{"id":"88da28e4-adf3-4ba2-94e4-c038a34faea1","arxiv_id":"1909.01538","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"In Eu-doped GaN, red emission efficiency is limited at low excitation by a competitive carrier trap and at high excitation by the small number and small cross-section of efficient Eu sites, with peak external quantum efficiency near 46 to 48 percent at low temperature.","lead":"This paper measures how efficiently europium atoms embedded in gallium nitride turn light or electricity into red emission, and identifies a defect that steals energy at low power and a scarcity of efficient sites at high power. The measured peak external quantum efficiency of about 46 to 48 percent at low temperature suggests practical red LEDs on a mature semiconductor platform.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The two-order-of-magnitude cross-section contrast rests on an unverified equal-efficiency assumption for minority and majority Eu sites; lifetime similarity alone does not establish equal radiative efficiencies.","rationale":"The reader's weakest_assumption correctly identifies the equal-emission-efficiency assumption as load-bearing, and I agree. The paper's central quantitative claim is the two-order-of-magnitude difference between the OMVPE8 and OMVPE4 excitation cross-sections, and the high-fluence QE ceiling depends on the limited number of high-efficiency minority sites. That entire edifice rests on the rescaling step described in the PL section. The concern is concrete: PL lifetime equality does not imply equal radiative efficiency, and the thermal-quenching data in Fig. 3 suggest that OMVPE4 and OMVPE8 respond differently to temperature, so their nonradiative rates likely differ. If the assumption fails, the extracted OMVPE4 cross-section is not physically meaningful and the central contrast is not established. This concern does not, however, undermine all of the paper: the OMVPE8 cross-section, the trap saturation behavior, and the absolute QE measurements are independent and remain plausible. Therefore the appropriate verdict stays CONDITIONAL, as the reader assigned. I do not see a more fundamental flaw that would demand rejection; the equal-efficiency assumption is testable with a straightforward refit of the published fluence data.","tokens_in":6925,"tokens_out":4522,"duration_ms":48860,"concrete_test":"Simultaneously fit the OMVPE8 and OMVPE4/7 fluence data with I_i = eta_i N_i (1 - exp(-sigma_i f)), using the known site abundances from SIMS and treating the efficiency ratio eta_4/eta_8 as a free parameter. If the best-fit ratio deviates from unity by more than a factor of two, the equal-efficiency assumption fails and the OMVPE4 cross-section must be re-derived without the rescaling.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the photoluminescence section, the authors fit Eq. (3) to the OMVPE8 saturation curve and extract 1.6e-15 cm2, then state: 'As the PL lifetime of all peaks have similar values, it can be assumed that the emission efficiency of the different sites is similar. This allows us to use the saturation value of OMVPE8, together with the relative abundance of the sites, to scale the excited fraction of the OMVPE4/7 peak.' This is the key step that calibrates the vertical axis of Fig. 1c and therefore sets the OMVPE4 cross-section to 1.2e-17 cm2. The rescaling converts measured PL intensity into an absolute excited fraction under the assumption that minority and majority sites have identical radiative efficiencies. If OMVPE4 has a different nonradiative rate, which is plausible given its much stronger thermal quenching shown in Fig. 3, then the ordinate of Fig. 1c is miscalibrated and the fitted OMVPE4 cross-section is not a true excitation cross-section. The two-order-of-magnitude contrast between sites, and the quantitative high-fluence QE ceiling that follows from the limited number of efficient sites, would then not be established. No independent measurement of the OMVPE4 emission efficiency is provided; PL lifetime similarity is insufficient because lifetimes combine radiative and nonradiative decay rates.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports on the excitation efficiency and external luminescence quantum efficiency (QE) of Eu3+ ions in GaN under optical and electrical excitation. By fitting pulsed-photoluminescence saturation curves with Eq. (3), the authors extract an effective excitation cross-section of 1.6 x 10^-15 cm^2 for the minority OMVPE8 site and 1.2 x 10^-17 cm^2 for the majority OMVPE4 site, suggesting a two-order-of-magnitude contrast. Under electroluminescence, the corresponding values are 3.0 x 10^-15 cm^2 and 6.0 x 10^-18 cm^2. The external QE is measured versus fluence and temperature, reaching about 0.46 at low temperature. The measurements are interpreted with a two-regime model: at low fluence, an efficient carrier trap (the H1 trap) reduces the QE until it is filled; at high