{"id":"1b06bf7c-2265-4a2e-9ac9-947ac7454ab7","arxiv_id":"2604.18995","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":3,"one_line_summary":"Voltage pulses create nonvolatile, multi-state metastable CDW phases in bulk EuTe4 at room temperature via out-of-plane phase switching in its moiré superstructure.","lead":"This paper reports electrically driven, room-temperature, nonvolatile metastable states in bulk EuTe4, a material with an innate moiré superlattice from stacked incommensurate charge density waves. Multiple discrete resistivity states are induced by voltage pulses and characterized by transport, ARPES, and XRD, pointing toward multi-bit memory applications.","discovery_kind":"unclear","skeptic_critique":{"model":"glm-5.2","headline":"Joule heating confound is the most load-bearing concern: pulse-duration dependence and acknowledged thermal contributions could mean the metastable states reflect partial thermal cycling within the known giant hysteresis loop rather than novel electric-field-driven CDW phase switching.","rationale":"The reader correctly identified that the mechanistic interpretation rests on indirect evidence and that Joule heating is acknowledged but not resolved. I partially agree: the indirect-evidence concern is real but secondary. The more load-bearing issue is the Joule heating confound, which strikes at the root of the central claim. If Joule heating drives the observed metastability, the phenomenon is thermal cycling within a known hysteresis loop — still potentially useful for memory applications, but not the novel electric-field-driven CDW phase switching the paper claims. The experimental observations (nonvolatile plateaus, multi-state behavior, reversibility) are solid and well-characterized. The XRD and ARPES data effectively rule out changes in in-plane CDW periodicity. But the specific mechanistic claim — out-of-plane CDW phase domain formation driven by electric fields — is not distinguished from the simpler thermal-cycling hypothesis by the presented data. The paper would be significantly strengthened by the constant-energy pulse test described above. CONDITIONAL remains appropriate: the experimental observation is sound, but the mechanistic claim should be treated as a hypothesis pending thermal confound resolution. The paper does provide code/data availability (Ref. 42) and uses multiple complementary probes (transport, ARPES, XRD), which are genuine strengths, but none of these directly address the Joule heating confound.","tokens_in":8656,"tokens_out":2477,"duration_ms":104617,"concrete_test":"Perform voltage pulse experiments at constant deposited energy (V²·t/R) but with different V–t combinations (e.g., 14V/10μs vs. 4.4V/100μs, same energy). If the resulting metastable resistance states are identical across equal-energy pulses, Joule heating is the primary mechanism and the electric-field-driven CDW switching claim weakens substantially. If states differ systematically with peak electric field (V/d) at fixed energy, the field-driven mechanism is supported. Additionally, directly measure or calibrate the sample temperature rise during pulsing using the known R(T) hysteresis curve as an in-situ thermometer.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim is that metastable states arise from 'electric-field-induced switching of out-of-plane CDW phases.' However, the paper itself acknowledges (Results section, near Fig. 1c) that 'thermal contributions associated with Joule heating cannot be ruled out,' and the data show a 'non-negligible dependence on pulse duration' (Fig. 4b). This is the load-bearing issue: EuTe4 already possesses a giant thermal hysteresis spanning 100–500 K. If Joule heating during voltage pulses raises the sample temperature within this hysteresis loop, the observed multi-state resistivity plateaus could simply reflect partial thermal cycling to different points on the known hysteresis curve — a mundane explanation that would not require any novel electric-field-driven CDW domain mechanism. The XRD evidence (intensity suppression, FWHM broadening of CDW satellites) is equally consistent with thermally-induced CDW amplitude reduction as with field-driven phase domain formation. The paper's conclusion states the states are 'highly sensitive to electric field strength rather than pulse duration,' but Fig. 4b explicitly shows duration dependence, creating an internal tension. Without quantitatively separating field-driven from thermally-driven effects, the mechanistic claim rests on an uncontrolled confound. The eight-state domain model (Fig. 4c–d) is a theoretical enumeration with no direct structural confirmation, but this is secondary to the primary Joule heating issue: if heating is the dominant driver, the entire mechanistic framework needs revision.","agreement_with_reader":"partial"},"referee_report":{"model":"glm-5.2","summary":"This manuscript reports electrically driven, nonvolatile metastable states in bulk EuTe4, a layered CDW material hosting an innate moiré superstructure from stacked incommensurate monolayer and bilayer CDWs. The authors combine in-situ transport, ARPES, and XRD measurements to show that voltage pulses produce discrete, reversible resistivity plateaus at 300–400 K, with preserved in-plane CDW