{"id":"2f5d7142-7734-42c6-95ea-f17cc4311206","arxiv_id":"1908.08849","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Mercury chalcogenide nanocrystals can be tuned from n-type semiconductor to metal by increasing size, and ligand chemistry controls doping across orders of magnitude, enabling a 2.5 µm HgTe photodiode with 20 mA/W response and 3×10^9 Jones detectivity.","lead":"This PhD thesis studies tiny mercury selenide and mercury telluride nanocrystals that absorb infrared light. It shows how crystal size and surface chemistry control their electronic doping, and builds a prototype 2.5 micrometer infrared photodiode with moderate but real performance.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"HgSe metallicity claim hinges on k.p-fitted 1De level; transport exponents do not independently establish per-particle metal since film remains semiconducting.","rationale":"The reader identified the fitted k.p parameters as the weakest assumption, and my independent read agrees: the electron-count ladder, and specifically the 18-electron count for 17 nm HgSe, is the quantitative backbone of the semiconductor-to-metal transition claim. The thesis does provide partial support from multiple techniques, including the internal consistency of the IR plasmon peak with a high carrier density and the decreasing temperature shift of the intraband/plasmon feature with size. However, the transport evidence is explicitly qualified by the author as showing only hopping regimes in a semiconducting film, so it cannot independently certify per-particle metallicity. The k.p parameters are fitted to a small set of intraband transitions and then applied to sizes well beyond the fitting range; no direct measurement of 1De is provided. Because the central claim would fail if the 1De placement is wrong, the conditional verdict is appropriate. My concern does not move the verdict because the reader already assigned CONDITIONAL on essentially this basis.","tokens_in":57589,"tokens_out":2814,"duration_ms":29690,"concrete_test":"Perform electrochemical gating (as described in Ch. 2, §I.1) on films of the same 4.5 nm and 17 nm HgSe nanocrystals, directly locating 1Se, 1Pe, and 1De from conductance maxima/minima (Pauli blockades) without invoking the k.p parametrization. If the measured 1De–1Se spacing differs from the Eq. (2.12) prediction by more than thermal broadening, or if 1De does not lie below the Fermi level for 17 nm nanocrystals, the 18-electron count and the semiconductor-to-metal transition claim should be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of a size-driven semiconductor-to-metal transition in HgSe (Ch. 2, §II.2–3) is quantitatively tied to the electron count per nanocrystal: 2 for 4.5 nm and 18 for 17 nm. That count requires the three-band k.p model of Eq. (2.12) with fitted parameters EΓ6-Γ8 = -0.2 eV and Ep = 10.3 eV (Ch. 2, §II.2.b.ii) to position the 1De level above the Fermi level for the largest nanocrystals. If these parameters are not size-independent, or if the fit to intraband transitions is non-unique, the 1De energy, the inferred 18-electron count, and hence the claimed metallicity are unsupported. The companion transport evidence is weaker than presented: the thesis itself states that at film level the resistance still decreases with increasing temperature, i.e. the film is semiconducting; the T^-1/4 (Mott) versus T^-1/2 (Efros-Shklovskii) exponents distinguish two hopping regimes, not per-particle metallic versus semiconducting behavior (Ch. 2, §II.3.b). The IR plasmon cross-check is order-of-magnitude only (6.3×10^18 cm^-3 versus 4.5×10^18 cm^-3) and uses the same effective-mass assumption. Thus the quantitative semiconductor-to-metal transition is not independently verified.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This PhD thesis, written in French, studies the optoelectronic properties of colloidal HgSe and HgTe nanocrystals for infrared detection. The central claims are that the Fermi level and doping level of HgSe nanocrystals can be controlled by quantum confinement and by surface ligand chemistry, that HgSe undergoes a size-driven semiconductor-to-metal transition with the electron count increasing from about 2 to 18 electrons per nanocrystal as the diameter grows from 4.5 to 17 nm, and that a HgTe-based photodiode assembled from nanocrystal inks detects 2.5 µm light with a photoresponse of 20 mA/W and a detectivity of 3x10^9 Jones. The manuscript reconstructs