{"id":"a1846c8f-7f60-4fea-8d7d-4b53000907f0","arxiv_id":"1908.01101","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":2,"one_line_summary":"The high silicon flux observed when doping beta-Ga2O3 by plasma-assisted MBE comes from volatile SiO produced by oxidation of the silicon source, not from silicon's own vapor pressure.","lead":"This paper shows that silicon doping during a crystal-growth method called plasma-assisted molecular beam epitaxy gets a boost because oxygen in the chamber reacts with the hot silicon source to form a gas called silicon monoxide. The result tells growers of beta-gallium oxide, a material for power electronics, why silicon doping depends on oxygen pressure and time rather than just on the source temperature.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The paper's quantitative support (Eq. 5 vs Eq. 6) rests entirely on the unvalidated assumption that SIMS Si3s equals delivered SiO fluence; a pressure- or flux-dependent sticking coefficient would distort the fitted exponent and temperature trend.","rationale":"I read the paper as an experimental attribution of anomalous Si flux to active oxidation forming volatile SiO. The multi-pronged evidence is coherent: the discrepancy with Si vapor pressure, the weak temperature dependence from 800 to 1000 C, the oxygen-pressure exponent near 0.5, the plasma-power suppression, and the extended-exposure passivation. The load-bearing quantitative connection is the match between Eq. (5) and Eq. (6), which converts measured Si3s values into source flux values. That conversion assumes unity sticking and perfect SIMS integration, and the paper provides no control experiment or independent flux measurement to support it. The reader identified exactly this weakest assumption. My proposed check would settle it by measuring the SiO flux directly. If the direct flux measurement reproduces the same exponents and trends, the mechanism is confirmed; if not, the quantitative claims are weakened even though the qualitative mechanism may still be correct. The Fig. 2 caption/text mismatch and missing error bars are real but secondary. Because the concern is specific and testable but does not by itself overturn the qualitative mechanism, the reader's CONDITIONAL verdict stands without change.","tokens_in":6790,"tokens_out":5984,"duration_ms":72712,"concrete_test":"Use a retractable line-of-sight quadrupole mass spectrometer (or a quartz crystal microbalance) placed at the substrate position to measure the actual SiO flux from the Si cell under the exact temperature and oxygen-pressure conditions of Fig. 2(b), without relying on SIMS. Compare the independently measured flux-vs-pressure exponent and temperature trend to Eq. (5). If the independent measurement reproduces P^0.44 and a 28% flux increase over 200 C, the SIMS integration assumption is validated and the central mechanism stands; if it yields a different exponent or a stronger T dependence, the paper's quantitative claim is an artifact of Si incorporation or desorption in the epitaxial layer rather than a property of the source.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative support for the SiO mechanism is the near-0.5 oxygen-pressure exponent in Eq. (5), which is interpreted through Eq. (6) as the stoichiometric signature of reaction (3). This interpretation is valid only if the SIMS-integrated sheet density Si3s measures the delivered SiO fluence with a constant, unity conversion factor. The paper explicitly assumes this ('all the Si atoms are expected to stick to the epitaxial layer irrespective of whether the growth is on or off', oxygen-pressure variation section), but no control verifies it. During the 3-s shutter pulses the substrate is held at 630 C with the Ga shutter closed; if Si desorbs, segregates, or incorporates with a pressure- or flux-dependent sticking coefficient, then the fitted 0.44 exponent and the weak temperature trend (28% over 200 C) are artifacts of the measurement rather than properties of the source chemistry. The qualitative story (weak T, strong P, plasma suppression, passivation) is plausible and well supported, but the quantitative match to P^0.5 is the load-bearing confirmation and it rests on an unvalidated metrological assumption. SIMS error bars and replicate runs are also absent, so the agreement between 0.44 and 0.5 cannot be assessed.