Review Reports

Comment 1: The reason that the choice of ITO-coated quartz substrates has been justified in terms of lattice matching, electrical properties, and potential influence on leakage current suppression should be more detailed.

Response 1: Thank you for pointing this out. We agree with this comment. Therefore, we have decided to add this part to the document.

The polycrystalline ITO layer, with its cubic lattice constant of approximately 10.12 Å, not only provides an intermediate template that bridges the lattice mismatch between amorphous quartz and perovskite BiFeO₃ (a ≃ 3.96 Å)—thereby reducing misfit dislocations and promoting high-quality crystallinity—but also functions as a transparent bottom electrode with a conductivity of around 10³ S·cm⁻¹, ensuring a uniform electric field during polarization cycling.
We have updated the manuscript accordingly—the changes can be found in lines 114–122.

Comments 2: How was the thickness ratio precisely controlled during pulsed laser deposition, and what was the error margin in determining the actual layer thicknesses?

Response 2: Previous experiments conducted by our group on BiFeO₃ and SrTiO₃ targets focused on quantifying deposition rates (nm/min) by systematically varying pulsed laser deposition parameters, particularly the laser fluence. After controlled ablation intervals, film thicknesses were measured using a profilometer, allowing for the precise determination of growth rates. In parallel, plasma diagnostics with a Langmuir probe provided real-time data on ion current density and electron temperature, which were directly correlated with the deposition kinetics. Through analysis of thickness profiles at different ablation durations, reproducible growth ratios were established, ensuring consistent film quality throughout all experimental runs.

The error margin in determining the actual layer thickness was on average +/- 2.59 nm.

We have updated the manuscript accordingly—the changes can be found in lines 109–113 and line 127.

Comment 3: Were the grain size variations correlated with the piezoelectric coefficient statistically significant?

Response 3: No. Although the Pearson correlation coefficient for our films is quite high, it indicates a strong linear trend between grain size and the piezoelectric coefficient, the associated p-value is above the common 0.05 threshold. In other words, we can’t reject the null hypothesis of no true correlation, so these results aren’t statistically significant, and more data are required.

Comment 4: Casting (Adv. Compos. Hybrid Mater. 2022, 5 (4), 2906-2920) is an important method for preparing functional films. These works should be compared with your preparation method to highlight the importance of your work.

Response 4: We appreciate the reviewer’s comment and the opportunity to clarify our methodological choice. In our introduction and literature review, we emphasized the body of work showing that pulsed laser ablation (PLD) consistently yields the highest-quality BiFeO₃ thin films. Prior studies have demonstrated its advantages in preserving stoichiometry, achieving high crystallinity, reducing surface roughness to the sub-nanometer scale, and enhancing functional responses such as ferroelectricity and piezoelectricity.

We compared different fabrication routes, including Ni-chain acrylate casting, which is suitable for composite films with EMI-shielding applications but lacks precise control over crystallinity and stoichiometry. In contrast, laser ablation directly transfers material from the BiFeO₃ target to the substrate under controlled conditions, allowing active tuning of film thickness, grain size, and orientation. This ensures phase-pure, dense oxide films free of polymer residues or secondary contamination.

Because our study focuses on BiFeO₃ as a functional oxide for device-level applications, the ability to reproducibly grow crystalline, stoichiometric films was critical. For these reasons, and supported by the strong consensus in the literature, we adopted laser ablation as the most reliable and effective fabrication method for our work.

Comment 5: How was the polarization fatigue behavior of the multilayer films?

Response 5: Thank you for highlighting the importance of polarization fatigue behavior. At this time, we have not yet performed fatigue measurements on these multilayer films and therefore cannot provide results in the current manuscript.

Comment 6: Could strain engineering effects arising from the lattice mismatch between BiFeO₃ and SrTiO₃ have contributed to the enhanced piezoelectric response?

Response 6: Yes. The effects of strain engineering induced by the reticular mismatch between BFO and STO can significantly contribute to an improved piezoelectric response, and there is experimental evidence that supports this statement. Figure 12 (b) demonstrates directly that the reticular mismatch imposes an epitaxial tension that translates into variations in c/a, and these, in turn, control the magnitude of the d_33. The graph shows that when c/a moves away from the bulk value d_33, it tends to increase.

Comment 1: Can authors elaborate more on why SETiO3 was used in this work?

Response 1: We thank the reviewer for the opportunity to clarify our substrate selection. SrTiO₃ (STO) was chosen for three principal reasons:

1. Lattice matching and strain engineering:

·       STO crystallizes in the cubic perovskite structure with a lattice parameter of 3.905 Å, nearly identical to the pseudocubic BiFeO₃ lattice (≈ 3.96 Å).

·       The resulting misfit of only ~1.4 % imposes a well‐controlled biaxial strain on the BiFeO₃ film, tuning its tetragonality (c/a) and thereby enhancing spontaneous polarization and piezoelectric response.

