{"id":"e7f860d2-2743-4c16-9e8d-a60953d222e8","arxiv_id":"2607.23456","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.5,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"High-concentration Fe doping of Cs2AgBiBr6 single crystals narrows the high-T band gap and suppresses strain disorder across the ~125 K phase transition.","lead":"Iron doping of lead-free Cs2AgBiBr6 single crystals narrows the optical gap above ~125 K and reduces lattice strain below the cubic-to-tetragonal transition. The dual effect is offered as a route to more stable low-temperature optoelectronic and spintronic devices.","discovery_kind":"extension","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"The dual-function claim rests on a nominal 50% Fe occupancy that is never measured — and Table I's own unchanged Bi/Fe–Br bond lengths argue against anything near 50% substitution — while \"band-gap narrowing\" is inferred solely from heavily Stokes-shifted self-trapped-exciton PL, not from any band-","rationale":"The reader's weakest_assumption correctly flagged the missing composition assay and the under-constrained significance of the tiny σ change — my concern agrees in kind but sharpens it in two ways the reader did not state: (1) the paper's own Table I bond lengths are quantitatively inconsistent with 50% Bi→Fe substitution given the ionic-radius argument the authors themselves invoke, turning an absent assay into an internal tension; and (2) the gap-narrowing leg depends on equating Stokes-shifted STE PL peak position with the band edge, with the Fe/defect-emission alternative left unexcluded and no absorption data. I keep the verdict at CONDITIONAL rather than moving to REJECT because the concern is exactly the class of checkable deficiency the reader already conditioned on (composition metrology, error bars), the multi-technique trends are genuinely internally consistent, the authors honestly report the PL quenching that tempers their device claims, and the proposed occupancy refinement can be executed on existing data at essentially zero cost. If that refinement confirms ≪50% Fe, the verdict should drop; if it confirms near-50%, the structural leg of the claim is substantially strengthened and only the optical-proxy issue for the gap would remain.","tokens_in":10777,"tokens_out":3259,"duration_ms":148900,"concrete_test":"Re-refine the already-collected 100 K and 225 K SC-XRD datasets with the B-site occupancy freed between Bi and Fe (and Fe occupancy reported with esd). Given the Z=83 vs Z=26 contrast this is decisive and requires no new data. If refined Fe occupancy comes out far below 0.5 — as the unchanged M–Br bond lengths in Table I predict — the nominal-stoichiometry framing and the magnitude attributed to both claimed effects must be rescaled, and a complementary absorption edge (diffuse reflectance/Tauc) measurement should then be required before the red-shift is called \"band-gap narrowing.\"","verdict_should_be":"UNCHANGED","load_bearing_attack":"Both legs of the central claim (gap narrowing above Ts; strain suppression below Ts) are framed as consequences of \"Cs2AgBi0.5Fe0.5Br6\", i.e., 50% Bi→Fe replacement. But composition is never assayed: no ICP-OES/EDS/XPS, and — most surprisingly — no occupancy refinement in the SC-XRD, even though Bi (Z=83) and Fe (Z=26) have enormous X-ray scattering contrast, making this the cheapest, most decisive test available from data already collected. Worse, Table I is internally inconsistent with 50% incorporation: high-spin Fe3+ (0.645 Å) replacing Bi3+ (1.03 Å) at half the B-sites should markedly shrink the average metal–halide bond (Fe–Br in FeBr6 octahedra is ~2.5 Å), yet the reported Bi/Fe–Br distances (2.812(6)–2.821 Å) are statistically indistinguishable from the pristine Bi–Br values (2.813(5)/2.823(5) Å). The text claims a \"concomitant shortening\" that its own table does not show. The celebrated \"reversed inequality\" (Br1 vs Br2) hinges on ~0.01 Å differences against esds of ~0.005–0.006 Å each — a ~1.5σ effect — and one entry carries an implausible esd of zero (\"2.821(0)\"). On the electronic side, CABB emission is self-trapped-exciton PL with a large Stokes shift; peak position tracks STE relaxation, not the band edge. The authors themselves report a 10× PL quench and 10× shorter τ2 from Fe-induced non-radiative centers, so the cubic-phase red-shift is at least as consistent with a new lower-energy Fe/defect emission channel as with true gap narrowing — and no absorption/reflectance measurement is