REVIEW 3 major objections 6 minor 13 references
Ferroelectricity in layered bismuth oxide down to 1 nanometer
T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash
Pith's one-line read The paper claims that a samarium-stabilized layered bismuth oxide film retains switchable ferroelectric polarization at a thickness of about one nanometer, demonstrated by a macroscopic polarization–electric field hysteresis loop with…
desk verdict A credible 1-nm ferroelectric loop with a genuine materials innovation, but the missing leakage/transient analysis and a composition gap between experiment and DFT need to be resolved before the record claim stands. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is a layered bismuth oxide built by deleting one bismuth layer from the fluorite structure, yielding repeat units of three Bi-O layers separated by wide gaps, in a tetragonal-like phase with in-plane parameter of about 3.94 Å and out-of-plane parameter of about 9.24 Å. Samarium substitution locks the oxygen sublattice in place and lowers the formation energy by 0.41 eV per atom. The polar ground state Pmm2 Bi6O9 contains four inequivalent bismuth sites; two of them, sitting inside O8 hexahedron cages, develop lobe-like electron localization from bismuth lone pairs on the way from the centrosymmetric Pmmm phase to the polar Pmm2 phase, and these lone-pair displacements are the microscopic engine of polarization. A double-well energy landscape with 24 meV per atom separation, together with a hybrid-DFT band gap of about 1.2 eV that opens to 2 eV with samarium substitution, establish an insulating switchable ground state. The positive-up-negative-down (PUND) pulse scheme subtracts non-ferroelectric switching contributions to isolate the true switched charge.
What would settle it
Measure the same 1 nm film in a capacitor with blocking electrodes and check that the PUND remanent polarization scales linearly with electrode area and vanishes on heating through the reported 493 K transition; if the switched charge is dominated by capacitive or resistive artifacts, or survives in a nonpolar polymorph of the same composition, the central claim would be refuted.
Extended reading notes
Core claim
On the paper's own terms, the discovery is that a layered bismuth oxide, Bi1.8Sm0.2O3, sustains macroscopic ferroelectric hysteresis at one-nanometer thickness, with remanent polarization 17 μC/cm², and that the polarization grows to 50 μC/cm² at 4.56 nm. The authors describe this as the first time a standard ferroelectric hysteresis loop—the direct electrical fingerprint—has been measured at this thickness; prior atomic-scale ferroelectric reports relied on microscopy, local piezoresponse, or resistance switching. The measured structure is a three Bi-O layer repeat that matches a predicted Pmm2 Bi6O9 polar phase with samarium substitution, and the polarization follows from stereochemically active bismuth lone pairs breaking inversion symmetry. The loop, domain writing, retention data, and calculated polarization are offered together as evidence that a practically useful out-of-plane ferroelectric can exist at one unit cell.
Load-bearing premise
The load-bearing premise is that the macroscopic PUND loop taken from the 1 nm film comes from ferroelectric switching of the uniform layered bismuth oxide phase across the whole electrode area, and not from leakage, trapped charge, electrode interfaces, or a minority nonpolar phase.
Editorial extensions
If this is right
- A 1 nm ferroelectric with out-of-plane polarization can be incorporated as a switchable dielectric in field-effect transistors and nonvolatile memories, where the read and write voltage scales down with thickness.
- Because the films grow by sol-gel deposition on sapphire, SrTiO3, and even Au/SiO2/Si substrates, the route is compatible with inexpensive, large-area processing rather than requiring molecular-beam epitaxy.
- The retention fit, with power-law decay exponent 0.047 at 1 nm, predicts polarization persisting for days, a prerequisite for memory operation.
- The measured 17 μC/cm² at 1 nm is comparable to or larger than conventional perovskite films several times thicker, indicating the design suppresses the usual critical-size collapse.
- The direct PUND loop at about 1 nm puts a quantitative electrical benchmark on scaling, allowing fair comparison with hafnium oxide and perovskite candidates.
Reading between the lines
- If the loop is truly intrinsic, the structural recipe of vacancy-ordered fluorite slabs plus a lone-pair-active cation could be tried in other bismuth- or antimony-based oxides to push direct ferroelectric hysteresis below one nanometer; the paper does not report such variants.
- A natural next experiment is to sweep electrode area and temperature while watching the switched charge: uniform scaling with area and disappearance near the reported 493 K transition would strongly support single-phase intrinsic switching.
- The authors do not separate how much of the thickness dependence comes from the depolarization field versus a low-dielectric-constant dead layer; if the dead layer dominates, even thinner films with improved interfaces might retain a loop.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript reports a layered bismuth oxide film, Bi1.8Sm0.2O3 (BSO), grown by sol-gel on sapphire and STO-based substrates, and claims ferroelectricity down to ~1 nm thickness based on macroscopic PUND hysteresis loops (remanent polarization 17 μC/cm2) plus PFM switching and retention. The structure is characterized by XRD, RSMs, and HAADF-STEM, and DFT structure prediction identifies a polar Pmm2 Bi6O9 phase, with Sm substitution modeled as Bi5SmO9, giving a computed polarization around 30 μC/cm2. The authors argue this is the first direct macroscopic ferroelectric loop at 1 nm thickness.
