The research team, working from outcrops near the village of Betzenstein, around five kilometers northwest of the site close to Obertrubach, combined traditional fieldwork with a battery of modern analytical techniques. They collected fifty-six rock samples from four outcrops at different elevations, including the twenty-five-meter-high Römerbrücke outcrop, which reveals a complete vertical story from nodular basinal carbonates at the bottom through marl-limestone alternations to a pervasively dolomitized reefal massive facies at the top. Thin sections stained with Alizarin-red-S allowed researchers to distinguish dolomite from calcite under the microscope, while X-ray diffraction with Rietveld refinement pinned down the precise crystal chemistry of the dolomite, and mass spectrometry delivered carbon and oxygen isotope ratios and trace element concentrations, including rare earth elements and yttrium measured by laser ablation ICP-MS.

The most striking pattern to emerge was vertical. Dolomite abundance rises steadily upward through the succession. The lowermost units, nodular carbonates and carbonate-marl alternations deposited in deeper basinal waters, contain only about six percent dolomite on average. The overlying bedded carbonates average a mere four percent. But the massive facies at the top, interpreted as sponge-algal reef complexes, reaches a mean dolomite content of ninety-one percent. Even within the partially dolomitized intervals, marl layers consistently hold more dolomite, around twenty-one to twenty-three percent, than the limestones above and below them. This stratigraphic fingerprint, the authors argue, is a direct clue to both the direction of fluid flow and the timing of the transformation.

Microscopy sharpened the picture further. In marls, dolomite crystals cluster in the sediment immediately surrounding lithified ammonite shells and inside intraparticle pores. In limestones and transitional lithologies, dolomite preferentially invades the muddy matrix and open or sediment-filled voids, while leaving rigid siliceous sponges, microbial carbonates, and calcitic fossils such as brachiopods largely untouched. The team realized this selective distribution encodes a precise time window. Dolomitization must have occurred after sponges and microbial carbonates had acquired early rigidity and after the fine sediment inside ammonite shells had hardened, but before the surrounding matrix and marls had fully lithified. Since limestone lithification is known to happen very early during diagenesis, often within the uppermost meters of burial, this means the dolomitizing fluids arrived at very shallow burial depth, a conclusion that fundamentally reshapes how these rocks should be interpreted.

Within the dolostones themselves, the researchers distinguished two rock types and four dolomite varieties. The lower, clay-rich dolostone type contains finely crystalline dolomite with crystals sixteen to sixty-four micrometers across, alongside a medium crystalline variety of sixty-four to two hundred fifty micrometers. X-ray diffraction showed both are comparatively calcium-rich, corresponding to roughly fifty-two point six to fifty-five point six mole percent calcium carbonate, and less well ordered crystallographically. The upper, clay-poor dolostone type contains a different medium crystalline dolomite that is closer to ideal stoichiometry at fifty point zero to fifty-two point six mole percent calcium carbonate, with sharper diffraction peaks and higher crystallinity. Fluorescence microscopy revealed crisp growth zones in the lower dolomite types but only faint, irregular zoning in the upper variety, suggesting the upper dolomite had undergone more extensive recrystallization as fluids passed repeatedly through the shallower, more permeable strata.

Geochemistry delivered the decisive verdict on the fluids themselves. Carbon isotope values between two point zero and three point two permil match the range of Late Jurassic seawater, while oxygen isotope values, more negative than expected, were likely reset during later burial diagenesis. More diagnostic are the trace elements. Rare earth element patterns normalized to post-Archean Australian shale mirror seawater signatures, with yttrium-to-holmium ratios of forty-three to sixty-seven, well above the threshold of thirty-six that characterizes modern seawater. A pronounced negative cerium anomaly indicates the water column was well oxygenated. Sodium concentrations in all dolomite types fall within published ranges for near-normal seawater dolostones, with only a slight overlap suggesting mildly evaporated water, while strontium concentrations sit comfortably within near-normal seawater ranges and far below the elevated values typical of hypersaline brines.

