{"id":"20a8cba3-0e5d-45cd-ba31-485466edef19","arxiv_id":"1908.03036","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Aluminium deposited on germanium and annealed at 350 degrees Celsius forms a p+ Al-Ge contact layer whose aluminium diffusion depth is roughly four times the deposited film thickness.","lead":"This paper shows that annealing a thin aluminium layer on a germanium crystal at 350 degrees Celsius creates a rugged p+ contact via solid-state regrowth. The authors map how aluminium and germanium interdiffuse with depth and how the doping level drops from the surface into the crystal.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"SIMS baseline interpretation in Section III.B is the load-bearing weak point: unannealed Al/Ge controls already show Al tails as deep as the film thickness, so the 'four times' diffusion scaling may be partly a measurement artifact.","rationale":"The reader's weakest_assumption correctly identifies the SIMS baseline interpretation in Section III.B as the key vulnerability. The unannealed samples show Al tails exactly equal to the Al film thickness, and the paper dismisses these as stray chamber Al without any control experiment. This matters because the headline quantitative claim, a roughly 4x relation between Al diffusion depth and Al film thickness, is read directly off annealed SIMS tails with no correction for the baseline effect. If the baseline is an artifact, the scaling is not quantitatively supported, and the supporting p+/p gradient inferred from diffusion depth and depth-resolved Hall data becomes uncertain. I agree with the reader's CONDITIONAL verdict: the qualitative phenomenon, Al-Ge solid-state regrowth producing a p-type doped region on Ge, is credible and consistent with the cited 1970s literature, and the Hall and sheet-resistance data do show a conductivity gradient. However, the paper's specific quantitative claims about diffusion length and near-surface concentration are not secure without controls, error bars, or calibration. No change to the reader's verdict is needed; the conditionality should stand.","tokens_in":8939,"tokens_out":6212,"duration_ms":74013,"concrete_test":"Re-analyze the raw SIMS data behind Figs. 2 and 3 by aligning the Al/Ge interface for each annealed sample and its paired unannealed control, then subtract the unannealed Al profile from the annealed Al profile at every depth. If the depth at which the excess Al signal exceeds background is approximately 3t rather than 4t, or if the unannealed tail persists on a Ge-only sample that received no Al evaporation, the 'four times' scaling and the inferred diffusion lengths must be revised downward.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section III.B infers thermal Al diffusion from raw SIMS tail depths in annealed samples (200 nm and 480 nm for 60 nm and 120 nm Al; 60 nm and 580 nm for 15 nm and 150 nm Al) and from this derives the central quantitative claim that Al diffusion into Ge extends to about four times the as-deposited Al film thickness. The unannealed controls, however, already show Al tails extending one full film thickness into Ge (60 nm and 120 nm), which the authors attribute to stray Al in the evaporation chamber rather than to room-temperature diffusion. This attribution is asserted without a control measurement on an Al-free Ge sample or a SIMS artifact assessment. If the unannealed tail is SIMS mixing, knock-on, or crater/interface roughening, the same artifact contributes to every annealed profile. The appropriate thermal-diffusion depth would then be the annealed-minus-unannealed excess, roughly 3t rather than 4t, or possibly not linear in film thickness at all. Because the thickness-controlled regrown-layer length is presented as the paper's new quantitative result and is used to support the p+/p gradient picture, this baseline interpretation is load-bearing. The internal inconsistency between '10^23/cc' in Section III.B and '10^20/cm3' in the Conclusion further indicates that the reported concentration values are not quantitatively reliable.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports fabrication of p+ contacts on p-type Ge crystals by thermal evaporation of Al followed by annealing at 350°C and slow cooling. The authors use SIMS depth profiling to argue that Ge diffuses through the entire Al layer and that Al diffuses into Ge to a depth of approximately four times the as-deposited Al film thickness. Hall measurements on progressively lapped Al-Ge layers are used to infer an Al concentration gradient and p+/p doping profile. The paper claims that the regrown contact layer is rugged and that its thickness can be controlled by the deposited Al thickness. The central quantitative claims are the four-times diffusion scaling and top-surface Al dopant concentrations near 