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes that the anomalously high Si doping flux observed in plasma-assisted MBE of beta-Ga2O3 arises from active oxidation of the Si source to volatile SiO, rather than from Si sublimation. The authors support this mechanism with SIMS-derived integrated Si sheet densities (Si3s) measured under variations of Si cell temperature, oxygen chamber pressure, plasma power, and shutter-open time. They report a weak temperature dependence (28% increase between 800 and 1000 °C), a power-law oxygen-pressure dependence with fitted exponent 0.44 close to the stoichiometric 0.5 expected for Si(s) + 1/2 O2(g) -> SiO(g), suppression by activated oxygen, and a time-dependent reduction of Si flux under extended shutter-open conditions due to SiO2 formation. The paper concludes that Si flux is controlled by oxidation chemistry at the source, a result of practical relevance for Si doping in oxide MBE.","tokens_in":6889,"tokens_out":5960,"duration_ms":59129,"significance":"If the proposed mechanism is correct, it provides a coherent explanation for why Si doping in oxide MBE requires much lower cell temperatures than in non-oxide MBE, and it identifies source oxidation chemistry, not Si sublimation, as the flux-limiting process. The qualitative picture is supported by several independent observables (temperature trend, pressure trend, plasma-power effect, time-dependent passivation), and the comparison with D'Evelyn's reaction-probability data provides a useful external anchor. The main weakness is that the quantitative confirmation rests on the unvalidated assumption that SIMS-integrated Si3s equals the delivered SiO fluence with a constant, unity conversion factor, and the pressure-exponent fit includes a plasma-off data point together with plasma-on points. These issues are addressable with additional experiments and analysis, so the central claim is defensible but needs revision.","major_comments":[{"comment":"The load-bearing quantitative claim is the fitted exponent 0.44, which is interpreted through Eq. (6) as the stoichiometric signature of reaction (3). This interpretation requires that Si3s is a faithful, constant-factor measure of the SiO flux delivered to the substrate. The manuscript states that 'all the Si atoms are expected to stick to the epitaxial layer irrespective of whether the growth is on or off' (oxygen-pressure variation section), but no experiment verifies this. If Si desorbs from the Ga2O3 surface at 630 °C, or if its sticking coefficient depends on oxygen pressure or on the surface condition during the 3-s pulse, the measured exponent and the temperature trend would be artifacts of the measurement rather than properties of the source chemistry. The authors should provide replicate runs, SIMS error bars, and a control experiment (e.g., varying the substrate temperature or the post-pulse dwell time) that tests the unity-sticking assumption.","section":"Oxygen-pressure variation (Eq. 5)"},{"comment":"The power-law fit of Eq. (5) includes the 5 × 10−10 Torr data point taken after 12 h of pumping with the plasma off, whereas all other pressure points have the plasma on. Because Fig. 2(c) shows that activated oxygen strongly suppresses SiO flux, the 5 × 10−10 Torr point represents a different oxidation regime from the plasma-on points. Inclusion of this point in the fit can bias the exponent; the authors should refit the data excluding this point, or explicitly model the plasma contribution, or measure the pressure dependence with the plasma off for all points.","section":"Oxygen-pressure variation (Eq. 5)"},{"comment":"The manuscript asserts that under equilibrium active oxidation, the Si flux should scale as P_O2^0.5, but no derivation is given and the MBE environment is not in equilibrium because SiO is continuously removed by pumping. The 0.5 exponent follows from stoichiometry only under an additional assumption, such as local equilibrium at the surface or a specific rate-limiting step (e.g., dissociative adsorption of O2). The authors should state that kinetic assumption explicitly or cite a kinetic model for active oxidation under non-equilibrium, plasma-assisted conditions. Without this, the agreement between the fitted 0.44 and the theoretical 0.5 is suggestive but not a quantitative confirmation.","section":"Eq. (6) and theoretical exponent"}],"minor_comments":[{"comment":"The caption of Fig. 2 lists panels as (a) oxygen chamber pressure, (b) Si cell temperature, and (c) plasma power, but the text refers to Fig. 2(a) as temperature and Fig. 2(b) as oxygen pressure. Please correct the panel labels or the in-text references.","section":"Fig. 2 caption and text"},{"comment":"Please specify the units of the fitting constant k and state the confidence interval or R2 for the fitted exponent 0.44, so readers can assess the closeness to 0.5.","section":"Eq. (5)"},{"comment":"The text states that 'group II elements like Si' provide shallow donor levels; Si is a group IV element. This appears to be a typo and should