{"id":"ca9ddb06-031c-43f5-883a-2d83d025545f","arxiv_id":"1908.04953","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A closed-form expression for short-time signal development in silicon diode sensors shows that impact-ionization gain in an idealized LGAD detector retains an advantage over PIN diodes at frame rates up to 10 GHz.","lead":"This paper derives a simple formula for how fast silicon diode X-ray detectors can generate their signal, comparing ordinary PIN diodes with LGAD diodes that amplify the signal through impact ionization. The authors argue that an idealized LGAD detector with a gain of 30 stays advantageous over a PIN diode even at frame rates near 10 billion frames per second.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 10 GHz claim presumes small space-charge effects, which can become comparable to the bias field at the high fluxes the paper targets.","rationale":"The paper is an honest idealized model; it explicitly states the small-space-charge assumption and warns that drift speeds are assumed saturated. Nevertheless, the headline claim is a quantitative boundary (10 GHz) drawn from a simulation that implements exactly that assumption. The boundary is meaningful only if the assumption is valid at the flux levels to which the boundary is applied. My estimate shows that the assumption can fail at fluxes that are plausible for the intended applications (high-flux XFEL beams), and the paper provides no cross-check. This is not an internal inconsistency but a missing quantitative domain-of-validity argument, so it does not warrant rejection. It does warrant the conditional acceptance the reader recommended, with the additional requirement that the flux/space-charge bound be supplied. I therefore leave the reader's verdict unchanged. I agree with the reader's weakest_assumption; the K-constant and code-sensitivity issues are secondary.","tokens_in":5035,"tokens_out":14992,"duration_ms":152620,"concrete_test":"Estimate the incident flux Φ_min required for a 10σ detection at τ=0.1 ns using the paper's series-noise scaling (Qmin ~ 1/√τ, signal ~ Eq. (11)). For that Φ_min, compute ρ0=Φ_min Eγ/(3.62 λ A d) and the space-charge field E_sc≈eρ0(v_e−v_h)τ/ε_Si. If E_sc exceeds 0.1×(2×10^4 V/cm), rerun the Sec. 3 simulation with a Poisson-corrected field and field-dependent carrier velocities; the 10 GHz claim survives only if the collected charge at τ=0.1 ns changes by less than ~10%.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The claim that a gain-30 LGAD retains an advantage over a PIN at 10 GHz is computed in Sec. 3 under constant saturated drift velocities (v_e=100 µm/ns, v_h=60 µm/ns) and a uniform 2×10^4 V/cm bulk field. The Introduction restricts the treatment to depositions for which 'the space-charge field that arises during the collection of these electron-hole pairs remains small relative to the field created by the reverse bias', but no quantitative criterion is given. In the saturated high-flux regime the paper targets, this is not automatic: for a pair density ρ0 and electron-hole separation L≈(v_e−v_h)τ, the induced space-charge field at the electrodes is E_sc≈eρ0L/ε_Si. At τ=0.1 ns (10 GHz), L≈4 µm. A flux of order 5×10^3 10-keV photons per (100 µm)^2 pixel per pulse yields ρ0≈3×10^13 cm^-3 and E_sc≈2×10^3 V/cm, already ~10% of the bias field; higher fluxes push the field below the saturation threshold. Since Eq. (11) makes the LGAD benefit scale as (1/2)K A_e v_e τ, any field-induced reduction of v_e (and of α_e in the gain layer) shrinks the quadratic term and shifts the 10 GHz boundary. The paper gives no estimate of the flux at which the assumption fails, so the central claim is unverified outside an unspecified flux window.