Significance. If the central claim is correct, the paper reports a remarkable result: a macroscopic ferroelectric hysteresis loop at 1 nm thickness, a regime where most ferroelectrics lose switchable polarization or can only be probed by local or indirect methods. The work combines chemical solution deposition, careful STEM imaging, and first-principles structure prediction, and the DFT analysis includes an energy double-well landscape, ELF lone-pair visualization, and Berry-phase polarization. The PFM data provide local support for switchable domains and retention. These strengths make the study potentially important for atomic-scale ferroelectric devices. However, the quantitative macroscopic claim rests on a PUND measurement whose leakage and transient behavior are not documented in the main text, and the DFT model composition does not match the measured film composition, so the current manuscript does not fully close the gap between evidence and claim.
major comments (3)
- [Ferroelectric hysteresis loop (Fig. 3A)] The central claim of a 1-nm ferroelectric loop relies on PUND data, but the main text reports no leakage current density, no pulse transient waveforms, no pulse-width or voltage dependence, and no explicit voltage/field axis for the loop. PUND subtraction is only valid when the non-switching P and U pulses have identical leakage and trapping behavior; at 1 nm thickness, direct tunneling and trap-mediated currents, as well as the Bi3+/Ti4+ interfacial charge transfer acknowledged in the DFT section, can produce apparent switched charge. The authors should present the raw PUND pulse data, leakage curves, and an analysis showing that the 17 μC/cm2 remanent polarization is not dominated by non-ferroelectric charge injection.
- [The design of layered structure / Theoretical calculation] The measured composition Bi:Sm:O = 1.8:0.2:3 (10% Sm on the Bi site) does not match the DFT model Bi5SmO9 (16.7% Sm) used for the predicted ferroelectric structure and polarization. Since the experimental composition is used to identify the phase, the authors should either perform DFT for the measured composition, show that the Bi5SmO9 model is representative within a robust range of Sm content, or explicitly justify why the 6.7% difference does not affect the structure and polarization conclusions.
- [Ferroelectric hysteresis loop (semi-empirical polarization estimate)] The manuscript states that a semi-empirical method (ref. 39) gives a spontaneous polarization of 49.8–53.4 μC/cm2 from a Bi displacement of about 0.22 Å, but the formula, parameters, and error estimate are not provided. This value is also substantially higher than the DFT Berry-phase polarization of about 30 μC/cm2, so the statement that the calculation is consistent with the measured 17–50 μC/cm2 range is too vague to be assessed. The authors should give the exact relation used and discuss the discrepancy between the two theoretical estimates.
minor comments (6)
- [Abstract] The phrase 'samarium bondage' appears to be a typo; it should likely be 'samarium bonding' or 'samarium binding', and the intended meaning should be clarified.
- [Fig. 3A and Fig. 3B] The hysteresis loop figures do not show the applied voltage or electric-field scale; adding an explicit axis with field values would allow readers to assess the coercive field and the plausibility of the measurement at 1 nm.
- [Ferroelectric hysteresis loop (PUND description)] The text refers to Figs. S28–S29 for PUND subtraction details, but the main text should at least summarize the pulse sequence, pulse widths, delay times, and the subtraction procedure so that the measurement can be evaluated without the supplement.
- [PFM characterization (Fig. 4)] The PFM phase lag for the 1 nm film is reported as 70°–80°, which is well below the 180° expected for full polarization reversal; the authors should explain whether this reflects partial switching, electrostatic artifacts, or a thickness-dependent effect.
- [General structures] The term 'T-like phase' is used without a definition; it should be defined (e.g., tetragonal-like as opposed to rhombohedral or other distortions) at first occurrence.
- [Theoretical calculation (reference formatting)] Reference 52 is missing the closing parenthesis in the year ('1996.' instead of '1996).'), and should be corrected.
Circularity Check
No significant circularity: the 1-nm ferroelectricity claim rests on direct PUND and PFM measurements, and the DFT polarization is an independent first-principles result not fitted to the loop.
full rationale
The central claim—macroscopic ferroelectric hysteresis at ~1 nm thickness—is established by direct PUND measurement (Fig. 3A–B) and by PFM writing and local butterfly loops (Fig. 4), not derived from a model parameter. The DFT portion is a separate calculation: the double-well energy landscape and the Berry-phase polarization (~30 μC/cm²) are computed from first principles and are only compared with the measured 17–50 μC/cm² as a consistency check, so no fitted parameter is renamed as a prediction. The only mild in-sample aspect is that the USPEX structure search is constrained by HAADF-STEM information and by the substrate in-plane lattice constant, and the resulting candidate is then compared with HAADF-STEM images as confirmation; however, this structural validation is not the load-bearing step for the ferroelectricity claim, which stands on the independent electrical and PFM data. The minor self-citations (refs 44–45, co-authored by Fang and Diéguez) are used for routine context—BTO energy comparison and lone-pair activity—and are not load-bearing. No equation in the paper reduces the claimed ferroelectricity to an input parameter, and the measured loop is not reconstructed from the DFT polarization or from the STEM-derived displacement values.
