Walnut Shells Turned Into Citric Acid Carbons That Strip Iron From Water

Every year, mining operations, metallurgical plants and farms discharge metal-laden effluents into rivers and groundwater, and unlike many organic pollutants, dissolved metals never biodegrade. Iron, though essential to life, becomes a nuisance when it exceeds recommended limits in drinking water, staining fixtures, fouling pipelines, encouraging bacterial growth and ruining taste. Now, a team of researchers in the Republic of Moldova has shown that one of the country’s most abundant agricultural wastes, walnut shells, can be transformed into chemically tailored carbon adsorbents capable of capturing dissolved iron, and that the secret lies in a delicate trade-off between heat and chemistry.

The study, published in the journal Results in Engineering, was carried out by Irina Ceban, Crina Vicol and Maria Mihet, who prepared two activated carbons from walnut shells using citric acid as a mild, biodegradable modifying agent. The shells were washed, dried at 150 degrees Celsius for 24 hours, crushed and sieved, then impregnated with a 50 percent citric acid solution before thermal treatment. The researchers deliberately chose two very different processing routes: one sample, designated AC450-A, was heated to 450 degrees Celsius in air, while the second, AC600-N, was heated to 600 degrees Celsius under a flow of nitrogen. After treatment, both materials were washed to constant filtrate pH and dried, yielding two carbons with strikingly different personalities.

The central question was deceptively simple: does it pay more to preserve the oxygen-rich functional groups that citric acid installs on a carbon surface, or to build a more carbonized, thermally stable structure? Citric acid is prized in adsorbent design because it grafts carboxylic, hydroxyl and carbonyl functionalities onto carbon, creating sites that can bind metal ions through ion exchange, electrostatic attraction and surface complexation. But heat works against this chemistry. As treatment temperature rises, the carbon framework condenses and aromatizes, becoming more stable, while thermally labile oxygen groups are progressively burned off or transformed. The Moldovan team set out to quantify exactly where that balance falls.

Characterization told the story in fine detail. Elemental analysis showed carbon content climbing from 81.37 percent in AC450-A to 84.90 percent in AC600-N, while hydrogen fell from 3.92 to 2.93 percent, a signature of deeper carbonization and structural condensation. Sulphur was undetectable in both samples, confirming the purity of the walnut shell precursor. Fourier-transform infrared spectroscopy revealed the chemical consequences: the lower-temperature sample retained stronger O-H, C=O and C-O bands, indicating abundant oxygen-containing surface groups, whereas the 600-degree sample showed enhanced aromatic C=C features and diminished oxygen signals, consistent with a more condensed, carbon-rich framework.

Perhaps the most surprising finding came from nitrogen adsorption measurements. Both carbons exhibited extremely low BET surface areas, just 0.35 square metres per gram for AC450-A and 0.98 square metres per gram for AC600-N, three orders of magnitude below conventional activated carbons, which typically range from several hundred to more than 1000 square metres per gram. The isotherms were Type II-III, indicating essentially nonporous or macroporous materials with negligible microporosity. Yet both samples removed iron efficiently. This apparent paradox, a carbon with almost no measurable porosity that still captures metal ions, points squarely at surface chemistry rather than physical surface area as the dominant retention mechanism, echoing classic work by Boehm and others showing that oxygen functionalities govern metal adsorption on oxidized carbons.

The point of zero charge reinforced this picture. Using potentiometric titration under argon with sodium chloride electrolytes at three ionic strengths, the team determined pHpzc values of 3.58 for AC450-A and 2.94 for AC600-N. Below these thresholds the surfaces are protonated and positively charged; above them, carboxylic and phenolic groups deprotonate, generating negatively charged sites that attract cationic iron species. Suspension pH measurements, both below 4.0, confirmed the strongly acidic character imparted by the citric acid treatment, and scanning electron microscopy showed AC450-A as a rough, cracked, defect-rich surface while AC600-N appeared denser and more homogeneous, with EDX detecting calcium in the latter, likely concentrated from the natural mineral fraction of the shells after high-temperature carbonization.

