Ferroelectric Switching in Hybrid Molecular Beam Epitaxy-Grown BaTiO3 Films

Anusha Kamath Manjeshwar; Bharat Jalan; Chin-Hsiang Liao; Jiaxuan Wen; Richard D. James; Steven J. Koester; Zhifei Yang

REVIEW 2 major objections 5 minor 5 references

Ferroelectric Switching in Hybrid Molecular Beam Epitaxy-Grown BaTiO3 Films

T0 review · 2 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper reports the first direct measurement of remnant polarization in MBE-grown BaTiO3 films, using an all-epitaxial SrRuO3/BaTiO3/SrRuO3 capacitor grown by hybrid MBE.

desk verdict First direct remnant polarization in MBE-grown BaTiO3 is reported with careful PUND methodology; the PUND subtraction is the main caveat, but the paper deserves a serious referee. read the letter →

arxiv 2505.01905 v1 pith:PSNYI5AX submitted 2025-05-03 cond-mat.mtrl-sci

Anusha Kamath Manjeshwar , Zhifei Yang , Chin-Hsiang Liao , Jiaxuan Wen , Steven J. Koester , Richard D. James , Bharat Jalan This is my paper · ORCID

classification cond-mat.mtrl-sci

keywords BaTiO3SrRuhybridmolecularbeamepitaxyremnantpolarizationPUNDferroelectricswitchingepitaxialheterostructurelead-free

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper reports an all-epitaxial capacitor, 16 nm SrRuO3 / 40 nm BaTiO3 / 16 nm SrRuO3, grown on Nb-doped SrTiO3 (001) by hybrid molecular beam epitaxy with metal-organic precursors. It claims that Positive-Up-Negative-Down (PUND) measurements isolate a genuine ferroelectric switching current, yielding hysteretic polarization–electric-field loops with a remnant polarization of about 15 µC/cm² at frequencies from 500 Hz to 20 kHz. This is presented as the first direct measurement of remnant polarization in MBE-grown BaTiO3, a synthesis route previously limited to indirect evidence such as piezoresponse force microscopy and electro-optic hysteresis. If correct, the result establishes hybrid MBE as a practical route to ferroelectric devices, not just to high-quality epitaxial films.

What carries the argument

The load-bearing object is the PUND pulse train, a five-pulse sequence (poling, Positive, Up, Negative, Down) in which the Positive and Negative pulses contain both ferroelectric switching and non-ferroelectric currents while the Up and Down pulses, arriving after a 10 µs delay, are assumed to contain only the non-ferroelectric parts. Subtracting Up from Positive and Down from Negative isolates the switching current without assuming a functional form for leakage or capacitance. The second piece of machinery is hybrid MBE growth of both SrRuO3 electrodes and the BaTiO3 layer using Ru(acac)3 and titanium tetraisopropoxide precursors, which the paper argues enables adsorption-controlled, stoichiometric growth with low defect densities.

What would settle it

Repeat the PUND measurement on the same devices with delay and wait times of 1, 10, and 100 µs while keeping rise and fall times fixed; if the extracted remnant polarization changes by more than the device-to-device scatter, the 10 µs subtraction does not fully separate ferroelectric from non-ferroelectric charge. Independently, measure hysteresis above the BaTiO3 Curie temperature: a genuine ferroelectric loop should collapse, while leakage artifacts would survive.

Watch

Extended reading notes

Core claim

The central claim is that an all-epitaxial SrRuO3/BaTiO3/SrRuO3 heterostructure grown by hybrid MBE on Nb:SrTiO3 (001) shows true ferroelectric switching. Using the PUND pulse train, which compares switching pulses (Positive and Negative) with non-switching pulses (Up and Down), the authors subtract leakage and capacitive currents and integrate the remaining current to obtain polarization–voltage hysteresis. They report a remnant polarization of about 15 µC/cm² that stays roughly constant between 500 Hz and 20 kHz, with coercive field and imprint voltage increasing with frequency. The asymmetry between up-to-down and down-to-up switching is attributed to structurally and chemically dissimilar electrodes, and the voltage-dependent leakage is hypothesized to come from gradients of defect dipoles, most likely oxygen or barium vacancies.