1. Electrical insulation under high fields:

·       STO exhibits extremely low leakage currents and high dielectric breakdown strength.

·       This enables the application of large electric fields across our BiFeO₃ films without parasitic conduction, ensuring accurate ferroelectric and piezoelectric measurements.

1. Chemical and structural compatibility

·       As a thermodynamically stable, lattice‐matched oxide, STO promotes the growth of phase‐pure BiFeO₃ with minimal secondary phases.

Together, these attributes make SrTiO₃ the optimal platform for exploring strain‐mediated enhancements in BiFeO₃’s functional properties.

Comment 2: Can authors explain more about the thickness ranges for both compounds?

Response 2: Our selections were driven by two goals: (i) to probe the critical strain state in BiFeO₃ and (ii) to ensure sufficient material volume for reliable structural and piezoelectric characterization.

1. BiFeO₃ layer (37–73 nm):

·       We bracketed the critical relaxation threshold (~50 nm) at which coherently strained BiFeO₃ begins to partially relax.

·       The lower bound (37 nm) lies well below this threshold, guaranteeing fully strained films for maximized tetragonality.

·       The upper bound (73 nm) extends above the threshold, permitting us to observe the onset of misfit dislocations and correlate them with changes in XRD, TEM, and piezoresponse force microscopy (PFM) signals.

·       This span thus captures the full transition from fully strained to relaxed regimes in a single study.

1. SrTiO₃ buffer layer (0–51 nm):

·       We varied STO thickness from zero (direct growth on bare substrate) up to 51 nm to systematically tune strain transfer into the overlying BiFeO₃.

·       Literature precedent (e.g., Auciello et al. [34], 0–48 nm) demonstrates that such buffer‐layer variations strongly modulate in‐plane lattice parameters and hence functional response.

·       Extending to 51 nm allowed us to slightly exceed prior ranges and confirm the plateau in strain‐mediated enhancement of BiFeO₃’s polarization and piezoelectric coefficient.

By selecting these complementary ranges, we balance the need for high‐fidelity strain control (near the 50 nm relaxation limit) with adequate film thickness for XRD peak analysis, cross‐sectional TEM imaging, and robust PFM amplitude and phase mapping.

Comment 3: For the characterization section, can authors add more details for each test? I also suggest that authors separate each test into subgroups, so it is easier for readers to follow.

Response 3: Thank you for the helpful suggestion to spell out each test’s procedure and add subgroup headings. We gave it a close look and feel that the current description already strikes the right balance: it covers all the essential details—instrument models, measurement conditions, and data-acquisition settings—while keeping the narrative concise and avoiding unnecessary repetition.
On the other hand we have taken into account the suggestions of adding subsections to make the clearest article for the reader. The following subsections were added within the manuscript body:

3.1.1 Rietveld Refinement Line 193

3.1.2 Grain Size Measurements Line 215

3.3 Behaviors Mechanisms Line 276

Comment 4: I noticed that all results presented in tables did not have +/-, which led to my concerns about the reliability of the results. Did you test the samples in repetition? If so, how many repetitions did you perform for each sample?

Response 4: We thank the reviewer for highlighting the need for statistical reliability in our tabulated results. In the revised manuscript, we will:

• Report numerical values with their associated ± standard deviations.
• Clarify that each data point represents the average of at least three independent measurements taken in different regions of the same sample.

These changes will be reflected in all relevant tables and noted explicitly in the Experimental Methods section.

Comment 5: Figures of plots are too small. Can authors improve them? It is very hard for us to read all the legends, axis title, etc...

Response 5: We thank the reviewer for highlighting the difficulty in reading our figures. In the revised manuscript, we have improved all plots to enhance legibility:

Increased each figure’s overall dimensions so that all elements scale appropriately.

Set axis titles, tick labels, legend text, and panel labels to a minimum of 12 pt font.

Updated color schemes to maximize contrast between data series and background.

We trust that these revisions make all legends, axis titles, and other annotations clear and easy to read.

Comment 1: The authors varied p between 1 and 3, which means that the total thickness of the resulting films is different for each sample. I recommend specifying the total thickness for each sample, as the functional properties of such systems strongly depend on total thickness. The effect of thickness variation on the properties should also be discussed.

Response 1: We thank the reviewer for underscoring the importance of clearly reporting total film thickness and its impact on functional properties. In the revised manuscript, we have addressed this in two ways:

1. Addition of Thickness Data:

·       We have expanded Table 1 to include, for each sample (p = 1–3), the individual layer thicknesses (n and m) as well as the total film thickness.

1. Emphasis on Thickness Ratio (r = n/m):

·       In Section 3.3, we now discuss how the piezoelectric and dielectric enhancements correlate more strongly with the ratio r = n/m than with absolute total thickness.

Comment 2: The properties are presented as a function of the ratio r, but the exact thickness of each BFO and STO layer is not given for each sample.

