{"id":"5b8a022f-c950-4bb6-9688-93fd33ed47a7","arxiv_id":"2506.07573","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Non-switchable polarization from defect-pinned domains, not fixed dipoles or internal bias alone, drives asymmetric pyroelectric and electrocaloric hysteresis in PZT films.","lead":"Measurements of pyroelectric current and electrocaloric temperature change in PZT capacitors show that non-switchable polarization from pinned ferroelectric domains creates asymmetric responses and voltage shifts. This points to defect engineering and poling protocols as ways to boost directional performance in solid-state cooling and energy harvesting.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The endpoint decomposition Y(±10 V)=±Ys+Yns is the load-bearing step: if ±10 V is not a fully saturated switchable state, the response-axis asymmetry attributed to non-switchable polarization is not uniquely identified.","rationale":"The reader identified the same fragile assumption: Sec. 2's decomposition Y(±10 V)=±Ys+Yns presupposes fully saturated, antipodal switchable states and a constant non-switchable offset. I agree, and this is the single load-bearing point rather than a minor calibration detail. Every downstream conclusion—existence of non-switchable polarization (Fig. 2b), conversion of non-switchable to switchable under bipolar pulses (Fig. 4a), correlation of Vsh with Yns/Ys (Fig. 4b,d), and repinning during aging (Fig. 4c)—is computed from endpoint half-sums and half-differences of the same loops. If the loops are not saturated at ±10 V, a pure internal-bias shift or asymmetric branch slope produces an apparent vertical offset (the paper's own Fig. 2a cautions about this for Vsh), so the vertical asymmetry would not uniquely imply non-switchable polarization. The Voffset=1.8 V measurement centers the coercive voltages but does not establish endpoint saturation.\n\nThe paper has compensating strengths: direct PEE and ECE hysteresis measurements, Maxwell-relation consistency between the two responses, error bars on the reported coefficients, and a clear qualitative demonstration that the residual vertical asymmetry survives Voffset compensation. These make the work worth conditional acceptance, not rejection. However, the central mechanistic conclusion would collapse if a saturation check shows Ys still rising above ±10 V or Yns varying with sweep amplitude. That is exactly the check proposed.\n\nTherefore the reader's CONDITIONAL verdict is appropriate and unchanged.","tokens_in":6935,"tokens_out":8709,"duration_ms":111399,"concrete_test":"Run the pulse-number and aging sequences while acquiring full Ip–V and P–V hysteresis loops at sweep amplitudes of ±5, ±10, ±15, and ±20 V under identical poling history. Plot the extracted |Ys| and |Yns| (using the Sec. 2 half-sum/half-difference rule) versus Vmax. If |Ys| continues to grow above ±10 V, or if |Yns| varies with Vmax or with sweep direction, the endpoints are not saturated and the central pulse, aging, and Vsh comparisons are biased. Re-extract trends from the saturated plateau and re-test whether the Vsh–|Yns/Ys| correlation survives. Also verify with P–V loops that the switchable polarization at ±10 V equals the saturated value at higher fields.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central mechanism rests on isolating a non-switchable polarization component from the response-axis asymmetry of Ip–V and ΔTsens–V loops. Section 2 defines Y(±10 V)=±Ys+Yns and extracts Ys and Yns as the half-difference and half-sum of the endpoint values. This is valid only if +10 V and −10 V are fully saturated, antipodal switchable states and if Yns is identical at both endpoints. No saturation check for the ±10 V window is reported; the loops in Fig. 3 are not shown to close at the endpoints, and ±10 V is only about three times the coercive voltage for a 1-μm PZT film. If the switchable branches are still rising at ±10 V, the half-sum contains an apparent vertical offset from branch curvature or slope asymmetry, and the half-difference underestimates the true switchable response. The Voffset=1.8 V correction in Fig. 2(b) removes only the coercive-voltage center shift; it does not guarantee endpoint saturation, so the residual Ip-axis asymmetry can still be an unsaturated-loop artifact rather than a genuine non-switchable polarization. Because the pulse-number dependence (Fig. 4a), the Vsh versus |Yns/Ys| comparison (Fig. 4b,d), and the aging interpretation (Fig. 4c) all use Ys and Yns from this same decomposition, a bias here propagates into every quantitative conclusion, including the claim that bipolar pulses convert pinned polarization into switchable response. The qualitative asymmetry is real, but its assignment to non-switchable polarization rather than to finite-voltage branch asymmetry is not yet established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports direct hysteresis measurements of the pyroelectric current (Ip) and the electrocaloric temperature change (ΔTsens) as functions