Load-bearing premise

The whole measurement rests on the assumption that, after a 10 microsecond delay, the Up and Down pulses carry exactly the same non-ferroelectric current as the Positive and Negative pulses; the paper's own I–V data (Figure 3g) show leakage that changes steeply with voltage and bias direction, so the cancellation could be incomplete.

Editorial extensions

If this is right

If correct, MBE-grown BaTiO3 joins the list of epitaxial ferroelectric films with directly verified non-zero remnant polarization, closing a long-standing gap between MBE growth studies and electrical characterization.
The reported value, about 15 µC/cm², is roughly five times the saturation polarization of the only earlier MBE-grown BaTiO3 film that showed P–E hysteresis, suggesting that electrode engineering rather than the growth technique was the main obstacle.
Because the remnant polarization is flat from 500 Hz to 20 kHz, the switched charge is not a low-frequency artifact and could persist at frequencies relevant to memory and capacitor applications.
The frequency-dependent coercive field and imprint voltage imply that switching from the up to the down state becomes progressively harder at higher frequencies, so device designs must budget for asymmetric switching kinetics.
The benchmark comparison indicates that scandate substrates and smaller capacitor areas are concrete levers that could raise the remnant polarization toward the values reported for PLD- and sputter-grown films.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A natural test of the subtraction logic is to vary the PUND delay and wait times from 1 µs to 100 µs; if the extracted remnant polarization drifts, the 10 µs choice is not neutral and the quoted number would need revision.
Because the paper's own static I–V data show strongly voltage-dependent leakage under negative bias, an independent method such as switched-charge measurement or temperature-dependent hysteresis could determine whether the cancellation is biased.
If the defect-dipole-gradient hypothesis is correct, oxygen annealing under an applied bias or at different temperatures should change the imprint voltage and leakage asymmetry in a predictable, testable way.
The same all-epitaxial platform could be extended to pyroelectric energy harvesting studies, since bulk BaTiO3's sharp Curie transition is already of interest and a direct remnant polarization gives a baseline for switchable charge.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The new thing here is the claim itself: a direct remanent polarization of about 15 uC/cm2 in an all-epitaxial SrRuO3/BaTiO3/SrRuO3 capacitor grown by hybrid MBE. If that claim holds, it fills a real gap: MBE-grown BTO has been characterized by PFM and electro-optics, but not by direct polarization hysteresis. The paper does solid work. It gives explicit PUND timing parameters, shows data from four devices with standard deviations, verifies phase-pure epitaxy by XRD, and uses RSM to document partial relaxation with a tetragonality of 1.022. The benchmarking against PLD and sputtered films is useful, and the authors are honest about unresolved features, notably the top SrRuO3 peak. The main soft spot is exactly what the stress-test note flags. The PUND subtraction assumes the Up and Down pulses reproduce the non-ferroelectric currents during the switching Positive and Negative pulses. The paper itself reports greater leakage and resistance degradation under negative bias (Figure 3g), which is the branch where the N-D subtraction matters. With only 10 us delay and wait, slow or state-dependent leakage could leave a residual that integrates into the reported Pr. The authors do not report a null measurement on a non-switching control or a leakage-compensated cross-check. That is a legitimate concern, but I would not call it a load-bearing flaw. The normalized switching peaks in Figure 3c are clear and look like ferroelectric switching, not just leakage; the impedance data support a capacitive response up to 40 kHz; and the claimed Pr is plausible for a 40 nm partially strained BTO film. The absence of a resolved top SRO peak is a minor structural gap, and the saturation-matching shift used in the integration is not fully justified, but these do not undermine the central result. For the ferroelectric and oxide MBE communities, this is a useful letter: it establishes that hybrid MBE can produce a working ferroelectric capacitor, and it gives a clear recipe for others to try. I would cite it if I were working on BTO thin films. It deserves a serious referee who can pressure-test the PUND analysis and ask for the additional control measurements; that is a normal part of the process. The verdict from the reading group, accept with moderate confidence, seems right to me. The paper is not overhyped and the claims are calibrated to what is shown.

