{"id":"ee83ad14-180f-4f88-9ee7-c76e329e2d51","arxiv_id":"2505.24566","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A 2D PZT MEMS scanner made with a three-mask process reaches 11.5° and 4.8° optical scan angles at 54.2 kHz and 3.6 kHz, with reported Q factors of 1050 and 750.","lead":"This paper reports a two-dimensional MEMS scanning mirror driven by thin-film PZT and fabricated with a simplified three-mask process. It achieves large horizontal and vertical scan angles at resonance on one compact chip, which matters for small laser projectors, OCT, and LiDAR.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline Q factors rest on an optical frequency sweep at 12 Vp-p, where scan-angle saturation and PZT nonlinearity can distort the resonance; the intrinsic device Q is more plausibly the lower 300/640 values from the Polytec measurement, and this discrepancy is not quantitatively resolved.","rationale":"The reader's weakest-assumption analysis correctly identifies the unresolved discrepancy between the two Q measurements as the main soft spot in the central performance claim. My review agrees with that assessment. The finding is load-bearing because Q is not a minor detail: it appears in the abstract, in Section 4's characterization, and in the concluding comparison to prior PZT MEMS scanners. The optical Q values also enter the paper's narrative of 'effective energy sustainability.' However, I do not think the concern invalidates the core proof-of-concept claim. The fabricated device demonstrably produces 1D and 2D Lissajous scans with measured optical angles of 4.8 degrees and 11.5 degrees at the stated resonance frequencies and 12 Vp-p drive, and the frequency-domain FEA agrees with the measured resonance locations. The fabrication flow is described in enough detail to be checked, and the three-mask approach is coherent. The weakest piece is the high-amplitude Q extraction, not the existence of the scanner. The appropriate disposition therefore remains CONDITIONAL, with the condition being a quantitative resolution of the Q discrepancy and preferably multi-device reproducibility data. My concrete test would settle the main alternative hypothesis, namely that the high reported Q is an artifact of nonlinear high-drive operation rather than a property of the device at its operating point. I see no reason to reject the paper or to demand formal verification at this stage. The concern is about overstatement of a performance metric, not about the basic feasibility of the architecture.","tokens_in":11956,"tokens_out":3088,"duration_ms":43681,"concrete_test":"Re-measure the frequency response of the same device at multiple drive amplitudes (1, 2, 5, 8, 12 Vp-p) with a calibrated optical setup, and fit a Lorentzian or a Duffing oscillator model to each response; report half-power bandwidth and inferred Q at each amplitude. If the low-voltage optical Q converges to the Polytec values near 300 and 640, and the apparent Q rises only under high drive because of nonlinear stiffening or saturation, then the intrinsic device Q is the lower value and the paper should report it as such, with the high-amplitude values labeled as effective nonlinear values. Also repeat the measurement on at least three nominally identical devices to assess sample-to-sample variation and to test whether the reported single-device values are representative.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central performance claim explicitly includes Q=750 (vertical) and 1050 (horizontal), and these values are used in the Abstract, Section 4, and Section 6 to position the scanner among high-performance PZT MEMS devices. The values come only from the optical setup's scan-angle-versus-frequency sweeps at 12 Vp-p (Figure 8b), while the Polytec MSA-600 measurement on the same device gives Q=300.56 and 642.76 (Figures 7c-d). The paper dismisses the lower values as artifacts of the tape/piezo-disk mounting and the low 2 V drive (Section 5), but offers no quantitative model or control experiment. The optical measurement is made in the same regime where the paper itself reports scan-angle saturation, hysteresis, and piezoelectric nonlinearity (Section 5); a Q extracted from a high-amplitude nonlinear resonance by a Lorentzian half-power fit is not necessarily the linear-device Q that the headline implies. If the intrinsic Q is closer to the Polytec values, the headline Q figures are overstated, and the comparison to prior scanners loses its quantitative basis. The direct scan-angle measurements at 3.6 kHz and 54.175 kHz are not called into question by this concern, but the device's true energy-loss performance is unresolved.