offered to discriminate. Notably, ref [14] (cited for the supporting intermediate-band DFT) found Fe doping did NOT alter the gap; the discrepancy is attributed to crystal quality, but the defect-emission alternative is never excluded. The 300 K PL \"exception\" is patched with an ad-hoc Raman argument. None of this makes the work wrong — the σ difference (~0.0048) is marginally significant given the stated lattice esds, and the qualitative trends are coherent — but the *m8","agreement_with_reader":"partial"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The manuscript reports hydrothermal growth of pristine and nominally 50% Fe-substituted Cs2AgBiBr6 single crystals and characterizes them by temperature-dependent single-crystal X-ray diffraction (100–225 K), steady-state and time-resolved photoluminescence (80–300 K), and near-room-temperature Raman spectroscopy. Both compositions undergo the known cubic (Fm-3m) to tetragonal (I4/m) transition near 125 K. The authors claim a dual function for Fe: (i) above Ts, a red-shifted PL peak is interpreted as band-gap narrowing, attributed to Fe-3d intermediate-band states; (ii) below Ts, a slightly smaller tetragonal strain parameter |σ| (0.3416 vs 0.3464 at 100 K), a reversal of the Bi/Fe–Br1 vs Bi/Fe–Br2 bond-length inequality, narrower PL linewidth, and the absence of the transient three-peak PL splitting seen in the pristine crystal are interpreted as suppression of detrimental phase-transition strain and antisite defects. The data are internally consistent across techniques in locating the ~125 K anomaly and in showing systematic differences between the two crystals. However, the central claims rest on a Fe content that is never measured, on a bond-length 'shortening' that Table I does not show, and on a gap-narrowing inference drawn from self-trapped-exciton PL without any absorption measurement.","tokens_in":11250,"tokens_out":4051,"duration_ms":25926,"significance":"If the central claim holds, the result is of genuine interest to the halide double-perovskite community: a dopant that simultaneously tunes the optical response in the cubic phase and reduces tetragonal strain through the 125 K transition would be a useful design handle for lead-free optoelectronics operating over wide temperature ranges. The study's strengths are the temperature-resolved, single-crystal (not thin-film) characterization: SC-XRD at six temperatures on both compositions, steady-state PL from 80–300 K with lineshape analysis, tri-exponential TRPL, and supporting Raman — a self-consistent dataset in which the 125 K anomaly appears independently in c-axis, PL intensity/peak/FWHM, and τ2. The synthesis protocol is described in sufficient detail to be reproducible. The paper is, however, observational rather than mechanistic: the load-bearing interpretive steps (actual Fe content, gap narrowing vs. new emission channel, antisite suppression) currently rest on inference rather than direct measurement.","major_comments":[{"comment":"The entire dual-function interpretation rests on the composition Cs2AgBi0.5Fe0.5Br6, i.e., 50% Bi→Fe replacement, but Fe content is never measured. There is no EDS, ICP-OES, or XPS, and — most importantly — no site-occupancy refinement of the SC-XRD data, even though Bi (Z=83) and Fe (Z=26) have very large X-ray scattering contrast, making an occupancy/refinement test essentially free on data already in hand. The only evidence offered for incorporation is a systematic c-axis contraction (Fig. 2), which is consistent with Fe doping but equally consistent with modest off-stoichiometry, vacancy disorder, or sample-to-sample variation. At minimum, the authors should (i) refine the Bi/Fe site occupancy at 100 K and 225 K and report the refined Fe fraction with esds, and (ii) provide an independent composition assay. If the refined/assayed Fe fraction is far below 50%, the stoichiometric formu","section":"§2.2.2, Table I, Methods"},{"comment":"The text claims a 'concomitant shortening of the average Bi/Fe–Br bond distance' as confirmation of Fe incorporation, but Table I shows the opposite of a statistically meaningful effect: the doped sample's Bi/Fe–Br1/Br2 distances are 2.821/2.812 Å versus 2.813(5)/2.823(5) Å for pristine — the averages (2.8165 vs 2.818 Å) are indistinguishable within the quoted esds, and substituting half the B-sites with high-spin Fe3+ (0.645 Å vs 1.03 Å for Bi3+) should produce a clearly resolvable contraction of the average metal–halide bond. Relatedly, the much-discussed 