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript develops a closed-form approximate expression for the short-time signal development of silicon diode sensors (PIN and LGAD) following instantaneous, longitudinally-uniform ionization. The derivation uses the Shockley-Ramo induced-charge framework with saturated drift velocities and a leading-order treatment of impact ionization, yielding a collected-charge expression, Eq. (11), with a term quadratic in the shaping time that encodes the gain contribution. An elemental Monte Carlo simulation, using all-orders gain-layer multiplication and literature values for drift velocities, compares relative collected charge for PIN and LGAD sensors of 50 μm thickness, with the LGAD gain-layer parameters tuned to give an overall gain of 30. The paper claims that, for such an idealized gain-30 LGAD, the impact-ionization gain provides an advantage over PIN diodes at frame rates up to about 10 GHz.","tokens_in":5321,"tokens_out":5628,"duration_ms":52796,"significance":"If the claims hold, the paper provides a simple analytic tool for estimating ultrafast signal development in silicon diodes and identifies a regime in which internal gain materially extends the achievable frame rate. The strengths include a derivation that follows transparently from the Shockley-Ramo theorem with clearly stated assumptions, a simulation that uses an all-orders treatment of gain-layer multiplication rather than the linearized analytic approximation, and explicit discussion of the granular limit. The specific prediction that a gain-30 LGAD retains an advantage at ~10 GHz is falsifiable and testable against device-level simulations or prototype measurements. The main limitations are the unspecified shaping constant K and the unquantified space-charge assumption, both of which affect the quantitative 10 GHz claim.","major_comments":[{"comment":"The constant K is introduced as a \"dimensionless constant of order 1 that relates the shaping time to the effective charge collection time,\" but it is never defined, given a value, or related to a specific shaper response. The central claim that a gain-30 LGAD is advantageous at frame rates up to 10 GHz depends on the conversion between the simulation's \"effective collection time\" and the shaping time τ, and hence on K. Without a specification of K (for example, its value for a CR-RC or other shaper), the mapping between the 0.1 ns collection time shown in Figure 1 and the 10 GHz frame-rate statement is not reproducible. Please define K, give its value for the assumed shaping, or express the simulation results directly in shaping time.","section":"Section 2, Eq. (11)"},{"comment":"The assumption that the space-charge field \"remains small relative to the field created by the reverse bias\" is load-bearing for the constant-saturated-velocity forms of Eqs. (3), (6), (10), and for the simulation's fixed drift speeds. No quantitative criterion is given for when this assumption holds. In the high-flux regime the paper targets, this is not automatic: for example, a 10 keV X-ray pulse of about 5×10^3 photons per (100 μm)^2 pixel in a 50 μm detector produces a pair density ρ0 ~ 3×10^13 cm^-3, and at τ ~ 0.1 ns the space-charge field is of order e ρ0 (v_e - v_h) τ / ε_Si ~ 2×10^3 V/cm, already ~10% of the assumed 2×10^4 V/cm bulk field. Since the gain term in Eq. (11) scales with the saturated electron velocity, any field-induced reduction of v_e or α_e would shrink the quadratic term and shift the 10 GHz boundary. Please provide an explicit validity estimate in terms of incident flux and pair density, and qualify the central claim accordingly.","section":"Introduction, first paragraph"}],"minor_comments":[{"comment":"The text states that the arrival time of a single X-ray deposition would be spread over \"between 0 and 0.5 psec,\" but for a 50 μm sensor with v_e = 100 μm/ns the transit time is 0.5 ns; this appears to be a typo and should read \"0.5 nsec.\"","section":"Section 3, granular limit paragraph"},{"comment":"The x-axis is labeled \"Effective Collection Time\" without units; the text indicates nanoseconds, but the units should be shown on the axis.","section":"Figure 1"},{"comment":"The denominator \"3.62 λ\" could be misread as 3.62 times the wavelength; explicitly write the units (3.62 eV) and note that λ is the attenuation length, with the expression giving the pair density in cm^-3.","section":"Section 2, Eq. (1)"},{"comment":"Reference [2] is a colloquium slide deck; consider citing a peer-reviewed description of the Dynamic Mesoscale Materials Science Capability or a similar source if one is available.","section":"References"},{"comment":"The phrase \"anode (holes) or cathode (electronics)\" is confusing; the parenthetical