the electronic structure using X-ray photoemission, infrared absorption, field-effect transport, and a three-band k.p model, and it also presents an original transient-photocurrent method for measuring the Urbach energy, a redox-doping study using polyoxometalate ligands, and a 100-pixel HgTe imager prototype.","tokens_in":57863,"tokens_out":3715,"duration_ms":44023,"significance":"If the electronic-structure reconstruction is valid, the thesis provides a coherent and useful picture of how confinement and surface chemistry tune doping in narrow-gap colloidal nanocrystals. The main experimental strengths are direct photoemission measurements of absolute level positions and work functions, systematic size- and ligand-dependent spectroscopy, a transient-photocurrent method for trap distributions, XPS evidence for electron transfer to polyoxometalate ligands, and a demonstrated SWIR photodiode with a multipixel array. These are concrete, falsifiable results that go beyond a purely phenomenological study. The weak point is that the quantitative doping counts and the semiconductor-to-metal transition rest on a k.p model fitted to the same class of intraband transitions that the model is then used to interpret; this is a parameterized consistency argument rather than an independent determination. The device-oriented chapters and the ligand-doping control are largely independent of that fragility and appear sound.","major_comments":[{"comment":"The electron counts per nanocrystal, and hence the claimed semiconductor-to-metal transition, are quantitatively set by the position of the 1De level calculated from the three-band k.p model. The parameters EΓ6-Γ8 = -0.2 eV and Ep = 10.3 eV are obtained by fitting intraband transition energies from the literature and from the author's own measurements. Using these fitted levels to interpret the same type of intraband data, and to infer 2, 8, or 18 electrons per nanocrystal, is a fit-to-data consistency check, not an independent prediction. The manuscript gives no uncertainty bars on the fitted parameters, no test of their size independence, and no alternative determination of the doping level. I therefore regard the quantitative 18-electron count and the precise location of the semiconductor-to-metal boundary as not fully supported, even though the qualitative trend of increasing n-type doping with decreasing confinement is plausible from the direct photoemission data.","section":"Ch. 2, §II.2.b.ii (Eq. 2.12)"},{"comment":"The transport measurements are presented as validating the semiconductor-to-metal transition, but the T^-1/4 (Mott) versus T^-1/2 (Efros-Shklovskii) analysis distinguishes two hopping regimes and does not by itself establish that the individual particles are metallic. The manuscript explicitly states that the film resistance still decreases with increasing temperature in both cases, i.e., the film remains semiconducting, so the metal-insulator transition is claimed at the per-particle level while the measured quantity is a film-level property. This evidence is therefore weaker than the text suggests and cannot independently compensate for the uncertainty in the k.p-based doping levels. The authors should either soften the transport-based confirmation, provide an explicit model showing how per-particle metallicity maps onto the observed exponent, or add independent per-particle evidence such as gate-dependent spectroscopy or single-nanocrystal measurements.","section":"Ch. 2, §II.3.b (Fig. 46)"},{"comment":"The infrared plasmon check is an order-of-magnitude consistency test, not an independent cross-check. The carrier density obtained from the plasmon frequency (6.3x10^18 cm^-3) is compared with the value inferred from the k.p-based level occupancy (4.5x10^18 cm^-3), but both estimates use the same effective mass and the same electronic-structure model. The agreement to within a factor of about 1.4 confirms that the measured intraband feature is compatible with a collective plasmon, but it does not add independent support for the 18-electron count. This should be stated more cautiously in the conclusions of Chapter 2.","section":"Ch. 2, §II.3.a (Eq. 2.13)"}],"minor_comments":[{"comment":"The heading contains a typo: 'Abréviatons' should be 'Abréviations'.","section":"Notations et abréviations (p. 5)"},{"comment":"The phrase 'dans le but de construire une caméra et donc, et donc d'être capable de reconstruire une image' contains a duplicated 