., Reference McIntosh, Elliott, Wampler and Tabor2023), the K–Ar age values of the untreated fractions reflect the mineralogical trends shown by XRD: the age values are least for Ilt-Sme-rich LSC-16N and greatest overall for HAM-18, which has the most prominent 10.0 Å peaks. The age values of the untreated clay fractions of each palaeosol increased with increasing grain size, and most distinctly so for CHA-83 (Fig. 2 & Table 2).
The picture that has emerged from the work of McIntosh et al. (Reference McIntosh, Elliott, Wampler and Tabor2023) and the present XRD and K–Ar results for untreated clay fractions is one of palaeosols whose K-bearing minerals are mixtures, in different proportions, of generally coarser, detrital, dioctahedral micaceous components older than 300 Ma and generally finer, diagenetic Ilt-Sme that is considerably younger, with some at least as young as ∼220 Ma. These results are consistent with the understanding of the provenance and post-depositional history of the Illinois Basin Pennsylvanian siliciclastic sediments as reviewed by McIntosh et al. (Reference McIntosh, Elliott, Wampler and Tabor2023). In brief, these sediments were predominantly mica detrital grains weathered and transported from Grenville (1300–980 Ma) and Appalachian (490–350 Ma) basement terranes and from reworked Mississippian–Lower Pennsylvanian clastic sediments (359–315 Ma) of the Illinois Basin (Kissock et al., Reference Kissock, Finzel, Malone and Craddock2018; Thomas et al., Reference Thomas, Gehrels, Sundell, Greb, Finzel and Clark2020). Detrital Ilt-Sme was not thought to be a significant component in these mica-rich Palaeozoic clastic sediments shed from the mid- and older Palaeozoic sources from the eastern Appalachian orogen (Aronson & Lewis, Reference Aronson and Lewis1994). Diagenetic Ilt-Sme can be connected to the estimated maximum burial of the Illinois Basin (∼270–160 Ma; Rowan et al., Reference Rowan, Goldhaber and Hatch2002). Diagenetic illite can be connected to prolonged fluid mineral interactions at low temperatures (<150°C) with either meteoric/porewater or hydrothermal brine (McIntosh et al., Reference McIntosh, Tabor and Montañez2025). The observed stacking orders also correlate with available thermal histories and burial (McIntosh et al., Reference McIntosh, Elliott, Wampler and Tabor2023). It is probable that diagenetic Ilt-Sme is more abundant in these palaeosols based primarily on K–Ar data. The HAM-18 palaeosol is more likely to have detrital Ilt-Sme and white mica due to its location in the Illinois Basin and the weak development of the palaeosol (McIntosh et al., Reference McIntosh, Elliott, Wampler and Tabor2023). The above-described conceptual framework allows discussion of how the mineralogical and geochronological measurements of the TPB-treated palaeosol clay fractions consistently support the hypothesis tested in the present study and thus correspond to expectations based on the foundational work of A.D. Scott’s research group.
Key findings from the XRD traces of the TPB-treated clay fractions
The XRD results provide information on how the Ilt-Sme and mica responded to exchange with the NaCl–NaBPh4 solution in the various palaeosol size fractions. Understanding these responses, particularly the Ilt-Sme response, to that exchange is made more complicated, however, by the variability in the original interlayer composition of the untreated clay fractions. This understanding is then further complicated by the unknown degree to which NH4+ replaced Na+ in the expanded layers of the treated clay fractions, both original and newly formed, during quenching, thereby producing an artificial interstratified complex. Interstratified material was dominant in the XRD traces of all of the treated fractions, regardless of the treatment duration, but the positions, shapes and intensities of the main peaks for such materials were different from those of the original Ilt-Sme, indicating substantial changes in the character of the interstratified materials. The present XRD results are not amenable, however, to establishing the detailed character of those changes.