of DC bias in 1-μm-thick Pb(Zr0.65Ti0.35)O3 capacitors. By decomposing the responses at ±10 V into switchable and non-switchable components, and by applying a DC voltage offset to compensate the internal-bias-induced voltage shift, the authors identify a residual response-axis asymmetry that they attribute to non-switchable polarization. Bipolar pulse cycling is shown to increase the switchable components while decreasing the non-switchable components, which is interpreted as depinning of previously pinned domains. After 100 pulses, time-dependent aging reduces the switchable component while the non-switchable component remains nearly constant, and the voltage-axis shift Vsh evolves in parallel with the ratio |Yns/Ys|. The central claim is that non-switchable polarization originates from defect-induced domain pinning, that this pinned polarization contributes to the voltage shift, and that controlled poling can convert it into switchable response for enhanced PEE and ECE.","tokens_in":7258,"tokens_out":4917,"duration_ms":66122,"significance":"If the central claim holds, the work provides a direct experimental route to separate switchable and non-switchable polarization contributions in pyroelectric and electrocaloric hysteresis, and it offers a concrete mechanism—domain pinning by defects—connecting response-axis asymmetry, voltage-axis shift, and poling history. The simultaneous measurement of PEE and ECE on the same device and the consistency of their pulse-cycling dependencies are notable strengths, as is the absence of fitted free parameters in the main decomposition beyond the experimentally determined offset voltage and heater calibration. The paper also makes a falsifiable prediction: bipolar pulse cycling should convert pinned polarization into switchable response, and aging should partially reverse this conversion. If the endpoint decomposition is validated, the conclusions would be of practical interest for directional poling strategies in ferroelectric coolers and energy harvesters.","major_comments":[{"comment":"The decomposition Y(±10 V) = ±Ys + Yns is the load-bearing step of the paper, but the manuscript does not establish that +10 V and −10 V are fully saturated, antipodal switchable states with a constant non-switchable contribution at both endpoints. The loops in Fig. 3 are not shown to close at ±10 V, and ±10 V is only about three times the coercive voltage for a 1-μm PZT film. If the switchable branches are still rising at ±10 V, or if Yns depends on voltage or history, then the half-sum and half-difference do not cleanly isolate Yns and Ys. Because the pulse-number dependence in Fig. 4(a), the Vsh-versus-ratio comparison in Figs. 4(b) and 4(d), and the aging interpretation in Fig. 4(c) all use Ys and Yns extracted from this same decomposition, a bias here propagates into every quantitative conclusion. Please report saturation checks—for example, loop closure at the endpoints or measurements extended to higher voltages—and quantify the sensitivity of Ys and Yns to the chosen endpoint voltage.","section":"Section 3.1, Fig. 2(b)"},{"comment":"This is a separate issue from endpoint saturation: even with perfectly saturated endpoints, the extraction of Yns by half-sum at ±10 V will contain an apparent vertical offset if the voltage-axis shift is not fully compensated. The paper needs to clarify whether Ys and Yns are computed from raw loops, from Voffset-compensated loops, or from loops with Voffset re-adjusted for each condition.","section":"Section 3.1, Fig. 2(b)"},{"comment":"Additionally, the 'agrees well' correlation in Fig. 4(d) is presented without a quantitative metric such as a correlation coefficient, slope, or uncertainty estimate. Given that Vsh and |Yns/Ys| are extracted from the same loops, a quantitative assessment is needed to support the claim that the voltage shift is governed by the non-switchable-to-switchable ratio rather than by an unrelated drift of the loop center.","section":"Section 3.3, Figs. 4(c) and 4(d)"},{"comment":"","section":"Section 4, Conclusions"}],"minor_comments":[{"comment":"","section":"Section 2, Eq. (2)"},{"comment":"","section":"Section 2, Eq. (1)"},{"comment":"","section":"Fig. 4"},{"comment":"","section":"Section 3.2, Fig. 3"},{"comment":"","section":"Section 1, References"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the scope of a specialty condensed-matter or applied-physics journal. The stress-test concern about endpoint saturation is genuine and load-bearing; however, it can be addressed with additional saturation checks and a clearer specification of the compensation protocol, so a major revision rather than rejection is appropriate. The manuscript would also benefit from a more quantitative treatment of the Vsh-versus-ratio correlation and from clarifying the time-evolution protocol for the aging data."