'et donc'; one occurrence should be removed.","section":"Ch. 4, heading and opening paragraph of Section I"},{"comment":"The figure captions do not state the k.p model parameters used to place the 1De level, nor the assumed degeneracies (2, 6, 10 electrons) underlying the doping counts. Adding this information in the captions would make the derivation easier to audit.","section":"Ch. 2, Fig. 41 and Fig. 42"},{"comment":"The thermal-activation expression n = 2 exp(-(E1Se - EF)/kBT) is introduced without derivation; the prefactor of 2 deserves a brief explanation (spin degeneracy or level degeneracy) so that the resulting 0.03-electron estimate is reproducible.","section":"Ch. 3, Eq. (3.2)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a PhD thesis rather than a standard journal article, and much of its content has already appeared in the author's published papers. If the intended venue is a preprint server, the format is appropriate; if it is an archival journal, the editor should weigh the incremental contribution and the thesis-style presentation. My main technical concern is the quantitative doping reconstruction in Chapter 2, which is load-bearing for the semiconductor-to-metal claim; the device work and ligand-doping experiments are more direct and were, in my view, convincing."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a PhD thesis that assembles work already published by the group into a coherent story about doping control and infrared detection in HgSe/HgTe colloidal nanocrystals. The reader's conditional verdict is fair. What the thesis does well: it gives a detailed reconstruction of electronic structure using photoemission, infrared spectroscopy, electrochemistry, and field-effect transistors, and the size-dependent doping trend in HgSe (from about 2 to 18 electrons per nanocrystal) is supported by multiple techniques. The ligand-based doping control via dipoles and POM oxidants is clearly demonstrated, including direct evidence of tungsten reduction by XPS. The 2.5 µm HgTe photodiode with 20 mA/W and 3e9 Jones is a reasonable benchmark for solution-processed detectors, and the 100-pixel imager is a useful proof of concept.\n\nThe main soft spot is the quantitative semiconductor-metal transition claim. The electron count per nanocrystal relies on the k.p model with EΓ6-Γ8 = -0.2 eV and Ep = 10.3 eV fitted to intraband transitions; those parameters then position 1De above the Fermi level for the largest particles. That is a fit, not an independent prediction, and the thesis gives no error bars on the electron counts. The transport evidence is weaker than it first appears: the thesis itself states that the film resistance still decreases with increasing temperature, so the film is semiconducting. The Mott vs Efros-Shklovskii exponents distinguish two hopping regimes in a disordered film, not per-particle metallic versus semiconducting behavior. The IR plasmon cross-check is order-of-magnitude only and uses the same effective mass assumption. So the central claim of a size-driven semiconductor-to-metal transition is plausible but not independently verified.\n\nThe circularity concern is real but not fatal. The main experimental observations—size-dependent Fermi level from photoemission, ligand-induced shifts, device response—rest on direct measurements. The specific claim of metallicity for 17 nm HgSe is the weakest link, and it would be better framed as 'strong n-type doping with delocalized electrons' unless single-particle measurements or a parameter-free calculation support true metallicity.\n\nWho is this for? Researchers working on colloidal infrared nanocrystals, particularly those interested in doping control or HgSe/HgTe devices. The thesis is a useful reference for the group's prior work, but a specialist will want to go to the original papers. It deserves a serious referee: the device work and doping control are valuable, and the metallicity claim should be scrutinized. I would recommend engagement rather than desk rejection, with a request to either strengthen or soften the metallicity conclusion.","headline":"A solid thesis compilation of HgTe/HgSe nanocrystal work, but the semiconductor-metal transition claim rests on fitted k.p parameters and film-level transport that does not independently prove per-particle metallicity.","tokens_in":58416,"tokens_out":3199,"would_cite":false,"duration_ms":29137,"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":"HgSe