Visible-Light Catalyst Strips 98% of Sulfur from Real Fuel in One Hour

Sulfur in fuels is one of the quiet saboteurs of modern life. When thiophene, benzothiophene, dibenzothiophene and their alkylated relatives burn inside an engine, they are converted into sulfur oxides, the toxic gases behind acid rain, corroded infrastructure and contaminated water and vegetation. Removing these stubborn compounds from oil before combustion is therefore one of the petroleum industry’s most persistent challenges, and one of its most expensive. Now a research team at Arak University in Iran reports that a simple nanocomposite made of cadmium oxide and graphitic carbon nitride can strip nearly all sulfur from real gas condensate fuel at room temperature and atmospheric pressure, using nothing more exotic than visible light and hydrogen peroxide.

The study, published in the journal Results in Optics, stands out for a reason that matters far beyond the laboratory bench. Most photocatalytic desulfurization studies published to date have tested their catalysts on model fuels containing a single, well-behaved sulfur compound dissolved in a clean hydrocarbon matrix. The Iranian team instead ran their experiments on genuine gas condensate collected from the South Pars Gas Complex along the Persian Gulf coast, a fuel containing 340 milligrams of sulfur per liter and a complex mixture of refractory organosulfur species, aromatics and other interfering compounds. Against this unforgiving backdrop, the CdO/g-C3N4 nanocomposite achieved a total sulfur removal of 98.4 percent, dropping the sulfur content from 340 to just 5.4 milligrams per liter while eliminating mercaptans entirely.

The catalyst itself is built from two materials that are individually promising but flawed. Graphitic carbon nitride, a nitrogen-rich polymer made simply by heating melamine powder to 600 degrees Celsius, has become one of the most studied visible-light photocatalysts of the past decade thanks to its suitable band gap, chemical stability and low cost. Its weakness is that photogenerated electrons and holes recombine rapidly, wasting the absorbed light energy before it can drive a chemical reaction. Cadmium oxide, an n-type semiconductor with a narrow direct band gap in the range of 2.2 to 2.7 electronvolts, absorbs visible light efficiently and conducts charge well, but performs poorly on its own. When the two are combined, the researchers hypothesized, the staggered alignment of their energy bands should push electrons and holes in opposite directions across the interface, keeping them separated and active.

The synthesis was deliberately kept simple. The team used a co-precipitation route in which cadmium nitrate was precipitated with ammonia in the presence of melamine, and the dried mixture was then calcined at 500 degrees Celsius for six hours. The melamine decomposed into graphitic carbon nitride while the cadmium hydroxide converted to cadmium oxide, yielding roughly five grams of a nanocomposite with a theoretical cadmium oxide loading of about 25 percent by weight. Compared with hydrothermal or solvothermal methods, co-precipitation avoids high-pressure autoclaves and organic solvents, and it promotes the intimate interfacial contact between the two semiconductors that heterojunction chemistry demands. A range of standard characterization techniques confirmed the result. Fourier-transform infrared spectroscopy showed the heptazine breathing vibration of carbon nitride shifting slightly from 807 to 803 wavenumbers, evidence of electronic interaction between the components rather than a mere physical mixture. X-ray diffraction confirmed that both crystal structures survived intact, with an average crystallite size of 16 nanometers estimated from the Scherrer equation.

Electron microscopy revealed a plate-like morphology, with most particles in the 100 to 200 nanometer range and cadmium oxide nanoparticles distributed across the layered carbon nitride sheets. Elemental mapping showed a homogeneous distribution of carbon, nitrogen, oxygen and cadmium with no detectable impurities. Nitrogen adsorption measurements added another piece to the puzzle: the nanocomposite possessed a specific surface area of 108 square meters per gram and a pore volume of 0.0436 cubic centimeters per gram, both substantially higher than pristine carbon nitride, whose low surface area is one of its known handicaps. The mesoporous structure, with pores averaging about 4 nanometers, should ease the diffusion of bulky sulfur-containing molecules toward active sites. Optical measurements put the composite’s band gap at 2.48 electronvolts, corresponding to an absorption edge at 456 nanometers, comfortably within the visible spectrum.

