{"id":"99314cee-ab80-4939-98ba-e5696ce0e790","arxiv_id":"1908.00950","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"An Associated Particle Imaging system using 14.1 MeV neutrons reconstructed centimeter-scale 3D distributions of carbon, silicon, and oxygen in a sand-and-carbon soil simulant.","lead":"A neutron-based imaging instrument maps carbon, silicon, and oxygen inside a prepared soil sample with centimeter-scale 3D resolution using tagged neutrons and gamma-ray detectors. It is a first step toward non-destructive, field-scale soil carbon measurement for climate and agriculture.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The <7 cm depth-resolution claim rests on a 1.3 ns timing FWHM from a forthcoming dry-sand calibration and is never validated against a known interface in the actual stacked soil-box geometry, so the central 3D-resolution claim is not yet supported.","rationale":"Good-faith reading: this is a first-results instrument paper, and the spatial ROI capability is demonstrated qualitatively by the graphite spectrum being dominated by the 4.44 MeV line. Lateral resolution is plausibly supported by the alpha detector position readout. The central weakness is that '3D resolution of several centimeters' is the headline result, but the depth component is inferred from a timing calibration that is not shown, and no reconstructed depth profile of a known target in the actual geometry is presented. The reader's CONDITIONAL verdict is appropriate; my concern is essentially the same, so no change. The concrete test would settle whether the timing transfer and the overall reconstruction actually deliver the claimed depth resolution. A secondary inconsistency worth reconciling is the conclusion's statement that all detectors show expected energy resolution, while Fig. 5 notes the LaBr3 resolution is about 3% rather than the previously measured ~1%.","tokens_in":5552,"tokens_out":9237,"duration_ms":97582,"concrete_test":"Perform a validation run in the same 2x3x3 stacked-box geometry used in Section 3: embed a thin graphite target (a few cm thick) at a known depth inside the sand-filled boxes, with the same source-to-surface distance, neutron output, CFD settings, and run length. Reconstruct the depth distribution of 4.44 MeV events selected in the graphite ROI and fit its centroid and FWHM. Accept the <7 cm depth-resolution claim only if the fitted FWHM is below 7 cm and the centroid is within ~2 cm of the true depth; report the timing FWHM measured in that same run. If the width exceeds 7 cm, the abstract's 3D-resolution claim must be revised or made conditional on the forthcoming timing measurement.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central quantitative claim is the 3D resolution, with depth resolution <7 cm. That number comes directly from a 1.3 ns FWHM time difference reported in Section 2.4, measured with a 1-inch-thick dry sand sample and said to be published in a forthcoming paper; the measurement includes both sample-extent and alpha-detector contributions. The stacked soil-box experiment in Section 3 is not accompanied by any depth calibration in that geometry, and the reconstructed y-z projection (Fig. 4) explicitly says the graphite and sand standards are not clearly visible, so there is no independent evidence that the depth coordinate is resolved at the 7 cm level. The 9-hour run also shows enough gain drift to degrade the LaBr3 energy resolution from ~1% to ~3% (Fig. 5), so assuming the timing behavior is unchanged from the short dry-sand calibration is not safe. If the effective timing FWHM in the reported measurement is larger than 1.3 ns, or if neutron multiple scattering and finite source spot bias the reconstructed depth, the claimed <7 cm depth resolution is unsupported by the data shown. This is load-bearing because the conclusion and abstract advertise a demonstrated 3D resolution that the paper does not actually validate in the measurement geometry.