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a 2D PZT MEMS resonant scanner fabricated with a three-mask SOI process. The device uses an outer frame driven by four PZT cantilevers for vertical scanning at 3.6 kHz and a wing-shaped middle frame for horizontal scanning at 54.2 kHz. The authors present FEA modal analysis, fabrication details, and optical and Polytec MSA-600 characterization. They claim optical scan angles of 4.8° and 11.5°, Q factors of 750 and 1050, and bandwidth-efficiency products of 24.2 and 623 deg·mm·kHz at 12 Vp-p, and position these as among the higher values for 2D PZT MEMS scanners.","tokens_in":12232,"tokens_out":4321,"duration_ms":48459,"significance":"If the performance claims hold, the work would be a useful proof-of-concept: the three-mask process is a genuine simplification, the fabricated device is real and measured, and the simulated resonance frequencies agree closely with measurements (3.718 vs 3.6 kHz; 54.504 vs 54.175 kHz). The main value is the compact dual-axis architecture with a single die, which is relevant for OCT, LiDAR, and display applications. However, the headline Q factors and the resulting comparison to prior scanners are not yet established, because the paper reports two inconsistent Q measurements and explains the discrepancy only qualitatively.","major_comments":[{"comment":"The paper reports Q values of 750 (vertical) and 1050 (horizontal) from the optical setup (Fig. 8b) and 300.56 and 642.76 from the Polytec MSA-600 setup (Figs. 7c-d). The discrepancy is attributed to the tape/piezo-disk mounting and the 2 V excitation, but no quantitative model, control experiment, or ring-down measurement is provided. Because the optical Q is extracted at 12 Vp-p in a regime the paper itself associates with scan-angle saturation and piezoelectric nonlinearity, the higher values could be an artifact of a nonlinear frequency response rather than the intrinsic device Q. Since these Q values are used in the Abstract, Section 4, and Section 6 to claim high performance, this point must be resolved.","section":"Section 5, Figs. 7 and 8"},{"comment":"The optical scan angles (4.8° and 11.5°) are reported without measurement uncertainty, number of repeated measurements, or a statement of how many devices were tested. The conversion from projected beam width to angle is described only qualitatively. Because the bandwidth-efficiency product in Section 5 and the comparison to prior scanners in Section 6 are computed from these single values, the precision and reproducibility of the quoted angles should be documented.","section":"Section 4, Fig. 8a"},{"comment":"The conclusion that the bandwidth-efficiency product \"places it among the higher 2D piezoelectric MEMS scanners reported\" is not supported by a quantified comparison. The paper cites prior scanners in the Introduction but does not tabulate their θ_opt, D, f, or θ·D·f values. Please add a comparison table or otherwise provide the quantitative basis for this ranking.","section":"Section 6"}],"minor_comments":[{"comment":"The die dimensions are given inconsistently: the Abstract says a 7 × 4.7 mm frame, Section 2A says a 1 cm × 1 cm die and movable components occupying 8 mm × 8 mm, and later the outer frame is 4 mm × 7.5 mm. Please reconcile these numbers.","section":"Section 2A"},{"comment":"Section 5 refers to \"as shown in Figure 6\" when discussing the Polytec results; the correct reference appears to be Figure 7.","section":"Section 5"},{"comment":"The text and Fig. 8b caption describe the Q extraction as \"calculated\" without specifying the fitting procedure; please state the Lorentzian fit range and the number of frequency points used.","section":"Section 4, Fig. 8b"},{"comment":"The middle frame is said to be connected to the outer frame via primary suspension flexures \"each measuring 220 µm in length and 520 µm in width,\" which seems like an unusual aspect ratio; please clarify which dimension is length and which is width.","section":"Section 2A"},{"comment":"The FEA section reports modal frequencies but no simulated Q or harmonic response amplitudes; since the experimental Q is a key claim, a brief FEA-based damping estimate would strengthen the comparison.","section":"Section 3"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a competent device demonstration, but the Q-factor evidence is the main barrier. If the authors can provide a voltage-dependent Q measurement or a ring-down measurement that reconciles the two reported values, I would be willing to reconsider the paper favorably. The paper would also benefit from an explicit comparison table with prior 2D PZT MEMS scanners."