likely should refer to the electrode that collects electrons, so a wording such as \"cathode (electron-collecting electrode)\" would clarify.","section":"Section 3, simulation description"}],"recommendation":"major_revision","confidential_remarks":"The paper is a readable technical contribution that fits the journal's scope. The two major concerns — the undefined K and the unquantified space-charge validity range — are both addressable in revision: K can be specified for a concrete shaper, and a quantitative flux/pair-density criterion can be added. If the authors also tighten the abstract's 10 GHz statement to explicitly carry the stated assumptions, the revised manuscript would be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe short version: this is a clean, honest idealized study of when impact-ionization gain helps ultrafast silicon X-ray sensors. Its main new piece is Eq. (11), a short-time closed form showing collected charge grows with a quadratic gain-onset term (1/2)K A_e v_e tau^2. The paper then uses that and an all-orders simulation to argue a gain-30 LGAD beats a PIN up to ~10 GHz frame rates under saturated drift conditions.\n\nWhat it does well: the derivation is straightforward Shockley-Ramo plus impact ionization, with literature drift velocities and ionization coefficients. The simulation treats gain to all orders, and the paper explicitly flags the granular limit (single X-ray arrival time limits to <2 GHz). The claim is honestly scoped as an idealized sensor. That is worth credit.\n\nSoft spots, in proportion: the shaping constant K is introduced as 'of order 1' but never bounded. Since the quadratic term is (1/2)K A_e v_e tau^2, the 10 GHz boundary moves with K; a factor of two in K shifts the onset time. Should be easy to pin down with a specific shaper. More importantly, the small-space-charge assumption is load-bearing but not quantified. The stress-test math shows that at the high fluxes the paper targets (say 5x10^3 10-keV photons per (100 um)^2 pixel per pulse), the induced space-charge field can reach ~10% of the 2x10^4 V/cm bias within 0.1 ns. That is enough to pull drift velocities off saturation and reduce alpha_e in the gain layer, shrinking the quadratic term. The paper doesn't say where that assumption breaks, so the 10 GHz number is conditional on an unspecified flux window. That's not a fatal flaw for an idealized design study, but it should be said out loud. Finally, no simulation code or sensitivity scan is provided; for a one-parameter tuning (alpha_e for gain 30), that's minor, but code would let readers test K and space-charge sensitivities.\n\nBottom line: the central derivation is internally consistent and the conclusion holds within the stated regime. The paper is for detector physicists planning XFEL/synchrotron diagnostics and for anyone wanting a compact formula for LGAD signal onset. It deserves a serious referee; a revision should add a bound on K and a quantitative criterion for the small-space-charge limit.\n\nRecommendation: send it to review, with the caveats above. It's a useful contribution to an active design space.","headline":"A clean idealized model of LGAD signal onset that supports a conditional 10 GHz advantage, with the main caveat being an unquantified small-space-charge limit.","tokens_in":5865,"tokens_out":1937,"would_cite":true,"duration_ms":19224,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["29.40.Wk"],"model":"deepseek-v4-flash","headline":"Impact ionization keeps an idealized gain-30 LGAD's X-ray signals above electronic noise at frame rates up to 10 GHz.","keywords":["LGAD sensors","impact ionization gain","ultrafast X-ray detection","series noise","signal development","saturated drift velocity","high frame rate","silicon diode sensors"],"falsifier":"Pulse a gain-30 LGAD with a sub-nanosecond X-ray pulse and record signal amplitude versus shaping time down to ~0.1 ns. Eq. (11) predicts the gain contribution grows as $\\tau^2$, so the LGAD-minus-PIN charge should increase by a factor of four when the shaping time doubles; if the advantage instead saturates or vanishes at high instantaneous flux, the saturated-velocity assumption or the quadratic gain term is