Changes in the intensity of the 10.0 Å mica peaks upon TPB treatment indicate removal of much of the K from muscovite (the 10.0 Å minerals recognized in these palaeosols by McIntosh et al., Reference McIntosh, Elliott, Wampler and Tabor2023) in the longer treatments, but the effect of 1 h of treatment cannot be discerned from the XRD traces because the background on which the 10 Å peaks are superimposed is quite different for the treated and untreated clay portions. Distinct 10.0 Å shoulders evident after 1 h of treatment of the coarser fractions of HAM-18 and CHA-83, however, suggest that a substantial fraction of the original mica remained. These shoulders were still present but much reduced in intensity upon 1 day of treatment; such changes in other XRD traces may exist but are not clearly discernible. This and the absence of any 10.0 Å reflection after the longer 22 day treatments indicate that most of the K in the <5 µm palaeosol muscovite can be extracted via the TPB method in 1 day, but somewhat more time is needed to complete the process.
Disappearance of the 10.0 Å reflections does not mean that all K and Ar had been removed from the mica. Scott (Reference Scott1968, figs 2, 6 & 7) showed that artificially comminuted (dry ground) 0.7–2.0 µm muscovite retained ∼10% of its K after TPB treatment sufficiently long to eliminate its 10.0 Å peaks; a finer fraction (0.2–0.7 µm) held even more (∼30%) of its K against extended TPB treatment. Although detrital clay-sized muscovite, which is naturally formed, may respond differently to such treatment than the artificially dry-ground muscovite did, the absence of 10.0 Å reflections is at present not sufficient to establish that all K was removed from muscovite in the studied clay fractions. Pattern G in Scott’s (Reference Scott1968) fig. 6 has no 10.0 Å peak but indicates some residual 10.0 Å layers interstratified with much more abundant 12.3 Å layers in the product of extended TPB treatment of the 0.7–2.0 µm muscovite.
According to the mechanisms considered by Scott (Reference Scott1968), the absence of the 10.0 Å reflections after the longest treatment of the present study could mean that the muscovite originally present suffered either virtually complete extraction of interlayer K by progressive ‘edge weathering’ or extraction of all of the K from enough individual interlayers (‘layer weathering’) of such minerals to convert the original mineral into an artificial interstratified phyllosilicate entity. For muscovite artificially comminuted by dry grinding and treated via the TPB method, layer weathering appeared to contribute substantially to K extraction from clay-sized grains but not to 10–20 µm grains (Scott, Reference Scott1968, fig. 6). This ‘small particle effect’, in which fine particles lose K by layer weathering of some but not all interlayers, was more strongly evident among the chemically dispersed natural particles of Grundite. The finest of that material (<0.08 µm) held just over half of its K against long TPB treatment.
K contents and K–Ar age values after TPB treatment
Observed changes in K content and K–Ar age value upon TPB treatment cannot directly show how much interlayer K or radiogenic Ar were extracted from any component. But the patterns of change, complemented by the evidence already presented that interlayer K is extracted very quickly from Ilt-Sme (e.g. K-bentonite; Fig. 2) but more slowly from mica, provide a clear picture of nearly complete removal of the K and Ar from detrital mica and retention of a substantial proportion of those elements by diagenetic Ilt-Sme. Six patterns of change in K–Ar age value and/or K content are notable:
(1) Distinct increases in K–Ar age value for some fractions of LSC-16N and CHA-83 upon 1 h of treatment indicate rapid K and Ar extraction from a considerable portion of the diagenetic Ilt-Sme, while relatively little K and Ar were extracted from the older detrital components.
(2) The K–Ar age value decreased (relative to that of the untreated clay) in all cases upon treatment for 22 days (27 days for the fine clay fractions). These changes in age value indicate greater loss of K and Ar from the detrital materials of each palaeosol than from its diagenetic materials. Among the medium and coarse clays, the decreases were distinctly greater for HAM-18 and CHA-83 than for LSC-16N, and for all palaeosols the change increased with increasing grain size. Given the evidence discussed earlier that the proportion of detrital muscovite increases with increasing grain size and increases generally from LSC-16N to CHA-83 to HAM-18, these changes confirm that K and Ar were removed more completely, although more slowly, from detrital muscovite than from other K-bearing minerals. It is noteworthy that Sears et al. (Reference Sears, Hesse, Vali, Elliott, Aronson and Martin1998) and Chaudhuri et al. (Reference Chaudhuri, Środoń and Clauer1999) both reported indications that intercalation of n-octadecylammonium ions was effective for extracting K and Ar from detrital mica.