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe useful core of this paper is a working 2D PZT MEMS scanner built with a genuinely simpler three-mask process. The device scans 4.8° vertical at 3.6 kHz and 11.5° horizontal at 54.2 kHz at 12 Vp-p, which are respectable numbers for a compact 7×4.7 mm die with a 1×1.4 mm mirror. The measured resonant frequencies agree with FEA to within a few percent, and the cascaded outer–middle–inner frame design is a sensible engineering choice. That is a credible proof of concept.\n\nThe soft spot is the quality-factor story. The optical frequency sweep at 12 Vp-p gives Q = 750 and 1050, but the Polytec measurement on the same device at 2 V gives 300 and 640. The authors attribute the gap to mounting and drive level, but they offer no quantitative model or control experiment. Worse, the 12 V data are taken in a regime where they themselves report scan-angle saturation and piezoelectric nonlinearity. Fitting a Lorentzian to a nonlinear resonance does not necessarily give the linear-device Q. Since the headline bandwidth–efficiency products and the comparison to earlier scanners are computed from the high Q values, the paper's central performance claim is presently supported only by the more favorable of two conflicting measurements.\n\nOther issues are minor by comparison. There is a single device, no error bars on scan-angle extraction (which is done by measuring the projected beam width on a screen), and the FEA on mirror deformation is plausible but not checked interferometrically. None of these kill the work; they just cap how much can be concluded.\n\nThe paper deserves serious refereeing. A good referee can push the authors to either defend the optical Q with a calibrated measurement or report the intrinsic Q honestly, to add repeatability data, and to soften the “among the highest” positioning. As it stands, I would not cite it as evidence of high-Q performance, but I would cite it as an example of a three-mask 2D PZT scanner if I were writing about fabrication-process reduction.\n\nNet: send it out, but expect revision.","headline":"Solid three-mask 2D PZT scanner demonstration, but the headline Q factors rest on the more favorable of two conflicting measurements and need a quantitative fix.","tokens_in":12750,"tokens_out":2492,"would_cite":false,"duration_ms":28290,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A two-axis PZT MEMS laser scanner made with only three lithography masks reaches optical scan angles of 11.5° and 4.8° at a 12 V peak-to-peak drive.","keywords":["laser scanning","piezoelectric MEMS","PZT thin film","2D micromirror","resonant scanner","microfabrication","three-mask process","optical coherence tomography"],"falsifier":"Drive the same chip in the optical setup at increasing voltages from below 2 V up to 12 V peak-to-peak, and separately repeat the laser-vibrometer frequency sweep with a stiff mount and in vacuum; if the extracted quality factor stays near 300 (vertical) and 643 (horizontal) rather than rising to 750 and 1050, then the headline quality factors and the bandwidth-efficiency products derived from them are measurement artifacts rather than intrinsic device properties.","tokens_in":11783,"feed_emoji":"🪞","tokens_out":9019,"duration_ms":88954,"temperature":0.7,"pith_summary":"This paper claims that a two-dimensional resonant laser scanner can be built on a single silicon chip with only three lithography masks, using thin-film PZT (lead zirconate titanate) actuators on both scan axes. The device sweeps a 1 × 1.4 mm oval mirror at 3.6 kHz vertically and 54.2 kHz horizontally, reaching optical scan angles of 4.8° and 11.5° at a 12 V peak-to-peak drive, with quality factors of 750 and 1050 measured in the optical setup. On the bandwidth-efficiency product that scanner papers use to compare devices, the reported values of 24.2 and 623 deg·mm·kHz place this design among the higher 2D PZT scanners in the literature it cites. If those figures hold, the design offers OCT, LiDAR, and display systems a low-voltage, compact scanning element at lower fabrication complexity than the six-mask process it descends from. The paper supports the claim with finite-element analysis, a detailed three-mask fabrication