nanocrystals become metallic as their size increases, and a HgTe diode detects 2.5 µm infrared light at 20 mA/W with 3×10^9 Jones detectivity.","keywords":["colloidal nanocrystals","HgSe","HgTe","electronic structure","self-doping","semiconductor-metal transition","infrared photodetector","ligand doping"],"falsifier":"Electrochemically titrate the electron count per HgSe nanocrystal as a function of diameter and compare with the predicted occupations of 2, 8, and 18 electrons; if the counts do not step at the claimed sizes, the k.p-derived phase diagram and the size-driven semiconductor-metal transition are unsupported.","tokens_in":57330,"feed_emoji":"📡","tokens_out":9451,"duration_ms":93685,"temperature":0.7,"pith_summary":"The thesis aims to replace expensive epitaxial infrared detectors with solution-grown mercury chalcogenide nanocrystals, and it establishes that the electronic properties of HgSe and HgTe nanocrystals can be mapped and controlled rather than taken as fixed. By combining photoemission, infrared absorption, and a three-band k.p model, the author reconstructs absolute energy levels and the Fermi level for different sizes and surface chemistries. The central result is that HgSe nanocrystals are self-doped, with the electron count rising from about 2 to about 18 per nanocrystal as the diameter grows from 4.5 to 17 nm, driving a semiconductor-to-metal transition confirmed by infrared absorption and low-temperature transport. The same band-engineering data are then used to build a HgTe-based photodiode that detects 2.5 µm light with a photoresponse of 20 mA/W and a detectivity of 3×$10^{9}$ Jones.","feed_headline":"Bigger mercury-selenide nanocrystals turn metallic","feed_subtitle":"A HgTe photodiode made from the same nanocrystal inks detects 2.5 µm light at 20 mA/W and 3×10^9 Jones.","key_machinery":"The load-bearing tool is the reconstructed absolute electronic structure of the nanocrystals, built from three complementary measurements plus one model. X-ray photoemission gives the Fermi level and work function; infrared absorption gives the 1Se–1Pe intraband and 1Sh–1Se interband transitions; and a three-band k.p model, with fitted parameters $E_{\\Gamma6-\\Gamma8} = -0.2$ eV and $E_p = 10.3$ eV for HgSe, places the 1De level. Because the conduction levels in these inverted-gap materials are approximately $\\hbar^2\\pi^2/(2m_e^* R^2)$ above the band edge, the level on which the Fermi energy sits directly counts electrons per nanocrystal (2, 8, and 18 for 1Se, 1Pe, and 1De). The same machinery yields a phase diagram of doping versus confinement and guides the diode design by providing band energies for contact selection.","core_discovery":"On the paper's own terms, the central discovery is a size- and surface-dependent electronic phase diagram for HgSe and HgTe nanocrystals. The author reconstructs the full electronic structure from photoemission (Fermi level, work function, valence-band edge), FTIR (interband and intraband transitions), and a three-band k.p calculation of the higher conduction levels. For HgSe, the Fermi level lies above the 1Se level for 4.5 nm particles, above 1Pe for 5.8 nm particles, and above 1De for 17 nm particles, corresponding to roughly 2, 8, and 18 electrons per nanocrystal. This is a semiconductor-to-metal transition driven by decreasing confinement rather than by added dopants, corroborated by the intraband peak becoming a temperature-insensitive plasmon and by the low-temperature hopping exponent switching from Mott ($T^{-1}$/4) to Efros-Shklovskii ($T^{-1}$/2). Surface ligands act as a second control knob: dipolar ligands such as S2− lower the Fermi level and switch interband transitions back on, and electron-accepting polyoxometalate ligands reduce the HgSe doping to 0.03 electron per nanocrystal. The device payoff is a HgTe photodiode in the extended short-wave infrared: a 100-pixel array images a laser profile, and single pixels give 20 mA/W response and 3×$10^{9}$ Jones detectivity at 2.5 µm. The author notes that even the most doped HgSe films remain semiconducting in their resistance-temperature behavior because interparticle transport is still hopping, so the metallicity is per-particle.","pith_inferences":["The same size-driven Fermi-level rise may occur in other zero-gap or inverted-gap nanocrystals; a testable extension is to map doping versus diameter in HgS and Ag2Se using the same photoemission-plus-infrared reconstruction.","The electron counts 