{"id":"77f72cf3-5c7b-4ec5-a779-717e8a7f2aeb","arxiv_id":"2411.18707","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"An all-passive upconversion system converts incoherent NIR light at ~10^-7 W/cm^2 to visible light with imaging resolution around 100 lp/mm.","lead":"This paper demonstrates a passive device that converts very dim near-infrared light, down to ~10^-7 W/cm^2, into visible light using organic molecules and nanoscale light-trapping structures. If the performance is confirmed, it could enable nighttime infrared imaging without electrical power.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The ~2 nW/cm2 low-intensity signal in Fig. 4b sits on a ~38 nW/cm2 unsubtracted baseline; without on/off subtraction or spectral identification, the 10^-7 W/cm2 claim is not secured.","rationale":"The reader's weakest assumption—that the ~2 nW/cm2 visible signal above the ~38 nW/cm2 baseline is genuine upconverted light rather than leakage or drift—is also the most load-bearing point I find. The paper is well organized, and the higher-intensity demonstrations (Fig. 1g, Fig. 2f, Fig. 3e, Fig. 4c) plus the measured enhancement factors are internally consistent and support a real, working all-passive upconversion imaging system. The headline claim, however, is specifically the extension to ~10^-7 W/cm2, and that rests on a single low-intensity data region in Fig. 4b where the signal is small relative to an unsubtracted baseline and no error bars are given. The SI S12 discrepancy reinforces the concern because the authors' own model predicts an order-of-magnitude lower efficiency; that discrepancy could be due to reasonable assumptions, but it removes independent quantitative support for the low-intensity point. Because the reader already assigned CONDITIONAL, my critique does not move the verdict. The conditional status should remain pending a direct background-subtracted, ideally spectrally resolved, measurement at the claimed lowest input intensity.","tokens_in":21956,"tokens_out":3489,"duration_ms":34139,"concrete_test":"Repeat the Fig. 4b minimum-intensity measurement with the broadband NIR LED chopped at ~1 kHz and a lock-in amplifier referenced to the chopper, recording the in-phase signal through a 610 ± 10 nm bandpass filter for bare glass, bare BHJ, and the fully integrated device at ~50 nW/cm2 input. If the device's on-minus-off signal does not exceed three standard deviations of the glass control, the claimed 2 nW/cm2 upconverted component is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim—visible emission of ~2 nW/cm2 at ~50 nW/cm2 NIR input, supporting the ~10^-7 W/cm2 headline—depends on isolating a signal that is only ~5% of the measured baseline. In the minimum-intensity measurement (Methods 6.3; Fig. 4b), the authors explicitly do not subtract the bare-glass control from the device curve, and the stated baseline is ~38 nW/cm2 from NIR filter leakage, ambient background, and dark current. Without an on/off subtraction or wavelength discrimination, the excess over baseline cannot be attributed to upconverted 610 nm light. The authors' own efficiency estimate in SI S12 predicts a system efficiency an order of magnitude below the reported value; while many modeling assumptions could explain part of that gap, at face value it means the low-intensity numbers are not corroborated by the independent efficiency model and are consistent with baseline contamination. The qualitative upconversion imaging at higher intensities is well supported, but the headline low-intensity point is the load-bearing quantitative result and it is not yet isolated from background.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a passive NIR-to-visible upconversion system based on triplet-triplet annihilation in a Y6/rubrene/DBP bulk heterojunction, enhanced by a dichroic backreflector and gold nanopillar plasmonic resonators, and integrated into a dual-wavelength Keplerian telescope for imaging. The authors claim upconversion of incoherent NIR light at input intensities down to ~10^-7 W/cm2, with ~2 nW/cm2 of visible output at ~50 nW/cm2 input, as well as imaging resolution near 100 lp/mm and demonstrations including flower images.","tokens_in":22077,"tokens_out":2759,"duration_ms":28108,"significance":"If the low-intensity claim is secure, this would be a notable advance for all-passive NIR imaging and sensing, roughly an