flow, and optical and vibrometer characterization.","feed_headline":"Three masks build a 2D laser scanner that hits 11.5° at 12 volts","feed_subtitle":"PZT MEMS mirror scans two axes with quality factors of 750 and 1050 at just 12 V peak-to-peak.","key_machinery":"The load-bearing mechanism is the mechanically coupled dual-axis architecture built from three cascaded frames. The outer frame is anchored to the substrate by torsional beams and driven by four PZT cantilevers through folded beams, producing the low-frequency vertical resonance. The middle frame is shaped like four wings, each carrying PZT drive electrodes, and its out-of-phase oscillation with the inner frame produces the high-frequency horizontal resonance. The inner frame carries the 1 × 1.4 mm oval mirror and mechanically isolates it from the torsional flexures, while a 175 µm D-D-shaped silicon rim on the backside stiffens both mirror and outer frame against dynamic deformation. The fabrication route is the three-mask SOI process itself: top-electrode patterning, PZT etch, and a backside deep reactive ion etch that releases the structure, which is the claimed reduction from the six-mask process used in the group's prior scanners.","core_discovery":"The central claim is that mechanically coupling the two scan axes in a three-cascaded-frame structure — an outer frame that resonates on torsional beams for the vertical sweep, a four-wing middle frame with PZT actuators that drives the inner frame for the horizontal sweep, and an inner frame that decouples the mirror from the flexures — enables a compact 2D PZT resonant scanner to be fabricated with a three-mask process. The paper reports measured optical scan angles of 4.8° (vertical) and 11.5° (horizontal) at resonances of 3.6 kHz and 54.175 kHz, driven with 12 Vp-p periodic pulses. Quality factors of 750 (vertical) and 1050 (horizontal) are extracted from the optical frequency sweep, giving bandwidth-efficiency products of 24.2 and 623 deg·mm·kHz that the authors state are among the higher values reported for 2D PZT-MEMS scanners. Finite-element modal analysis places the vertical mode at 3.718 kHz and the horizontal mode at 54.504 kHz, in close agreement with measurement, with stresses well below the fracture strength of single-crystal silicon.","pith_inferences":["If the lower vibrometer quality factors (about 301 and 643) are closer to the intrinsic device values than the optical ones, the actual bandwidth-efficiency products would be roughly two to three times smaller, and the claim of ranking among the higher 2D PZT scanners would need re-benchmarking.","The scan-angle saturation near 12 Vp-p suggests the reported angles are near the practical ceiling for this actuator geometry, so the authors' listed future changes (thicker device layer, modified folded beams) are the more plausible route to larger angles than raising drive voltage.","Because the paper cites competing piezoelectric materials such as AlScN and KNN, the same three-frame geometry could serve as a test bed: building the identical structure with a different piezoelectric film would separate the material's contribution from the architecture's contribution to scan angle and quality factor.","The decoupling inner frame and stiffening rim, validated by finite-element analysis for low mirror deformation, are the design features most transferable to other resonant scanner platforms, since they solve the general problem of keeping a mirror flat while allowing large angular motion."],"forward_implications":["Reducing the process from six masks to three lowers fabrication cost and alignment risk, which matters for moving resonant scanners from lab demonstrations toward production.","A 12 V peak-to-peak drive voltage is compatible with compact display and handheld imaging electronics, since no high-voltage driver stage is needed.","The horizontal bandwidth-efficiency product of 623 deg·mm·kHz places the device in the range the paper's cited reviews associate with high-resolution laser projection and imaging scanners.","The close agreement between simulated resonances (3.718 kHz and 54.504 kHz) and measured ones (3.6 kHz and 54.175 kHz) indicates the mechanical model is predictive enough to guide scaling of the design."],"supporting_citations":[{"why":"The six-mask PZT-MEMS process from which the paper's three-mask flow is derived; supplies the fabrication baseline.","marker":"[41]"},{"why":"The group's prior 1D resonant PZT scanner that this 2D device extends, raising the fast-axis