fluence, the limited number of high-efficiency minority sites and the small cross-section of the majority site limit the QE. The paper concludes that Eu:GaN is promising for low-current-density applications such as micro-LEDs.","tokens_in":7208,"tokens_out":7871,"duration_ms":73922,"significance":"If the reported cross-section contrast is correct, the work provides quantitative guidance for operating Eu-doped GaN as an efficient red emitter, particularly in micro-LEDs where low current densities are typical. The experimental approach, including absolute QE measurements in an integrating sphere and the decomposition into minority and majority site contributions, is useful and goes beyond earlier qualitative work. The two-regime QE model is plausible and connects the fluence-dependent QE to a saturable trap and to site-specific excitation cross-sections. However, the central quantitative claim rests on an unverified equal-efficiency assumption for the two sites, and the trap-concentration estimate contains internal inconsistencies. The significance is therefore conditional on resolving these issues.","major_comments":[{"comment":"The ordinate of Fig. 1c is calibrated by converting measured PL intensity into an absolute excited fraction under the assumption that the emission efficiency of OMVPE8 and OMVPE4/7 is similar because their PL lifetimes are similar. This is not a sufficient condition: the radiative efficiency is the ratio of the radiative rate to the total decay rate, and similar lifetimes do not imply equal radiative rates. The paper's own Fig. 3 shows that the majority site is thermally quenched much more strongly than OMVPE8, which is direct evidence that the nonradiative rates differ. Therefore the fitted OMVPE4 cross-section of 1.2 x 10^-17 cm^2 is not an unambiguous excitation cross-section, and the two-order-of-magnitude contrast between the two sites is not established. The authors should provide an independent calibration of the per-site emission efficiency, for example by resonant excitation or by per-site absolute quantum efficiency measurements, or they should explicitly quantify the systematic uncertainty this assumption introduces.","section":"Effective excitation cross-section of photoluminescence"},{"comment":"The isolation of the OMVPE4 contribution from the OMVPE4/7 peak assumes that OMVPE7 has exactly the same fluence dependence as OMVPE8, differing only by a scaling factor, and that it saturates at the same fluence. Although the manuscript states this is observed, no evidence is shown for the relationship between OMVPE7 and OMVPE8 intensities, and no deconvolution details are provided. Given that OMVPE7 and OMVPE8 are distinct charge states of the same defect configuration (Ref. 8), their capture cross-sections could in principle differ. The authors should justify this scaling or estimate the resulting systematic error in the derived OMVPE4 cross-section.","section":"Effective excitation cross-section of photoluminescence and Figure 1"},{"comment":"The derivation of the trap-concentration upper limit NT = 2.0 x 10^17 cm^-3 is not transparent. Using the values quoted in the same paragraph (Eu* = 2.4 x 10^17 cm^-3 at 77 K and QE = 0.46), the nonradiative recombination density is about 2.8 x 10^17 cm^-3, not 2.0 x 10^17 cm^-3. In addition, the assertion that 'all of the efficient trapping centers will be occupied' at the optimum fluence is not supported by any independent measurement. The trap assignment is also internally inconsistent: the Discussion concludes that the competing trap is more likely related to the Ga vacancy, while the Conclusion states that the authors 'related' the trap to unintentionally doped carbon. These inconsistencies need to be resolved because they are load-bearing for the proposed low-fluence loss mechanism.","section":"Discussion"},{"comment":"Table 1 lists the majority-site concentration as 5 x 10^19 cm^-3, but the Discussion text in the same section states a concentration of 9 x 10^19 cm^-3, and the Introduction says OMVPE4 comprises about 90% of the total 1 x 10^20 cm^-3 Eu concentration. This factor-of-two inconsistency changes the estimated total excited Eu3+ concentration and propagates directly into the trap-concentration estimate. The text and table should be reconciled.","section":"Discussion and Table 1"}],"minor_comments":[{"comment":"The phrase 'The growth temperature of the of the optical active layer' contains a duplicated article; it should read 'The growth temperature of the optical active layer.'","section":"Experimental"},{"comment":"The capture cross-section of the H1 trap is written as '10^-13 cm^-2' in two places; the unit should be cm^2, not cm^-2.","section":"Discussion"},{"comment":"The abstract and the temperature-dependence section report a maximum external QE of 46%, while the Conclusion states 48% at 77 K. These values should be made consistent.","section":"Abstract, main text, and Conclusion"},{"comment":"The Conclusion attributes the competing carrier trap to 'unintentionally doped carbon,' which contradicts the Discussion's conclusion favoring Ga vacancies; this needs to be aligned with the main text.","section":"Conclusion"},{"comment":"The sentence 'A laser repetition rate of 1 kHz ensured that all excited Eu3+ ions relaxed to the ground state before the next pulse arrived' is grammatically awkward; consider rewriting, for example, 'A laser repetition rate of 1 kHz ensures that...'