order of magnitude above nightglow and far below previous passive upconversion demonstrations. The paper has several strengths: the imaging concept that preserves ray directionality is clearly explained; the qualitative upconversion imaging results are supported by multiple samples and control comparisons at higher intensities; and the FDTD simulations for the individual nanophotonic components are consistent with the measured factor-of-two enhancements. The main quantitative claim, however, currently rests on a background-limited measurement and an efficiency model that disagrees with the measurement by an order of magnitude, so the significance of the headline number is not yet established.","major_comments":[{"comment":"The central low-intensity result is not isolated from background. The text states that the fully integrated upconverter emits about 2 nW/cm2 at the lowest NIR intensity near 50 nW/cm2, but also states that the baseline from NIR filter leakage, ambient background, and dark current is about 38 nW/cm2, and that the bare-glass reading is not subtracted. The claimed signal is therefore only about 5% of the measured baseline. Without an on/off subtraction, a wavelength-resolved measurement, or a chopping/ lock-in scheme, the excess over the baseline cannot be attributed to upconverted 610 nm light. Please provide background-subtracted data with uncertainties, or an independent spectral or temporal discrimination of the upconverted signal at this intensity.","section":"Section 4 and Methods 6.3, Fig. 4b"},{"comment":"The expected-efficiency calculation in S12 predicts a system efficiency of 0.0003% at 0.1 mW/cm2 on the upconverter, whereas the measured value is stated as about 0.003%, an order-of-magnitude discrepancy. The manuscript attributes this to setup differences and uncertainties, but this is a load-bearing point: the model does not corroborate the measured low-intensity output, and the discrepancy is in the direction consistent with background contamination inflating the measured signal. Please either refine the model with the actual LED spectrum, the measured intensity dependence, and the integrated-device geometry, or present a quantitative uncertainty budget that explains the factor-of-ten gap.","section":"Supporting Information S12"},{"comment":"The claimed total enhancement of about 4 for the fully integrated upconverter is obtained by comparing a measured enhancement of 3.9 with FDTD simulations over varying air-gap thickness and inferring an air gap of about 0.25 micrometers. This air-gap thickness is a free parameter that is not independently measured, and the inference uses a single data point to select it. Since the enhancement underpins the low-intensity performance claim, please provide a direct measurement of the gap (for example, cross-sectional imaging or interferometry) or, failing that, explicitly state that the 3.9x enhancement is a fit-dependent estimate and show the sensitivity of the low-intensity conclusion to this assumption.","section":"Supporting Information S9"}],"minor_comments":[{"comment":"The sentence beginning 'the resolution in our experiment...' starts with a lowercase letter after a full stop; please capitalize.","section":"Section 2, paragraph after Fig. 1g"},{"comment":"The word 'Periody' appears in the caption; this appears to be a typo for 'Period'.","section":"Supporting Information Fig. S4 caption"},{"comment":"Please clarify in the Fig. 4b axes or caption that the plotted visible intensity includes the unsubtracted baseline, and consider overlaying the subtracted curve (P_VIS - P_BG) with error bars, since the current presentation makes the 2 nW/cm2 claim difficult to evaluate.","section":"Methods 6.3, minimum intensity measurement"},{"comment":"The phrase 'state-of-the-art TTA materials' is not referenced; please cite the relevant benchmark works or point to Table S2 for the comparison basis.","section":"Section 2, first paragraph"}],"recommendation":"major_revision","confidential_remarks":"The self-citation of the authors' prior EQE work (ref. 35) is an appropriate independent benchmark and does not by itself raise a circularity concern. However, the order-of-magnitude disagreement in S12 between the expected and measured efficiencies is a serious internal inconsistency that needs to be resolved before the low-intensity claim can be accepted. The paper is likely to be of interest to the optics community if the authors can provide a background-subtracted low-intensity measurement and a credible quantitative model."