Referee Report

2 major / 5 minor

Summary. The manuscript reports the hybrid molecular beam epitaxy (MBE) growth of an all-epitaxial SrRuO3 (16 nm)/BaTiO3 (40 nm)/SrRuO3 (16 nm) heterostructure on Nb-doped SrTiO3 (001) substrates and its ferroelectric characterization. The central claim is that Positive-Up-Negative-Down (PUND) measurements yield hysteretic polarization-electric field loops with a remnant polarization Pr ~ 15 uC/cm2 that is stable from 500 Hz to 20 kHz, which the authors present as the first direct measurement of remnant polarization in MBE-grown BaTiO3 films. Supporting evidence includes phase-pure epitaxial X-ray diffraction, reciprocal space mapping showing a partially relaxed BTO layer with tetragonality of about 1.022, impedance spectroscopy showing near-ideal capacitive behavior up to about 40 kHz, and a benchmarking comparison with PLD- and sputtered-grown films. The paper also proposes defect-dipole origins for the observed switching asymmetry and leakage.

Significance. If the central claim holds, this is an important step for MBE-grown ferroelectrics: it provides the missing electrical verification of ferroelectric switching in a synthesis technique that is otherwise well established for epitaxial oxide heterostructures, and it demonstrates a practical hybrid MBE route with adsorption-controlled growth of both the electrodes and the ferroelectric layer. The paper is commendable for reporting averaged results over four devices with standard deviations, giving complete PUND timing parameters, and being transparent about the unresolved top-electrode peak and the leaky negative-bias branch. The structural and dielectric data (XRD, RSM, impedance phase near -90 degrees) are consistent with a high-quality ferroelectric capacitor. However, the electrical claim rests on the validity of the PUND subtraction, which warrants careful scrutiny.

major comments (2)

[Section 4.4, Figure 3c and 3g] The PUND subtraction of the 'Up' and 'Down' currents from the 'Positive' and 'Negative' currents is valid only if the non-ferroelectric current is identical in the switching and non-switching pulses. Figure 3g shows a strong leakage current and resistance degradation under negative bias, and the N-D branch is precisely that bias region. The stated checks (t_d = t_w = 10 us and no back-switching, Supplementary Figures S2-S3) address transient decay and imprint-driven back-switching, but not polarization-state-dependent leakage (e.g., a ferroelectric diode effect) or leakage components with time constants longer than 10 us that would not fully decay during the delay. Because no control measurement on a non-switching or paraelectric capacitor is reported, the residual non-ferroelectric charge integrated into the PUND loops is uncalibrated and could bias the reported 15 uC/cm2. This is a load-bearing issue for the paper's central claim.
[Section 4.4, paragraph defining Pr, Ec, and V_offset] The sentence 'We shift the polarization values such that the magnitudes of the saturation polarization at both peak voltages are similar' describes an unspecified adjustment of the integration constant. If this shift is chosen to force matching of the saturation magnitudes, it can impose a symmetry on the P-V loop that partly predetermines the extracted Pr and imprint voltage. Moreover, the equations defining Pr, Ec, and V_offset are garbled in the manuscript, so it is unclear whether Pr is computed from the difference of the two remanent values (which would be independent of a constant shift) or from the shifted zero-field intercept (which would not). The authors should state the exact definitions and report the magnitude of the shift; without this, the reported Pr is not uniquely defined.

minor comments (5)

