{"id":"f3ee6ce0-9cba-4dd2-a9d8-9c9e769ee5b4","arxiv_id":"2504.15183","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Dynamical decoupling enabled high-sensitivity multiple quantum NMR measurements, revealing that the width of spin cluster size distributions grows quadratically with evolution time.","lead":"Dynamical decoupling pulses are used to boost NMR signal and reconstruct how many spins cluster together during multiple quantum dynamics. The paper reports a new empirical scaling law: the spread of cluster sizes grows quadratically with evolution time.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The quadratic Δs(t_n) law rests on Tikhonov-inverted cluster-size distributions; absent synthetic validation and regularization parameters, the growth could be a smoothing artifact as peaks move and decay.","rationale":"Good-faith reading: the paper makes a useful methodological advance (DD-enhanced MQC acquisition) and supports it with a direct FID/DD comparison at n=6,9; the inversion scheme is a reasonable way to go beyond Gaussian fits. My concern is not with the DD method but with the step that converts coherence-order spectra into a cluster-size distribution. Eq. (3) is a Fredholm integral equation of the first kind; without regularization constraints, its solution is unstable. Teal–Eccles is a well-regarded algorithm, but the paper does not state the chosen regularization level or show that the inferred f_n(s) is stable against noise and smoothing. Because the central claim is a scaling exponent (quadratic) extracted from the width of a peak that moves and decays, the regularizer's tendency to broaden low-amplitude peaks is a concrete, plausible mechanism for the observed law. The reader's weakest-assumption analysis identifies the same point; I agree. A synthetic-data inversion test would settle it: if the algorithm recovers known non-quadratic widths as quadratic, the headline claim is not yet established. If it recovers the true widths, the empirical law is substantially strengthened. The paper's own admission that no model exists makes this validation more important, not less. I recommend keeping the verdict conditional: the result is plausible and worth reporting, but the quadratic law should be labeled provisional until the inversion's robustness is demonstrated.","tokens_in":8794,"tokens_out":6373,"duration_ms":59820,"concrete_test":"Generate synthetic MQC signals S_n,k from a known bimodal f_n(s) with large-cluster widths held constant (or growing linearly), using the same n-values, coherence-order grid, and noise level as Fig. 4; invert with the same Teal–Eccles algorithm and extract the FWHM of the larger population. If the recovered Δs(t_n) is quadratic, the reported scaling is a regularization artifact. Also re-analyze the experimental data with two or three regularization strengths and with a direct two-Gaussian fit to S_n,k, and report error bars on Δs from bootstrap resampling.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's only new quantitative claim is the empirical quadratic growth of the dispersion Δs of the large-cluster population (Fig. 4d). This quantity is not directly measured: it is the FWHM of one component of f_n(s), obtained by numerically inverting the ill-posed linear system in Eq. (3) using a Tikhonov/adaptive-truncation algorithm (Teal–Eccles [42]). No regularization parameter, noise level, or convergence criterion is reported, and no synthetic-data test is shown. Tikhonov-type smoothing strongly biases solutions toward broad, smooth distributions; as the large-cluster peak moves to larger s and its amplitude decays with n, the regularizer will preferentially broaden it, plausibly generating a t^2 growth of the FWHM even when the true width is constant or grows more slowly. The only validation offered is the qualitative FID-vs-DD comparison at n=6 and n=9 (Fig. 3e), which checks the detection block, not the inversion, and does not cover the long-time regime (n=12 and beyond) on which the scaling law rests. The paper itself states no model underlies the observation, which is acceptable only if the observable is shown to be robust to the inversion procedure. As it stands, the quadratic law is not empirically secured.