The optimization experiments revealed how sensitive the process is to operating conditions. Raising the catalyst dosage from 0.015 to 0.025 grams per liter improved sulfur removal by more than half, but pushing further actually hurt performance, likely because particle aggregation and turbidity blocked light from penetrating the slurry. Hydrogen peroxide, the oxidant of choice, worked best at a two-to-one ratio with the fuel; beyond that, excess radicals simply recombined into water. Acetic acid played a chemically interesting dual role, providing an acidic medium while reacting with hydroxyl radicals to form peracetic acid, a far more aggressive oxidant against refractory sulfur compounds like dibenzothiophene. The final piece was acetonitrile, added at a one-to-one ratio with the fuel. The solvent does more than dissolve; it extracts the polar sulfoxides and sulfones produced by oxidation out of the fuel phase, driving the reaction forward. Under these optimized conditions, the nanocomposite reached 98.8 percent sulfur removal within 90 minutes, with a shorter 60-minute exposure achieving 98.4 percent.

The kinetics told a satisfying story. The process followed pseudo-first-order behavior with an apparent reaction order of 1.06 and a rate constant of 0.075 per minute, with a coefficient of determination of 0.9971 indicating an excellent fit. Compared with recently reported photocatalysts, the composite’s rate constant was more than three times higher than that of a cerium oxide attapulgite carbon nitride system and fifteen times higher than titanium and metal-organic framework alternatives, although a palladium oxide barium titanate catalyst posted a faster rate. The control experiments were equally telling. Without catalyst, light alone removed only 10 percent of the sulfur; in the dark, the catalyst removed 15 percent through adsorption alone; and light without catalyst managed a mere 5 percent. Only the combination of visible light and the heterojunction photocatalyst delivered the full performance, confirming that genuine photocatalysis, not adsorption or photolysis, was doing the work.

The mechanism the authors propose is the classic heterojunction playbook. Visible light excites electrons in both semiconductors, and the favorable band alignment separates the charge carriers at the interface. Electrons react with dissolved oxygen to generate superoxide radicals, while hydrogen peroxide acts as an electron acceptor and is converted into hydroxyl radicals. The photogenerated holes either oxidize adsorbed sulfur compounds directly or produce additional hydroxyl radicals from water. Together, these reactive oxygen species convert dibenzothiophene-type molecules into sulfoxides and sulfones, whose increased polarity allows them to be extracted into the acetonitrile phase. The contrast with the individual components was stark: pristine cadmium oxide removed only 25.7 percent of the sulfur and pristine carbon nitride 36.4 percent, while the composite’s performance was roughly 16 percentage points higher than carbon nitride alone.

Practicality was addressed on several fronts. The catalyst survived five consecutive reuse cycles with only a 12 percent decline in efficiency, dropping to about 86.6 percent, and post-use analysis showed the crystal structure essentially preserved with surface area falling only negligibly. The loss is attributed to temporary fouling by oxidation products rather than structural degradation, and simple thermal regeneration and ethanol washing restored the material for each cycle. Notably, the treated fuel actually emerged with improved properties: the octane number rose from 78 to 89, and olefin and aromatic contents decreased. Sunlight activation, while less efficient than the mercury lamp, still achieved 28.8 percent sulfur removal, hinting at a route to electricity-free operation. The authors are candid about the caveats. Cadmium raises legitimate toxicity concerns, and the team recommends future leaching studies using ICP techniques before any industrial deployment. Pilot-scale validation, testing on diesel and gasoline, and techno-economic assessment remain necessary hurdles. Still, by demonstrating deep desulfurization of a real, messy industrial fuel under ambient conditions, without high-pressure hydrogen or the 350-degree temperatures that conventional hydrodesulfurization demands, the work marks a meaningful step toward cleaner combustion chemistry.

Subject of Research: Photocatalytic oxidative desulfurization of real gas condensate fuel using a CdO/graphitic carbon nitride heterojunction nanocomposite

Article Title: Photocatalytic desulfurization of real fuel by the synthesis of CdO/g-C 3 N 4 nanocomposite as an efficient catalyst

Article References: Azizi, A., Shariati, M., Shahbazy, Z., & Roozbahani, P. A. (2026). Photocatalytic desulfurization of real fuel by the synthesis of CdO/g-C3N4 nanocomposite as an efficient catalyst. Results in Optics, Article 101162. https://doi.org/10.1016/j.rio.2026.101162

Image Credits: AI Generated

DOI: 10.1016/j.rio.2026.101162

Keywords: photocatalysis, desulfurization, graphitic carbon nitride, cadmium oxide, heterojunction, gas condensate, hydrogen peroxide, reactive oxygen species, sulfur removal, clean fuels, visible light, nanocomposite