[Section 2.3, typos] The word 'thaat' appears in the sentence 'We conclude thaat growth on the more closely lattice-matched scandate substrates...' and should be corrected to 'that'.
[Abstract and Section 2.3] The abstract states Pr ~ 15 uC/cm2, while Section 2.3 reports 'Pr ~ 15.2 uC/cm2'; these values should be made consistent.
[Section 4.4, pulse-train definition] The equivalence between the PUND pulse train and a four-quadrant triangular sweep is stated as f = 1/[2(t_r + t_f)], but the actual waveform includes wait times t_w = 10 us, making the pulses rectangular. The authors should clarify how the frequency f is defined for the rectangular PUND train and how this affects the interpretation of the frequency-dependent data in Figure 3e.
[General, measurement conditions] The measurement temperature is not explicitly stated; the authors should confirm that all electrical measurements were performed at room temperature.
[Supplementary Information] Several supporting claims (Figures S2-S3, Tables S2-S5) are made with reference to supplementary material; if the supplementary information is not part of the review package, the authors should ensure the main text is self-contained or provide the supplementary material.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the ferroelectric claim is a direct PUND measurement benchmarked against external literature; self-citations support growth parameters only.

full rationale

The paper's central claim is an experimental measurement: alloy-epitaxial SrRuO3/BaTiO3/SrRuO3 capacitor grown by hybrid MBE shows hysteretic P-E loops with Pr ~15 uC/cm2 extracted via the Positive-Up-Negative-Down method. No parameter is fitted to the data, no theoretical quantity is derived from an input that already contains the output, and no prediction is generated from a model defined in terms of the target result. The PUND subtraction in Section 4.4 is a standard measurement protocol: U and D pulses capture the non-ferroelectric contribution, and P-U and N-D isolate the ferroelectric current. The validity of this subtraction depends on an empirical assumption about transient decay, which the paper addresses with stated delay/wait times and supplementary checks; this is a measurement-validity concern, not a circularity by construction. The self-citations in the growth sections (refs 88-93 for hybrid MBE growth windows of SrRuO3 and BaTiO3) are used only to justify precursor choices and adsorption-controlled growth conditions; they do not bear on the ferroelectric switching result itself, which is compared against external benchmarks from PLD/sputtered films in Figure 4. No quoted equation reduces to its own input, and no uniqueness theorem or ansatz is imported to force the conclusion. The finding is therefore self-contained as an experimental report, and no circular step is exhibited.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

No free parameters are fitted to data; the ferroelectric result is a measurement, not a model. Three background assumptions are listed: the PUND subtraction identity, the current-to-polarization integration with a constant shift, and the Poisson-ratio estimate of the unresolved top SrRuO3 peak. No invented entities are proposed; defect-related explanations are explicitly hypotheses.

assumptions (3)

domain assumption The Up and Down pulses of the PUND train measure the exact non-ferroelectric background present during the Positive and Negative pulses.
The whole remnant-polarization extraction relies on this equality; if leakage changed with voltage history, subtraction would be imperfect (Sections 4.4, Figure 3c).
standard math Integrating the background-subtracted current density over time yields the polarization, and the integration constant is fixed by equating saturation polarization magnitudes at the two peak voltages.
Standard Maxwell relation used in ferroelectric measurements; the constant-shift step is a processing convention from Section 4.4.
domain assumption The top SrRuO3 (103) peak position is estimated assuming coherent strain to BTO and a Poisson ratio of 0.3.
The top SrRuO3 peak is not resolved in the reciprocal space map; this estimate underpins the discussion of dissimilar electrodes and imprint (Section 2.3, Figure 2c).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Ferroelectric Switching in Hybrid Molecular Beam Epitaxy-Grown BaTiO3 Films." pith.science (2026). https://pith.science/paper/PSNYI5AX