Assumptions & free parameters
free parameters (2)
- Sm content (Bi1.8Sm0.2O3) =
20 at% of cation site according to formula; measured Bi:Sm:O=1.8:0.2:3; DFT uses Bi5SmO9 (approximately 16.7 at% Sm)
- PFM decay exponent alpha =
0.047
assumptions (4)
- domain assumption DFT with PBE/HSE and USPEX structure search correctly identifies the ground state of Bi6O9.
- domain assumption The PUND hysteresis loop measures true ferroelectric polarization in the 1 nm film under the reported electrode geometry.
- domain assumption The 1 nm film is structurally uniform and single-phase across the entire capacitor area.
- ad hoc to paper The bulk DFT structure persists at 1 nm thickness with Sm substitution and substrate interfaces.
invented entities (1)
-
Layered Pmm2 Bi6O9 (and Sm-substituted Bi5SmO9) ferroelectric phase
independent evidence
Cite this review
Pith. "Pith review of Ferroelectricity in layered bismuth oxide down to 1 nanometer." pith.science (2026). https://pith.science/paper/ZDHYVX6B
@misc{pith2026250109549,
author = {Pith},
title = {Pith review of: Ferroelectricity in layered bismuth oxide down to 1 nanometer},
year = {2026},
howpublished = {\url{https://pith.science/paper/ZDHYVX6B}},
note = {Machine review of arXiv:2501.09549}
}
read the original abstract
Atomic-scale ferroelectrics are of great interest for high-density electronics, particularly field-effect transistors, low-power logic, and nonvolatile memories. We devised a film with a layered structure of bismuth oxide that can stabilize the ferroelectric state down to 1 nanometer through samarium bondage. This film can be grown on a variety of substrates with a cost-effective chemical solution deposition. We observed a standard ferroelectric hysteresis loop down to a thickness of ~1 nanometer. The thin films with thicknesses that range from 1 to 4.56 nanometers possess a relatively large remanent polarization from 17 to 50 microcoulombs per square centimeter. We verified the structure with first-principles calculations, which also pointed to the material being a lone pair-driven ferroelectric material. The structure design of the ultrathin ferroelectric films has great potential for the manufacturing of atomic-scale electronic devices.
Reference graph
Works this paper leans on
-
[1]
Beijing Advanced Innovation Center for Materials Genome Engineering, Institute for Advanced Materials and Technology, University of Science and Technology Beijing, 100083, China
-
[2]
Institute of Microstructure and Property of Advanced Materials, Faculty of Materials and Manufacturing, Beijing University of Technology, Beijing, 100124, China
-
[3]
Centro de Física de Materiales (CSIC-UPV/EHU), Manuel de Lardizabal Pasealekua 5, 20018 Donostia/San Sebastián, Spain
-
[4]
Fisika Aplikatua Saila, Gipuzkoako Ingeniaritza Eskola , University of the Basque Country (UPV/EHU), Europa Plaza 1, 20018 Donostia/San Sebastián, Spain
-
[5]
University of Michigan –Shanghai Jiao Tong University Joint Institute, Shanghai Jiao Tong University, Shanghai, China
-
[6]
Department of Physical Chemistry, University of Science and Technology Beijing, Beijing, 100083, China
-
[7]
Department of Materials Science and Engineering, The Iby and Aladar Fleischman Faculty of Engineering, The Raymond and Beverly Sackler Center for Computa tional Molecular and Materials Science, Tel Aviv University, Tel Aviv, Israel
-
[8]
Institute of High Energy Physics, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Beijing 100049, P. R. China
Show all 13 references
-
[9]
Physical Science and Engineering Division, King Abdullah University of Science and Technology (KAUST), Thuwal 23955–6900, Saudi Arabia
-
[10]
National Synchrotron Radiation Laboratory, University of Science and Technology of China, Hefei 230026, China. 2
-
[11]
dead layer
Institute of Solid State Chemistry, University of Science and Technology Beijing, Beijing, 100083, China. *Correspondence to: linxingzhang@ustb.edu.cn, tianjianjun@mater.ustb.edu.cn, luyue@bjut.edu.cn #These authors contributed equally to this work. Abstract: Atomic-scale ferr...
-
[12]
The simulation model is shown in Fig
and [210], which are highly consistent and used to support the accuracy of the predicted structure. The simulation model is shown in Fig. S50. In addition, we observed that matching between BSO film and STO substrate is achieved by Bi-O and Ti-O layer of substrate. Therefore, ...
2003
-
[51]
4, 1–8 (2019)
Koohfar, et al., npj Quantum Mater. 4, 1–8 (2019). 52 G. Kresse, J. Furthmüller, J Comput. Mat. Sci. 6, 15 (1996. 53 G. Kresse, J. Furthmüller, Phys. Rev. B 54, 11169 (1996). 54 J.P. Perdew et al., Phys. Rev. Lett. 100, 136406 (2008). 55 P. E. Blöchl, Phys. Rev. B. 50, 17953 (...
2019 arXiv
Reviewed August 10, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.