wavevectors but suppressed CDW satellite intensity and broadened FWHM. They interpret these observations as arising from electric-field-induced switching of out-of-plane CDW phases, proposing an eight-state domain model. The experimental observations themselves—plateaus, nonvolatility, reversibility, and the XRD/ARPES signatures—are clearly presented and internally consistent.","tokens_in":8905,"tokens_out":2229,"duration_ms":102029,"significance":"Room-temperature, nonvolatile, multi-bit CDW memory in a bulk crystal is a practically significant result. The multi-messenger approach integrating transport, ARPES, and XRD with in-situ pulsed excitation is a strength, as is the systematic characterization across a wide temperature window. The observation that in-plane periodicity is robustly preserved (joint lock-in) while out-of-plane order is modulated is a notable materials-physics finding. The distinction drawn between bulk and thin-flake behavior (Sec. near p. 7–8) adds physical insight.","major_comments":[{"comment":"Joule heating confound (Fig. 1c, Fig. 4b, and Supplementary Information): The manuscript acknowledges that 'thermal contributions associated with Joule heating cannot be ruled out' (Results, near Fig. 1c) and that pulse-duration dependence is 'non-negligible' (Fig. 4b). Given that EuTe4 exhibits a giant thermal hysteresis spanning 100–500 K, Joule heating during voltage pulses could raise the sample temperature within this hysteresis loop, and the observed multi-state plateaus could partially reflect thermal cycling to different points on the known hysteresis curve rather than purely field-driven CDW domain switching. The XRD evidence (intensity suppression, FWHM broadening of CDW satellites, Fig. 3c–d) is equally consistent with thermally-induced CDW amplitude reduction as with field-driven phase domain formation. This is load-bearing for the central mechanistic claim. The Supplementary","section":null},{"comment":"Information analysis is referenced but not summarized in the main text; a quantitative bound on the temperature rise during pulsing, or a control experiment (e.g., varying pulse duty cycle at fixed energy, or comparing field-driven vs. furnace-heated trajectories through the hysteresis loop), would substantially strengthen the claim that the mechanism is field-driven rather than thermal.","section":null},{"comment":"Internal tension between conclusion point (2) and Fig. 4b: The conclusion states the metastable states are 'highly sensitive to electric field strength rather than pulse duration' (p. 8, point 2). However, Fig. 4b explicitly shows a pronounced dependence on pulse duration for longer pulses (10 µs to 100 µs), and Fig. 1c (blue regime) demonstrates a 'further decrease in resistance with the application of longer pulses.' These observations directly contradict the stated conclusion. The authors should either reconcile this tension with quantitative analysis (e.g., showing that field strength dominates when energy is held constant) or revise the conclusion to accurately reflect the role of pulse duration.","section":null},{"comment":"Eight-state CDW domain model (Fig. 4c–d): The model enumerates eight possible CDW states defined by relative phases (θ1, θ2, θ3) between Te layers, but no direct structural measurement (e.g., real-space imaging or layer-resolved diffraction) confirms which specific phase combinations are realized. The model is used to interpret the XRD intensity suppression and FWHM broadening, but these observables are also consistent with simpler explanations (e.g., reduced CDW amplitude and correlation length without invoking discrete domain states). The authors should either provide additional evidence supporting the discrete eight-state picture or more clearly frame this as a proposed interpretation rather than an established mechanism.","section":null}],"minor_comments":[{"comment":"The abstract and title reference arXiv:2604.18995 (cs.CL) about diffusion LLMs, but the manuscript content is entirely about EuTe4 CDW physics (arXiv:2604.18998, cond-mat.str-el). This appears to be a metadata error that should be corrected.","section":null},{"comment":"Fig. 2 is referenced in the text (p. 7, 'exceeding a 50% drop, see Fig. 2a') but is not included in the provided manuscript text. Please ensure all figures are present.","section":null},{"comment":"The phrase 'andin-situtransport' (Abstract/Introduction) has a missing space; similar formatting issues appear elsewhere (e.g., 'andthe' in the Introduction).","section":null},{"comment":"The ARPES results are mentioned only briefly in the main text and deferred to the Supplementary Information. A short summary of the key ARPES findings (gap changes, linear voltage response) in the main text would improve readability.","section":null},{"comment":"The distinction between bulk and thin-flake behavior (p. 7–8) is important but somewhat buried. Consider a dedicated subsection or a summary table comparing the two regimes.","section":null}],"recommendation":"major_revision","confidential_remarks":"The metadata mismatch (cs.CL abstract vs. cond-mat physics content) should be flagged to the authors immediately, as it may indicate a submission error. Regarding the Joule heating issue: the authors' own acknowledgment is honest, but the current treatment is insufficient for the mechanistic claim to stand. If the authors can provide a quantitative thermal estimate or a control experiment, the paper would be substantially strengthened; if not, the mechanistic claim should be appropriately hedged. The experimental observations (plateaus, nonvolatility, reversibility, preserved q-vectors) are sound regardless of mechanism and would remain a valuable contribution."