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper is a solid system-level demonstration of all-passive NIR-to-visible upconversion imaging. The integration of TTA in Y6/rubrene/DBP with plasmonic nanopillars and a dichroic backreflector works: they show roughly 4x total enhancement, resolution around 100 lp/mm, flower images under diffuse NIR are convincing, and the enhancements track their simulations. Credit also for transparency: the authors explicitly disclose the baseline issue in Fig. 4b instead of hiding it.\n\nThe soft spot is the quantitative low-intensity claim. The ~2 nW/cm^2 of visible output at ~50 nW/cm^2 input sits on a ~38 nW/cm^2 baseline, and the excess is only about 5% of that. There are no error bars, no on/off subtraction, and no spectral verification that the excess is actually 610 nm upconverted light rather than NIR leakage or drift. The authors' own efficiency estimate in SI S12 is off by an order of magnitude from the measured value; they attribute it to setup differences and slope uncertainty, but a 10x gap means the model does not corroborate the absolute output. For a claim resting on a small difference between two large readings, that is not enough.\n\nAlso, the abstract says the output reaches intensities 'perceptible by the human eye,' but that is a calculation from measured power and scotopic threshold, not a demonstration with an observer. Minor, but it should be phrased as an estimate.\n\nNone of this undercuts the qualitative result: the system clearly upconverts passive NIR images at higher intensities, and the engineering is meaningful. The low-intensity claim is what makes the paper significant, though, and it needs stronger evidence: repeated measurements with the glass control subtracted, a detection limit, and ideally a spectrum at the lowest input to identify the 610 nm peak.\n\nI would send this to peer review; the engineering and the main demonstration deserve referee time. But the referee should put real pressure on the baseline question before the 10^-7 number is accepted.","headline":"The imaging and engineering are credible and publishable; the 10^-7 W/cm^2 headline number is a small, unsecured excess over baseline.","tokens_in":22785,"tokens_out":3005,"would_cite":true,"duration_ms":40303,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"An all-passive upconverter using triplet-triplet annihilation with plasmonic and dichroic enhancement converts incoherent near-infrared light near $10^{-7}$ W/cm$^2$ into the visible, making passive night-vision-scale imaging plausible.","keywords":["triplet-triplet annihilation","photon upconversion","bulk heterojunction","near-infrared imaging","plasmonic enhancement","dichroic thin-film coating","passive night vision","incoherent light"],"falsifier":"Falsification would come from a spectral or action-spectrum test at the lowest intensity: chop the $750$–$930$ nm input on and off, replace it with equal-power light outside the Y6 absorption band, and subtract a bare-glass control with the same filter stack; if the $\\sim2$ nW/cm$^2$ visible excess does not vanish or scale with on-resonance input, the claim that upconversion operates at $50$ nW/cm$^2$ is not supported.","tokens_in":21668,"feed_emoji":"🌙","tokens_out":16616,"duration_ms":134632,"temperature":0.7,"pith_summary":"The paper claims that incoherent near-infrared light at irradiances down to about $10^{-7}$ W/cm$^2$ can be shifted into the visible with no electrical or optical power input, using triplet-triplet annihilation in an organic semiconductor film. The demonstrated device—a Y6/rubrene/DBP bulk heterojunction sandwiched between gold nanopillars that enhance NIR absorption and a dichroic thin-film stack that sends backward-emitted visible light forward—produces roughly $2$ nW/cm$^2$ of visible light over a $0.005$ cm$^2$ output beam at a measured input near $50$ nW/cm$^2$. Because the upconverter sits at an intermediate image plane and preserves ray angles, the same setup resolves about $100$–$110$ line pairs per millimeter and can image scattering objects under diffuse NIR light. If the result holds, it would be the lowest reported input intensity for a power-free incoherent upconversion imaging system, within an