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript describes an Associated Particle Imaging (API) system for non-destructive soil-carbon measurements and reports first experimental results from a premixed sand sample containing 4% carbon by weight. A DT neutron generator with a position-sensitive alpha detector, LaBr3 and NaI gamma detectors, and a Pixie-16 digital DAQ are used to reconstruct the 3D location of neutron inelastic-scattering interactions from the alpha-particle position and the time difference between alpha and gamma detection. The authors present reconstructed x-y and y-z projections of a stacked soil-box phantom with graphite and sand standards, gamma-ray spectra extracted from selected spatial regions, and a table of identified inelastic lines. They claim a lateral resolution below 5 cm at 60 cm from the neutron source and a depth resolution below 7 cm, based on a 1.3 ns FWHM timing measurement, and they state that 3D resolution of several centimeters in all dimensions has been demonstrated. The paper also outlines planned improvements to increase count rate and reduce measurement time.","tokens_in":5785,"tokens_out":3637,"duration_ms":40796,"significance":"If the resolution and quantification claims are substantiated, the instrument would offer a genuinely useful capability: non-destructive, three-dimensional mapping of carbon and other light elements in soil with centimeter-scale resolution, on much larger representative volumes than conventional core sampling. The approach is physically direct, uses no fitted parameters, and the paper shows a clean demonstration of region-of-interest gamma spectroscopy with clear separation of carbon, silicon, and oxygen lines. The authors are also candid about limitations, including the weak visibility of the phantom in the y-z projection and energy-resolution degradation due to gain drift. However, the central resolution claims are not yet fully supported by the data shown: the depth-resolution claim depends on a calibration measurement not presented in the manuscript, and the lateral-resolution claim is inferred from geometry rather than measured in the reconstructed images. The quantitative connection to the known 4% carbon content is also not established. These gaps are fixable and do not invalidate the approach, but they are load-bearing for the paper's main conclusions.","major_comments":[{"comment":"The central claim of a depth resolution below 7 cm rests entirely on the 1.3 ns FWHM timing value reported in Section 2.4, which is said to come from a forthcoming publication using a 1-inch-thick dry sand sample. The timing measurement itself is not shown, and no depth calibration is performed in the stacked soil-box geometry actually used in the experiment. Figure 4 explicitly states that the graphite and sand boxes are not clearly visible in the y-z projection, so the data do not provide independent evidence that the depth coordinate is resolved at the 7 cm level in the reported measurement. Please include the timing distribution, validate the depth scale against a known interface or step in the measured geometry, or soften the depth-resolution claim accordingly.","section":"Section 2.4, Fig. 4"},{"comment":"The claimed lateral resolution of <5 cm at 60 cm is derived from the stated <1 mm position accuracy of the alpha detector through simple geometry, rather than from a measured edge response or line-spread function in the reconstructed image. Figure 3 shows the x-y projection but does not provide a quantitative measure of how sharply the soil, sand, and graphite boundaries are reconstructed. Please report a lateral-resolution measurement obtained from the data, with uncertainties, or explicitly label the current value as an expected geometric limit rather than a demonstrated resolution.","section":"Section 2.2, Fig. 3"},{"comment":"The title and abstract promise the ability to 'measure carbon in soil,' yet the paper provides no quantitative comparison of the reconstructed carbon signal to the known 4% carbon content of the soil mixture. The gamma spectra in Fig. 5 are normalized and displayed by eye, and no peak-area or carbon-yield analysis is presented. To support the soil-carbon application, please fit the 4.44 MeV carbon peak in the soil spectrum, account for the nearby 28Si line at 4.497 MeV, compare the carbon signal in the soil and graphite regions with the sand blank, and report the resulting carbon mass or density with an uncertainty. Without this step, the quantitative claim of carbon measurement is not demonstrated.","section":"Section 3, Fig. 5"},{"comment":"The paper attributes the degradation of LaBr3 energy resolution from about 1% to about 3% at 4.4 MeV to gain drift during the 9-hour run. Since the depth-resolution claim relies on a timing FWHM measured under different, presumably shorter, conditions, the stability of the timing response over the long measurement is not established. Please provide evidence that the 1.3 ns timing FWHM is valid over the full 9-hour acquisition, or quantify the possible drift in the timing path and its effect on the claimed <7 cm depth resolution.","section":"Section 3, Fig. 5"}],"minor_comments":[{"comment":"There is a typo in 'Measurements ... where used to obtain' which