},"author_rebuttal":{"model":"glm-5.2","summary":"We thank the referee for a careful and constructive report. The referee raises three major points: (1) the Joule heating confound and whether the mechanism is field-driven versus thermal, (2) an internal tension between conclusion point (2) and the pulse-duration dependence shown in Fig. 4b, and (3) the evidentiary status of the eight-state CDW domain model. We address each below. We agree that points (1) and (3) require revision to the manuscript, and that point (2) requires a correction of the stated conclusion. We also note one standing limitation regarding direct layer-resolved structural confirmation of the domain model.","responses":[{"response":"We agree that the Joule heating question is load-bearing for the mechanistic claim and that the manuscript does not currently present sufficient quantitative analysis in the main text. We will address this in revision. Specifically, we will (i) summarize the Supplementary Information Joule heating analysis in the main text, including an order-of-magnitude estimate of the temperature rise during pulsing based on the measured current, pulse duration, sample geometry, and thermal conductivity of EuTe4; (ii) add a discussion of why the asynchronous behavior between resistivity and XRD peak intensity (Fig. 3c) disfavors a purely thermal explanation—Joule heating would be expected to suppress CDW amplitude and resistivity in concert, whereas we observe substantial resistivity changes with only weak CDW peak intensity changes at low voltages; and (iii) explicitly acknowledge in the main text that a furnace-heated control trajectory through the hysteresis loop has not been performed and that this remains a limitation. We will also add a control experiment varying pulse duty cycle at fixed total energy, which we can perform on the existing device geometry. We agree that the current phrasing ('cannot be ruled out') is insufficient and will revise the mechanistic discussion to clearly delineate what the evidence supports (field-driven domain formation) from what remains unexcluded (thermal contributions to amplitude reduction).","revision_made":"partial","referee_comment":"Joule heating confound: the manuscript acknowledges thermal contributions cannot be ruled out, and the XRD evidence is equally consistent with thermally-induced CDW amplitude reduction. A quantitative bound on temperature rise or a control experiment is needed."},{"response":"The referee is correct that the stated conclusion is too strong and is contradicted by the data shown in Fig. 4b and Fig. 1c. We will revise conclusion point (2) to accurately reflect the data. The intended meaning was that the threshold voltage for initiating switching is primarily field-controlled (i.e., below a critical field, no switching occurs regardless of pulse duration), but once above threshold, the degree of switching does depend on pulse duration, particularly for longer pulses. The current phrasing 'rather than pulse duration' misrepresents this. We will revise to state that metastable state formation requires a threshold electric field strength, but the extent of switching exhibits a non-negligible dependence on pulse duration, particularly for pulses exceeding 10 µs. We thank the referee for catching this inconsistency.","revision_made":"yes","referee_comment":"Internal tension between conclusion point (2) and Fig. 4b: the conclusion states metastable states are 'highly sensitive to electric field strength rather than pulse duration,' but Fig. 4b shows pronounced pulse-duration dependence for 10–100 µs pulses, and Fig. 1c (blue regime) shows further resistance decrease with longer pulses."},{"response":"We agree that the eight-state model is a proposed interpretation rather than an established mechanism, and that the current XRD observables (intensity suppression, FWHM broadening) do not uniquely distinguish between discrete domain states and a simpler picture of reduced amplitude and correlation length. We will revise the manuscript to frame the eight-state model explicitly as a proposed interpretation. That said, we note two points in defense of the domain picture that we will also incorporate: (i) the asynchronous behavior between resistivity and CDW peak intensity at low voltages is not naturally explained by a uniform amplitude reduction, but is consistent with domain formation preceding amplitude suppression; and (ii) the multi-step, discrete nature of the resistance plateaus is more naturally explained by discrete domain configurations than by a continuous amplitude reduction. However, we acknowledge that these arguments are indirect. We do not have layer-resolved diffraction or real-space imaging confirming specific (θ1, θ2, θ3) phase combinations, and we will state this limitation explicitly. We will also note that the eight-state enumeration follows from the symmetry of the stacking geometry (three independent relative phases, each binary) and is thus the natural set of candidate states, but that direct confirmation requires experiments beyond the scope of this work.","revision_made":"partial","referee_comment":"Eight-state CDW domain model: no direct structural measurement confirms which specific phase combinations are realized, and the XRD observables are also consistent with simpler explanations (reduced CDW amplitude and correlation length without discrete domain states)."