{"id":"b3ac88a5-126d-40ee-9dcc-252e386c93dc","arxiv_id":"2508.10814","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A suspended 4H-SiC membrane platform is demonstrated by patterning devices after photoelectrochemical undercut, yielding 1D photonic cavities with Q in the thousands and a TFLN-on-SiC phononic cavity.","lead":"This paper shows a way to make suspended 4H-silicon carbide devices: first suspend a thin membrane with a photoelectrochemical etch, then pattern photonic or acoustic structures on it. It demonstrates free-standing photonic crystal cavities and a lithium-niobate-on-SiC resonator, though no color-center defect is integrated yet.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proximity-effect advantage in Fig. 1d is confounded by dose selection and lacks statistics; central claim under-supported.","rationale":"The reader's verdict (CONDITIONAL) already identifies the central weakness: the proximity-effect comparison in Fig. 1d is confounded by independent dose optimization and lacks statistical support. I agree that this is the most load-bearing concern because it directly undermines the abstract's key claim of decreased fabrication errors. The paper credibly demonstrates that suspended membranes can be patterned and yield functional devices, so the fabrication sequence itself is validated; however, the advantage claim is not. The recommended check—a controlled dose-series experiment with error bars—would directly resolve whether the effect is real or an artifact of dose calibration. The acoustic resonator and high-temperature annealing claims are also under-supported but are secondary to the paper's main selling point. Thus, no change to the reader's conditional verdict is necessary.","tokens_in":10521,"tokens_out":4775,"duration_ms":58263,"concrete_test":"Perform an electron-beam lithography dose series on identical test patterns (e.g., hole arrays) on suspended and nonsuspended 4H-SiC substrates, using the same resist, thickness, and no proximity-effect correction. For each dose, fabricate at least 20 devices and measure the fabricated hole diameters by SEM; plot mean percent error ± standard deviation versus dose for both conditions. If the minimum achievable percent errors overlap within the measurement uncertainty, or if the dose at which the minimum occurs differs by more than the expected effect of removed backscatter, then the claimed fabrication-error reduction is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central advantage claim—patterning on suspended membranes reduces secondary electron production and proximity-effect-induced fabrication errors—rests entirely on Fig. 1d, which compares percent error for features written without proximity-effect correction on suspended versus nonsuspended substrates. This comparison is not sufficient to establish the claimed mechanism for three reasons. First, no error bars, sample counts, or statistical tests are reported, so the difference between the two conditions could be within measurement noise. Second, the doses were 'chosen independently for suspended and nonsuspended devices to minimize percent error.' When each condition is independently tuned to its optimal dose, the comparison measures the quality of dose calibration rather than the physical effect of reduced backscattering; the suspended case might simply have had a more favorable dose-target relationship. Third, the metric (hole dimensions in a 1D PhC) is not validated as a proxy for device-level performance (e.g., cavity Q or resonance wavelength) that matters for defect integration. The physical mechanism is plausible—a 500 nm membrane with vacuum underneath should reduce long-range backscattering relative to a bulk substrate—but Fig. 1d does not isolate that mechanism from other differences such as resist adhesion, charging, or mechanical vibration. If this advantage is not established, the paper's primary differentiation from prior undercut-after-patterning approaches is weakened, though the fabrication sequence itself would still be demonstrated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a fabrication approach for suspended 4H-SiC devices: first, unintentionally doped epitaxial membranes are suspended using a previously reported photoelectrochemical (PEC) undercut method (with two added biases), and then photonic/phononic devices are patterned directly onto the suspended membranes via electron-beam lithography and dry etching. The authors claim that patterning on suspended membranes reduces secondary-electron backscattering and proximity-effect-induced dimensional errors. They demonstrate 1D photonic crystal cavities (free-space and tapered-waveguide-coupled) with room-temperature Q factors of a few thousand, and a thin-film lithium niobate (TFLN) on SiC phononic-crystal structure. They also claim the monolithic platform withstands high-temperature annealing and allows heterogeneous integration.","tokens_in":10845,"tokens_out":3807,"duration_ms":42272,"significance":"If the proximity-effect advantage and the device demonstrations are borne out, the platform offers a practical route to suspended 4H-SiC quantum devices without wafer bonding, grinding, or post-etch undercutting. The paper provides detailed fabrication parameters and simulation repositories, and the cavity Q values are consistent with prior 4H-SiC PhC work. The use of the authors' own prior PEC method and simulation repositories is a normal continuation of previous work, not circular, since the central claim is tested by direct fabrication and measurement. However, the central differentiating claim—reduced