wrong.","tokens_in":4813,"feed_emoji":"⚡","tokens_out":13058,"duration_ms":123325,"temperature":0.7,"pith_summary":"This paper develops a closed-form approximation for the first instants of signal formation in a silicon diode after an instantaneous, longitudinally uniform X-ray deposition. In a PIN diode the collected charge grows only linearly with the electronic shaping time, while in an LGAD the impact-ionization gain adds a term quadratic in that time. Because series readout noise grows as the inverse square root of the shaping time, the quadratic term is what lets an idealized gain-30 LGAD hold a signal-to-noise advantage over a PIN diode at frame rates up to 10 GHz. A simple Monte Carlo simulation of a 50 $\\mu$m sensor with a 2 $\\mu$m gain layer places the boundary at an effective collection time near 0.1 ns. The 10 GHz claim applies to high-flux saturated depositions; for single X-ray quanta the depth-dependent onset of gain limits the frame rate to below 2 GHz.","feed_headline":"LGAD impact-ionization gain beats PIN diodes to 10 GHz frame rates","feed_subtitle":"At sub-nanosecond shaping times a quadratic gain term keeps LGAD signals above electronic noise, where PINs fade.","key_machinery":"The load-bearing identity is Eq. (11), built from the parallel-plate induced-charge picture under the assumption of constant saturated drift speeds. Its new content is the quadratic-in-$\\tau$ term $\\frac{1}{2}K A_e v_e^s \\tau$ multiplying the linear collection term: carrier multiplication makes the carrier density grow during collection, so the induced charge accumulates faster than linearly immediately after deposition. The paper pairs the analytic expression with an elemental simulation of a 50 $\\mu$m sensor containing a 2 $\\mu$m gain layer tuned to an overall gain of 30, confirming a gain-onset time scale near 0.1 ns. The argument also uses the standard series-noise scaling, in which the minimal detectable charge worsens at least as fast as the $3/2$ power of the frame rate when the collected charge falls linearly with shaping time; the quadratic term is exactly what counteracts that degradation.","core_discovery":"The paper's central object is an equation for the short-time collected charge, Eq. (11). After a shaping time $\\tau$, the collected charge for a PIN diode is proportional to $\\tau$, while for an LGAD it carries an extra factor $1 + \\frac{1}{2}K A_e v_e^s \\tau$, so the gain contribution grows as $\\tau^2$. Here $A_e$ is the thickness-averaged electron impact-ionization coefficient and $v_e^s$ is the saturated electron drift velocity, so the quadratic term defines the time scale on which gain develops. The paper claims that for an idealized LGAD with overall gain 30, this quadratic term keeps the collected charge above the series-noise floor for effective collection times down to about 0.1 ns, corresponding to frame rates near 10 GHz, where a same-thickness PIN diode would fall below threshold. It also states that this advantage does not hold in the granular single-quantum limit, where the spread in arrival times of carriers at the gain layer limits the frame rate to below 2 GHz.","pith_inferences":["A direct experimental extension would measure LGAD pulse height versus shaping time on a pulsed X-ray source: Eq. (11) predicts the gain contribution doubles when $\\tau$ doubles, so observing saturation of that growth would pinpoint where the approximation breaks.","The same quadratic-gain argument should transfer to other saturated-drift multiplication sensors, such as thin avalanche photodiodes, with the frame-rate boundary set by their own impact-ionization coefficients and drift speeds.","The paper does not quantify when space-charge screening slows carriers; the first testable signature of that breakdown would be a flattening of the gain-onset time scale at high deposited flux density.","The stated granular-limit delay of 0.5 ps is inconsistent with the stated 2 GHz bound unless the delay is read as the ~0.5 ns electron transit time across a 50 $\\mu$m bulk; the 2 GHz figure is consistent with the longer value."],"forward_implications":["If Eq. (11) is correct, LGAD sensors remain useful for high-flux X-ray imaging at frame rates near 10 GHz, while PIN diodes lose signal to readout noise at