(3) Large differences in age value (50 Myr or more) are evident between the 1 day-treated and 22 day-treated clays of three fractions: HAM-18 1–2 µm and 2–5 µm and CHA-83 2–5 µm. The K and Ar extracted after the first day should have come mostly from a relatively coarse component, and the large decreases in age values suggest a detrital component substantially older than 300 Ma, which fits well with prior evidence of substantial amounts of detrital muscovite in these palaeosols. The mass changes of the clay fractions with the 22 day treatment were probably approximately the same as those observed with the 1 day treatment (Supplementary File 2). If so, the rather large amounts of K and radiogenic Ar extracted from these three clay fractions after 1 day can be calculated to a good approximation. Such calculations indicate the extraction of K and Ar from material with K–Ar age values from ∼380 to 460 Myr (Supplementary File 1), which fits well with the current understanding of the sources of detrital muscovite in the palaeosols.
(4) After loss of a large fraction of its K during 1 h of reaction, the age value of the HAM-18 fine fraction was not any larger than that of the untreated material. Unlike the fine clay of CHA-83, for which an age value increase indicated extraction of K and Ar mainly from diagenetic Ilt-Sme in the first hour, at least half of the K and Ar extracted from the HAM-18 fine clay in the first hour must have come from material older than 300 Ma. The large decrease in K content of this fraction in 1 h (which far outweighed the possible effect of clay mass change) indicates an older material dominated by something finer and more abundant than detrital muscovite in the <1 µm clay. With the complementary XRD evidence that Ilt-Sme is predominant in the HAM-18 <1 µm clay, the dominant material is inferred to be detrital or possibly pedogenic Ilt-Sme.
(5) The much smaller K contents of the HAM-18 and CHA-83 fine clay fractions show that a large proportion of the K was extracted quickly from those fractions after 1 h of treatment despite the uncertainty in estimating the amount of K extracted from differences in K content due to clay mass change. The radiogenic Ar content of these fractions also decreased by large amounts. The large and rapid release of a large proportion of the original K was observed from the illite-rich ordered Ilt-Sme and also from the K-bentonite clay. By contrast, no such relatively large, rapid loss of K is evident for the LSC-16N fine clay, which is predominantly R0 Ilt-Sme. The contrast in relative amounts of K extracted by TPB treatment from the low-K, R0 Ilt-Sme of LSC-16N and the several higher-K, ordered Ilt-Sme – a contrast supported by a similar contrast in relative amounts of radiogenic Ar extracted – is a strong indication that Scott (Reference Scott1968) was right to infer that K is extracted from fine illite by ‘layer weathering’ and that when ‘some of the layers are separated … the bonding in the adjacent layers that still contain K becomes stronger, as suggested by Bassett (Reference Bassett1959)’. Lastly, the considerable loss of K counters arguments that K is being resorbed into the muscovite or Ilt-Sme lattices after exchange.
(6) With increasing treatment duration, the age values of all clay fractions converge towards 247 Myr, which is the mean of the tightly grouped age values of all clay fractions treated for nearly 1 month. A parsimonious explanation of this observation would be that the long treatments took virtually all of the interlayer K and Ar from detrital components and that the mean age of the diagenetic component is ∼247 Ma for each palaeosol, thus supporting the tested hypothesis. Such an explanation must be considered an approximation, however, for reasons addressed in the following subsection.