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports a dynamical-decoupling-enhanced multiple-quantum (MQ) NMR study of polycrystalline adamantane. The authors insert a dynamical decoupling (DD) detection block into the standard MQC experiment, improving the signal-to-noise ratio by roughly a factor of 22 in scan count, and show that the DD-acquired coherence distributions reproduce the conventional FID-based distributions at n=6 and n=9 cycles. Using the improved data, they invert Eq. (3) with the Teal–Eccles adaptive-truncation algorithm to obtain cluster-size distributions f_n(s). From these distributions they extract three scaling features: the largest cluster size grows linearly with evolution time, the 97% cumulative propagation front grows as t^3 in agreement with earlier ballistic estimates, and, as the central new claim, the dispersion Δs of the larger-cluster population grows quadratically with evolution time, Δs ∝ t_n^2. The authors state that they lack a model for this quadratic law and present it as a first empirical observation.","tokens_in":9113,"tokens_out":3735,"duration_ms":36733,"significance":"The dynamical decoupling detection block is a practical experimental contribution with a concrete, quantitative demonstration of SNR gain and a useful validation against standard acquisition at n=6 and n=9. If the quadratic width law were established robustly, it would constitute a new empirical scaling result for many-body clustering under double-quantum dynamics in a strongly interacting spin system, of interest to the NMR and quantum-scrambling communities. The significance of the manuscript therefore hinges on whether the extracted Δs(t_n) is a faithful property of the physical cluster-size distribution. The central weakness is that this quantity is obtained from an ill-posed numerical inversion whose regularization parameters, synthetic-data validation, and error propagation are not reported, so the reality of the quadratic growth is not yet empirically secured.","major_comments":[{"comment":"The inversion that produces f_n(s) from the coherence distributions is described as ill-posed, and the Teal–Eccles algorithm is referenced, but no regularization parameter, noise level, or convergence criterion is reported, and no synthetic-data test is shown. Tikhonov-type smoothing can preferentially broaden peaks as they move toward larger s and lose amplitude, so the quadratic growth of the FWHM in Fig. 4d could be generated by the regularizer even when the true width is constant or grows more slowly. To make the central claim load-bearing, the authors must report the regularization parameter and demonstrate with a synthetic benchmark (e.g., a known constant-width moving Gaussian) that the inversion recovers the correct width across the relevant range of amplitudes and positions.","section":"Eq. (3) and Fig. 4b–4d"},{"comment":"The 'larger cluster' population whose FWHM defines Δs is not operationally defined. The text says the width is represented by shadows and that the 97% cumulative value is used for the propagation front, but it does not specify how the two populations in the bimodal f_n(s) are separated, how the FWHM is computed for overlapping or asymmetric peaks, or how sensitive the extracted Δs is to these choices. A precise algorithm is needed, along with a stability check showing that the quadratic scaling persists under reasonable variations of the population-selection threshold and the cumulative level.","section":"Fig. 4c–4d"},{"comment":"The data points in Fig. 4d are shown without error bars, and the fit to Δs ∝ t_n^2 is presented without a goodness-of-fit measure or a comparison to alternative scaling laws (for example, linear or t^β with a fitted β). Because Δs is a derived quantity obtained from an inversion of noisy experimental data, its uncertainty must be propagated through Eq. (3) and the extraction procedure. Without such uncertainty quantification, the claim that the growth is specifically quadratic is not quantitatively supported.","section":"Fig. 4d"},{"comment":"The DD detection block is validated against FID acquisition only at n=6 and n=9 (Fig. 3e), while the quadratic law in Fig. 4d relies on later times, including n=12 and beyond, where no FID reference is shown. The comparison at n=12 in Fig. 4a uses different numbers of scans for DD and FID and is only qualitative. The authors should demonstrate that the DD detection block does not distort the coherence distribution at long times, for instance by comparing DD against a high-scan-count FID at n=12, or by simulating the effect of the detection block on a known distribution.","section":"Figs. 3e and 4a"}],"minor_comments":[{"comment":"There are several typographical errors, including 'plycristalline' for 'polycrystalline', 'ara' for 'are', 'prethermatization' for 