Fluid inclusion salinities published for dolomites from adjacent regions reinforce this interpretation. Samples from the Bavarian Molasse Basin record salinities of about forty grams per liter, roughly four mass percent sodium chloride equivalent, and dolomites from the Swabian Alb range from about four point five to fifteen point five mass percent. These values overlap with estimates for Late Jurassic seawater, thought to have been between about thirty-seven and forty-two parts per thousand, and with slightly to moderately evaporated seawater categories. The picture is of seawater that had been modestly concentrated by evaporation in shallow, sun-exposed parts of the platform, but never became the true hypersaline brine associated with sabkha evaporite environments. Notably, no Upper Jurassic evaporites are found in the Franconian Alb, ruling out strong evaporation there.

Piecing these strands together, the authors propose an elegant hydrological model. Slightly concentrated seawater in shallower-water areas of the carbonate system descended under gravity and migrated basinward, replacing precursor carbonates through a coupled dissolution-reprecipitation mechanism at the moving replacement front. Because the flow was directed downward and outward from the platform interior, dolomitization intensifies upward in the succession toward the shallow massive facies and weakens basinward, which neatly explains why the bedded carbonates of Unit 3 remain almost untouched even where they overlap in elevation with dolomitized strata. As regional shallowing continued and the massive facies expanded through the Kimmeridgian, the zone of pervasive dolomitization advanced, producing the clay-rich dolostone below and the more recrystallized, clay-poor dolostone above.

The findings resolve a debate stretching back to 1891, when von Gümbel first proposed direct precipitation of dolomite from seawater, and they undercut the once-popular mixing-zone model, which invoked brackish groundwater but struggles against evidence that the entire Jurassic succession remained submarine. Burial-compaction models proposed for the neighboring Swabian Alb also falter, because supplying enough magnesium to build massive dolostones from compacting pore waters is notoriously difficult. By anchoring dolomitization to early diagenesis and near-normal seawater, the new study sidesteps both problems. It also places the Franconian Alb firmly within a global pattern: Upper Jurassic dolostones from the Gulf Coast to Saudi Arabia, Iran, Turkey, Lebanon, and Switzerland were commonly formed in shallow-water settings by evaporation-related fluids, with the Northwestern Tethys realm, spanning modern Western Europe to Turkey, favoring near-normal seawater with slightly elevated salinity under warm, semi-arid conditions that intensified during the Kimmeridgian.

Beyond settling an old argument, the work carries practical weight, because dolostones of this kind are important reservoir rocks for oil, gas, and groundwater, and understanding how seawater alone can convert limestone to dolomite at shallow burial refines predictions of where such reservoirs form. It also whispers something larger about the deep past. The dolomite problem exists partly because modern oceans seem reluctant to precipitate dolomite, yet here was an epicontinental sea at the northern margin of the Tethys Ocean doing exactly that, on a grand scale, using nothing more exotic than sun-warmed, gently evaporated seawater. In the limestones and dolostones of the Franconian Alb, the slow chemistry of a vanished Jurassic sea is still readable, one crystal at a time.

Subject of Research: Early diagenetic seawater dolomitization of Upper Jurassic carbonates in the Franconian Alb, southern Germany

Article Title: Near-normal seawater dolomitization of Upper Jurassic carbonates in the Franconian Alb, southern Germany

Article References: Hao, R., Mu, N., Jansen, D., She, M., Merkel, A., & Munnecke, A. (2026). Near-normal seawater dolomitization of Upper Jurassic carbonates in the Franconian Alb, southern Germany. Swiss Journal of Geosciences, 119, 111-131. https://doi.org/10.3897/sjg.119.193770

Image Credits: AI Generated

DOI: 10.3897/sjg.119.193770

Keywords: dolomite, dolostone, Franconian Alb, Upper Jurassic, Kimmeridgian, seawater dolomitization, early diagenesis, geochemistry, stable isotopes, trace elements, carbonate platform, Southern Germany