2, 8, and 18 correspond to shell filling, so Coulomb-blockade or electrochemical charging experiments could test whether the transition steps through discrete shell occupations, refining the phase diagram beyond the continuum k.p model.","Oxidative polyoxometalate ligands could be generalized to produce stable p-type films from other self-doped nanocrystals; a concrete test is to combine them with a strongly dipolar n-type ligand on the same batch to form a ligand-defined p-n junction.","The transient-photocurrent Urbach measurement is faster than optical sub-gap absorption and, if it transfers to other materials, gives a practical passivation-quality metric for nanocrystal films."],"forward_implications":["HgSe nanocrystals of controlled size offer tunable mid- to far-infrared plasmonic resonances with carrier densities near 10^18 cm^-3, lower than in doped oxide nanocrystals, so the same material can serve as an intraband absorber or a plasmonic element.","Surface-chemistry doping gives a post-synthesis route to p-type or ambipolar mercury-chalcogenide films, useful for reducing dark current and for building junctions without introducing impurity atoms.","The measured absolute band energies and work functions for HgTe allow rational electrode and barrier selection, which the author uses to reach a photoresponse of 20 mA/W and a detectivity of 3×10^9 Jones at 2.5 µm.","The ink-based phase-transfer ligand exchange produces thicker absorbing layers (around 500 nm) than layer-by-layer film exchange, simplifying fabrication of vertical photodiodes.","A 100-pixel HgTe nanocrystal array can reconstruct an infrared laser intensity profile, a proof-of-concept step toward nanocrystal-based infrared cameras."],"supporting_citations":[{"why":"Supplies the HgSe size series, intraband absorption, and work-function values that anchor the electron-count reconstruction.","marker":"(22)"},{"why":"Supplies the three-band k.p model and the electrochemical method used to place the 1De level and read electron counts.","marker":"(38, 49, 50)"},{"why":"Provide the literature intraband transition energies across HgSe sizes used to fit the k.p parameters.","marker":"(49, 51)"},{"why":"Quantifies ligand dipole strengths and the DDT-to-S2- doping shift on HgSe, the basis for ligand-controlled doping.","marker":"(48)"},{"why":"Provides the HgTe synthesis route used to make the five size series for the band-energy measurements.","marker":"(47)"},{"why":"Report HgTe transfer-curve doping character versus size, supporting the Fermi-level positions in HgTe.","marker":"(11, 45)"},{"why":"Gives the power-law-to-Urbach-energy relation used to measure trap distributions in HgTe films.","marker":"(65)"},{"why":"Shows ligand-induced p/n doping in PbS nanocrystals, the comparison frame for the ligand-doping strategy.","marker":"(82)"}],"fun_headline_variants":["Size flips HgSe nanocrystals from semiconductor to metal","Growing HgSe nanocrystals turns them metallic","Size-driven metal transition in HgSe nanocrystals","Nanocrystal size controls HgSe metal transition"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative doping levels and the semiconductor-metal transition rest on the fitted three-band k.p parameters for HgSe ($E_{\\Gamma6-\\Gamma8} = -0.2$ eV and $E_p = 10.3$ eV) being size-independent, and on reading per-particle metallicity from film-level hopping exponents.","fun_headline_variants_meta":{"raw":{"variants":["Size flips HgSe nanocrystals from semiconductor to metal","Growing HgSe nanocrystals turns them metallic","Size-driven metal transition in HgSe nanocrystals","Nanocrystal size controls HgSe metal transition"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000269,"raw_usage":{"total_tokens":1743,"prompt_tokens":1191,"completion_tokens":552,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":807,"completion_tokens_details":{"reasoning_tokens":492}},"tokens_in":807,"tokens_out":552,"duration_ms":5838,"temperature":1.0,"reasoning_tokens":492,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:27:30.681635+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Electrochemically titrate the electron count per HgSe nanocrystal as a function of diameter and compare with the predicted occupations of 2, 8, and 18 electrons; if the counts do not step at the claimed sizes, the k.p-derived phase diagram and the size-driven semiconductor-metal transition are unsupported.","supporting_citations":[],"review_version":1}