'prethermalization', and 'the the temporal increase' in the introduction; these should be corrected.","section":"Throughout"},{"comment":"The wording 'The saturated scale in the last plot' is unclear; the authors should state explicitly that the color scale saturates above 4000 transients and that this saturation is an artifact of the display range.","section":"Fig. 2c"},{"comment":"The subscripts F and B for the forward and backward DQ blocks are used without being defined in the figure caption; they should be defined at first use in the text or caption.","section":"Fig. 3b"},{"comment":"The abstract mentions 'decay times greater than 1000-fold', which appears to refer to earlier DD work rather than to the present measurements; the body reports a ~22-fold reduction in scan number. The abstract should clarify which quantity is enhanced 1000-fold.","section":"Abstract and §I"},{"comment":"The notation uses both s_j for the kernel widths and s as the continuous cluster size; the relationship between the discrete grid s_j and the argument of f_n(s) should be stated explicitly, including the number of grid points and the range of s used in the inversion.","section":"Eq. (3)"}],"recommendation":"major_revision","confidential_remarks":"The paper is a reasonable experimental report with a genuinely useful DD detection scheme, but the central novel claim—the quadratic growth of the cluster-size dispersion—rests on an unvalidated inversion and a post hoc FWHM extraction. The absence of regularization parameters, synthetic tests, and error bars is the main obstacle; these are standard requirements for reporting results from ill-posed inversions. I would recommend that the editor request a revised version that supplies those details and a stability analysis, rather than accept the current form. The t^3 propagation front is used as a consistency check with previous work, which is acceptable but not independent validation. The novelty claim should be softened unless the quadratic law survives the requested robustness tests."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Sequeiros et al. report two connected things: a dynamical-decoupling detection block that dramatically improves SNR in multiple-quantum coherence (MQC) experiments on adamantane, and a claim that the dispersion of the spin-cluster-size distribution grows quadratically with evolution time. The first is real. The comparison with FID acquisition at n=6 and n=9 matches well, and at n=12 the DD version resolves structure that FID cannot see even with twice the scans. That is a useful experimental contribution. The inversion of the MQC line shape into a cluster-size distribution via Eq. (3) is a sensible extension of earlier Gaussian or stretched-exponential analyses, and the algorithm is referenced.\n\nThe second claim is not yet supported. The quadratic law is extracted from the FWHM of the large-cluster population of f_n(s), and f_n(s) is obtained by an ill-posed regularized inversion. The paper gives no regularization parameter, no noise level, no convergence criterion, and no synthetic-data reconstruction. Tikhonov-style smoothing will broaden a peak that is moving and decaying; the t^2 growth in Fig. 4d could be that artifact. The only validation shown (Fig. 3e) checks the DD detection block, not the inversion, and it stops at n=9 while the quadratic law relies on later times. The authors also do not specify how the 'larger cluster' population is selected at short times, which is another degree of freedom. They are explicit that no model underlies the observation; that is fine in principle, but an empirical law needs to be robust to the procedure that produced it. Right now it isn't.\n\nThe paper is honestly written and the citations to their own earlier scale results are legitimate context, not padding. The t^3 propagation front is presented as consistency with their prior ballistic result, which is a reasonable cross-check rather than circularity.