For decades, engineers and researchers have optimized the Pidgeon process by studying heat transfer, reaction kinetics, furnace configuration, and thermal field distribution. Yet nearly all of this work shares a hidden assumption: that calcined dolomite behaves as an ideal binary mixture of magnesium oxide and calcium oxide, produced by the complete decomposition of the mineral dolomite, CaMg(CO3)2. Under that ideal framework, the silicothermic reaction requires a one-to-one molar ratio between the two oxides, and the maximum magnesium producible is set by whichever oxide is the limiting reagent. The calcium silicate larnite, Ca2SiO4, is understood as the thermodynamic sink that forms during reduction itself, driving the equilibrium toward metallic magnesium release.

A new study published in Results in Engineering by Aldemir de Melo Sotero and colleagues at the Universidade Federal do Pará in Brazil challenges that assumption in a way that could reshape how magnesium producers evaluate their raw materials. The team investigated two natural dolostones from Brazil with sharply contrasting clay contents: the Alcântara Dolostone, drawn from a Cretaceous coastal formation in the state of Maranhão, which contains the hydrated magnesium-aluminum clay mineral palygorskite at levels of 2.7 and 11.1 weight percent in two variants, and a Silurian dolostone more than 400 million years old that is essentially pure dolomite at 99.8 weight percent. The pure sample served as a mineralogical reference, while the clay-bearing rocks exposed a yield-destroying chemistry that bulk analysis cannot see.

The problem begins with what palygorskite does when heated. This phyllosilicate loses adsorbed and channel water below about 300 degrees Celsius, dehydroxylates between roughly 400 and 500 degrees Celsius, and then undergoes structural collapse above approximately 800 degrees Celsius, releasing poorly ordered, chemically reactive silica. That temperature range overlaps almost exactly with the calcination stage of the Pidgeon process, in which dolomite decomposes into lime and periclase. The freshly released reactive silica does not wait for the reduction stage. It reacts immediately with free calcium oxide, and to a lesser extent magnesium oxide, forming calcium and magnesium silicate phases before any reducing agent is introduced. The same silicate chemistry that is productive and mechanistically necessary during reduction becomes a yield-limiting constraint when it occurs during calcination, because it permanently removes oxides from the pool available for reduction.

To quantify this sequestration, the researchers designed a full 2³ factorial experiment varying charge mass from 1 to 10 grams, calcination temperature from 900 to 1100 degrees Celsius, and residence time from 1 to 8 hours, with center-point replicates, for a total of 12 runs per sample. They characterized the raw materials and calcined products using wavelength-dispersive X-ray fluorescence, simultaneous differential scanning calorimetry and thermogravimetry, and powder X-ray diffraction coupled with Rietveld quantitative phase analysis. The Rietveld method, a full-pattern fitting approach that iteratively refines a calculated diffraction profile against the measured pattern, delivered weight fractions for every crystalline phase with refinement agreement indices showing weighted profile R-factors between 5.4 and 9.0 percent and goodness-of-fit values close to unity.

The phase assemblages revealed a far more complex silicate chemistry than previously reported for clay-bearing dolostones. In the moderately clay-rich sample, larnite formed under every experimental condition, reaching weight fractions up to 43.8 percent, while diopside, MgCaSi2O6, gehlenite, Ca2Al2SiO7, and a magnesium silicate with Mg3Si2O9 stoichiometry appeared at higher temperatures and longer residence times. The gehlenite confirmed that aluminum released from palygorskite structural collapse also partitions into silicate phases. The Mg3Si2O9 phase, which is thermodynamically unstable at calcination temperatures, was interpreted as a retrogressive product crystallizing during cooling, when freshly calcined periclase reacts with residual amorphous silica. In the pure Silurian reference sample, by contrast, the calcined products consisted exclusively of periclase and lime, with no silicate phases detected in any run, establishing a clean baseline.

