The precursor was prepared by grinding Bi(NO3)3·5H2O with sorbitol, C6H14O6, until a homogeneous, transparent, viscous mass formed. The pentahydrate’s crystallization water was sufficient for this mixing process, which took about 7 to 10 minutes; no additional water was required. Optional ethanol, added at 2.5 mL per 10 g of precursor, served as a polar protic solvent, improved the distribution of components, limited local overheating and promoted more stable, reproducible pyrolysis. The resulting complex was soluble in polar organic solvents and water and remained stable in solution for at least one month, according to the study’s description of the precursor chemistry.

During synthesis, the precursor was dried in air for one hour and then heated in stages. An initial ramp of about 10–15 degrees Celsius per minute brought the material to 100 degrees Celsius, where the organic component oxidized vigorously and released gases without an open flame. Most gas evolution occurred within one to two minutes. A slower second ramp of 3–5 degrees Celsius per minute removed remaining gaseous products and formed the target phases. Because nitrate supplied oxygen within the precursor, much of the sorbitol was oxidized during the self-sustaining decomposition, reducing the need for a separate carbon-removal treatment. The researchers described the approach as a “soft” low-temperature method intended to limit sintering and preserve both the developed morphology and metastable phases of the products.

X-ray diffraction showed that the precursor ratio divided the products into four broad regions. At a Bi(NO3)3·5H2O-to-sorbitol molar ratio of 0.4 to 1.5, metallic bismuth predominated alongside α-Bi2O3. The largest metal fraction, 60 wt.%, occurred at the highest sorbitol content, represented by a ratio of 0.4; the metallic fraction fell to no more than 15 wt.% at 0.7 and about 3.5 wt.% at 1.3. Microscopy showed rounded bismuth particles surrounded by agglomerates of elongated and irregular α-Bi2O3 particles. At ratios from 1.5 to 2.3, α-Bi2O3 became dominant, and a ratio of 1.9 produced individual α-Bi2O3 without detected metallic bismuth or appreciable β-Bi2O3.

A ratio of 2.3 to 4.2 shifted the dominant product toward β-Bi2O3, with neighboring phases appearing near the interval boundaries and a more disordered β structure at its upper end. A ratio of 2.8 yielded β-Bi2O3 with an ordered crystal structure and a porous, honeycomb-like morphology. Its measured specific surface area was 6.4 m²/g, and its band gap was 2.35–2.41 eV across the tested holding temperatures, consistent with the visible-light responsiveness of this polymorph. At ratios of 4.3 or above, the products became X-ray amorphous. These samples had large cellular macropores and a specific surface area of 12.3 m²/g, the highest among those examined, although the authors emphasized that this remained moderate because dense, partially sintered regions limited access to internal porosity.

Synchronous thermal analysis and mass spectrometry helped explain how the precursor ratio changed the heat released and the final phase. In the sorbitol-rich region, a 1.1 ratio produced three main stages between 40 and 470 degrees Celsius. The first, from 40 to 158 degrees Celsius, coincided with a 20.8% mass loss and released nitrogen oxides, water and carbon dioxide. A second stage, from 158 to 350 degrees Celsius, produced another 28.8% loss and included a weak endothermic feature attributed to melting metallic bismuth. The final stage, from 350 to 470 degrees Celsius, involved an 8.3% loss as residual organic material and carbon underwent afterburning, with bismuth oxide apparently acting as a catalyst. Gas analysis and thermal effects suggested that unstable sorbitol nitrate derivatives participated in the initial decomposition, but the researchers did not directly isolate or identify these intermediates.

At a ratio of 1.9, close to the nominal balance between oxidizer and reductant, the first stage from 40 to 180 degrees Celsius was faster and produced a 36.7% mass loss, with a narrow exothermic maximum at 141 degrees Celsius. A second stage extended to 485 degrees Celsius and accounted for a further 17.3% loss. At the β-Bi2O3-favoring ratio of 2.8, the initial 40–180-degree-Celsius stage produced a larger 47.3% loss, while the later stage to 440 degrees Celsius was smaller at 9.5%. In the strongly oxidizer-rich condition, represented by a ratio of 4.2, decomposition was more endothermic and prolonged. It began with a 34.2% loss to 155 degrees Celsius, followed by a combined 20.3% loss as bismuth oxynitrates decomposed up to 500 degrees Celsius. These observations indicate that excess nitrate shifted much of the initial decomposition to lower temperatures, a condition favorable to retaining β-Bi2O3.