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe short version: this paper reports direct hysteresis loops of pyroelectric current and electrocaloric temperature change in 1-µm PZT capacitors and splits each response into switchable and non-switchable parts using the half-difference and half-sum of the ±10 V endpoints. The qualitative story is that the non-switchable component comes from defect-pinned domains, not fixed dipoles, because bipolar pulse cycling converts non-switchable into switchable response, and the voltage shift tracks the ratio of non-switchable to switchable components. That is a concrete and testable claim, and it is new relative to the prior indirect inference from poling-direction asymmetry.\n\nWhat the paper does well: two independent probes (PEE and ECE) give consistent pulse-number and aging trends, which is reassuring because they have different noise sources and systematic corrections. The offset-compensation step (Voffset = 1.8 V) is a clean way to separate an internal-bias voltage shift from a genuine response-axis shift, and the residual Ip asymmetry they find after compensation is a straightforward observation. The distinction between fixed dipoles and domain pinning via pulse cycling is also well motivated, and the simultaneous growth of switchable and shrinkage of non-switchable components is a nice fingerprint of depinning.\n\nThe soft spots are real but not fatal. The decomposition assumes that ±10 V corresponds to fully saturated, antipodal switchable states and a constant non-switchable contribution. No saturation check is reported, and ±10 V is only about three times the coercive voltage for a 1-µm film. If the switchable branches are still rising at the endpoints, the half-sum contains a slope-curvature artifact and the half-difference underestimates Ys. That bias would propagate into the pulse-number trends, the Vsh-versus-ratio comparison, and the aging interpretation, all of which are the paper's quantitative payload. The stress-test note is right about this. That said, the qualitative asymmetry is real, and the direction of the trends would likely survive a more careful saturation check; the issue is the clean quantitative separation, not the whole mechanism. The other soft spot is the lack of error bars or multi-sample data on the central trends; Fig 4 would be more convincing with a second sample and uncertainty estimates. The Voffset choice is also somewhat circular in that it centers the coercive voltages, but the residual vertical offset is then the evidence, so the circularity is mild.\n\nWho is this for: anyone working on pyroelectric energy harvesting, electrocaloric cooling, or imprint in ferroelectric films. The paper is a solid experimental contribution, and the mechanism is plausible.\n\nI would send it to review. A good referee should ask for a saturation check (loop closure at ±10 V, or a measurement at ±15 V), error bars on the extracted components, and a discussion of how the decomposition would change under partial saturation. With those, the paper would be much stronger.","headline":"Direct PEE/ECE hysteresis measurements that separate switchable and non-switchable polarization in PZT, with a plausible defect-pinning mechanism that is somewhat more model-dependent than the paper suggests.","tokens_in":7789,"tokens_out":3456,"would_cite":true,"duration_ms":39884,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["77.70.-a","77.80.Dj"],"model":"deepseek-v4-flash","headline":"Defect-pinned domains, not internal bias alone, produce the non-switchable polarization that makes pyroelectric and electrocaloric hysteresis loops asymmetric in PZT capacitors.","keywords":["pyroelectric effect","electrocaloric effect","ferroelectric thin films","non-switchable polarization","domain pinning","imprint voltage shift","PZT capacitors","bipolar pulse poling"],"falsifier":"Measure the pyroelectric hysteresis loop at several maximum voltages beyond ±10 V (for example ±15 V and ±20 V) on the same sample after identical poling, and test whether the extracted non-switchable component $Y_{\\rm ns}$ stays constant while the switchable component $Y_{\\rm s}$ saturates. If $Y_{\\rm ns}$ changes with the sweep range or with sweep history, the assumed decomposition $Y(\\pm 10\\,\\mathrm{V}) = \\pm Y_{\\rm s} + Y_{\\rm ns}$ is not valid and the reported ratio-to-$V_{\\rm sh}$ correlation could be an artifact.","tokens_in":6743,"feed_emoji":"⚡","tokens_out":8572,"duration_ms":91118,"temperature":0.7,"pith_summary":"This paper reports direct measurements of the pyroelectric current and the electrocaloric temperature change as hysteresis loops against bias voltage in 1-micrometer-thick lead zirconate titanate (PZT) capacitors. Both loops are asymmetric along the voltage and response axes, and the paper argues that the response-axis asymmetry cannot be explained by an internal bias alone. Using the values at +10 V and −10 V, the authors isolate a non-switchable polarization component and show that bipolar pulse poling converts it into switchable polarization, identifying defect-induced domain pinning as its origin. They also show that the voltage-axis shift tracks the ratio of non-switchable