.","section":"Experimental"},{"comment":"The phrase 'of which a fraction of f * sigma_ex = 0.1 is excited' uses notation that could be confused with a product of two variables; clarify that this is the dimensionless product sigma_ex * f = 0.1.","section":"Discussion"}],"recommendation":"major_revision","confidential_remarks":"The equal-efficiency assumption is the central issue. The authors should be asked to supply either per-site quantum-efficiency measurements or a quantitatively justified argument that the radiative efficiencies are equal. The inconsistencies in the trap concentration and the Discussion/Conclusion trap assignment also need careful fixing. If these points are addressed, the paper could make a solid contribution to the field."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is worth a read. What's actually new: site-resolved effective excitation cross-sections for Eu3+ in GaN under both optical and electrical pumping, and a fluence-dependent external QE that peaks at ~0.29 at RT and ~0.46–0.48 at 77 K. The authors separate the two dominant Eu sites and show the minority site has a cross-section roughly two orders of magnitude larger than the majority site, and that a large-cross-section trap limits QE at low fluence while saturation of the high-efficiency sites limits it at high fluence. That story is coherent and the fits are internally consistent.\n\nThe paper does several things well: top-hat excitation profile, integrating sphere for absolute QE, spectral deconvolution, and the use of pulsed excitation to avoid cw trap dynamics. The saturating-trap model is a natural explanation for the low-fluence behavior, and the comparison to prior low QE values is fair.\n\nNow the soft spots, in proportion. The most serious is the trap identity: the Discussion says the carbon concentration is too low (1.2e16 cm-3) and concludes it's likely a Ga vacancy, while the Conclusion says the limiting trap 'we have related to unintentionally doped carbon.' That's a direct contradiction. The numbers also don't line up: the Abstract says 46%, the Conclusion says 48%; the table lists the majority-site concentration as 5e19 cm-3 while the text says 9e19 cm-3. No error bars anywhere. The carbon concentration figure is given without a reference or measurement description.\n\nI don't fully buy the stress-test's claim that the two-order-of-magnitude cross-section contrast rests on the equal-efficiency assumption. The cross-sections come from the exponential curvature of the saturation curves, so a constant efficiency offset scales the ordinate but not the fitted exponent. The assumption matters, though, for the absolute excited fractions used to estimate the total excited Eu concentration and the trap upper limit, and for the statement about the fraction of minority sites excited. That part should be flagged, but not as a fatal flaw.\n\nAll in all, the paper deserves a serious referee. It needs a revision that reconciles the trap conclusion, fixes the numerical inconsistencies, and adds error bars. A reader in nitride optoelectronics or rare-earth doping will get real value from the cross-section data. I'd engage with it.","headline":"Solid quantitative study of Eu:GaN excitation cross-sections and QE limits; the cross-section contrast is likely robust, but inconsistencies in trap identity and numbers need fixing.","tokens_in":7805,"tokens_out":5119,"would_cite":true,"duration_ms":49663,"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":"The paper identifies two europium sites in GaN whose very different excitation cross-sections set both the peak and the ceiling of red emission efficiency.","keywords":["Eu-doped GaN","red luminescence","photoluminescence quantum efficiency","excitation cross-section","rare-earth doping","carrier trap","micro-LED","gallium nitride"],"falsifier":"Perform site-selective resonant excitation or calibrated decay measurements to determine the radiative efficiency of the majority OMVPE4 site independently of the equal-lifetime assumption. If its true radiative efficiency is much lower than that of the minority site, the rescaled excited-fraction curves and the claimed two-order-of-magnitude cross-section contrast, and with them the high-fluence efficiency ceiling, would need to be revised.","tokens_in":6716,"feed_emoji":"🔴","tokens_out":9021,"duration_ms":81358,"temperature":0.7,"pith_summary":"This paper sets out to explain what controls the efficiency of red