resonance toward 54 kHz; supplies the performance baseline for comparison.","marker":"[42]"},{"why":"The review that defines the optical bandwidth-efficiency product and the range of reported scanners used to claim the device ranks among the higher 2D PZT scanners.","marker":"[22]"},{"why":"The LiDAR-mirror review used to justify the trade-off between fast-axis and slow-axis resonance frequencies and their role in resolution and frame rate.","marker":"[4]"},{"why":"An earlier PZT-actuated 2D Lissajous scanner; serves as a comparison baseline for prior dual-axis PZT devices.","marker":"[35]"},{"why":"The reported fracture strength of single-crystal silicon used to argue that simulated stress levels preserve mechanical integrity.","marker":"[43]"},{"why":"The study of oxide-interface effects on resonator quality factor used to explain the Q discrepancy and the role of the buried oxide layer.","marker":"[47]"}],"fun_headline_variants":["Three-mask process yields dual-axis PZT scanner at 12V","Compact PZT MEMS mirror scans 2 axes at 12V, 11.5°","Two-axis scanning with a three-mask PZT MEMS design","PZT MEMS scanner: 11.5° horizontal at 54 kHz, 12V","Mechanically coupled frames achieve dual-axis scanning"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The headline performance numbers assume that the quality factors measured in the optical setup (750 and 1050) are the device's true values, and that the roughly two-to-three-times-lower values from the separate laser measurement (about 301 and 643) are only artifacts of the tape mount and the weak 2 V drive used there, a dismissal the paper does not back with quantitative evidence.","fun_headline_variants_meta":{"raw":{"variants":["Three-mask process yields dual-axis PZT scanner at 12V","Compact PZT MEMS mirror scans 2 axes at 12V, 11.5°","Two-axis scanning with a three-mask PZT MEMS design","PZT MEMS scanner: 11.5° horizontal at 54 kHz, 12V","Mechanically coupled frames achieve dual-axis scanning"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000755,"raw_usage":{"total_tokens":3429,"prompt_tokens":1086,"completion_tokens":2343,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":702,"completion_tokens_details":{"reasoning_tokens":2241}},"tokens_in":702,"tokens_out":2343,"duration_ms":17477,"temperature":1.0,"reasoning_tokens":2241,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T12:17:47.091436+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Drive the same chip in the optical setup at increasing voltages from below 2 V up to 12 V peak-to-peak, and separately repeat the laser-vibrometer frequency sweep with a stiff mount and in vacuum; if the extracted quality factor stays near 300 (vertical) and 643 (horizontal) rather than rising to 750 and 1050, then the headline quality factors and the bandwidth-efficiency products derived from them are measurement artifacts rather than intrinsic device properties.","supporting_citations":[{"cited_title":"Wang, and P","cited_arxiv_id":null,"evidence_quote":"An earlier PZT-actuated 2D Lissajous scanner; serves as a comparison baseline for prior dual-axis PZT devices."},{"cited_title":"Journal of microelectromechanical systems, 2012","cited_arxiv_id":null,"evidence_quote":"The six-mask PZT-MEMS process from which the paper's three-mask flow is derived; supplies the fabrication baseline."},{"cited_title":"Resonant PZT MEMS scanners with integrated angle sensors","cited_arxiv_id":null,"evidence_quote":"The group's prior 1D resonant PZT scanner that this 2D device extends, raising the fast-axis resonance toward 54 kHz; supplies the performance baseline for comparison."},{"cited_title":"Baran, and H","cited_arxiv_id":null,"evidence_quote":"The review that defines the optical bandwidth-efficiency product and the range of reported scanners used to claim the device ranks among the higher 2D PZT scanners."},{"cited_title":"Watkins, and H","cited_arxiv_id":null,"evidence_quote":"The LiDAR-mirror review used to justify the trade-off between fast-axis and slow-axis resonance frequencies and their role in resolution and frame rate."},{"cited_title":"Journal of Materials Research, 2011","cited_arxiv_id":null,"evidence_quote":"The reported fracture strength of single-crystal silicon used to argue that simulated stress levels preserve mechanical integrity."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The study of oxide-interface effects on resonator quality factor used to explain the Q discrepancy and the role of the buried oxide layer."}],"review_version":1}