10^20/cm3.","tokens_in":9081,"tokens_out":3487,"duration_ms":36868,"significance":"If the results hold, the paper offers a simple, low-temperature route to p+ contacts on Ge, relevant to detector and electronic device applications. The systematic variation of Al film thickness and the combination of SIMS and Hall measurements are strengths. The paper builds on a known solid-phase regrowth approach and adds a claimed thickness control law. However, the quantitative claims rest on a small number of single measurements and on a SIMS baseline interpretation that is not adequately controlled. The qualitative picture of interdiffusion and p-type doping is plausible, but the load-bearing 'four times thickness' scaling and the reported concentration values need stronger support.","major_comments":[{"comment":"The unannealed Al/Ge SIMS profiles already show Al tails extending 60 nm and 120 nm into Ge, which the authors attribute to stray Al in the evaporation chamber rather than to room-temperature diffusion. This attribution is asserted without a control measurement on an Al-free Ge sample or an assessment of SIMS artifacts such as ion-beam mixing, knock-on, or crater/interface roughening. Because the annealed diffusion depths (200 nm and 480 nm for 60 nm and 120 nm Al, respectively; 60 nm and 580 nm for 15 nm and 150 nm Al) are read from the same kind of raw tail, any such artifact would contribute to every annealed profile. The appropriate thermal diffusion depth would then be the annealed-minus-unannealed excess, roughly 3t rather than 4t, and the linear scaling with film thickness would not be established. This directly affects the central claim that the regrown layer length is controlled by the as-deposited Al thickness.","section":"III.B, Figs. 2 and 3"},{"comment":"The text states that the Al concentration on the top surface of the regrown Al-Ge layer is 'as high as 10^23/cc' for 120 nm Al, while the Conclusion states that 'Dopant concentration at the top surface ... can be as high as 10^20/cm3.' These values differ by three orders of magnitude. The Hall data in Fig. 4(b) additionally give carrier concentrations near 10^19/cm3 for the 15 ohm sheet resistance layer. The manuscript must reconcile these numbers; as written, the reported Al concentration values are internally inconsistent and the claimed p+ doping level is not reliably quantified.","section":"III.B, Fig. 5(b) and Conclusion"},{"comment":"The depth calibration for the lapped Hall measurements uses sheet resistance values: 'Duration of the CML is calibrated using measured sheet resistance values.' Since sheet resistance is then reported as the dependent variable versus depth, using the same values to assign depths risks circularity. The total lapped depth is stated as approximately 150 nm, but individual depth values plotted in Fig. 4(a) are not independently verified by profilometry or another technique. Please provide independent depth measurements at each CML step or state clearly that the depth axis is inferred from the sheet resistance calibration.","section":"III.B, Hall measurement and depth calibration"},{"comment":"No error bars, replicate measurements, or uncertainty analysis are provided for the SIMS profiles, diffusion lengths, sheet resistances, carrier concentrations, or Hall coefficients. All quantitative conclusions rest on single measurements per condition. Given the observed variability in electrical data (e.g., 1.5 ohm for 120 nm Al versus 7 ohm for 30 nm Al), the paper should report statistics from at least three samples per condition or explicitly justify why a single measurement is representative.","section":"III.B, Figs. 2-5"}],"minor_comments":[{"comment":"The statement 'Al can be found (above 10 counts) up to 60 nm and 120 nm in Ge' does not define the threshold or its units; please specify the detection limit or background level used to define the diffusion depth.","section":"III.B, SIMS threshold"},{"comment":"The text contains inconsistent notation, such as '0C' instead of '°C', and inconsistent use of 'Al/Ge' for unannealed and 'Al-Ge' for annealed samples; please standardize these conventions.","section":"Throughout"},{"comment":"The discussion of the extrinsic-to-intrinsic transition and its shift with Al concentration is difficult to follow; a figure or a more explicit explanation of the carrier compensation and mobility effects would improve clarity.","section":"III.B, conductivity transition discussion"},{"comment":"The reference to 'A1' (Al) and 'AI' appears multiple times; please correct these typographical errors.","section":"I, Introduction"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of the journal and addresses a practical contact fabrication problem. The qualitative picture is plausible and supported by existing literature, and the systematic thickness variation is a