@misc{pith2026250501905,
  author       = {Pith},
  title        = {Pith review of: Ferroelectric Switching in Hybrid Molecular Beam Epitaxy-Grown BaTiO3 Films},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PSNYI5AX}},
  note         = {Machine review of arXiv:2505.01905}
}

read the original abstract

Molecular beam epitaxy (MBE) is a promising synthesis technique for both heterostructure growth and epitaxial integration of ferroelectric BaTiO3. However, a direct measurement of the remnant polarization (P_r) has not been previously reported in MBE-grown BaTiO3 films. We report the in-situ growth of an all-epitaxial SrRuO3/BaTiO3/SrRuO3 heterostructure on Nb-doped SrTiO3 (001) substrates by hybrid MBE using metal-organic precursors. This capacitor structure consisting of 16 nm SrRuO3/40 nm BaTiO3/16 nm SrRuO3 shows hysteretic polarization-electric field (P-E) curves with P_r = 15 {\mu}C cm-2 at frequencies ranging from 500 Hz to 20 kHz, after isolating the intrinsic ferroelectric response from non-ferroelectric contributions using the Positive-Up-Negative-Down (PUND) method. We hypothesize that the asymmetry in switching behavior and current leakage has origins in structural defects.

Figures

Figures reproduced from arXiv: 2505.01905 by the authors.

**Figure 1.** Electrode candidates demonstrated in metal-ferroelectric BaTiO3-metal heterostructures[18,25,37–71]. (a) The number of reports (noted above each bar) of hysteretic P-E curves (solid bar, left) and those with a non-zero remnant polarization (𝑃$, solid bar shaded with diagonal lines, right) reported in metal-ferroelectric BaTiO3-metal heterostructures by synthesis method. (b) Number of reports of different electrode c… view at source ↗

**Figure 2.** (a) Schematic showing the structure of the defined capacitor devices (thicknesses exaggerated for clarity) fabricated from the ~16 nm SRO/~ 40 nm BTO/ ~16 nm SRO/Nb:STO (001) heterostructure by reactive ion etching and Pt deposition for top and bottom contacts by sputtering. The devices were annealed in 760 Torr of oxygen at 500 °C for 1 hour. (b) 2θ-ω coupled X-ray diffraction scans for the as-grown heterostructure… view at source ↗

**Figure 3.** (a) Schematic of a PUND pulse train with five pulses: a negative poling pulse, “Positive” (P), “Up” (U), “Negative” (N), and “Down” (D) pulses in order. A pulse (inset) can be defined by four characteristic times – a delay time (𝑡/), a rise time (𝑡$), a fall time (𝑡5), and a wait time (𝑡0). A voltage bias 𝑉1234 is used to shift the center of the pulse train to account for the horizontal shift in the center of the re… view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Benchmarking ferroelectric properties of SRO/BTO/SRO heterostructures by PVD technique and substrate[18,25,38,40,44,45,48,52,60,62,63,68]. Remnant polarization 𝑃$ vs. BTO thickness 𝑡'() is plotted for (a) different PVD techniques and (b) different substrates, with the …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

5 extracted references · 5 canonical work pages

[1]

stubborn

Introduction Barium titanate (BaTiO3, BTO) is a leading lead-free high-κ dielectric and ferroelectric candidate[1] for applications in non-volatile memories, capacitors, sensors, and electro-optic devices. Bulk BTO has a tetragonal crystal structure at room temperature with lattice constants of a = b = 3.992 Å and c = 4.036 Å and a Curie temperature, Tc ~...

work page 2000
[28]

F. Niu, B. W. Wessels, Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures Processing, Measurement, and Phenomena 2007, 25, 1053. [29] L. Mazet, R. Bachelet, L. Louahadj, D. Albertini, B. Gautier, R. Cours, S. Schamm-Chardon, G. Saint-Girons, C. Dubourdieu, Journal of Applied Physics 2014, 116, 214102. [30] K. J. Kormondy, ...

work page 2007
[54]