Implications for improved dating of diagenetic illite
The results of the present study confirm the tested hypothesis for clastic sediments of the Illinois Basin that became soils and then palaeosols. Nearly all of the interlayer K and radiogenic Ar of detrital material in the clay fractions of these rocks were extracted via the TPB method while some of the K and associated radiogenic Ar of the diagenetic Ilt-Sme were left. Evidence for the virtually complete extraction of K and Ar from the muscovite is that the extended TPB treatments brought the coarser clay fractions of HAM-18, which by all indications were relatively rich in detrital muscovite, to age values far smaller than that of the muscovite and nearly the same as those of similarly treated palaeosol clay fractions that originally had much less muscovite. This result was favoured by the inability of muscovite to retain much K unless it is finer than 1 µm (Scott, Reference Scott1968) and by the typically steep decrease in detrital muscovite abundance with decreasing particle size. It is not evidence, however, of the virtually complete extraction of these elements from all detrital material.
Scott (Reference Scott1968) showed clearly that some interlayer K of both sub-micrometre muscovite and fine Ilt-Sme resist extraction via the TPB method. Detrital Ilt-Sme, similarly to Scott’s fine Ilt-Sme and the diagenetic Ilt-Sme of the K-bentonite, presumably holds some of its K against extended TPB treatment. The age value of the HAM-18 fine clay after extended treatment, at 257 ± 5 Myr, indicates that extraction of K and Ar from the detrital material in it was less complete than in all other clay fractions treated for 22 or 27 days. This is consistent with the evidence that the HAM-18 palaeosol, but not the two others, has substantial detrital Ilt-Sme, that its fine clay is richer in Ilt-Sme than its coarser clay fractions and that Ilt-Sme holds a small but substantial fraction of its K against such treatment. As mentioned, HAM-18 is liable to have more detrital Ilt-Sme given its position in the Illinois Basin and weaker palaeosol development.
Two other aspects of the palaeosols studied indicate that the parsimonious explanation (given earlier) of the convergence of age value towards 247 Myr for all three palaeosols’ clay fractions is inaccurate. First, it is not evident that diagenetic illite from locations hundreds of kilometres apart across Illinois that experienced different maximum burial depths (∼1 km in the north to ∼3 km in the south) should have the same mean age. McIntosh et al. (Reference McIntosh, Elliott, Wampler and Tabor2023) suggested that the mean age of the LSC-16N diagenetic Ilt-Sme is less than that from the two other palaeosols due to slower but more protracted diagenesis. A protracted diagenesis is supported by recent clumped isotope geothermometry from pedogenic carbonates of the same palaeosols of the LSC core, whose elevated temperatures (mean = 45°C) suggest bond reordering during sustained burial diagenesis (McIntosh et al., Reference McIntosh, Tabor and Montañez2025). Second, feldspar is not affected by the treatment. Substantial XRD peaks at 3.19 Å indicated the presence of more than trace feldspar contents in the coarser clay fractions of LSC-16N but not in its fine clay. The age value of that fine clay, at 237 ± 8 Myr, indicates that the mean age of the LSC-16N diagenetic Ilt-Sme is less than 247 Ma. The presence of detrital feldspar in the two coarser fractions could account for their slightly larger age values.
For dating diagenetic Ilt-Sme, the finding that K and Ar were not disproportionately extracted from diagenetic Ilt-Sme by the TPB method is of equal importance to the finding that nearly all K and radiogenic Ar can be extracted from clay-sized detrital muscovite of palaeosols by the method. The former is based entirely on the results obtained from the treated K-bentonite diagenetic clay fraction. That decrease in the K–Ar age for the treated bentonite clay fraction was significant for only the two longer treatments. One possible reason for the small decrease includes the presence of a very fine later generation of diagenetic Ilt-Sme that is younger or apparently younger than the coarser Ilt-Sme.
Palaeosols are especially favourable for the nearly complete extraction of K from detrital components because feldspars are more susceptible to chemical weathering than dioctahedral mica. To confidently find the K–Ar age of diagenetic illite in feldspar-rich mudstones and fault gouges, a way to chemically extract all K from phyllosilicates at a temperature low enough to preserve the radiogenic argon in clay-sized feldspar is needed but is still unexplored. In sum, TPB shows promise for the removal of muscovite from mica–diagenetic illite complex mixtures, yielding diagenetic Ilt-Sme for K–Ar geochronologic methods. Understanding the impact of TPB exchange on both detrital Ilt-Sme and diagenetic Ilt-Sme would enable further delineation of which phase is releasing K and radiogenic Ar during TPB exchange, as well as whether the remaining 2:1 layers were held by stronger bonds.