luminescence from europium-doped gallium nitride, a candidate red emitter on the GaN platform. It reports that the few europium ions sitting in a minority site are about two orders of magnitude easier to excite than the abundant majority-site ions, so they dominate at low pump fluence while the majority site only begins to contribute under strong pumping. The external quantum efficiency is limited at low fluence by an efficient carrier trap that steals excitation energy, and at high fluence by saturation of the scarce efficient sites combined with the tiny cross-section of the majority site. Under pulsed excitation at low temperature, the external quantum efficiency reaches 46%, which makes the material attractive for red micro-LEDs operated at low current density.","feed_headline":"Europium in GaN reaches 46% red efficiency","feed_subtitle":"Two europium sites, one far easier to excite, explain why red emission peaks at a specific pump strength.","key_machinery":"The argument is carried by a two-level saturable-excitation kinetics model. For a short excitation pulse the fraction of excited Eu ions obeys $Eu^* = Eu_{tot}(1-e^{-\\sigma_{ex}f})$, and for steady current injection the excited fraction saturates as $Eu^* = Eu_{tot}\\sigma_{ex}\\tau j/q / (1+\\sigma_{ex}\\tau j/q)$. Spectral deconvolution separates the minority OMVPE8 peak from the overlapping OMVPE4/7 peak; because the PL lifetimes are similar across sites, the paper assumes similar emission efficiencies and uses the OMVPE8 saturation curve to rescale and subtract the OMVPE7 contribution, isolating the majority OMVPE4 response. The model then attributes the low-fluence rise of quantum efficiency to filling of a competing carrier trap and the high-fluence fall to saturation of the efficient minority sites.","core_discovery":"The central claim is a quantitative account of excitation in GaN:Eu. Using a saturating-exponential model, $Eu^* = Eu_{tot}(1-e^{-\\sigma_{ex}f})$, for pulsed above-bandgap excitation, the paper derives effective excitation cross-sections of $\\sigma_{ex} \\approx 1.6\\times10^{-15}$ cm$^2$ for the minority OMVPE8 site and $\\sigma_{ex} \\approx 1.2\\times10^{-17}$ cm$^2$ for the majority OMVPE4 site; under current injection the corresponding values are $3.0\\times10^{-15}$ cm$^2$ and $6.0\\times10^{-18}$ cm$^2$. Because the minority sites, though comprising under ten percent of the europium ions, are so much easier to excite, they dominate the emission at low fluence. The external quantum efficiency is then governed by two competing limits: a carrier trap with an effective cross-section around $10^{-13}$ cm$^2$ suppresses the low-fluence efficiency until it is filled, and the scarcity of the efficient minority sites plus the small cross-section of the majority sites caps the high-fluence efficiency. With this behaviour the external quantum efficiency peaks near a photon fluence of $7\\times10^{13}$ cm$^{-2}$, reaching about 0.29 at room temperature and 0.46 at 77 K.","pith_inferences":["The same saturable-trap logic implies that pulsed or modulated drive at constant average power should outperform continuous-wave drive in Eu:GaN, because cw excitation lets the competing trap recapture carriers many times during a europium decay; the paper notes this direction but does not measure a duty-cycle series.","If the limiting low-fluence trap is a gallium vacancy, then samples grown or annealed to vary the III/V ratio should show a directly proportional change in the low-fluence quantum efficiency; this is a testable prediction the paper does not perform.","The two-order-of-magnitude cross-section gap suggests that converting majority sites into minority-like configurations, through co-doping or growth conditions that change the local defect environment, could lift the high-fluence efficiency ceiling substantially, a route the authors leave unexplored.","A direct site-selective measurement of the majority site's radiative efficiency would settle whether the reported cross-section contrast survives without the equal-efficiency assumption, and would turn the apparent two-site model into a firmly measured one."],"forward_implications":["Eu-doped GaN red LEDs should be designed to run at the pump fluence or current density where the carrier trap is filled but the minority sites are not yet saturated, rather than at the highest possible drive.","The measured 46% external quantum efficiency at low temperature is a benchmark for red emitters on GaN; operating in the pulsed, low-fluence regime used here should be part of any comparison.","Reducing the density of the large-cross-section carrier trap (the paper argues it is likely gallium-vacancy-related) should raise the low-fluence quantum efficiency directly.","Because the majority OMVPE4 site has an excitation cross-section near $10^{-17}$ cm$^2$, increasing the europium concentration alone will not raise high-fluence output unless the majority site's excitation