'reversed inequality' d'(Bi/Fe–Br1) > d'(Bi/Fe–Br2) rests on a 0.009 Å difference against combined esds of ~0.006 Å — roughly a 1.5σ effect — and one entry carries an implausible esd of zero, '2.821(0)'. Either the bond-length discussion should be removed, or the authors must demonstrate with proper error propagation that the differen","section":"§2.2.2, Table I"},{"comment":"The claim that Fe 'narrows the band gap' in the cubic phase is inferred solely from a red-shift of the photoluminescence peak above Ts. In Cs2AgBiBr6 the emission is a strongly Stokes-shifted self-trapped-exciton band (as the authors themselves note, §2.3.3, citing [19,20]); its peak position tracks STE relaxation energetics and defect channels, not the band edge. The authors' own data supply an obvious competing explanation: Fe introduces non-radiative centers (10× integrated-intensity quench, Fig. 4a; 10× reduction of τ2, Fig. 6), so a new lower-energy Fe/defect emission channel would produce exactly the observed red-shift without any gap change. This is also the finding of ref. [14], which the authors cite for the DFT intermediate-band picture while disagreeing with its experimental conclusion; the disagreement is attributed to crystal quality without independent evidence. A temperatu","section":"§2.3.1–2.3.2, Fig. 4b"},{"comment":"Two structural-mechanistic claims exceed the data. (a) 'The equivalence of the Ag–Br bond lengths between the two samples confirms that Fe substitution occurs exclusively at the Bi site, thereby suppressing the formation of Ag–Bi anti-site defects': with esds of 0.004–0.006 Å, bond-length equivalence constrains nothing about antisite occupancy, which would require refinement of mixed Ag/Bi site occupancies or a probe such as solid-state NMR (cf. ref. [9], which the authors cite for antisites). (b) 'Fe preferentially occupying Bi sites along the c-axis' is asserted from the reversed Br1/Br2 inequality (see comment 2) and offered to explain the strain reduction; in the I4/m double-perovskite structure the B-site sublattice does not have crystallographically distinct 'along-c' Bi positions in the Fm-3m-derived model, so this needs either a refinement demonstrating anisotropic Fe distributio","section":"§2.2.2"},{"comment":"The reduction in |σ| from 0.3464 to 0.3416 is reported without error bars. Propagating the lattice-parameter esds in Table I (a: 0.002–0.006 Å; c: 0.005–0.008 Å) gives an uncertainty on each σ of order 0.001, so the difference is only marginally significant (~3σ under generous assumptions, less if the larger esds apply to both), and no information is given on whether the two crystals were measured on the same instrument/run protocol or whether the result is reproducible across multiple crystals. Since the strain reduction is the quantitative core of the 'cryogenic stability' claim, the authors should report σ with propagated uncertainties, state the number of crystals measured, and ideally show the full σ(T) curve from 100–125 K rather than a single 100 K point, especially given that the PL 'transient splitting' it is meant to explain occurs over 90–125 K.","section":"§2.2.2, Table I, Fig. 2"}],"minor_comments":[{"comment":"Methods state data were collected at 'a steady temperature of 99 K' while the text and Table I refer to 100 K; please reconcile. Also, 'Crystals refinement software' (§2.2.1 and Methods) presumably means CrysAlisPro; please use the correct product name.","section":"Methods, §2.2.1"},{"comment":"Table I, Bi/Fe–Br1 entry '2.821(0)': an esd of exactly zero is not credible for a refined bond length; presumably a rounding or transcription error.","section":"Table I"},{"comment":"Figure 1 is described in its caption as a 'phase diagram'; it is a furnace temperature program, not a phase diagram. Please relabel.","section":"Fig. 1"},{"comment":"τ2 is described as the 'long-lived' component, yet three components τ1 < τ2 < τ3 (or some ordering) are fitted and the ordering is never stated explicitly; please give the typical magnitude and ordering of the three lifetimes and clarify why the intermediate component, rather than τ3, carries the radiative STE assignment.","section":"§2.3.4, Fig. 6"},{"comment":"The abstract and Conclusion state that optical and X-ray studies 'establish' Fe doping and a narrowed gap; given the comments above, more cautious wording ('consistent with', 'indicates') is appropriate unless the additional measurements are provided.","section":"Abstract, Conclusion"},{"comment":"The single-point 80 K anomaly