Conclusions
Illite-rich, diagenetic Ilt-Sme loses more than half of its K in 1 h of TPB treatment, but a substantial fraction of the original K is not susceptible to TPB treatment, remaining in the solid phase after extended (1 month) treatment. To a good first approximation, the K–Ar age value of the K-depleted Ilt-Sme is the same as it was for the untreated material, indicating extraction of the associated radiogenic Ar, but no other, with the extracted K. This conclusion, based on results from a monomineralic, illite-rich, diagenetic Ilt-Sme from a K-bentonite, was essential for the interpretation of the observed changes in K contents and K–Ar age values of palaeosol clay fractions during TPB treatments of varying duration.
Patterns of change in K contents and K–Ar age values upon TPB treatment of palaeosol clay fractions of various sizes confirm that interlayer K is extracted quickly but not completely from fine illitic material, while also showing that K is extracted more slowly but more completely from coarser clay-sized muscovite. TPB treatment for 3 weeks is sufficient to remove nearly all of the K and radiogenic Ar from <5 µm detrital muscovite. Changes in K and radiogenic Ar contents after 1 day of treatment can be used to calculate an approximate age value for detrital muscovite if it was originally abundant, because Ilt-Sme’s contributions to change are minimal after 1 day.
Because of the rapid response of fine illitic material to TPB treatment, the change in K–Ar age value of fine clay upon 1 h of TPB treatment reveals information regarding the sources of such material in mudstones. The age value will be substantially larger after 1 h of treatment if the fine material was predominantly diagenetic but not if it was predominantly detrital.
The proportions of K and radiogenic Ar extracted by TPB treatment are greater for ordered, illite-rich Ilt-Sme than for R0 Ilt-Sme with low illite content. The available data show that as much as three-quarters of the K and Ar can be extracted from illite-rich R1 Ilt-Sme and indicate relatively much smaller losses from R0 Ilt-Sme, but the latter could not be established quantitatively because of changes in the mass of the clay fractions upon TPB treatment.
The nearly complete removal of K and Ar from detrital muscovite, and of at least a large part of the K and Ar from a detrital illitic component of one of the palaeosol samples studied, caused the K–Ar age values of all clay fractions treated for several weeks to converge into a narrow range (mean value 247 Myr), consistent with residual K and Ar remaining predominantly in contracted interlayers that originally were in diagenetic Ilt-Sme. Further research is needed on the duration as well as on more bentonite clay fractions to establish how well the proportions of K and radiogenic Ar represent the mean age of diagenetic Ilt-Sme. These results will enable further understanding of this exchange on clay fractions that have both detrital and diagenetic illite, as is seen in the palaeosols in this study.
Supplementary material
The supplementary materials for this article may be found at https://doi.org/10.1180/clm.2026.10042.
Acknowledgements
We thank the Illinois State Geological Survey, specifically Scott Elrick and John Nelson, for their insights into Illinois Basin stratigraphy and access to drill cores. We thank Nicholas Rosenau for access to samples from the LSC core. We thank Jesús Solé, Arkadiusz Derkowski and Tyler Kane for constructive comments that improved this manuscript. John Tougas prepared the figures for publication. Dr J. Marion Wampler, now deceased, suggested using NaTPB to measure the ages of diagenetic illite in complex mixtures with muscovite via K–Ar geochronology. We will miss his insights and creative thinking.
Financial support
The Department of Chemistry provided a stipend to Arya Shahbazi-Asl.
Competing interests
The authors declare none.
Disclaimer
Any use of trade, firm or product names is for descriptive purposes only and does not imply endorsement by the US Government.
Open access