to switchable polarization, and that aging after pulsing partially re-pins the domains. If this picture is right, pyroelectric and electrocaloric performance can be directionally optimized through controlled poling and defect engineering.","feed_headline":"Pinned domains, not bias alone, shape ferroelectric response","feed_subtitle":"Direct hysteresis loops show defect-pinned domains create non-switchable polarization and voltage shifts in PZT films.","key_machinery":"The load-bearing object is the two-component decomposition of the measured response, $Y(\\pm 10\\,\\mathrm{V}) = \\pm Y_{\\rm s} + Y_{\\rm ns}$, applied to $Y = I_{\\rm p}$ (pyroelectric current) and $Y = \\Delta T_{\\rm sens}$ (electrocaloric temperature modulation). The switchable part $Y_{\\rm s}$ is the half-difference of the responses at the two sweep endpoints and the non-switchable part $Y_{\\rm ns}$ is the half-sum. The paper combines this decomposition with two operations: applying a DC offset voltage to cancel the voltage-axis shift, which reveals the residual response-axis asymmetry, and applying bipolar triangular pulses, which redistribute charges and remove defect dipoles. The correlated evolution of the ratio $|Y_{\\rm ns}/Y_{\\rm s}|$ with the voltage-axis shift $V_{\\rm sh}$ under pulsing and aging is the evidence that ties the voltage shift to pinned domains.","core_discovery":"The central claim is that the polarization-axis asymmetry in pyroelectric and electrocaloric hysteresis in PZT thin films is caused by a non-switchable polarization component, and that this component arises from defect-induced domain pinning rather than from fixed dipoles or from a pure internal bias. The authors decompose the measured response $Y$ at the sweep endpoints as $Y(\\pm 10\\,\\mathrm{V}) = \\pm Y_{\\rm s} + Y_{\\rm ns}$, where $Y_{\\rm s}$ is the switchable contribution and $Y_{\\rm ns}$ is the non-switchable one. Under repeated bipolar triangular pulses, $|Y_{\\rm s}|$ grows while $|Y_{\\rm ns}|$ shrinks by comparable amounts, which is the signature of pinned domains being released and becoming switchable. The voltage-axis shift $V_{\\rm sh}$ changes together with the ratio $|Y_{\\rm ns}/Y_{\\rm s}|$ both during pulsing and during subsequent aging, so the same pinned domains that create the non-switchable response also create the imprint-like voltage shift. The paper concludes that pyroelectric and electrocaloric output can be enhanced in a preferred direction by depinning domains through controlled poling and by engineering defects.","pith_inferences":["An extension the paper leaves implicit is that the same $\\pm V$ decomposition could be applied to other ferroelectric systems, such as hafnia-based films, to test whether their wake-up effects involve convertible non-switchable polarization; the paper notes that hafnia shows pyroelectric enhancement without electrocaloric enhancement, so simultaneous measurement of both signals would discriminate ","A testable consequence of the defect-pinning picture is that the non-switchable fraction should increase when oxygen vacancies or acceptor dopants are introduced deliberately, and should decrease after field cycling; measuring $|Y_{\\rm ns}/Y_{\\rm s}|$ under controlled defect concentrations would tie the ratio to a specific microscopic quantity.","Because the decomposition assumes full saturation at $\\pm 10\\,\\mathrm{V}$, a consistency check is to repeat the extraction at different sweep amplitudes; if $Y_{\\rm ns}$ stays constant while $Y_{\\rm s}$ saturates, the method would generalize to other thicknesses and compositions, and if not, the quantitative conclusions would need a voltage-dependent correction."],"forward_implications":["Direct pyroelectric and electrocaloric hysteresis measurements can separate switchable from non-switchable polarization contributions in ferroelectric films, something conventional polarization-voltage loops cannot do because integration hides the non-switchable part.","Bipolar pulse poling converts pinned, non-switchable domains into switchable response, increasing both the pyroelectric and electrocaloric signals together, as required by Maxwell's relation.","The voltage-axis shift in hysteresis loops can originate from pinned domains rather than purely from interfacial screening, so interpreting imprint as an internal bias alone can be misleading.","Aging after pulse poling partially re-pins domains, so the switchable contribution decays while the non-switchable component stays nearly constant, meaning device performance will drift with time.","Pyroelectric measurements can serve as a reliable probe of polarization dynamics because their pulse-number and time dependences match the electrocaloric response."],"supporting_citations":[{"why":"Reports the direction-dependent pyroelectric enhancement that this paper sets out to explain; supplies the previous observation and the poling-time correlation.","marker":"[10]"},{"why":"Proposes the interfacial screening model for imprint, the main alternative explanation