}],"tokens_in":8703,"tokens_out":1103,"duration_ms":139365,"standing_objections":["We do not have layer-resolved structural measurements (e.g., real-space STM imaging of domain configurations or layer-resolved diffraction) that could directly confirm which specific phase combinations among the eight candidate states are realized. This is an experimental limitation that cannot be resolved within the scope of the current manuscript."]},"desk_editor":{"model":"glm-5.2","letter":"The paper reports something genuinely new: electrically driven, nonvolatile, multi-state metastability in bulk EuTe4 at 300–400 K. This is distinct from prior work on thin flakes (refs. 29–30), which showed single hidden states. The combination of in-situ transport with ARPES and XRD on the same material is well executed, and the core experimental observations — discrete resistivity plateaus, nonvolatility over ~9 hours, reversibility via thermal annealing, and preservation of in-plane CDW wavevectors — are solid and internally consistent. The XRD evidence (unchanged q-vectors, reduced intensity, broadened FWHM) is clean and the multi-messenger approach earns credit. Code and data are available (ref. 42), which is good practice. The paper is also appropriately careful in distinguishing bulk behavior from thin-flake results and gives a physical argument for why they differ (limited vs. extensive out-of-plane CDW units). That argument is reasonable. The soft spot is the Joule heating issue, and it is real. The paper itself acknowledges that thermal contributions cannot be ruled out, and Fig. 4b shows clear pulse-duration dependence — which directly contradicts the claim in the conclusion that states are 'highly sensitive to electric field strength rather than pulse duration.' Given that EuTe4 has a giant thermal hysteresis (100–500 K), Joule heating during pulses could place the sample at different points on the known hysteresis curve without any novel field-driven domain mechanism. The XRD changes (intensity suppression, FWHM broadening) are equally consistent with thermally-induced CDW amplitude reduction as with field-driven phase domain formation. The paper does not quantitatively separate field-driven from thermally-driven effects. The eight-state domain model (Fig. 4c) is a theoretical enumeration with no direct structural confirmation — this is secondary, but it means the mechanistic framework is a well-motivated hypothesis, not a confirmed result. The reader's CONDITIONAL verdict is about right. The experimental observations stand on their own; the mechanistic interpretation needs either quantitative Joule heating analysis or direct real-space domain imaging to be convincing. This paper is for condensed matter physicists working on CDW memory or moiré superstructures. It deserves a serious referee who can push on the Joule heating quantification and the internal tension between Fig. 4b and the conclusion.","headline":"First report of electrically driven, room-temperature, nonvolatile metastable CDW states in bulk EuTe4 — a real advance, but the Joule heating confound is not fully resolved and the mechanistic claim is stronger than the evidence supports.","tokens_in":9726,"tokens_out":585,"would_cite":true,"duration_ms":74828,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["71.45.Lr","72.15.Nj","64.70.Kb"],"model":"glm-5.2","headline":"Voltage pulses switch CDW phases in EuTe₄ at room temperature","keywords":["EuTe4","charge density wave","moiré superstructure","metastable states","nonvolatile memory","electric-field switching","thermal hysteresis","CDW domains"],"falsifier":"If direct structural imaging (e.g., scanning tunneling microscopy or nanobeam diffraction) of the pulsed state fails to reveal domains with the predicted phase combinations, or if controlled heating experiments reproduce the same resistivity plateaus without an electric field, the domain-switching mechanism would be undermined in favor of a thermal or defect-migration explanation.","tokens_in":8735,"feed_emoji":"⚡","tokens_out":1472,"duration_ms":47164,"temperature":0.7,"pith_summary":"This paper reports that bulk EuTe₄—a layered crystal hosting an innate moiré superlattice formed by two incommensurate charge density waves (CDWs)—exhibits electrically driven, nonvolatile metastable states at and above room temperature (300–400 K). Applying voltage pulses to the material produces discrete, step-like drops in electrical resistance, yielding multiple distinguishable resistivity plateaus that persist for hours and are fully reversible by thermal annealing. Through combined transport, angle-resolved photoemission spectroscopy (ARPES), and X-ray diffraction (XRD) measurements, the