\n\nBottom line: this deserves peer review, because the method is promising and the community should see it. But the referee should insist on synthetic-data validation, reported regularization parameters, and error bars on the widths and exponent. If those checks come out clean, the quadratic law becomes an interesting empirical observation; until then it is a provisional hint. I'd bring it to a reading group for a methodological discussion, but I wouldn't cite the quadratic law in my own work yet.","headline":"A genuinely useful DD-based detection scheme for MQC, attached to a headline quadratic clustering law that is not yet secured because the underlying inversion is unvalidated.","tokens_in":9614,"tokens_out":2872,"would_cite":false,"duration_ms":27412,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Spin cluster sizes disperse quadratically as they grow in time","keywords":["multiple quantum coherences","dynamical decoupling","spin cluster size distribution","Tikhonov regularization","Loschmidt echo","out-of-time-order correlators","Floquet prethermalization","nuclear magnetic resonance"],"falsifier":"Generate synthetic multiple-quantum coherence data from a known cluster-size distribution whose true dispersion grows linearly (or not at all) with time, add realistic noise, run the paper's inversion pipeline, and check whether the recovered $\\Delta s$ still grows as $t_n^2$. A quadratic output in that test would show the law is produced by the inversion; a non-quadratic output would support the physical claim.","tokens_in":8636,"feed_emoji":"⚛️","tokens_out":7570,"duration_ms":65438,"temperature":0.7,"pith_summary":"This paper establishes a new use of dynamical decoupling (DD)—a periodic pulse train usually employed to protect quantum states—as a sensitive detector for multiple-quantum NMR experiments on strongly coupled nuclear spins. With the boosted signal-to-noise ratio, the authors numerically invert the measured coherence spectra to recover the full distribution of spin-cluster sizes, not just its average. Their central observation is that the dispersion of the cluster-size distribution grows as $t_n^2$ with multiple-quantum evolution time, even though the cluster front itself propagates ballistically as $t_n^3$. The finding matters because cluster-size heterogeneity is a quantity that standard Gaussian fits cannot resolve, and any future model of many-body spin dynamics will need to reproduce this quadratic widening.","feed_headline":"Spin cluster sizes disperse quadratically in time","feed_subtitle":"Dynamical decoupling boosts NMR signal enough to map cluster-size distributions and reveal their t-squared widening.","key_machinery":"The load-bearing machinery is the combination of a double-quantum Hamiltonian $H_{\\mathrm{DQ}} = -\\frac{1}{2}\\sum_{i<j} d_{ij}(I_i^+ I_j^+ + I_i^- I_j^-)$, engineered by a pulse sequence, with a detection stage built from a dynamical-decoupling Floquet train. The experiment measures a generalized echo $S_{n,\\phi}$ for many reversion phases $\\phi$; a Fourier transform over $\\phi$ gives the distribution of even coherence orders $\\tilde{S}_{n,k}$. That distribution is inverted through $\\tilde{S}_{n,k} = \\sum_j \\exp(-k^2/s_j)\\, f_n(s_j) + \\epsilon_{n,k}$, with Tikhonov regularization, yielding the cluster-size distribution $f_n(s)$. The DD detection block is what supplies enough signal-to-noise for the numerically unstable inversion to work at long evolution times.","core_discovery":"On the paper's own terms, the discovery is an empirical law for how spin clusters spread in size during double-quantum evolution in a dense three-dimensional dipolar system. Using adamantane as a highly interconnected spin network, the authors measure multiple-quantum coherence distributions at increasing evolution times and invert them—modeling each coherence profile as a weighted sum of Gaussians over cluster sizes—to obtain $f_n(s)$, the probability that a cluster contains $s$ correlated spins. They find that the distribution becomes bimodal at longer times, and that the full width at half maximum of the larger-cluster population, $\\Delta s$, grows as $t_n^2$, while the fastest cluster front (the 97% cumulative point) grows as $t_n^3$. The authors state explicitly that they lack a microscopic model for the quadratic law; the claim is the empirical observation itself, made accessible by the improved sensitivity of DD-based acquisition.","pith_inferences":["If the quadratic law is physical, cluster growth is not described by a single growing size: the width of the cluster-size distribution carries independent information about scrambling, possibly a second velocity characterizing roughening of the operator front.","A synthetic-data validation of the inversion—feeding a known $f_n(s)$ into equation (3), adding noise, and checking the recovered $\\Delta s$—would determine whether the $t_n^2$ law survives outside the regularizer. The paper does not report this test.","The same DD-boosted protocol could be applied to lower-dimensional or disordered spin systems to test