order of magnitude of nightglow and above the dark-adapted eye threshold for a pupil-collimated view.","feed_headline":"Upconverter turns 10^-7 W/cm2 NIR light visible, no power needed","feed_subtitle":"Triplet-triplet annihilation plus plasmonic and dichroic optics images scenes far dimmer than prior power-free systems.","key_machinery":"The load-bearing mechanism is triplet-triplet annihilation in a Y6/rubrene/DBP bulk heterojunction, where Y6 absorbs a NIR photon and converts it through interfacial charge-transfer states into rubrene triplets; two triplets fuse into a high-energy singlet that emits from DBP around $610$ nm. Around that film, periodic gold nanopillars (period $340$ nm, diameter $170$ nm, height $40$ nm, with a $7$ nm SiO$_2$ passivation layer) create a localized plasmon resonance near $850$ nm that roughly doubles absorption in the heterojunction, and a $12$-layer Nb$_2$O$_5$/SiO$_2$ dichroic stack transmits NIR while reflecting visible light within a $30°$ half-angle to recover backward emission. The imaging architecture places the upconverter at the mutual focal plane of a NIR lens and a visible lens, so incident ray angles map to positions on the film and back to output angles; because exciton diffusion lengths are below $40$ nm, the film does not blur the image.","core_discovery":"The central claim is that a single all-passive element can upconvert broadband, incoherent light in the $750$–$930$ nm band into visible light near $610$ nm at input intensities down to about $50$ nW/cm$^2$ (approximately $10^{-7}$ W/cm$^2$), with no external voltage or pump beam. The paper reports that the fully integrated upconverter emits around $2$ nW/cm$^2$ of visible light at that lowest input intensity, and that collimating this emission onto a fully dilated dark-adapted pupil would give about $0.02$ nW/cm$^2$, above the scotopic threshold near $0.005$ nW/cm$^2$. It also reports that the element, placed at the shared focal plane of NIR and visible lenses, preserves ray directionality and achieves upconversion imaging at $100$–$110$ lp/mm, with the limit set by relay-lens chromatic aberration rather than by the film or its nanophotonic overlay.","pith_inferences":["The measured total enhancement is about $4\\times$, short of the simulated product because of an estimated $0.25$ µm air gap; removing the gap by monolithic fabrication should recover part of the missing factor and may lower the operating floor.","The paper's own Supporting Information S12 expected-efficiency estimate is an order of magnitude below the measured output; a direct EQE-versus-intensity measurement of the integrated stack would settle which number is off.","The observed roughly $35\\%$ reduction of the quadratic-to-linear threshold suggests plasmonic near-fields act on triplet generation, not just absorption; resonators optimized for triplet density rather than absorption could push the operating floor lower.","The shared-focal-plane imaging geometry is material-agnostic; porting it to other incoherent upconverters, such as rare-earth or perovskite materials, would test whether the passivity and resolution advantages transfer."],"forward_implications":["A passive, battery-free NIR imager could operate at input intensities near $10^{-7}$ W/cm$^2$, orders of magnitude below parametric wave-mixing systems and below voltage-driven organic upconversion devices.","The measured output at the lowest input exceeds the estimated scotopic threshold when collimated onto a dilated pupil, so a directly eye-viewed night-vision eyepiece is a concrete near-term target.","Because the system works with broadband incoherent light, it can image reflective or scattering scenes illuminated by diffuse NIR, not only transmissive targets.","Resolution of $100$–$110$ lp/mm is limited by the NIR relay lens's chromatic aberration, so upgrading the relay optics should improve resolution without altering the upconverter.","The quadratic-to-linear threshold of TTA is reduced by roughly $35\\%$ with the plasmonic resonator, which may allow the same film to operate efficiently at still lower intensities."],"supporting_citations":[{"why":"Supplies the Y6/rubrene/DBP bulk-heterojunction TTA material system and the prior bare-film EQE (0.02–0.04%) used to benchmark the film and estimate expected system efficiency.","marker":"[35]"},{"why":"Establishes the solid-state triplet-triplet