useful contribution. However, the quantitative 'four times thickness' scaling and the top-surface concentration claim are load-bearing and currently rest on an uncontrolled SIMS baseline and an internal numerical inconsistency. The required fixes—control experiments, SIMS artifact assessment, independent depth calibration, replicate statistics, and reconciliation of concentration values—are feasible within a revision and do not require a fundamentally new approach. I therefore recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a careful, modest experimental follow-up on a contact technique that has been around since the 1970s. The new bit is a systematic Al-thickness scan (15–150 nm) with SIMS and Hall depth profiling, and the claim that the Al diffusion depth into Ge is roughly four times the as-deposited Al film thickness. The qualitative picture—Ge out-diffuses through the Al layer, Al in-diffuses into Ge, forming a p+ gradient—is supported by the data and consistent with the older literature. The authors are upfront that regrowth itself is known; they are adding process control data.\n\nWhat's genuinely good: the experiment is simple and the characterization is mostly appropriate. The SIMS profiles of annealed versus unannealed samples show clear Ge diffusion to the top of the Al layer. The Hall data at different lapped depths show the expected monotonic decrease in carrier concentration, and the sheet-resistance trend is internally coherent. The paper ships no code or machine-checked proofs, but the data are the product here.\n\nThe soft spots are real but not fatal. The biggest one is the SIMS baseline. Unannealed Al/Ge controls already show Al tails one full film thickness into Ge (e.g., 60 nm and 120 nm). The authors call this stray Al from the evaporation chamber, but they offer no control—no Al-free Ge sample, no check for SIMS knock-on or crater roughness. If that tail is a measurement artifact, it contributes to every annealed profile, and the central quantitative claim (“four times the Al thickness”) becomes annealed-minus-unannealed excess, roughly 3t, or possibly not a clean linear scaling at all. This is load-bearing, because the thickness scaling is the main new result.\n\nSecondary issues: no replicate statistics or error bars on a single-sample-per-thickness dataset; the CML depth calibration uses the same sheet-resistance values being reported, which is mildly circular (though the total depth is known from profilometry); and there is a numerical inconsistency between 10^23/cc in the text and 10^20/cm3 in the conclusion. Also, the detector contact framing is not backed by any diode I–V or detector measurement, so the practical payoff is asserted, not shown.\n\nWho should read it: process engineers working on Ge detectors or Al-Ge contacts will get a useful parameter scan and a clear reminder of an old trick. It is not a new mechanism or device concept. The paper deserves a serious referee, and a good referee should ask for a baseline control, uncertainty quantification, and reconciliation of the concentration numbers before publication.","headline":"Solid process-engineering follow-up on a 1970s contact technique, with a plausible but under-verified thickness-scaling claim.","tokens_in":9777,"tokens_out":2204,"would_cite":false,"duration_ms":22797,"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":"Annealing a thermally deposited aluminum film on p-type germanium at 350°C produces a rugged p+ contact by solid-state regrowth, with aluminum penetrating to roughly four times the film thickness.","keywords":["germanium crystal","p-type contact","aluminum regrowth","solid state regrowth","SIMS depth profile","Hall effect","sheet resistance","carrier concentration"],"falsifier":"A cross-sectional transmission electron microscope scan with elemental mapping through the annealed interface would settle it: if the Al-rich region does not extend roughly four times the Al film thickness into the Ge crystal, the SIMS-based claim is wrong; repeating the SIMS on unannealed samples at lower sputter energy would reveal whether the apparent Al tail is an artifact.","tokens_in":8596,"feed_emoji":"⚡","tokens_out":9834,"duration_ms":95978,"temperature":0.7,"pith_summary":"This paper reports a low-temperature route to a p+ electrical contact on p-type germanium: evaporate aluminum onto the crystal, anneal at 350°C, and cool slowly. The authors' central claim is that solid-state regrowth of Al on Ge produces a rugged Al-Ge layer in which germanium moves through the entire aluminum film and aluminum penetrates into the underlying crystal to roughly four times the deposited film thickness. Depth profiles by secondary-ion mass spectrometry and Hall measurements show a heavily aluminum-doped p+ surface, with carrier concentration near $10^{20}\\,\\mathrm{cm^{-3}}$, that