L. Qiao, X. Bi, Physica Status Solidi (a) 2010, 207, 2511. [55] Y. Watanabe, Y. Matsumoto, H. Kunitomo, M. T. Mitsuru Tanamura, E. N. Erina Nishimoto, Jpn. J. Appl. Phys. 1994, 33, 5182. [56] B. Wague, N. Baboux, P. R. Romeo, Y. Robach, B. Vilquin, JMP 2020, 11, 509. [57] R. Kullmer, Applied Physics A: Materials Science & Processing 1997, 65, 273. [58] N....

work page 2010
[82]

K. D. Fredrickson, A. B. Posadas, A. A. Demkov, C. Dubourdieu, J. Bruley, Journal of Applied Physics 2013, 113, 184102. [83] D. P. Cann, J.-P. Maria, C. A. Randall, Journal of Materials Science 2001, 36, 4969. [84] H. P. Nair, Y. Liu, J. P. Ruf, N. J. Schreiber, S.-L. Shang, D. J. Baek, B. H. Goodge, L. F. Kourkoutis, Z.-K. Liu, K. M. Shen, D. G. Schlom, ...

work page 2013
[109]

,’() [Å] 𝑎∥,’() [Å] 𝑎

J. W. Arblaster, Platinum Metals Review 2007, 51, 130. [110] R. Bechmann, The Journal of the Acoustical Society of America 1956, 28, 347. Figure 1: Electrode candidates demonstrated in metal-ferroelectric BaTiO3-metal heterostructures[18,25,37–71]. (a) The number of reports (noted above each bar) of hysteretic P-E curves (solid bar, left) and those with a...

work page 2007

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [1]

stubborn

Introduction Barium titanate (BaTiO3, BTO) is a leading lead-free high-κ dielectric and ferroelectric candidate[1] for applications in non-volatile memories, capacitors, sensors, and electro-optic devices. Bulk BTO has a tetragonal crystal structure at room temperature with lattice constants of a = b = 3.992 Å and c = 4.036 Å and a Curie temperature, Tc ~...

work page 2000

[2] [28]

F. Niu, B. W. Wessels, Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures Processing, Measurement, and Phenomena 2007, 25, 1053. [29] L. Mazet, R. Bachelet, L. Louahadj, D. Albertini, B. Gautier, R. Cours, S. Schamm-Chardon, G. Saint-Girons, C. Dubourdieu, Journal of Applied Physics 2014, 116, 214102. [30] K. J. Kormondy, ...

work page 2007

[3] [54]

L. Qiao, X. Bi, Physica Status Solidi (a) 2010, 207, 2511. [55] Y. Watanabe, Y. Matsumoto, H. Kunitomo, M. T. Mitsuru Tanamura, E. N. Erina Nishimoto, Jpn. J. Appl. Phys. 1994, 33, 5182. [56] B. Wague, N. Baboux, P. R. Romeo, Y. Robach, B. Vilquin, JMP 2020, 11, 509. [57] R. Kullmer, Applied Physics A: Materials Science & Processing 1997, 65, 273. [58] N....

work page 2010

[4] [82]

K. D. Fredrickson, A. B. Posadas, A. A. Demkov, C. Dubourdieu, J. Bruley, Journal of Applied Physics 2013, 113, 184102. [83] D. P. Cann, J.-P. Maria, C. A. Randall, Journal of Materials Science 2001, 36, 4969. [84] H. P. Nair, Y. Liu, J. P. Ruf, N. J. Schreiber, S.-L. Shang, D. J. Baek, B. H. Goodge, L. F. Kourkoutis, Z.-K. Liu, K. M. Shen, D. G. Schlom, ...

work page 2013

[5] [109]

,’() [Å] 𝑎∥,’() [Å] 𝑎

J. W. Arblaster, Platinum Metals Review 2007, 51, 130. [110] R. Bechmann, The Journal of the Acoustical Society of America 1956, 28, 347. Figure 1: Electrode candidates demonstrated in metal-ferroelectric BaTiO3-metal heterostructures[18,25,37–71]. (a) The number of reports (noted above each bar) of hysteretic P-E curves (solid bar, left) and those with a...

work page 2007