proximity-effect-induced fabrication error on suspended membranes—rests on a single comparison (Fig. 1d) with a confounded dose protocol and no statistics. The acoustic-cavity demonstration lacks measured mechanical response. These points need to be strengthened before the stated advantages are accepted.","major_comments":[{"comment":"The central claim that patterning on suspended membranes reduces proximity-effect-induced fabrication errors is not established by the presented data. The doses were independently chosen for each condition to minimize percent error (Fig. 1d caption), which confounds the physical backscattering effect with dose calibration. A fair test would use identical dose recipes or sweep dose across both conditions. Additionally, no error bars, number of devices/holes, or statistical tests are reported, so the observed difference could be within run-to-run variation. This is load-bearing because the claimed reduced proximity effect is the paper's primary differentiation from undercut-after-patterning methods.","section":"Fig. 1d and 'One major benefit' paragraph"},{"comment":"The percent-error metric is based on hole dimensions in 1D PhCs after fabrication. This does not necessarily reflect device-level performance (cavity Q, resonance wavelength, collection efficiency) that matters for defect integration. The authors should either demonstrate that the reduced dimensional error translates into improved device metrics, or qualify the claim as only about lithographic dimensional control. The statement about 'reduced sensitivity to variations in feature size and orientation' is also not quantified (e.g., no slope or variance comparison).","section":"Fig. 1d and device-level relevance"},{"comment":"The TFLN-on-SiC acoustic cavity is presented as a demonstration, but only SEM images are provided. No mechanical resonance spectrum, Q factor, or frequency response is shown. Without a measured mechanical mode, the structure is a fabricated phononic crystal, not a demonstrated cavity. The claim of obtaining 'fully suspended TFLN-on-SiC' is fine as a structural result, but the broader implication of an acoustic cavity should be explicitly limited to the fabricated geometry unless a resonance measurement is included.","section":"Fig. 4 and 'To this end' paragraph"},{"comment":"The cavity Q values are reported from single measurements without uncertainty, fit residuals, or number of devices. Since Q is used to argue the platform's viability (e.g., 'comparable with other 4H-SiC PhCs'), the lack of error bars and sample counts weakens the comparison. At minimum, report standard errors from repeated measurements or clarify that these are representative single devices.","section":"Figs. 2b and 3b"}],"minor_comments":[{"comment":"The abstract claims the platform 'support[s] high temperature annealing,' but the manuscript does not actually expose the suspended membranes to high temperatures post-suspension. This is presented as a projected advantage, not a demonstrated result. Please clarify this distinction.","section":"Abstract and conclusion"},{"comment":"The caption says 'data taken by measuring holes from 1D PhCs' but does not specify the number of holes, devices, or independent fabrication runs. This information should be added to support the comparison.","section":"Fig. 1d caption"},{"comment":"The two external biases are described as additions to the protocol in [31]. It would be helpful to state explicitly which aspects of the PEC process are new relative to [31], to avoid confusion about prior work.","section":"Methods, PEC etching"},{"comment":"The claim of 'first reported demonstrations of tapered waveguide cavities in 4H-SiCs' is appropriately qualified, but given reference [49] recently demonstrated waveguide fiber interfaces in SiC, please double-check that this novelty claim is accurate and clearly scoped to 'tapered waveguide cavities.'","section":"Conclusion, first reported tapered cavities"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within scope for a fabrication-focused applied physics journal. The main risk is that the central proximity-effect advantage is over-sold relative to the evidence; the authors should be given the opportunity to add measurements or moderate the claims. The TFLN-on-SiC structure is a nice proof-of-concept but needs at least a basic mechanical characterization to be called an acoustic cavity."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What you should know: this is a fabrication methods paper, not a defect-integration demonstration. The core move—undercut first, pattern afterward—is real, useful, and largely new. It avoids wafer bonding, grinding, and post-hoc undercutting, and it is compatible with high-temperature annealing and a heterogeneously stacked TFLN-on-SiC resonator. The cavity work is preliminary but functional: Q of a few thousand is comparable to earlier 4H-SiC photonic crystals, and the claim to the first tapered waveguide cavity in 4H-SiC looks fair—[49] demonstrates a waveguide-fiber interface, not a tapered cavity.\n\nThe paper is also honest in the right places. Process details are concrete, and the SI candidly walks through resist adhesion failures before settling on the carrier-wafer trick. The simulation repos are cited, and the PEC undercut method is their own prior work but cited openly; that is self-reference, not circularity.