be corrected.","section":"Introduction"},{"comment":"The sentence listing 'seven different oxygen pressures' actually lists eight values (3e-6, 5e-6, 1e-5, 1.25e-5, 1.5e-5, 1.7e-5, 1.9e-5, and 5e-10 Torr). Please correct the count or clarify that the base-pressure point is additional.","section":"Oxygen-pressure variation"},{"comment":"Reference [27] is not cited in the text; either cite it where relevant or remove it from the reference list.","section":"References"},{"comment":"In Eq. (2), the symbol rendered as π for the atomic mass should be m (or the appropriate mass variable), and the equation formatting should be cleaned up to avoid confusion.","section":"Eq. (2)"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and addresses a topic of practical importance. The central mechanism is plausible and qualitatively well supported, but the quantitative confirmation of the P^0.5 law depends on an untested sticking-coherence assumption and on a fit that mixes plasma-on and plasma-off conditions. I recommend major revision rather than rejection because the issues are fixable with additional control experiments and refitting, and the qualitative story is strong."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core claim holds up: the high Si flux in plasma-assisted MBE of beta-Ga2O3 is best explained by active oxidation of the Si source to volatile SiO, not by Si sublimation. The multi-pronged evidence is genuinely good—the four-order-of-magnitude discrepancy from vapor pressure, the weak temperature dependence (28% over 200 C, matching D'Evelyn), the pressure exponent near 0.5, the suppression by plasma power, and the time-dependent passivation with SiO2 all point in the same direction. That is a solid piece of experimental mechanism work, and it gives the community useful knobs: oxygen pressure and plasma power, not just cell temperature.\n\nWhat is new here is the systematic SIMS-based demonstration in an actual MBE growth environment. Prior reports noted the anomalously low cell temperatures, but this paper connects them to a specific reaction path and tests it across pressure, temperature, plasma power, and time. I think the mechanism is probably right.\n\nThe soft spots are real but not fatal. The biggest one is the metrological assumption that integrated Si3s from SIMS equals delivered SiO fluence with unity, constant conversion. The paper states this explicitly at 630 C, but it is not directly verified. If the Si sticking coefficient or incorporation efficiency changes with oxygen pressure, the fitted exponent 0.44 could be a measurement artifact. The linear shutter-time series in Fig. 3(a) helps, because a constant flux over short pulses supports the integration idea, but it does not rule out pressure-dependent sticking. This is the load-bearing quantitative confirmation (Eq. 5 vs 6), so it deserves at least one control or a clear discussion. Also, there are no error bars or replicates on the SIMS data, and the power-law fit is reported without uncertainty, so we cannot tell whether 0.44 is meaningfully different from 0.5. A minor but annoying issue: the Fig. 2 caption and the text disagree on which panel is temperature and which is pressure; that needs fixing.\n\nI also think the paper would be stronger if it cited [27] (Krishnamoorthy et al., 2017) where prior flux values appear to be discussed; that does not weaken the mechanism, but it would put the new data in context.\n\nOverall: the qualitative story is convincing, and the quantitative support is plausible but less solid than the paper implies. A serious referee should see this, not a desk reject. The right outcome is a conditional accept requiring error analysis, a check on the SIMS-sticking assumption, and a fix to the figure/text mismatch. I would cite this paper if I were working on oxide MBE doping.","headline":"Convinces me that SiO active oxidation explains the anomalous Si flux in beta-Ga2O3 MBE, but the quantitative case rests on an unvalidated SIMS-fluent assumption and missing uncertainties.","tokens_in":7624,"tokens_out":1725,"would_cite":true,"duration_ms":19069,"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 establishes that silicon doping in plasma-assisted molecular beam epitaxy of beta-Ga2O3 is controlled by the formation of volatile SiO at the silicon source, not by silicon's sublimation vapor pressure.","keywords":["Si doping","beta-Ga2O3","plasma-assisted MBE","volatile SiO","active oxidation","SIMS","molecular beam epitaxy","oxide semiconductors"],"falsifier":"Direct line-of-sight mass spectrometry of the species leaving the silicon cell during oxygen exposure: if mass 44 (SiO) flux does not rise with oxygen pressure roughly as $P_{O_2}^{0.5}$ and remain nearly flat with cell temperature, the active-oxidation mechanism is wrong. A simpler check is to pre-oxidize the cell and show that no silicon doping occurs until the >1100 degrees Celsius refresh step is performed.","tokens_in":6451,"feed_emoji":"🧪","tokens_out":5295,"duration_ms":52831,"temperature":0.7,"pith_summary":"The paper argues that the anomalously high silicon flux seen when doping beta-Ga2O3 by plasma-assisted MBE comes from a chemical reaction at the silicon cell: oxygen in the chamber oxidizes solid silicon into volatile SiO gas, which then delivers the dopant. This explains why the measured doping depends weakly on cell temperature but strongly on oxygen pressure, following a power law with exponent near 0.5. It also explains why long shutter openings and higher plasma power reduce the flux: extended exposure to activated oxygen converts the silicon surface to solid SiO2, shutting off the SiO supply until the cell is refreshed at high temperature. The result matters because it turns Si doping in oxide MBE from a sublimation problem into an oxidation-chemistry problem, with oxygen pressure and plasma conditions as the practical control knobs.","feed_headline":"Silicon dopes beta-Ga2O3 by escaping as volatile SiO","feed_subtitle":"SIMS data show the doping flux tracks oxygen pressure as P^0.44, not silicon's vapor pressure.","key_machinery":"The load-bearing mechanism is the Si-O pressure-temperature diagram, specifically the boundary between active oxidation, which produces volatile SiO and leaves a clean Si surface, and passive oxidation, which forms solid SiO2. The three-step surface process (oxygen adsorption, surface reaction to SiO, and SiO desorption) converts the silicon source into a chemical reactor whose output flux is set by the oxygen partial pressure through the stoichiometric relation $Si_{3s} = k P_{O_2}^{0.5}$. The paper uses SIMS-derived silicon sheet density from short shutter pulses as the operational measure of that flux.","core_discovery":"The central claim is that in plasma-assisted MBE growth of $\\beta$-Ga2O3, the silicon dopant flux is not limited by the vapor pressure of elemental silicon but by the formation of volatile silicon monoxide, SiO. In the Si-O pressure-temperature regime of the MBE chamber, solid silicon undergoes active oxidation $\\mathrm{Si(s)} + \\frac{1}{2}\\mathrm{O_2(g)} \\rightarrow \\mathrm{SiO(g)}$, and this reaction sets the delivered flux. Evidence comes from integrated SIMS sheet densities from 3-second Si shutter pulses: the sheet density changed only 28% as the cell temperature went from 800 to 1000 degrees Celsius, whereas sublimation would predict about three orders of magnitude; the same sheet density scaled with oxygen chamber pressure as $P_{O_2}^{0.44}$, close to the stoichiometric exponent 0.5; and raising plasma power (more activated oxygen) suppressed the flux by pushing the surface toward passive oxidation and SiO2 formation. Extended shutter-open time likewise quenched the flux after about 15 minutes, consistent with growth of a solid SiO2 layer, and heating the cell above 1100 degrees Celsius restored it via the disproportionation reaction $\\mathrm{SiO_2(s)} + \\mathrm{Si(s)} \\rightarrow 2\\mathrm{SiO(g)}$.","pith_inferences":["If the active-oxidation mechanism is general, other dopant sources that form volatile suboxides, such as germanium (GeO) and tin (SnO), should show analogous oxygen-pressure-dependent fluxes in oxide MBE; this is a testable prediction for future work.","The near-0.5 pressure exponent suggests that doping calibrations in oxide MBE could be expressed as universal functions of oxygen pressure and plasma power rather than cell temperature, potentially simplifying transfer of recipes between different MBE systems.","A direct line-of-sight mass spectrometry measurement of mass 44 (SiO) during oxygen exposure would cleanly confirm the mechanism and quantify the active-oxidation probability, which the SIMS integration approach only infers indirectly.","The finite lifetime of the SiO source before SiO2 passivation implies that the Si cell has a limited total dopant dose per refresh cycle; this may set practical design constraints for high-dose or long-duration doping profiles."],"forward_implications":["Si cell temperature is a weak control knob over a wide range, so reproducible doping in oxide MBE must be controlled primarily through oxygen pressure and plasma conditions.","Oxygen pressure stability during growth becomes a doping-stability requirement: fluctuations in chamber pressure will translate directly into fluctuations in delivered silicon flux.","Sustained doping runs with the shutter open for tens of minutes