that must be distinguished from domain pinning.","marker":"[11]"},{"why":"Presents the defect-dipole model of domain stabilization in perovskite ferroelectrics, a basis for the pinning mechanism invoked here.","marker":"[12]"},{"why":"Shows alignment of defect dipoles in polycrystalline ferroelectrics, supporting the idea that defect dipoles can stabilize a preferred polarization direction.","marker":"[13]"},{"why":"Shows that oriented defects clamp domain walls, providing the mechanism by which pinned domains raise the switching field and create a voltage shift.","marker":"[16]"},{"why":"Supplies the direct measurement method for pyroelectric and electrocaloric effects in thin films that the present experiments are built on.","marker":"[18]"},{"why":"Shows that electric field cycling de-ages Fe-doped PZT by releasing pinned domains, justifying the bipolar pulse treatment used here.","marker":"[19]"},{"why":"Documents how field cycling evolves electromechanical properties through defect redistribution, supporting the pulse-number dependence observed in PEE and ECE.","marker":"[20]"}],"fun_headline_variants":["Pinned domains dictate pyroelectric and electrocaloric asymmetry","Defect-pinned domains drive PZT pyroelectric and electrocaloric response","Non-switchable polarization from pinned domains drives PZT response","Pinned domains, not bias, control pyroelectric and electrocaloric effects"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative conclusions assume that at +10 V and −10 V the switchable response is fully saturated and linear, and that the non-switchable component is constant across the voltage sweep; if either assumption fails, the extracted split and its correlation with the voltage shift would be biased.","fun_headline_variants_meta":{"raw":{"variants":["Pinned domains dictate pyroelectric and electrocaloric asymmetry","Defect-pinned domains drive PZT pyroelectric and electrocaloric response","Non-switchable polarization from pinned domains drives PZT response","Pinned domains, not bias, control pyroelectric and electrocaloric effects"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001478,"raw_usage":{"total_tokens":6012,"prompt_tokens":1091,"completion_tokens":4921,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":707,"completion_tokens_details":{"reasoning_tokens":4842}},"tokens_in":707,"tokens_out":4921,"duration_ms":40428,"temperature":1.0,"reasoning_tokens":4842,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T05:30:03.209458+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the pyroelectric hysteresis loop at several maximum voltages beyond ±10 V (for example ±15 V and ±20 V) on the same sample after identical poling, and test whether the extracted non-switchable component $Y_{\\rm ns}$ stays constant while the switchable component $Y_{\\rm s}$ saturates. If $Y_{\\rm ns}$ changes with the sweep range or with sweep history, the assumed decomposition $Y(\\pm 10\\,\\mathrm{V}) = \\pm Y_{\\rm s} + Y_{\\rm ns}$ is not valid and the reported ratio-to-$V_{\\rm sh}$ correlation could be an artifact.","supporting_citations":[{"cited_title":"Usami, Y","cited_arxiv_id":null,"evidence_quote":"Reports the direction-dependent pyroelectric enhancement that this paper sets out to explain; supplies the previous observation and the poling-time correlation."},{"cited_title":"Grossmann, O","cited_arxiv_id":null,"evidence_quote":"Proposes the interfacial screening model for imprint, the main alternative explanation that must be distinguished from domain pinning."},{"cited_title":"Lambeck and G.H","cited_arxiv_id":null,"evidence_quote":"Presents the defect-dipole model of domain stabilization in perovskite ferroelectrics, a basis for the pinning mechanism invoked here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows alignment of defect dipoles in polycrystalline ferroelectrics, supporting the idea that defect dipoles can stabilize a preferred polarization direction."},{"cited_title":"Robels and G","cited_arxiv_id":null,"evidence_quote":"Shows that oriented defects clamp domain walls, providing the mechanism by which pinned domains raise the switching field and create a voltage shift."},{"cited_title":"Direct Measurement of Pyroelectric and Electrocaloric Effects in Thin Films","cited_arxiv_id":null,"evidence_quote":"Supplies the direct measurement method for pyroelectric and electrocaloric effects in thin films that the present experiments are built on."},{"cited_title":"Genenko, Hans Kungl, Ljubomira Ana Schmitt, and Torsten Granzow","cited_arxiv_id":null,"evidence_quote":"Shows that electric field cycling de-ages Fe-doped PZT by releasing pinned domains, justifying the bipolar pulse treatment used here."},{"cited_title":"Evolution of electromechanical properties in Fe-doped (Pb,Sr)(Zr,Ti)O 3 piezoceramics","cited_arxiv_id":null,"evidence_quote":"Documents how field cycling evolves electromechanical properties through defect redistribution, supporting the pulse-number dependence observed in PEE and ECE."}],"review_version":1}