\n\nNow the soft spots, in proportion. The proximity-effect advantage (Fig. 1d) is the weakest link. The comparison is confounded: doses were independently chosen for each condition to minimize error, there are no error bars or sample counts, and the metric is hole dimension rather than device-level performance. A skeptic could argue the suspended case simply got a better dose calibration. The physical mechanism is plausible—a 500 nm membrane with vacuum underneath should reduce long-range backscattering—but the figure does not isolate that mechanism from resist adhesion, charging, or vibration. That said, the paper does not fall apart if this comparison is weak: the fabrication sequence stands on its own.\n\nThe TFLN-on-SiC acoustic cavity is shown only by SEM with no measured mechanical response; if the pitch is spin-phonon coupling, that is a significant gap. And the title promises defect integration, but no defect is implanted, annealed, or measured anywhere in the paper. That is an overreach, not a fatal one, but the abstract and title should match the content.\n\nWho gets value: process engineers working on SiC and people building suspended quantum devices in wide-bandgap materials. It deserves a serious referee—the platform is useful and the claims are testable with straightforward follow-up experiments: proper statistics for Fig. 1d, a mechanical characterization of the phononic cavity, and a title that says what was actually done. I would engage with it, and I would expect heavy revision before acceptance.","headline":"A solid fabrication paper whose central structural claim is credible and useful; the proximity-effect advantage is plausible but not nailed down, and the title overpromises defect integration.","tokens_in":11316,"tokens_out":1559,"would_cite":true,"duration_ms":20168,"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 new 4H-SiC fabrication route suspends a monolithic membrane first and then patterns devices on it, cutting electron-beam proximity-effect errors and enabling hybrid quantum structures without wafer bonding or grinding.","keywords":["4H-silicon carbide","suspended membranes","photoelectrochemical etching","photonic crystal cavities","thin-film lithium niobate","phononic cavities","electron-beam proximity effect","quantum defect integration"],"falsifier":"Write identical test features with identical electron-beam dose and no proximity-effect correction on suspended and nonsuspended 4H-SiC films, measuring feature dimensions across many samples with reported error bars; if percent error is statistically identical, the claimed proximity-effect reduction is not the cause of the observed improvement.","tokens_in":1563,"feed_emoji":"🔬","tokens_out":1681,"duration_ms":62896,"temperature":0.7,"pith_summary":"This paper tries to establish a better way to build suspended 4H-SiC quantum devices: instead of patterning devices in bulk material and undercutting them later, first suspend a thin membrane from a monolithic wafer using photoelectrochemical etching, then fabricate devices directly on the suspended film. The authors argue that patterning on a suspended membrane reduces electron backscattering and therefore lowers proximity-effect fabrication errors, simplifying lithography and improving feature fidelity. If correct, this approach removes the need for wafer bonding, grinding, and chemical-mechanical polishing, while also surviving high-temperature annealing and allowing heterogeneous integration with thin-film lithium niobate. As proof of concept, they fabricate 1D photonic crystal cavities, tapered-fiber-coupled cavities, and a TFLN-on-SiC acoustic resonator, showing quality factors in the low thousands and compatibility with subsequent color-center integration. The broader significance is a more flexible, scalable fabrication route for point-defect-based quantum technologies in silicon carbide.","feed_headline":"Pattern devices after suspension to cut SiC fabrication error","feed_subtitle":"Suspended-first 4H-SiC platform yields photonic cavities plus hybrid acoustic devices for quantum color centers.","key_machinery":"The central object is the suspended 4H-SiC membrane formed by photoelectrochemical (PEC) etching: dopant-dependent band bending sends photogenerated holes to the n-type substrate and preserves the unintentionally doped epilayer, while two external biases (a substrate/electrolyte bias and an epilayer/substrate reverse bias) improve undercutting. Patterning devices directly on this already-suspended film removes the underlying backscattering target for the electron beam, which is the mechanism claimed to reduce proximity-effect-induced fabrication errors and to make lithography optimization simpler.","core_discovery":"The central claim is that fabricating devices after suspending a 4H-SiC membrane—rather than defining devices in bulk and undercutting them later—reduces electron-beam proximity effects during lithography. Because the suspended film has no underlying substrate to backscatter electrons, features patterned into the membrane show lower percent error and less sensitivity to feature size and orientation than identical patterning on nonsuspended material. The paper supports this with percent-error measurements comparing suspended and nonsuspended films, and demonstrates the platform with free-space scattering and tapered-waveguide 1D photonic crystal cavities, reaching quality factors of a few tho","pith_inferences":["The same suspend-then-pattern logic should transfer to other hard-to-etch wide-bandgap or refractory materials, because reduced backscattering is a geometric effect rather than a chemical one, though this remains to be demonstrated.","If the proximity-effect benefit is real, the platform should also improve feature fidelity at larger write fields and multi-scale devices, a consequence the paper