Batch adsorption experiments across the pH 2 to 10 range revealed strongly pH-dependent behaviour. At pH 2, apparent removal was modest, roughly 5.2 milligrams per gram for AC450-A and 6.3 for AC600-N, because protonated surfaces and a flood of competing hydrogen ions shut down binding sites. As pH rose above the points of zero charge, deprotonation opened the door to electrostatic attraction and complexation. The highest apparent removal, 23 to 26 milligrams per gram for AC450-A and 27 to 31 for AC600-N, occurred at pH 9 to 10. Crucially, the authors caution that these high-pH values cannot be read as true adsorption capacities: at alkaline pH, ferrous iron hydrolyzes, and in the presence of dissolved oxygen it oxidizes to Fe(III), which precipitates as sparingly soluble hydroxides. Because no adsorbent-free precipitation blanks were run, adsorption, oxidation, hydrolysis and precipitation could not be quantitatively separated, and the results are framed honestly as overall iron removal.

Kinetic modelling added an empirical layer. The pseudo-second-order model of Ho and McKay clearly outperformed the Lagergren pseudo-first-order equation, with coefficients of determination reaching 0.997 for AC450-A and 0.999 for AC600-N, and calculated equilibrium capacities closely matching experimental values at most concentrations. The team, however, resists over-interpretation: a good fit to a pseudo-second-order equation is not proof of chemisorption, and the kinetic data were interpreted alongside the spectroscopic and surface-charge evidence rather than in isolation. Equilibrium data were fitted with four isotherm models, and the Langmuir equation won decisively, with R-squared values up to 0.998 for AC450-A and 0.999 for AC600-N. Model-derived monolayer capacities rose with temperature, from 4.47 to 6.76 milligrams per gram for AC450-A and from 4.62 to 9.27 for AC600-N between 20 and 40 degrees Celsius, and separation factors between zero and one confirmed favourable uptake across the entire concentration range.

Thermodynamics painted a nuanced picture. Gibbs free energies were negative at all three temperatures for both carbons, confirming spontaneous retention, but the van’t Hoff analysis yielded an apparent endothermic enthalpy of about +32.6 kilojoules per mole for AC450-A against a strongly exothermic -95.9 kilojoules per mole for AC600-N, with entropy changes of +207.9 and -217.2 joules per mole-kelvin respectively. With only three temperatures and limited linearity, the authors treat these numbers cautiously, refusing to assign a mechanism from them alone. The same discipline applies to the Dubinin-Radushkevich adsorption energies, which were explicitly not used to distinguish physical from chemical binding, a refreshing departure from common practice in the adsorption literature.

The broader significance of the work lies in its integrated approach and its honesty. Rather than chasing a headline capacity number, the study demonstrates that iron removal by these walnut-shell carbons is governed by a pH-dependent interplay of surface chemistry, iron speciation and thermal history, and that a carbon with almost no BET surface area can still perform meaningfully if its surface is rich in the right oxygen functionalities. The trade-off is clear: AC450-A preserves the carboxyl-rich chemistry that favours metal binding, while AC600-N gains carbonization, thermal stability and a modestly larger surface. For regions like Moldova, where walnut cultivation generates vast shell residues, the work offers a template for converting agricultural waste into water-treatment materials using a benign, food-grade acid. The authors note that model single-solute solutions were used, and that regeneration, competitive adsorption and performance in real waters remain open questions, but the foundation is laid: sometimes the smartest adsorbent is not the most porous one, but the one whose surface chemistry is preserved with care.

Subject of Research: Citric acid-modified activated carbons from walnut shells for Fe(II) adsorption from aqueous solutions

Article Title: Preparation and characterisation of citric acid-modified activated carbons for the adsorption of Fe(II) ions from aqueous solutions

Article References: Ceban, I., Vicol, C., & Mihet, M. (2026). Preparation and characterisation of citric acid-modified activated carbons for the adsorption of Fe(II) ions from aqueous solutions. Results in Engineering, 32, Article 113432. https://doi.org/10.1016/j.rineng.2026.113432

Image Credits: AI Generated

DOI: 10.1016/j.rineng.2026.113432

Keywords: activated carbon, walnut shells, citric acid modification, Fe(II) adsorption, water treatment, heavy metals, biomass valorisation, point of zero charge, Langmuir isotherm, adsorption kinetics, surface functional groups, Republic of Moldova