should read 'were used to obtain'; similarly, 'horizontal extend of the sample' should be 'horizontal extent of the sample.'","section":"Section 2.4"},{"comment":"The sentence 'First experimental results and characterization of capabilities where obtained' should read 'were obtained.'","section":"Section 4"},{"comment":"The phrase 'which it hard to resolve' is a typo; it should be 'which is hard to resolve.'","section":"Section 3, Table 1 caption"},{"comment":"The statement that the alpha-detector position accuracy is '<1 mm on the alpha-detector, corresponding to <5 cm at 60 cm' would benefit from a brief geometrical explanation or a reference to the previous publication, since the conversion depends on the distance from the alpha detector to the neutron target and the flight path to the sample.","section":"Section 2.2"},{"comment":"The y-z projection in Fig. 4 is difficult to interpret because the axis labeling and color scale are not fully described; please add a scale bar, color-bar units, and a description of the coordinate origin relative to the neutron generator and the soil-box array.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a credible instrument paper with a clear and honest presentation of first results, but the central resolution claims are currently supported by a calibration that is deferred to a forthcoming publication and by geometric estimates rather than by measurements in the reported geometry. The editor may wish to insist that the timing calibration data and a quantitative depth validation in the actual phantom geometry be included in the revised manuscript, as the archival value of the paper depends on these data being available to the reader."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"My quick take: this is a real first-results paper for a genuinely new application—associated particle imaging applied to soil carbon mapping. The experimental data are concrete, not simulations, and the ability to pull region-selected gamma spectra from soil, sand, and a graphite brick is demonstrated convincingly. The x-y image clearly separates the boxes, and the spectra show distinct C, Si, and O features, even if the 4.4 and 4.5 MeV peaks are close. Credit where due: the instrument is built, the measurements exist, and the authors are appropriately cautious about describing it as first results.\n\nThe soft spots are exactly where the reader and stress-test note point. The abstract and conclusion advertise a demonstrated 3D resolution of several centimeters, but the depth resolution (<7 cm) rests entirely on a 1.3 ns FWHM timing resolution measured with a 1-inch dry sand sample in a forthcoming publication. That timing measurement is not shown in this paper, and it is never validated in the actual stacked soil-box geometry. The y-z projection in Fig. 4 explicitly says the graphite and sand boxes are not clearly visible, so there is no direct evidence that the depth coordinate is actually resolved at the 7 cm level in the reported measurement. The gain drift noted in Fig. 5 (LaBr3 resolution degrading from ~1% to ~3% over the long run) also makes it unsafe to assume the timing behavior is unchanged from the short calibration. And resolution values are quoted without uncertainties, which is a minor but fixable omission.\n\nThese are load-bearing flaws because the central claim is the 3D resolution. But they are not fatal for an instrument paper at this stage—they are missing validation steps, not contradictions. The physics is standard tagged-neutron work, the detection scheme is sound, and the authors are honest about future work (quantitative carbon extraction, faster readout). The citation pattern is appropriate; the self-cited alpha detector paper is the right reference for component details.\n\nWho this is for: nuclear instrumentation and applied neutron-sensing readers, especially anyone interested in soil monitoring or non-destructive element mapping. It deserves a serious referee—the experimental effort is real and the application is novel. My recommendation: send to peer review. Require the timing calibration data from the forthcoming paper, a depth calibration in the actual geometry, and error bars on the resolution claims before publication. If the authors can show a known interface resolved in the y-z projection, the paper will be solid.","headline":"Genuine first demonstration of API for soil carbon, but the headline 3D resolution claim is not yet supported by the data shown; the paper deserves a serious referee with major revision.","tokens_in":6316,"tokens_out":1690,"would_cite":false,"duration_ms":18737,"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 