in peak position and FWHM (Fig. 4b,c) is acknowledged as unexplained; since the multi-peak analysis (Fig. 5 insets) covers 90–125 K, please clarify whether the 80 K spectrum refits to a different component structure, which bears on the 'suppression of splitting' argument.","section":"§2.3.1, §2.3.3"},{"comment":"The statement that τ2 of the Fe-doped sample 'remains comparable to, or even exceeds, the values reported in the literature' is vague; please give a numerical comparison with the cited works [2, 9, 23].","section":"§2.3.4"},{"comment":"Reference [15] ('W. N. et al., Sci. Adv. 6, 262 (2020)') appears to have a malformed author list and possibly wrong page/article number; please check.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript notes it is already published in Solid State Communications 409, 116319 (2026), which the editor may wish to factor into how revision is handled for the arXiv version. The interpretation outruns the characterization in a way that routine measurements (EDS/ICP, occupancy refinement, absorption) would resolve; I would view a revision that adds these as substantially strengthening the paper, and one that merely softens language as acceptable but less valuable. Citation pattern and novelty disclosure look unremarkable."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"Punchline: they improved the hydrothermal growth, got cleaner crystals than the prior Fe-doping report, and show a consistent PL red-shift above ~125 K plus less low-T peak splitting. That comparative outcome is new. The dual-function story (gap narrowing + cryogenic strain suppression at nominal 50 % Fe) is still under-supported.\n\nWhat they did well. Temperature-dependent SC-XRD, steady-state PL, TRPL and near-RT Raman all track the same ~125 K cubic-to-tetragonal transition in both crystals. The multi-Gaussian PL work is useful: pristine shows transient three-component splitting just below Ts that the doped crystal lacks. They are frank that Fe cuts integrated PL and τ2 by roughly tenfold and that the material is not for high-efficiency PV. The growth protocol (pre-melt, long anneal, very slow cool) is concrete enough to reproduce.\n\nSoft spots, sized to the evidence. Composition is never assayed—no EDS/ICP/XPS and, oddly, no Bi/Fe occupancy refinement even though the X-ray contrast is huge and the data were already collected. Table I’s Bi/Fe–Br distances sit at 2.81–2.82 Å, statistically the same as pristine Bi–Br; that does not look like 50 % high-spin Fe3+. The celebrated “reversed” Br1/Br2 inequality is a ~0.01 Å shift on ~0.005 Å esds, and one esd is listed as zero. The tetragonal strain drop (|σ| 0.3464 → 0.3416) is real in direction but tiny and lacks lattice-constant error bars. On the electronic side, emission is self-trapped-exciton PL with a large Stokes shift; peak position is not a band-edge measurement, and no absorption/reflectance edge is shown. A lower-energy Fe/defect channel is therefore not excluded—especially given the strong non-radiative quench they themselves report. Device-stability language runs ahead of any cycling or device data.\n\nWho it is for: groups already deep in Cs2AgBiBr6 doping and low-T phase behavior. It is not a field-shifting result. It still deserves a serious referee who will demand composition metrology, proper esds, and an optical gap measurement before the dual-function claim stands. I would send it out, not desk-reject.","headline":"Real single-crystal trends under a better growth recipe, but the dual-function claim outruns unmeasured Fe occupancy and STE-only optics.","tokens_in":12219,"tokens_out":595,"would_cite":false,"duration_ms":27878,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["61.50.Ks","78.55.-m","61.72.U-","81.10.Dn"],"model":"grok-4.5","headline":"Fe doping of Cs2AgBiBr6 single crystals narrows the high-temperature band gap and damps cryogenic phase-transition strain.","keywords":["Cs2AgBiBr6","Fe doping","double perovskite","band-gap narrowing","phase transition","tetragonal strain","photoluminescence","single crystals"],"falsifier":"A direct compositional assay (e.g., ICP or EDS) confirming actual Fe occupancy near 50 %, combined with thermal-cycling strain or device-level measurements that either reproduce or fail to show the claimed stability gain across the 125 K transition.","tokens_in":11833,"feed_emoji":"❄️","tokens_out":768,"duration_ms":15168,"temperature":0.7,"pith_summary":"Cs2AgBiBr6 is a lead-free double perovskite whose usefulness is limited by a wide