authors establish that these metastable states do not arise from new ordered phases or changes in the in-plane CDW periodicity, which remains locked. Instead, the CDW amplitude is suppressed and the correlation length is reduced, consistent with electric-field-induced switching between different out-of-plane CDW phase configurations within the moiré stacking. The paper proposes an eight-state model enumerating possible relative phases between adjacent Te layers, and interprets the multi-step resistivity changes as gradual formation of metastable CDW domains of varying three-dimensional configurations. The authors position EuTe₄ as a platform for room-temperature, multi-bit, nonvolatile memory devices.","feed_headline":"Voltage pulses switch CDW phases in EuTe₄ at room temperature","feed_subtitle":"Bulk crystal with innate moiré CDW superlattice shows nonvolatile, multi-step resistive states from 300–400 K — a candidate for multi-bit","key_machinery":"The moiré superstructure of EuTe₄, formed by the stacking of two incommensurate CDWs—a monolayer CDW with wavevector q₁ = 0.644(5)b* and a bilayer CDW with wavevector q₂ = 0.678(5)b* + 0.5c*—which are jointly locked to the lattice. The metastable states are interpreted as different out-of-plane phase arrangements (θ₁, θ₂, θ₃) of these CDW layers, with eight possible configurations enumerated in the paper's model.","core_discovery":"The central discovery is that voltage pulses can drive bulk EuTe₄ into multiple discrete, nonvolatile metastable states at 300–400 K, and that these states arise not from changes in in-plane CDW ordering but from electric-field-induced switching of out-of-plane CDW phases in the moiré superstructure. The in-plane CDW wavevectors remain locked by a joint commensuration condition (q₁ + 2q₂ = 2b), while the CDW amplitude weakens and the diffraction peaks broaden, pointing to reduced correlation length and domain formation among the eight enumerated phase configurations of the stacked monolayer and bilayer CDWs.","pith_inferences":[],"forward_implications":["EuTe₄ could serve as a room-temperature multi-bit memory material, with each resistivity plateau representing a distinct stored state, operating across a 100 K window without cryogenic cooling.","The moiré CDW stacking mechanism suggests a new design principle for nonvolatile memory: seek materials with stacked incommensurate orders where out-of-plane phase switching is accessible but in-plane periodicity is protected by joint commensuration.","The eight-state phase model, if confirmed by direct real-space imaging, would provide a combinatorial framework for predicting the number and resistance values of accessible metastable states in other stacked CDW systems.","The difference between bulk and thin-flake behavior (bulk shows multiple low-resistance states; flakes show hidden high-resistance states) implies that device geometry can be tuned to select different metastable regimes, expanding the engineering space for CDW-based memory."],"fun_headline_variants":["Electric-field switching of CDW phases in bulk EuTe4","Voltage pulses drive nonvolatile CDW states in EuTe4","EuTe4 shows multi-state CDW switching at room temperature","Out-of-plane CDW phases switched in bulk EuTe4","Room-temp voltage switches out-of-plane CDW phases in EuTe4"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The claim that metastable states arise specifically from out-of-plane CDW phase domain formation rests on indirect evidence—XRD peak intensity suppression and broadening—rather than direct real-space imaging of the proposed domains. The eight-state model is a theoretical enumeration of possible phase combinations; no structural measurement confirms which specific configurations are actually realized. The paper also acknowledges that Joule heating cannot be fully excluded as a","fun_headline_variants_meta":{"raw":{"variants":["Electric-field switching of CDW phases in bulk EuTe4","Voltage pulses drive nonvolatile CDW states in EuTe4","EuTe4 shows multi-state CDW switching at room temperature","Out-of-plane CDW phases switched in bulk EuTe4","Room-temp voltage switches out-of-plane CDW phases in EuTe4","Slashing diffusion LLM decoding steps by up to 88%","Reducing spatio-temporal redundancy accelerates diffusion LLMs","Training-free rules cut diffusion LLM decoding latency","R2-dLLM cuts redundant decoding steps for diffusion language models","Faster diffusion LLMs via decoding redundancy reduction"]},"model":"glm-5.2","effort":"high","cost_usd":0.0,"raw_usage":{"total_tokens":2001,"prompt_tokens":593,"completion_tokens":1408,"prompt_tokens_details":null},"tokens_in":593,"tokens_out":1408,"duration_ms":49039,"temperature":1.0,"reasoning_tokens":1261,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-05T10:07:08.691137+00:00","model_set":{"reader":"glm-5.2"},"falsifier":"If direct structural imaging (e.g., scanning tunneling microscopy or nanobeam diffraction) of the pulsed state fails to reveal domains with the predicted phase combinations, or if controlled heating experiments reproduce the same resistivity plateaus without an electric field, the domain-switching mechanism would be undermined in favor of a thermal or defect-migration explanation.","supporting_citations":[],"review_version":2}