will suffer a decaying flux as the cell surface passivates, so long doping layers may need pulsed shutter operation or a mid-growth refresh step.","Higher oxygen plasma power, often desirable for oxide growth quality, suppresses SiO formation and therefore reduces silicon doping efficiency, creating a trade-off that growers must manage.","A high-temperature (>1100 degrees Celsius) silicon refresh step is an effective, practical way to restore a passivated Si source, and should be included in oxide MBE procedures for reproducible doping."],"supporting_citations":[{"why":"Supplies the vapor pressure data for silicon used to compute the expected sublimation flux that the paper's data contradict.","marker":"[26]"},{"why":"Provides the Si-O pressure-temperature diagram distinguishing active oxidation (SiO) from passive oxidation (SiO2), and the disproportionation reaction used for cell refresh.","marker":"[28]"},{"why":"Reports measured active-oxidation reaction probability versus temperature, whose ~30% increase between 767 and 987 degrees Celsius matches the observed 28% flux increase.","marker":"[30]"},{"why":"Gives the Clausius-Clapeyron relation that defines the baseline vapor-pressure-limited flux the paper argues against.","marker":"[24]"},{"why":"Provides the effusion-cell flux equation used to convert silicon vapor pressure to the expected substrate flux.","marker":"[25]"},{"why":"Earlier demonstrations of Si-doped beta-Ga2O3 at unexpectedly low cell temperatures, which establish the anomaly the paper's mechanism explains.","marker":"[9, 22, 23]"}],"fun_headline_variants":["Silicon dopes via volatile SiO, not silicon vapor","Si flux in beta-Ga2O3: SiO, not Si vapor, is the culprit","Oxygen pressure governs Si doping in beta-Ga2O3 MBE","SiO strips Si from the cell: a new doping mechanism","Beta-Ga2O3 MBE: Si doping is SiO-driven, not sublimation"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The integrated SIMS silicon sheet density from a 3-second shutter pulse is a faithful measure of the SiO flux delivered by the cell; if silicon atoms desorb, migrate, or incorporate unevenly during the growth interrupts, the fitted 0.44 pressure exponent and the weak temperature trend would misrepresent the source chemistry.","fun_headline_variants_meta":{"raw":{"variants":["Silicon dopes via volatile SiO, not silicon vapor","Si flux in beta-Ga2O3: SiO, not Si vapor, is the culprit","Oxygen pressure governs Si doping in beta-Ga2O3 MBE","SiO strips Si from the cell: a new doping mechanism","Beta-Ga2O3 MBE: Si doping is SiO-driven, not sublimation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000379,"raw_usage":{"total_tokens":2015,"prompt_tokens":944,"completion_tokens":1071,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":560,"completion_tokens_details":{"reasoning_tokens":971}},"tokens_in":560,"tokens_out":1071,"duration_ms":10640,"temperature":1.0,"reasoning_tokens":971,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:24:02.414982+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Direct line-of-sight mass spectrometry of the species leaving the silicon cell during oxygen exposure: if mass 44 (SiO) flux does not rise with oxygen pressure roughly as $P_{O_2}^{0.5}$ and remain nearly flat with cell temperature, the active-oxidation mechanism is wrong. A simpler check is to pre-oxidize the cell and show that no silicon doping occurs until the >1100 degrees Celsius refresh step is performed.","supporting_citations":[{"cited_title":"Journal of physical and chemical reference data 15.3: 967-983","cited_arxiv_id":null,"evidence_quote":"Supplies the vapor pressure data for silicon used to compute the expected sublimation flux that the paper's data contradict."},{"cited_title":"and Gustafsson, T., Surface Review and Letters, 6(01), pp.45-52","cited_arxiv_id":null,"evidence_quote":"Provides the Si-O pressure-temperature diagram distinguishing active oxidation (SiO) from passive oxidation (SiO2), and the disproportionation reaction used for cell refresh."},{"cited_title":"and Engel, T., 1987","cited_arxiv_id":null,"evidence_quote":"Reports measured active-oxidation reaction probability versus temperature, whose ~30% increase between 767 and 987 degrees Celsius matches the observed 28% flux increase."},{"cited_title":"Journal of Chemical Education 28.8: 428","cited_arxiv_id":null,"evidence_quote":"Gives the Clausius-Clapeyron relation that defines the baseline vapor-pressure-limited flux the paper argues against."},{"cited_title":"Molecular Beam Epitaxy","cited_arxiv_id":null,"evidence_quote":"Provides the effusion-cell flux equation used to convert silicon vapor pressure to the expected substrate flux."}],"review_version":1}