The research team, led by Liuhua Yang and Haikun Liu, worked with cemented paste backfill, a material that mines routinely pump underground as a mixture of tailings and Portland cement. Their tailings, sourced from the Nanchang Copper Mine in Jiangxi Province, China, were rich in reactive calcium — containing 24.05 percent calcium oxide alongside abundant silica, iron and aluminum oxides. Rather than testing a handful of discrete conditions, the researchers designed a continuous water-content gradient spanning five groups, from completely dry samples held at 55 percent relative humidity, through liquid-to-solid ratios of 0.4 and 1, all the way to fully suspended slurries at ratios of 3 and 20. Every sample was carbonated under identical conditions: pure carbon dioxide at 200 milliliters per minute, ambient temperature and atmospheric pressure, sampled at 6, 12 and 24 hours.

What emerged from this systematic sweep is a story of two fundamentally different crystallization regimes. In the water-starved samples, derivative thermogravimetry revealed bimodal decomposition profiles and broadened shoulder peaks — thermal fingerprints of amorphous calcium carbonate and poorly crystalline vaterite, the metastable precursors that normally transform into stable calcite but here remained kinetically frozen. A persistent portlandite peak at around 460 degrees Celsius showed that without free water, calcium transport channels were severed, confining the reaction to the outermost particle layer. The nanometric water films that remained simply could not carry ions fast enough: diffusion coefficients under such conditions fall two to three orders of magnitude below those in bulk solution.

As water availability increased, the thermal signatures transformed. The low-temperature shoulder vanished, the main decomposition peak narrowed and sharpened, and by the highest liquid-to-solid ratio the entire carbonate signal coalesced into a single intense peak near 800 degrees Celsius — the hallmark of highly crystalline calcite. Intriguingly, the total amount of carbonate barely changed across the entire gradient: after 24 hours, apparent calcium carbonate content varied only from 30.4 to 32.4 percent by weight, a relative difference of roughly 6.6 percent. Water, it turns out, is not primarily a lever on how much carbon gets captured, but on what form that carbon takes and how cleanly it forms.

X-ray diffraction and infrared spectroscopy told the same story from different angles. In dry samples, the principal calcite reflection at 29.4 degrees remained weak and broad even after a full day of reaction, while vaterite peaks lingered — evidence of a crystallization pathway stuck in first gear. In the wettest samples, the calcite peak became the dominant feature and vaterite disappeared entirely. Infrared spectra traced the carbonate vibrational modes sharpening and intensifying with water content, while the silicate framework signals shifted in parallel, revealing depolymerization and repolymerization of the aluminosilicate matrix as water enabled dissolved species to migrate and reorganize.

Perhaps the most visually striking evidence came from electron microscopy. Under water-deficient conditions, tailings particle surfaces became blanketed in a dense, gel-like encapsulation — flocculent, edgeless material formed when explosive local supersaturation triggers a storm of tiny nuclei that can neither grow nor ripen. This coating acts as a physical barrier, blocking carbon dioxide from penetrating and calcium from escaping, which explains why carbonation efficiency in the driest group stagnated almost immediately. At intermediate moisture, the surfaces sprouted spheroidal, cauliflower-like aggregates with rounded edges — crystals beginning to emerge but still crowded and intergrown. Only in the water-rich slurries did the researchers observe the prize: abundant rhombohedral calcite crystals with smooth faces, sharp edges and well-defined cleavage planes, grown to micron scale through solution-mediated dissolution and reprecipitation.

The elemental maps added a crucial twist about impurities. In dry samples, calcium signals overlapped intimately with silicon, aluminum, magnesium and iron, indicating that carbonate precipitated entangled with the unreacted matrix while magnesium ions became trapped at lattice growth sites — a form of crystal poisoning that distorts calcite morphology. In water-rich samples, the maps showed dramatic elemental segregation: discrete calcium-rich zones with impurity signals at background levels. Real-time solution chemistry explained why. Magnesium concentrations in the slurry rose steadily as the reaction proceeded, showing that magnesium was eluted into the liquid phase rather than incorporated into the solid. The researchers attribute this to hydration energetics — magnesium’s hydration free energy of roughly minus 1,921 kilojoules per mole substantially exceeds calcium’s minus 1,577, making dehydration and lattice incorporation less favorable. Alkali metal behavior added a further nuance: at the moderate liquid-to-solid ratio of 3, higher ionic strength compressed the electrical double layer around growing nuclei, encouraging impurity adsorption, whereas the dilute conditions at ratio 20 fostered a cleaner crystallization environment.

Solid-state nuclear magnetic resonance revealed that the silicate skeleton itself undergoes a water-driven reconstruction. Dry samples retained low-polymerization silicate dominated by Q1 and Q2 units — dimers and short chains left stranded after calcium leaching. As water content rose, the spectra shifted upfield, and the wettest samples showed a broad resonance spanning minus 95 to minus 115 parts per million, deconvolving into Q3 and Q4 sites characteristic of layered and three-dimensional framework structures. In other words, abundant water allowed detached silicate fragments to repolymerize into a porous, highly connected amorphous silica network — one that, unlike the dense passivation films of dry carbonation, appears to serve as an ion-transport conduit rather than a barrier, consistent with the complete consumption of portlandite observed in the wettest group.

The pH trajectory of the aqueous experiments delineated three kinetic stages: an initial alkaline plateau near 12.5 to 13 buffered by portlandite and calcium silicate hydrates, a rapid decline as those phases were consumed, and a final stabilization near 6.0 to 6.5 during which carbonate precipitation and calcite growth continued. Calcium concentrations traced a textbook dissolution-precipitation cycle — surging as minerals dissolved, plummeting during massive carbonate precipitation, and settling at equilibrium with calcite’s solubility product. Together these measurements sketch a self-purifying mechanism in which water physically separates the desired product from the contaminants that would otherwise sabotage it.

The practical implications are considerable but carefully caveated. The authors identify three water regimes: an unfavorable deficient domain up to a liquid-to-solid ratio of 0.4, a transitional domain at ratio 1, and a favorable water-sufficient domain at ratios of 3 and above — while stressing that these thresholds are composition-specific rather than universal. More water is not automatically better: carbon dioxide diffuses through liquid roughly four orders of magnitude more slowly than through gas, so excessive slurry depth could throttle the gas supply to reactive surfaces, and high water volumes carry real costs in reactor size, agitation energy, dewatering and wastewater treatment — particularly since the self-purification mechanism transfers magnesium and other dissolved ions into the process liquor. The engineering optimum, the authors suggest, lies near the minimum water content that can trigger dissolution-reprecipitation and impurity exclusion. Within that window, however, the message is transformative: a variable as simple as moisture can be tuned to control the size, purity and morphology of carbonate products, turning one of mining’s largest waste streams into a precision instrument for permanent carbon removal.

Subject of Research: Water-content-dependent mineral carbonation of cemented mine tailings for CO2 sequestration

Article Title: Water-content-dependent carbonate formation in cemented paste backfill: Phase assemblage, morphology, and ion partitioning

Article References: Yang, L., Liu, H., Li, T., Jiao, H., Niu, H., & Chen, S. (2026). Water-content-dependent carbonate formation in cemented paste backfill: Phase assemblage, morphology, and ion partitioning. Results in Engineering, 32, Article 113319. https://doi.org/10.1016/j.rineng.2026.113319

Image Credits: AI Generated

DOI: 10.1016/j.rineng.2026.113319

Keywords: mineral carbonation, CO2 sequestration, mine tailings, calcite, cemented paste backfill, water content, crystallization, passivation layer, magnesium inhibition, silicate polymerization, carbon capture, Ostwald ripening