grades down to the bulk p-type crystal, forming a p+/p structure. This matters because germanium diode detectors need thin p+ contacts that are usually made by boron ion implantation, and the paper argues the annealed aluminum contact offers a simpler equivalent.","feed_headline":"350°C aluminum anneal makes p+ contacts on germanium","feed_subtitle":"A single 350°C anneal drives aluminum four film-depths into germanium, replacing boron implants for detector contacts.","key_machinery":"The central mechanism is solid-state regrowth by layer exchange. At 350°C Ge diffuses into the Al film (solid solubility about 0.7%) and, on slow cooling, the supersaturated Ge regrows on the Ge crystal while carrying Al into the crystal. The as-deposited Al thickness acts as the control parameter for the final contact: the paper's SIMS data indicate the Al penetration depth into Ge is roughly four times that thickness. To map the concentration gradient, the paper combines SIMS depth profiling with Hall measurements on surfaces exposed by repeated chemical-mechanical lapping.","core_discovery":"On its own terms, the paper establishes that thermally depositing Al on p-type Ge and annealing at 350°C, below the Al-Ge eutectic of 424°C, then cooling slowly, gives a recrystallized Al-Ge layer with a p+/p doping profile. SIMS depth profiles show Ge present through the whole Al layer after annealing but absent from unannealed films, which the authors read as layer exchange and solid-phase regrowth. The same profiles show Al inside the Ge crystal to depths of about four times the as-deposited Al film thickness, for example roughly 200 nm for 60 nm Al and 480 nm for 120 nm Al. Hall measurements after successive lapping show the top surface is metallic, with Al concentrations near $10^{20}\\,\\mathrm{cm^{-3}}$ falling to about $10^{11}\\,\\mathrm{cm^{-3}}$ toward the bulk, so the contact is p+ at the surface and p-type below.","pith_inferences":["The four-times scaling was tested for Al films from 15 nm to 150 nm with a fixed 40-minute anneal; whether it holds for thicker films or longer anneal times is untested and would be a natural next measurement.","If the same regrowth works on n-type Ge, the p+ contact could serve as the hole-blocking or anode contact in a full detector diode, extending the method beyond p-type substrates.","The paper's interpretation of the unannealed Al tail as stray chamber contamination is directly testable by comparing SIMS profiles taken at different sputter energies; a persistent tail would point to ion-beam mixing instead.","A cross-sectional microscopy measurement of the annealed interface could independently confirm both the depth and the shape of the Al distribution implied by the SIMS profiles."],"forward_implications":["Ge diode detectors could be fitted with p+ contacts by evaporation and annealing, replacing boron ion implantation and the associated apparatus.","The depth of the p+ region can be tuned by choosing the aluminum film thickness, since the paper reports Al penetration scales as about four times that thickness.","The top surface of the regrown Al-Ge layer has low sheet resistance, so the contact can serve directly for electrical readout in a detector.","Because the doping falls off from p+ to p-type, the contact forms a graded hole-selective region rather than an abrupt junction, with carrier concentrations from about $10^{19}$–$10^{20}\\,\\mathrm{cm^{-3}}$ down to about $10^{11}\\,\\mathrm{cm^{-3}}$.","Since the anneal stays below the Al-Ge eutectic, the contact forms without melting, which keeps the process compatible with other low-temperature device steps."],"supporting_citations":[{"why":"Shows that n- and p-type layers can be formed in Ge below 300°C by solid-solid reactions with evaporated metal films, the route this work follows.","marker":"[4]"},{"why":"Establishes solid-phase epitaxial growth of Ge layers from Al and the p-type conductivity of the regrown layer.","marker":"[12]"},{"why":"Shows solid Al can serve as the medium from which Ge regrows on a crystalline Ge substrate, with the regrown layer p-type.","marker":"[13]"},{"why":"Demonstrates a Ge alpha particle detector with p+ contacts made by evaporating Al, annealing, and slow cooling.","marker":"[14]"},{"why":"Shows high-purity germanium gamma-ray spectrometers can use solid-phase regrown p+ contacts, the application motivating this work.","marker":"[15]"},{"why":"Supplies the solubility basis (0.7% at 350°C, near zero at room temperature) for supersaturation and regrowth during slow cooling.","marker":"[16]"},{"why":"Supports the metal-induced layer exchange crystallization mechanism that the paper invokes for Ge mixing through the Al film.","marker":"[17]"},{"why":"Provides earlier evidence of crystallization and mixing of Ge in metal films used to interpret the