sub-nanosecond shaping times.","Because the gain term is quadratic in $\\tau$, its relative contribution grows as the shaping time shrinks, so gain becomes more valuable as the frame rate rises.","The closed-form expression gives sensor designers a direct estimate of the gain-onset time scale $\\frac{1}{2}K A_e v_e^s$ without running full simulations.","In the granular single-quantum regime, the same picture predicts a lower frame-rate ceiling below 2 GHz, set by the spread of electron arrival times at the gain layer.","The treatment extends to white-beam or particle-stream depositions, since its main geometric requirement is a longitudinally uniform initial pair distribution."],"supporting_citations":[{"why":"supplies the 3.62 eV mean energy per electron-hole pair in silicon used in the deposition density.","marker":"[3]"},{"why":"gives the parallel-plate induced-charge relation behind the charge-collection rate expressions.","marker":"[4]"},{"why":"underpins the series-noise scaling that makes the quadratic gain term matter at short shaping times.","marker":"[5]"},{"why":"supplies the saturated electron and hole drift velocities (100 and 60 $\\mu$m/ns) used in the simulation.","marker":"[6]"},{"why":"provides electron impact-ionization coefficients used to choose the gain-layer mean free path of 0.61 $\\mu$m.","marker":"[7]"}],"fun_headline_variants":["LGAD quadratic gain beats PIN at 10 GHz","Quadratic gain term lifts LGAD past PIN at 10 GHz","LGAD gain grows quadratically, beating PIN to 10 GHz","Quadratic gain gives LGAD 10 GHz edge over PIN"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The formulas and the 10 GHz boundary assume that the space-charge field from the deposited electron-hole plasma stays small relative to the bias field, so electrons and holes keep constant saturated drift speeds throughout collection; if high-flux deposits screen the field and slow the carriers, the gain-onset time scale and the frame-rate boundary shift.","fun_headline_variants_meta":{"raw":{"variants":["LGAD quadratic gain beats PIN at 10 GHz","Quadratic gain term lifts LGAD past PIN at 10 GHz","LGAD gain grows quadratically, beating PIN to 10 GHz","Quadratic gain gives LGAD 10 GHz edge over PIN"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001163,"raw_usage":{"total_tokens":4802,"prompt_tokens":923,"completion_tokens":3879,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":539,"completion_tokens_details":{"reasoning_tokens":3807}},"tokens_in":539,"tokens_out":3879,"duration_ms":26405,"temperature":1.0,"reasoning_tokens":3807,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:27:39.158787+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Pulse a gain-30 LGAD with a sub-nanosecond X-ray pulse and record signal amplitude versus shaping time down to ~0.1 ns. Eq. (11) predicts the gain contribution grows as $\\tau^2$, so the LGAD-minus-PIN charge should increase by a factor of four when the shaping time doubles; if the advantage instead saturates or vanishes at high instantaneous flux, the saturated-velocity assumption or the quadratic gain term is wrong.","supporting_citations":[{"cited_title":"Canali, M","cited_arxiv_id":null,"evidence_quote":"supplies the 3.62 eV mean energy per electron-hole pair in silicon used in the deposition density."},{"cited_title":"Cavalleri, G","cited_arxiv_id":null,"evidence_quote":"gives the parallel-plate induced-charge relation behind the charge-collection rate expressions."},{"cited_title":"Spieler, Semiconductor Detector Systems, Oxford University Press, Oxford, U.K., (2005)","cited_arxiv_id":null,"evidence_quote":"underpins the series-noise scaling that makes the quadratic gain term matter at short shaping times."},{"cited_title":"Jacoboni, C","cited_arxiv_id":null,"evidence_quote":"supplies the saturated electron and hole drift velocities (100 and 60 $\\mu$m/ns) used in the simulation."},{"cited_title":"Robbins et al., Electron and hole impact ionization coefficients in (100) and in (111) Si, Journal of Applied Physics 58, 4614 (1985); https://doi.org/10.1063/1.336229","cited_arxiv_id":null,"evidence_quote":"provides electron impact-ionization coefficients used to choose the gain-layer mean free path of 0.61 $\\mu$m."}],"review_version":1}