The pressed nanopowders described in the study showed broad, large dielectric responses, with reported diffuse maxima reaching 105 in the real component and 108 in the imaginary component between 300 K and 400 K. The result was highly dependent on the test conditions: both components rose strongly at higher temperatures, and the response at 100 Hz was at least an order of magnitude larger than at 100 kHz. The authors interpret the pattern as an interplay between a possible diffuse ferroelectric-paraelectric transition and Maxwell-Wagner-Sillars polarization at material interfaces. The data, however, do not uniquely separate those proposed contributions from the intrinsic response of the nanoparticles.
The signal changed with the test conditions
The study compared five nominal compositions, x = 0, 0.05, 0.1, 0.15 and 0.2. The powders were prepared by solution combustion and calcined for five hours at 750 °C. For the electrical tests, the material was pressed into PTFE-cell disks 4 mm across and 0.2 mm thick under about 2.5 MPa. Temperature scans were run at 100 Hz and 100 kHz, while room-temperature frequency scans covered 4 Hz to 500 kHz. Because the samples were pressed disks, the reported values are effective responses of the powder assembly, including the packing and interfaces present in that setup.
At 100 Hz, both effective-permittivity components changed only weakly from 20 °C to 300 °C, then rose strongly from 300 °C to 400 °C. At 100 kHz, the corresponding boundary shifted to about 250 °C, but the response was at least one order of magnitude lower. The temperature of the maximum also varied with nominal Sm content at both frequencies. Below x = 0.15, the real and imaginary trends were similar; at higher Sm content, they tended to oppose one another.
Sm content did not produce a simple trend
Sm content also produced a nonmonotonic pattern at room temperature. At 4 Hz, the permittivity reached a local maximum near x = 0.05 and a local minimum near x = 0.15, so increasing the nominal substitution did not move the response steadily in one direction across the tested series. That comparison is complicated by changes in the reported phase mixtures. X-ray diffraction found 97% R3c, 2% Bi2Fe4O9 and 1% Bi25FeO40 at x = 0.05; 99% R3c and 1% Bi25FeO40 at x = 0.10; 7% R3c, 92% Pbnm and 1% Bi25FeO40 at x = 0.15; and 99% Pbnm with 1% Bi25FeO40 at x = 0.20. The nominal Sm series therefore does not isolate composition from structural variation.
Interfaces offered one explanation
Frequency plots added evidence that more than one relaxation process may be present. For x = 0, 5% and 10%, Cole-Cole plots showed a strongly deformed arc at high frequency and a sharp upward branch at low frequency. The paper classifies the low-frequency screening-charge relaxation as Maxwell-Wagner-type interfacial polarization, meaning in this account that charge effects build up at boundaries between materials and at the nanoparticle-air interface. It treats that process as consistent with barrier-layer interpretations. But the supplied analysis says the interfacial contribution was not independently separated from the nanoparticles' intrinsic response.
The powder was not a uniform set of particles
The particle population was not uniform. Transmission electron microscopy showed particles from 50 nm to 500 nm and large agglomerates. Individual nanoparticles were described as irregular, while the theoretical treatment represented them as effective spheres. The authors used effective-medium processing to analyze the nanoparticle response separately from the effective response of the pressed powder. That step is an attempt to disentangle levels of the measurement, not a direct observation of a single particle's dielectric behavior.
The theoretical side of the work used a modified LGDSH+FSM approach. As nominal Sm content rose from 0% to 20%, it predicted a sequence of modeled states labelled FE, FEI, QFEI, AFE and NP. Simulations compared 50 nm and 500 nm particles, reporting negligible size effects above 300 K for the 50 nm case and virtually absent effects for the 500 nm case. These are model scenarios rather than direct phase measurements for every particle, so they support the proposed reading without proving that each powder grain follows the same sequence.
The central caution is attribution
The study's central result is therefore a measured pattern within the reported laboratory setup, not a settled explanation of a material property. The supplied analysis reports no replicate-based uncertainty, error bars, confidence intervals or inferential statistical analysis. It also says full fit parameters and goodness-of-fit measures were not reported for the effective-medium treatment or for separating diffuse-transition and interfacial contributions. The authors' combined explanation remains model-based, with open questions about reproducibility across independently prepared batches and about how much of the response comes from intrinsic nanoparticles, interfaces, packing and phase impurities.
Paper data and sources
Original title: A colossal dielectric response of Bi1-xSmxFeO3 nanopowders
Authors: Vladyslav O. Kolupaiev, Olexander S. Pylypchuk, Vladimir N. Poroshin et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-25
DOI: Not available
Original paper · Full text