demonstrates that an associated-particle imaging system based on a DT neutron generator can reconstruct neutron scattering sites in soil in three dimensions, with lateral resolution under 5 cm and depth resolution under 7 cm…","keywords":["Associated Particle Imaging","soil carbon","14 MeV neutrons","DT neutron generator","inelastic neutron scattering","gamma-ray spectroscopy","time-of-flight tomography","3D reconstruction"],"falsifier":"Measure the FWHM of the alpha-to-LaBr3 time difference on the actual stacked soil-box array with the same CFD settings; if it is larger than the 1.3 ns obtained on the 1-inch dry sand sample, the <7 cm depth-resolution claim would not be supported by the data. A complementary check is to place a graphite slab of known thickness at a known depth and verify that the reconstructed 4.44 MeV depth profile has an edge width no larger than about 7 cm.","tokens_in":5378,"feed_emoji":"🌱","tokens_out":7231,"duration_ms":73446,"temperature":0.7,"pith_summary":"The paper reports first experimental results for an Associated Particle Imaging (API) system aimed at non-destructive measurement of carbon in soil. The central claim is that by detecting the alpha particle emitted back-to-back with each 14.1 MeV neutron from a DT generator, the system can locate neutron scattering centers in three dimensions, and by timing the resulting gamma rays it can resolve depth to within 7 cm and lateral position to within 5 cm. On a prepared soil of sand mixed with 4% carbon by weight, the reconstructed images separate pure sand, soil, and graphite volumes, and gated gamma spectra show the characteristic carbon (4.44 MeV), silicon, and oxygen lines. If these numbers hold, the method would let soil carbon be mapped and re-mapped in place at centimeter scale, without digging cores.","feed_headline":"3D soil-carbon maps: <5 cm across, <7 cm deep","feed_subtitle":"Alpha-tagged neutrons turn gamma-ray spectra into a 3D soil-carbon map without digging.","key_machinery":"The carrying mechanism is alpha-tagged neutron time-of-flight, also called associated particle imaging. Each D+T fusion emits a 3.5 MeV alpha particle and a 14.1 MeV neutron in opposite directions; a thin position-sensitive yttrium aluminum perovskite (YAP) scintillator records the alpha's position and arrival time, which gives the neutron's direction and starts a clock. When that neutron scatters inelastically off a nucleus in the soil, the nucleus emits a characteristic gamma ray (for example 4.44 MeV from 12C), and a fast scintillator (LaBr3 or NaI) records the gamma's arrival time and energy. The time difference converts to distance traveled by the neutron, so each detected gamma is assigned a 3D position. A 500-MHz digital DAQ with constant-fraction discrimination supplies the timing, and energy gating of the gamma spectrum selects which element produced each event.","core_discovery":"The discovery is that a compact, sealed DT neutron generator equipped with a position-sensitive alpha detector and fast gamma-ray detectors can do 3D element-specific imaging of soil. The paper demonstrates this on a 2x3x3 stack of aluminum boxes filled with a homogeneous sand-worm-casting mixture (4% carbon by weight), with a graphite brick and a pure-sand box as standards. Reconstructed x-y and y-z projections show the separated samples, and gamma-ray spectra selected from each spatial region show the expected inelastic-scattering lines. The paper claims lateral resolution of <5 cm at 60 cm from the source and depth resolution of <7 cm, and it shows that the 4.497 MeV silicon line sits close enough to the 4.44 MeV carbon line that it must be accounted for in carbon quantification.","pith_inferences":["A practical carbon-quantification workflow will need to deconvolve the 28Si 4.5 MeV line from the 12C 4.44 MeV line; the paper presents spectra but not a calibrated carbon density map, so the quantitative accuracy of the method remains open.","If the timing resolution can be pushed toward the ~1 ns the authors speculate is possible, depth resolution would improve to roughly 5 cm, which would make finer soil stratification visible; the paper does not demonstrate this improvement.","The same region-of-interest gating logic could be extended to the 6.13 MeV oxygen line, potentially mapping soil moisture or oxide content in the same scan, an application the paper does not explore.","The depth-resolution claim rests on a single timing measurement made on a 1-inch dry sand sample; a direct test on the real soil-box geometry would either confirm