pathway is also made more efficient.","Micro-LEDs, which typically run at low current densities, align with the regime where the efficient minority sites dominate, making GaN:Eu a promising red emitter for that application."],"supporting_citations":[{"why":"Earlier absolute external photoluminescence quantum efficiency measurements under cw and pulsed excitation; the paper contrasts its higher pulsed QE values and explains the discrepancy.","marker":"[17]"},{"why":"Pulsed-laser re-excitation study that established the high-carrier-density Auger-type de-excitation channel and the 1 kHz pulse-rate condition used here.","marker":"[9]"},{"why":"Identifies OMVPE7 and OMVPE8 as the same defect configuration in different charge states, the basis for separating and rescaling the minority-site contribution.","marker":"[8]"},{"why":"Provides the classification of OMVPE4 (Eu1) and minority centers and reports sub-bandgap excitation of the majority site's associated defect.","marker":"[5]"},{"why":"Shows donor-acceptor pair recombination excites the majority site, used to explain why its optical and electrical cross-sections differ.","marker":"[21]"},{"why":"Documents the H1 hole trap and its large capture cross-section, the candidate competing trap invoked in the low-fluence regime.","marker":"[13]"},{"why":"Carbon-related deep-level data used to argue, together with the measured carbon concentration, that the competing trap is more likely linked to gallium vacancies.","marker":"[16]"}],"fun_headline_variants":["Two europium sites set GaN red emission limit","Trap and scarce sites cap Eu:GaN efficiency at 46%","Eu in GaN hits 46% but only at right pump strength","GaN:Eu efficiency pinned by minority site cross-section","Exciting europium in GaN: easy sites win at low fluence"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole two-order-of-magnitude contrast between the minority and majority europium sites rests on the assumption that all sites emit with the same intrinsic efficiency because their photoluminescence lifetimes look alike; if the rare bright sites actually emit light more efficiently than the abundant sites, the rescaled excited-fraction curves and the predicted efficiency ceiling would not be valid.","fun_headline_variants_meta":{"raw":{"variants":["Two europium sites set GaN red emission limit","Trap and scarce sites cap Eu:GaN efficiency at 46%","Eu in GaN hits 46% but only at right pump strength","GaN:Eu efficiency pinned by minority site cross-section","Exciting europium in GaN: easy sites win at low fluence"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000413,"raw_usage":{"total_tokens":2145,"prompt_tokens":965,"completion_tokens":1180,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":581,"completion_tokens_details":{"reasoning_tokens":1089}},"tokens_in":581,"tokens_out":1180,"duration_ms":8360,"temperature":1.0,"reasoning_tokens":1089,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:14:11.916168+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Perform site-selective resonant excitation or calibrated decay measurements to determine the radiative efficiency of the majority OMVPE4 site independently of the equal-lifetime assumption. If its true radiative efficiency is much lower than that of the minority site, the rescaled excited-fraction curves and the claimed two-order-of-magnitude cross-section contrast, and with them the high-fluence efficiency ceiling, would need to be revised.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier absolute external photoluminescence quantum efficiency measurements under cw and pulsed excitation; the paper contrasts its higher pulsed QE values and explains the discrepancy."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Pulsed-laser re-excitation study that established the high-carrier-density Auger-type de-excitation channel and the 1 kHz pulse-rate condition used here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Identifies OMVPE7 and OMVPE8 as the same defect configuration in different charge states, the basis for separating and rescaling the minority-site contribution."},{"cited_title":"& Dierolf, V","cited_arxiv_id":null,"evidence_quote":"Provides the classification of OMVPE4 (Eu1) and minority centers and reports sub-bandgap excitation of the majority site's associated defect."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows donor-acceptor pair recombination excites the majority site, used to explain why its optical and electrical cross-sections differ."},{"cited_title":"& Shiojima, K","cited_arxiv_id":null,"evidence_quote":"Documents the H1 hole trap and its large capture cross-section, the candidate competing trap invoked in the low-fluence regime."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Carbon-related deep-level data used to argue, together with the measured carbon concentration, that the competing trap is more likely linked to gallium vacancies."}],"review_version":1}