annihilation mechanism at an organic semiconductor interface on which the upconverter is built.","marker":"[32]"},{"why":"Provides the night-sky spectral radiance data used to place the demonstrated 50 nW/cm2 input within an order of magnitude of nightglow.","marker":"[17]"},{"why":"Gives the external-voltage-driven organic upconversion device result whose input intensity this passive system is an order of magnitude below.","marker":"[24]"},{"why":"Reviews parametric wave-mixing upconversion imaging whose high pump intensities define the powered baseline this work avoids.","marker":"[25]"},{"why":"Precedent that surface-plasmon field enhancement reduces the TTA quadratic-to-linear threshold, supporting the observed threshold reduction.","marker":"[51]"},{"why":"Demonstrates passive image upconversion using a shared focal-plane geometry, the imaging approach adapted here for NIR-to-visible operation.","marker":"[40]"}],"fun_headline_variants":["All-passive NIR upconversion works at 10^-7 W/cm2","No-power NIR-to-visible upconversion at ultra-low light","Ultra-low-light NIR to visible via passive upconversion","Passive upconversion: NIR to visible at 0.1 µW/cm2","Upconverter sees 10^-7 W/cm2 NIR with zero power"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"At the lowest input intensity, the visible excess is only about $2$ nW/cm$^2$ sitting on a baseline of roughly $38$ nW/cm$^2$, so the load-bearing premise is that this small excess is genuine upconverted light rather than near-infrared leakage through the blocking filters or slow background drift.","fun_headline_variants_meta":{"raw":{"variants":["All-passive NIR upconversion works at 10^-7 W/cm2","No-power NIR-to-visible upconversion at ultra-low light","Ultra-low-light NIR to visible via passive upconversion","Passive upconversion: NIR to visible at 0.1 µW/cm2","Upconverter sees 10^-7 W/cm2 NIR with zero power"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000973,"raw_usage":{"total_tokens":4158,"prompt_tokens":987,"completion_tokens":3171,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":603,"completion_tokens_details":{"reasoning_tokens":3080}},"tokens_in":603,"tokens_out":3171,"duration_ms":41004,"temperature":1.0,"reasoning_tokens":3080,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T10:57:29.772968+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Falsification would come from a spectral or action-spectrum test at the lowest intensity: chop the $750$–$930$ nm input on and off, replace it with equal-power light outside the Y6 absorption band, and subtract a bare-glass control with the same filter stack; if the $\\sim2$ nW/cm$^2$ visible excess does not vanish or scale with on-resonance input, the claim that upconversion operates at $50$ nW/cm$^2$ is not supported.","supporting_citations":[{"cited_title":"H., Kats, M","cited_arxiv_id":null,"evidence_quote":"Supplies the Y6/rubrene/DBP bulk-heterojunction TTA material system and the prior bare-film EQE (0.02–0.04%) used to benchmark the film and estimate expected system efficiency."},{"cited_title":"J., Meng, L., Pedersen, C","cited_arxiv_id":null,"evidence_quote":"Establishes the solid-state triplet-triplet annihilation mechanism at an organic semiconductor interface on which the upconverter is built."},{"cited_title":"& Yang, X","cited_arxiv_id":null,"evidence_quote":"Provides the night-sky spectral radiance data used to place the demonstrated 50 nW/cm2 input within an order of magnitude of nightglow."},{"cited_title":"Atmospheric Optical Environment","cited_arxiv_id":null,"evidence_quote":"Gives the external-voltage-driven organic upconversion device result whose input intensity this passive system is an order of magnitude below."},{"cited_title":"Seeing Beyond the Visible","cited_arxiv_id":null,"evidence_quote":"Reviews parametric wave-mixing upconversion imaging whose high pump intensities define the powered baseline this work avoids."},{"cited_title":"Review of night vision technology","cited_arxiv_id":null,"evidence_quote":"Precedent that surface-plasmon field enhancement reduces the TTA quadratic-to-linear threshold, supporting the observed threshold reduction."},{"cited_title":"N., Wilson, M","cited_arxiv_id":null,"evidence_quote":"Demonstrates passive image upconversion using a shared focal-plane geometry, the imaging approach adapted here for NIR-to-visible operation."}],"review_version":1}