the 7 cm figure or reveal its limit."],"forward_implications":["Soil carbon can be mapped in three dimensions non-destructively, with voxels of roughly a few centimeters, removing the need for destructive coring to see depth changes.","Because gamma spectra can be gated to specific regions, unwanted counts from structural materials and shielding are strongly reduced, so cleaner spectra can be obtained from buried volumes.","The same measurement yields gamma lines from carbon, silicon, oxygen, and aluminum, so one scan can report several soil constituents simultaneously.","Repeated measurements of the same location become possible, allowing carbon stock changes over time to be tracked without disturbing the soil.","At full neutron output with a faster alpha readout, the 9-hour acquisition could drop below 15 minutes, making the technique practical for field-scale surveys."],"supporting_citations":[{"why":"Describes the position-sensitive alpha detector and four-corner readout that set the lateral resolution.","marker":"[10]"},{"why":"Documents the digital DAQ whose constant-fraction timing and coincidence logic produce the time-of-flight data.","marker":"[11]"},{"why":"Standard reference for constant-fraction discrimination, the timing method behind the depth resolution.","marker":"[12]"},{"why":"Supplies the sealed-tube DT neutron generator and its small beam spot, which determine the spatial resolution.","marker":"[6]"},{"why":"Shows the use of associated particle imaging to take gamma spectra from selected regions in a high-background planetary setting, the precedent for region-of-interest spectroscopy.","marker":"[13]"}],"fun_headline_variants":["Neutrons map soil carbon in 3D without digging","Alpha-tagged neutrons reveal soil carbon in 3D","Compact neutron imager sees soil carbon 3D","3D soil carbon maps from neutron+gamma imaging","Tagged neutrons measure soil carbon 3D"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claimed <7 cm depth resolution depends on a 1.3 ns timing FWHM measured on a separate 1-inch dry sand sample and reported as part of a forthcoming publication; the paper does not directly show that the same timing holds for the stacked soil-box geometry and the 9-hour run.","fun_headline_variants_meta":{"raw":{"variants":["Neutrons map soil carbon in 3D without digging","Alpha-tagged neutrons reveal soil carbon in 3D","Compact neutron imager sees soil carbon 3D","3D soil carbon maps from neutron+gamma imaging","Tagged neutrons measure soil carbon 3D"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.0007,"raw_usage":{"total_tokens":3078,"prompt_tokens":777,"completion_tokens":2301,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":393,"completion_tokens_details":{"reasoning_tokens":2237}},"tokens_in":393,"tokens_out":2301,"duration_ms":17979,"temperature":1.0,"reasoning_tokens":2237,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:26:51.422631+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the FWHM of the alpha-to-LaBr3 time difference on the actual stacked soil-box array with the same CFD settings; if it is larger than the 1.3 ns obtained on the 1-inch dry sand sample, the <7 cm depth-resolution claim would not be supported by the data. A complementary check is to place a graphite slab of known thickness at a known depth and verify that the reconstructed 4.44 MeV depth profile has an edge width no larger than about 7 cm.","supporting_citations":[{"cited_title":"Position Sensitive Alpha Detector for an Associate Particle Imaging System","cited_arxiv_id":"1811.08591","evidence_quote":"Describes the position-sensitive alpha detector and four-corner readout that set the lateral resolution."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents the digital DAQ whose constant-fraction timing and coincidence logic produce the time-of-flight data."},{"cited_title":"Knoll, Radiation Detection and Measurement , Wiley, 3rd ed., 2010","cited_arxiv_id":null,"evidence_quote":"Standard reference for constant-fraction discrimination, the timing method behind the depth resolution."},{"cited_title":"http://adelphitech.com/, 2018","cited_arxiv_id":null,"evidence_quote":"Supplies the sealed-tube DT neutron generator and its small beam spot, which determine the spatial resolution."},{"cited_title":"As- sociated particle imaging instrumentation for future planetary surface missions,","cited_arxiv_id":null,"evidence_quote":"Shows the use of associated particle imaging to take gamma spectra from selected regions in a high-background planetary setting, the precedent for region-of-interest spectroscopy."}],"review_version":1}