The team then built a Rietveld-based stoichiometric model that converts crystalline phase weight fractions into molar quantities per 100 grams of calcined material, performs elemental balances to separate reducible from sequestered oxide fractions, and simulates the silicothermic reduction with commercial 75-grade ferrosilicon. The results were striking. In the pure dolostone, available calcium oxide and magnesium oxide appeared in nearly equimolar proportions, matching the ideal Pidgeon stoichiometry. In the low-clay sample, available calcium oxide ranged from 22.5 to 35.5 weight percent, and theoretical magnesium yield ranged from 0.401 to 0.626 moles per 100 grams. In the high-clay sample, available calcium oxide collapsed to as little as 3.0 weight percent under severe conditions, leaving more than 90 percent of the periclase without its calcium counterpart and driving theoretical magnesium yield down to 0.053 moles per 100 grams, nearly an order of magnitude below the pure reference. Bulk oxide chemistry would have overestimated recoverable magnesium by about 42 percent for the low-clay sample and by up to roughly 1200 percent for the clay-rich sample under the harshest conditions.

Statistical analysis of the factorial design added a second layer of insight. For the low-clay sample, residence time was the only significant factor for both magnesium yield and total silicate formation, with opposing effects that confirmed the mineralogical trade-off: longer heating simultaneously increases silicate formation and decreases magnesium availability. For the clay-rich sample, temperature became the dominant factor by a wide margin, with residence time second and a significant temperature-by-time interaction reflecting the self-limiting nature of silicate growth once available calcium oxide is exhausted. Composite desirability optimization identified low temperature, short residence time, and low charge mass as the optimal calcination window for both samples, but the operating window was far narrower for the clay-rich rock. Even at their respective optima, the clay-rich sample yielded about 12 percent less theoretical magnesium and generated about 24 percent more slag than the low-clay sample, a penalty attributable to raw material mineralogy that process adjustment can mitigate but never eliminate.

The practical implications reach directly into industrial operations. The authors argue that ferrosilicon dosage should be calculated on the basis of effectively available calcium oxide rather than total periclase content, since dosing against total magnesium oxide in a clay-bearing feedstock would produce a substantial and costly excess of reducing agent. They also recommend routine X-ray diffraction characterization of dolomite raw materials as an integral component of quality control, replacing reliance on bulk oxide chemistry alone, because two feedstocks with identical bulk compositions can differ dramatically in reducible magnesium depending on how much of their oxide inventory is already locked into silicates. The mechanistic framework extends beyond palygorskite to other phyllosilicate impurities such as kaolinite, smectites, chlorites, and illite, each of which releases reactive silica at calcination temperatures. The theoretical yields reported here represent a mineralogical ceiling on recoverable magnesium, and the authors note that experimental validation through representative silicothermic reduction trials is planned as a dedicated follow-up study. For an industry that produces nearly a million tonnes of magnesium a year under thin margins, the message is clear: the clay hidden in the rock can quietly consume the metal before the furnace ever gets its chance to make it.

Subject of Research: Quantitative assessment of clay-derived reactive silica sequestration of CaO and MgO during dolostone calcination and its impact on theoretical magnesium yield in the Pidgeon silicothermic reduction process

Article Title: Magnesium yield estimation based on Rietveld refinement of clay-bearing dolostone: implications for silicothermic reduction

Article References: Sotero, A. D. M., Oliveira, R. L. R., Brito, C. E. C., Albuquerque, A. R. L., Paz, S. P. A., & Angélica, R. S. (2026). Magnesium yield estimation based on Rietveld refinement of clay-bearing dolostone: implications for silicothermic reduction. Results in Engineering, 32, Article 113155. https://doi.org/10.1016/j.rineng.2026.113155

Image Credits: AI Generated

DOI: 10.1016/j.rineng.2026.113155

Keywords: magnesium, Pidgeon process, silicothermic reduction, dolostone, palygorskite, Rietveld refinement, calcination, larnite, reactive silica, X-ray diffraction, ferrosilicon, mineralogy