Response 2: We thank the reviewer for underscoring the importance of reporting total film thickness. We have updated Table 1 to list, for each sample p = 1–3, the individual thickness of the BiFeO₃ (n) and SrTiO₃ (m) layers alongside the previously reported ratio r = n/m.

While total film thickness is now fully specified, we retain our emphasis on the thickness ratio r = n/m as the primary driver of the observed enhancements in piezoelectric and dielectric properties.

Comment 3: As the samples are polycrystalline, XRD measurements alone are insufficient to confirm the fabrication of well-defined multilayers and to rule out the formation of a BFO–STO solid solution. I suggest adding SEM or TEM analyses.

Response 3: We thank the reviewer for pointing out the limitations of XRD in distinguishing well-defined multilayers from a potential BFO–STO solid solution.

We will certainly aim to perform cross-sectional transmission electron microscopy (TEM) over the next few months. We hope to directly visualize the individual BiFeO₃ and SrTiO₃ layers, confirm sharp interfaces, and rule out any interdiffusion. The resulting micrographs and compositional profiles will be incorporated into a future revision of the manuscript to demonstrate the integrity of the multilayer architecture.

Comment 4: The authors assume that all samples exhibit a rhombohedral structure. However, since the samples consist of alternating BFO and STO layers, it should be clarified whether the reported symmetry corresponds to the BFO layers, the STO layers, or both.

Response 4: The rhombohedral R3c structure we report applies exclusively to the BiFeO₃ layers.

• The SrTiO₃ buffer layers retain their cubic Pm–3m symmetry throughout, and their diffraction reflections have been indexed accordingly.

Comment 5: BFO and STO have different bulk lattice parameters, and the samples are polycrystalline. It should therefore be clarified to which layer type the observed diffraction peaks and calculated c/a ratio correspond.

Response 5: We appreciate the reviewer’s attention to the distinction between BFO and STO diffraction signatures in our polycrystalline multilayers. In the revised manuscript, we have clarified that all reported lattice parameters and the derived c/a ratios refer exclusively to the BiFeO₃ layers.

Comment 1: The novelty is moderate because similar BFO/STO multilayer strategies have been reported previously. The manuscript could more explicitly clarify what is fundamentally new, is it the specific thickness ratio/periodicity optimization, or a new observed trend?

Response 1: In our revision, we will make it clearer that, while BFO/STO multilayer systems have indeed been reported before, the novelty of our work lies in three aspects. First, we perform a systematic r–p optimization, varying thickness ratios continuously (r = n/m = 0–1.16) and periodicities (p = 1–3), rather than limiting ourselves to isolated design points. This dense mapping uncovers previously unreported non-monotonic behaviors, such as the optimal leakage suppression at r = 0.30, p = 1, and the lowest coercive field at r = 0.45, p = 3. Second, we establish direct structure–property correlations by linking XRD-derived c/a distortions, AFM grain sizes, and PFM-measured d₃₃, showing how STO-induced interfacial stress and templating jointly tune rhombohedral distortion and grain growth. These correlations provide a predictive framework for engineering lead-free multilayers. Third, we construct a conduction-mechanism phase diagram, mapping the evolution from Ohmic/SCLC to Fowler–Nordheim tunneling as a function of r and p—an approach not previously reported for these heterostructures. Together, these contributions distinguish our study from prior reports and advance both the fundamental understanding and design strategy of BFO/STO multilayers.

We have updated the manuscript accordingly—the changes can be found in lines 422–428.

Comment 2: The introduction would benefit from a more in-depth discussion of previous work on BFO/STO heterostructures, to better position the contribution.

Response 2: Thank you for this insightful suggestion. We will conduct a more comprehensive literature survey of previous BFO/STO heterostructure studies covering multilayer architectures, interface coupling mechanisms, strain‐modulation effects, and functional property reports, and integrate an expanded discussion into the Introduction in our revised manuscript.

Comment 3: The rationale behind the chosen r and p ranges should be explained—were these based on prior literature, preliminary results, or equipment limitations?

Response 3: The selected r and p ranges were primarily guided by prior literature, as noted in the introduction, to ensure comparability with earlier studies and to explore parameter windows previously identified as relevant for functional optimization. At the same time, our own experimental observations revealed a practical limitation: as the number of periods increased, it became increasingly difficult to maintain stable deposition conditions for SrTiO₃, since the required ablation times grew significantly. This dual rationale, literature precedent and experimental feasibility, therefore, defined the chosen range of r and p values.

Comment 4: The statistical reliability of the piezoelectric coefficient (*d₃₃*) measurements is not addressed. Including standard deviations or repeated measurements would improve confidence.

Response 4: We are currently working on repeated d₃₃ measurements to assess their statistical reliability and will include the corresponding standard deviations in a future version of this work.

Comment 5: The manuscript does not discuss potential influence of substrate strain relaxation or interface intermixing; both could strongly affect the results.

Response 5: We will carefully review the potential effects of substrate strain relaxation and interface intermixing and incorporate a corresponding discussion in the revised manuscript.