band gap and by a structural transition near 125 K that can introduce lattice strain. This work grows high-quality single crystals with nominal half-substitution of Bi by Fe and shows that the same dopant does two things at once. Above the transition it red-shifts the photoluminescence, indicating a narrower gap; below the transition it reduces tetragonal strain and eliminates the transient PL peak splitting seen in the undoped crystal. The authors therefore present Fe incorporation, enabled by a modified hydrothermal growth protocol, as a practical route to a more stable, lower-gap material for low-temperature optoelectronics and spintronics.","feed_headline":"Fe doping narrows Cs2AgBiBr6 gap and calms its 125 K transition","feed_subtitle":"Same dopant red-shifts PL above the transition and cuts tetragonal strain below it in single crystals","key_machinery":"Site-selective Fe occupancy on the Bi site (preferentially along the c-axis in the I4/m phase), which contracts the lattice, reverses the Bi/Fe–Br bond-length inequality relative to the pristine crystal, and lowers the absolute tetragonal strain parameter |σ|.","core_discovery":"Nominal 50 % Fe-for-Bi substitution in Cs2AgBiBr6 single crystals simultaneously narrows the optical gap in the cubic phase above ~125 K and suppresses the strain and spectral disorder that accompany the cubic-to-tetragonal transition below that temperature, without destroying the host structure or the transition itself.","pith_inferences":["If the modest |σ| reduction truly tracks lower defect density, Fe doping may also lengthen carrier diffusion lengths enough to reopen thin-film photovoltaic interest despite the added non-radiative centers.","The preferential c-axis Fe alignment suggests a possible magnetically ordered ground state at still lower temperature that the present optical and XRD data do not probe.","The same growth protocol could be used to test whether other transition-metal substituents produce analogous dual electronic–structural benefits."],"forward_implications":["Fe-doped CABB single crystals become a more practical platform for cryogenic optoelectronic and spintronic devices that must survive repeated passage through 125 K.","Band-gap tuning in the cubic phase can be achieved without destroying the double-perovskite framework or introducing a new structural distortion.","The same modified hydrothermal protocol may be transferable to other B-site dopants that target both electronic and strain properties.","Low-temperature X-ray detectors or spin-based sensors could exploit the reduced non-radiative pathways relative to earlier doped films while retaining structural robustness."],"fun_headline_variants":["Fe doping narrows Cs2AgBiBr6 gap and calms its 125 K transition","50% Fe-for-Bi shrinks cubic-phase gap and tames cryogenic strain","Fe substitution narrows gap above 125 K and cuts tetragonal disorder","Cs2AgBiBr6: Fe doping tunes band gap while stabilizing low-T phase","Fe doping alters Cs2AgBiBr6 bands and suppresses 125 K instability"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The small drop in tetragonal strain and the cleaner low-temperature PL lineshape are taken as sufficient proof that the phase-transition instability has been meaningfully suppressed for real devices.","fun_headline_variants_meta":{"raw":{"variants":["Fe doping narrows Cs2AgBiBr6 gap and calms its 125 K transition","50% Fe-for-Bi shrinks cubic-phase gap and tames cryogenic strain","Fe substitution narrows gap above 125 K and cuts tetragonal disorder","Cs2AgBiBr6: Fe doping tunes band gap while stabilizing low-T phase","Fe doping alters Cs2AgBiBr6 bands and suppresses 125 K instability"]},"model":"grok-4.5","effort":"low","cost_usd":0.00515,"raw_usage":{"total_tokens":1352,"prompt_tokens":686,"num_sources_used":0,"completion_tokens":114,"cost_in_usd_ticks":51504000,"prompt_tokens_details":{"text_tokens":686,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":552,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":686,"tokens_out":114,"duration_ms":9386,"temperature":1.0,"reasoning_tokens":552,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-30T21:41:51.014075+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"A direct compositional assay (e.g., ICP or EDS) confirming actual Fe occupancy near 50 %, combined with thermal-cycling strain or device-level measurements that either reproduce or fail to show the claimed stability gain across the 125 K transition.","supporting_citations":[],"review_version":1}