does not directly test.","The combination of high-temperature annealing tolerance and TFLN integration points toward a single chip with telecom-emitting vanadium qubits acoustically coupled to lithium niobate transducers, an integration the paper motivates but does not yet realize.","A direct measurement of backscattered electron yield from suspended versus bulk membranes would isolate whether the fabrication benefit comes from the suspension itself or from other process differences; such a measurement is not reported."],"forward_implications":["Suspended 4H-SiC devices can be made without wafer bonding, grinding, or chemical-mechanical polishing, simplifying the fabrication of quantum-grade membranes and devices.","Patterning on suspended films reduces electron-beam proximity effects, so target feature sizes are achieved with lower percent error and less orientation bias, reducing calibration iterations.","The demonstrated 1D photonic crystal cavities reach quality factors of a few thousand, comparable to previously reported 4H-SiC PhCs, validating the etch fidelity of the suspended-first approach.","Monolithic membranes tolerate high-temperature annealing, opening paths to dopant activation and high-temperature activation of telecom emitters such as vanadium defects without thermal-mismatch damage.","Heterogeneous integration with thin-film lithium niobate produces fully suspended TFLN-on-SiC phononic cavities, offering a route to piezoelectric spin readout and acoustic spin control.","The platform supports waveguide and tapered-fiber interfaces, enabling scalable readout, photon transport, and entanglement between color-center qubits."],"supporting_citations":[{"why":"Supplies the photoelectrochemical etching method for selectively undercutting undoped 4H-SiC membranes that this platform directly extends.","marker":"[31]"},{"why":"Provides the base fabrication steps (cleaning, nickel contact, ITO hard mask, e-beam lithography, dry etching) on which the membrane process is built.","marker":"[32]"},{"why":"Prior 4H-SiC nanobeam photonic crystal fabrication that serves as the baseline for comparing the quality factors achieved here.","marker":"[26]"},{"why":"Alternative 4H-SiC-on-insulator platform that this suspended-membrane approach aims to avoid or complement.","marker":"[23]"},{"why":"Gives simulation details for the 1D photonic crystal cavities designed to resonate with surface states.","marker":"[36]"},{"why":"Provides simulation details for the tapered-fiber-coupled 1D photonic crystal cavity design.","marker":"[38]"},{"why":"Supplies the PDMS transfer process used to integrate thin-film lithium niobate onto the suspended SiC membrane.","marker":"[45]"}],"fun_headline_variants":["Suspend SiC first to cut patterning errors for quantum devices","Fabricate after suspending to reduce SiC device pattern errors","Suspended-first 4H-SiC platform sharpens device fabrication","Pattern suspended SiC membranes to avoid proximity effects"],"cache_read_input_tokens":13184,"weakest_assumption_plain":"The claimed fabrication-error reduction rests on a single comparison in which electron-beam doses were chosen independently for suspended and nonsuspended devices, so if the lower error comes from dose calibration rather than reduced backscattering, the platform's key advantage is not established.","fun_headline_variants_meta":{"raw":{"variants":["Suspend SiC first to cut patterning errors for quantum devices","Fabricate after suspending to reduce SiC device pattern errors","Suspended-first 4H-SiC platform sharpens device fabrication","Pattern suspended SiC membranes to avoid proximity effects"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000819,"raw_usage":{"total_tokens":3395,"prompt_tokens":689,"completion_tokens":2706,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":433,"completion_tokens_details":{"reasoning_tokens":2636}},"tokens_in":433,"tokens_out":2706,"duration_ms":23768,"temperature":1.0,"reasoning_tokens":2636,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T20:12:51.601716+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Write identical test features with identical electron-beam dose and no proximity-effect correction on suspended and nonsuspended 4H-SiC films, measuring feature dimensions across many samples with reported error bars; if percent error is statistically identical, the claimed proximity-effect reduction is not the cause of the observed improvement.","supporting_citations":[{"cited_title":"Tsukimoto, T","cited_arxiv_id":null,"evidence_quote":"Supplies the photoelectrochemical etching method for selectively undercutting undoped 4H-SiC membranes that this platform directly extends."},{"cited_title":"Kozlov, V","cited_arxiv_id":null,"evidence_quote":"Provides the base fabrication steps (cleaning, nickel contact, ITO hard mask, e-beam lithography, dry etching) on which the membrane process is built."},{"cited_title":"Majety, V","cited_arxiv_id":null,"evidence_quote":"Prior 4H-SiC nanobeam photonic crystal fabrication that serves as the baseline for comparing the quality factors achieved here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Alternative 4H-SiC-on-insulator platform that this suspended-membrane approach aims to avoid or complement."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives simulation details for the 1D photonic crystal cavities designed to resonate with surface states."},{"cited_title":"Soykal, P","cited_arxiv_id":null,"evidence_quote":"Supplies the PDMS transfer process used to integrate thin-film lithium niobate onto the suspended SiC membrane."}],"review_version":1}