SIMS depth profiles.","marker":"[18]"},{"why":"Underpins the interpretation of the extrinsic-to-intrinsic transition and temperature-dependent carrier behavior in the doped Ge layers.","marker":"[19]"},{"why":"Gives the electrically active Al concentration achieved by ion implantation as the benchmark for the $10^{20}\\,\\mathrm{cm^{-3}}$ surface doping reported here.","marker":"[20]"}],"fun_headline_variants":["350°C anneal makes p+ Ge contacts via Al solid-state regrowth","Al annealed at 350°C creeps four film-thicknesses into Ge","Solid-state Al regrowth gives p+ Ge contacts at 350°C","Annealed Al on Ge forms p+ contacts without boron implants","Low-temp Al anneal on Ge yields p+ contacts to 4x film depth"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative diffusion-length claim rests on reading the Al signal in unannealed SIMS profiles as stray contamination rather than as a measurement artifact; if that baseline is wrong, the reported 'four times the Al film thickness' depth is not supported.","fun_headline_variants_meta":{"raw":{"variants":["350°C anneal makes p+ Ge contacts via Al solid-state regrowth","Al annealed at 350°C creeps four film-thicknesses into Ge","Solid-state Al regrowth gives p+ Ge contacts at 350°C","Annealed Al on Ge forms p+ contacts without boron implants","Low-temp Al anneal on Ge yields p+ contacts to 4x film depth"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000743,"raw_usage":{"total_tokens":3305,"prompt_tokens":929,"completion_tokens":2376,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":545,"completion_tokens_details":{"reasoning_tokens":2272}},"tokens_in":545,"tokens_out":2376,"duration_ms":17157,"temperature":1.0,"reasoning_tokens":2272,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:26:13.910325+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A cross-sectional transmission electron microscope scan with elemental mapping through the annealed interface would settle it: if the Al-rich region does not extend roughly four times the Al film thickness into the Ge crystal, the SIMS-based claim is wrong; repeating the SIMS on unannealed samples at lower sputter energy would reveal whether the apparent Al tail is an artifact.","supporting_citations":[{"cited_title":"Formation of injecting and blocking contacts on high-resistivity germanium,","cited_arxiv_id":null,"evidence_quote":"Shows that n- and p-type layers can be formed in Ge below 300°C by solid-solid reactions with evaporated metal films, the route this work follows."},{"cited_title":"Solid-phase epitaxial growth of Ge layers,","cited_arxiv_id":null,"evidence_quote":"Establishes solid-phase epitaxial growth of Ge layers from Al and the p-type conductivity of the regrown layer."},{"cited_title":"Solid-phase growth of Ge from evaporated Al Layer,","cited_arxiv_id":null,"evidence_quote":"Shows solid Al can serve as the medium from which Ge regrows on a crystalline Ge substrate, with the regrown layer p-type."},{"cited_title":"Germanium alpha particle detectors with aluminum regrowth p+ contacts,","cited_arxiv_id":null,"evidence_quote":"Demonstrates a Ge alpha particle detector with p+ contacts made by evaporating Al, annealing, and slow cooling."},{"cited_title":"High purity germanium gamma-ray spectrometers with regrowth p+ contacts,","cited_arxiv_id":null,"evidence_quote":"Shows high-purity germanium gamma-ray spectrometers can use solid-phase regrown p+ contacts, the application motivating this work."},{"cited_title":"Some aspects of Ge epitaxial growth by solid solution,","cited_arxiv_id":null,"evidence_quote":"Supplies the solubility basis (0.7% at 350°C, near zero at room temperature) for supersaturation and regrowth during slow cooling."},{"cited_title":"Nucleation and growth kinetics during metal-induced layer exchange crystallization of Ge thin films at low temperatures,","cited_arxiv_id":null,"evidence_quote":"Supports the metal-induced layer exchange crystallization mechanism that the paper invokes for Ge mixing through the Al film."},{"cited_title":"Sigurd, G","cited_arxiv_id":null,"evidence_quote":"Provides earlier evidence of crystallization and mixing of Ge in metal films used to interpret the SIMS depth profiles."},{"cited_title":"Impurity concentration dependent electrical conduction in germanium crystal at low temperatures","cited_arxiv_id":"1907.05067","evidence_quote":"Underpins the interpretation of the extrinsic-to-intrinsic transition and temperature-dependent carrier behavior in the doped Ge layers."},{"cited_title":"Aluminium implantation in germanium: Uphill diffusion, electrical activation, and trapping,","cited_arxiv_id":null,"evidence_quote":"Gives the electrically active Al concentration achieved by ion implantation as the benchmark for the $10^{20}\\,\\mathrm{cm^{-3}}$ surface doping reported here."}],"review_version":1}