Thermodynamic calculations were used to interpret the experimentally observed phase boundaries. Five representative reactions were normalized to 26 moles of bismuth nitrate pentahydrate, and their standard reaction enthalpies, Gibbs free energies and theoretical adiabatic temperatures were compared with thermal-analysis data. In the metal-forming region, oxidation of sorbitol carbon to carbon dioxide rather than carbon monoxide produced an exothermicity 2.1 times greater and an adiabatic temperature 2.4 times higher than the corresponding carbon-monoxide reaction. The calculated preference was consistent with an initial mass-loss rate 3.1 times higher in practice. Adiabatic temperature was a theoretical upper limit that ignored heat loss, but comparisons among samples pyrolyzed under identical conditions remained useful.

The calculations associated β-Bi2O3 formation with a combination of rapid decomposition and moderate heat release. At a ratio corresponding to reaction IV, the precursor contained excess bismuth nitrate, and the calculated reaction heat was 2.1 times lower and the adiabatic temperature 1.7 times lower than for the α-Bi2O3-favoring reaction III. The extra nitrate acted as thermodynamic ballast: decomposition of surplus nitrate to nitrogen dioxide consumed part of the heat generated by sorbitol oxidation. The experimentally estimated initial mass-loss rate was nevertheless 49% per minute for reaction IV, slightly above the 42% per minute for reaction III. The authors propose that the lower total heat release allowed the β lattice to form without enough energy for relaxation into stable α-Bi2O3, whereas the greater heat of reaction III favored the stable α structure. At the largest nitrate excess, calculated to be ratio 4.33 in reaction V, heat release was considered insufficient for crystal-lattice formation, consistent with an X-ray-amorphous product. This remains a model-based interpretation rather than direct proof of every intermediate pathway.

The gas atmosphere widened the available phase windows. In argon, the region in which β-Bi2O3 predominated expanded and the α-Bi2O3 fraction decreased, although pure α-Bi2O3 was not obtained and the maximum measured α content was 53%. The maximum metallic-bismuth fraction increased from 63% in air to 79% in argon because carbon monoxide and organic intermediates were not oxidized by atmospheric oxygen. This overlap between the metal and β-phase regions nevertheless enabled a one-step β-Bi2O3/Bi composite at ratios of 3.1 to 3.7, containing 4.1–7.7 mol.% Bi. In carbon dioxide, the maximum metal content reached 69%, while the maximum α-Bi2O3 content was 48% and the phase (BiO)2CO3 reached 11%. A ratio of 2.8 to 3.4 produced a β-Bi2O3/(BiO)2CO3 composite containing 2.5–6.2 mol.% (BiO)2CO3. Such materials are candidates for photocatalysis and sensing because β-Bi2O3 is a narrow-bandgap semiconductor and metallic bismuth can support plasmonic effects, but the present study did not experimentally demonstrate those proposed mechanisms or applications.

Temperature treatment also changed particle dimensions and optical properties while leaving the underlying α-Bi2O3 crystal structure intact. Raising the isothermal holding temperature from 300 to 500 degrees Celsius increased the α-phase crystallite size from 49.6 to 91.9 nm and reduced its specific surface area from 4.50 to 2.40 m²/g. Its band gap remained nearly constant at about 2.79–2.82 eV. β-Bi2O3 showed greater temperature sensitivity, with band gaps from 2.35 to 2.41 eV and crystallite sizes from 18.5 to 30.4 nm. Raman spectroscopy detected high-frequency shifts of 5–9 cm⁻¹ in key modes, while refinement produced unit-cell parameters smaller than those of bulk β-Bi2O3. The authors associated this contraction with mechanical stress in the nanocrystalline material, while acknowledging that a size contribution could not be excluded. X-ray diffraction, Raman spectroscopy and differential scanning calorimetry also found no significant phase changes after β-Bi2O3 and amorphous Bi2O3 had been stored in air for 1.5 years, although a systematic ambient-transformation study was not performed.

Subject of Research: Experimental control of bismuth oxide phases and composites through sorbitol-nitrate pyrolysis

Article Title: Sorbitol-nitrate pyrolysis as a method for phase composition control of bismuth oxides and composites based thereon

Article References: Kaminsky, O., Kirichenko, E., & Makarevich, K. (2026). Sorbitol-nitrate pyrolysis as a method for phase composition control of bismuth oxides and composites based thereon. Frontiers in Physics, 14, 1982647. https://doi.org/10.3389/fphy.2026.1982647

Image Credits: AI Generated

DOI: 10.3389/fphy.2026.1982647

Keywords: bismuth oxide, sorbitol, oxidative pyrolysis, phase control, metastable phases, bismuth composites, thermodynamic analysis, green chemistry, photocatalysis, Sorbitol-nitrate, pyrolysis, method