whether quadratic dispersion is universal or specific to the three-dimensional dipolar network."],"forward_implications":["DD acquisition reaches the same signal-to-noise as conventional free-induction-decay acquisition with roughly 22 times fewer scans at the longest probed evolution time, making otherwise impractical measurements routine.","Cluster-size distributions can be recovered at evolution times where the Loschmidt echo has decayed close to the noise floor, revealing correlations among hundreds of spins.","The observed $t_n^2$ widening of the dispersion, together with the $t_n^3$ ballistic front, gives a quantitative target for microscopic theories of double-quantum spin dynamics.","The method confirms the earlier ballistic-propagation picture for a three-dimensional dipolar spin network such as adamantane."],"supporting_citations":[{"why":"Supplies the standard Gaussian model for multiple-quantum coherence distributions that the paper's equation (3) generalizes.","marker":"[33]"},{"why":"Establishes the NMR clustering picture that motivates interpreting coherence distributions in terms of spin clusters of different sizes.","marker":"[34]"},{"why":"Provides the constrained regularization approach for inverting ill-posed data, cited as the parsimony strategy behind the inversion.","marker":"[40]"},{"why":"Names Tikhonov regularization, the stabilizing method used for the numerical inversion of equation (3).","marker":"[41]"},{"why":"Supplies the specific one-dimensional inversion algorithm used to extract the cluster-size distribution $f_n(s)$.","marker":"[42]"},{"why":"Provides the earlier ballistic $t^3$ propagation result for the double-quantum Hamiltonian in adamantane against which the 97% cluster front is compared.","marker":"[12]"},{"why":"Connects the number of correlated spins, as the second moment of the coherence distribution, to out-of-time-order correlators and places the cluster measurement in the scrambling framework.","marker":"[2]"},{"why":"Defines the Loschmidt echo and generalized echo formalism used for the measured signal $S_{n,\\phi}$.","marker":"[36]"}],"fun_headline_variants":["Spin cluster sizes widen quadratically under DD","Dynamical decoupling reveals t-squared spin-cluster growth","Quadratic widening of spin clusters observed","Spin clusters broaden quadratically with time"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the cluster-size distribution recovered by Tikhonov inversion faithfully represents the physical distribution, rather than being shaped by the regularizer; if the smoothing broadens the larger-cluster peak as it moves and flattens, the observed quadratic widening could be an artifact.","fun_headline_variants_meta":{"raw":{"variants":["Spin cluster sizes widen quadratically under DD","Dynamical decoupling reveals t-squared spin-cluster growth","Quadratic widening of spin clusters observed","Spin clusters broaden quadratically with time"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000702,"raw_usage":{"total_tokens":3166,"prompt_tokens":942,"completion_tokens":2224,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":558,"completion_tokens_details":{"reasoning_tokens":2174}},"tokens_in":558,"tokens_out":2224,"duration_ms":15775,"temperature":1.0,"reasoning_tokens":2174,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T11:30:52.182522+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Generate synthetic multiple-quantum coherence data from a known cluster-size distribution whose true dispersion grows linearly (or not at all) with time, add realistic noise, run the paper's inversion pipeline, and check whether the recovered $\\Delta s$ still grows as $t_n^2$. A quadratic output in that test would show the law is produced by the inversion; a non-quadratic output would support the physical claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the standard Gaussian model for multiple-quantum coherence distributions that the paper's equation (3) generalizes."},{"cited_title":"Baum and A","cited_arxiv_id":null,"evidence_quote":"Establishes the NMR clustering picture that motivates interpreting coherence distributions in terms of spin clusters of different sizes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the constrained regularization approach for inverting ill-posed data, cited as the parsimony strategy behind the inversion."},{"cited_title":"Tikhonov and V","cited_arxiv_id":null,"evidence_quote":"Names Tikhonov regularization, the stabilizing method used for the numerical inversion of equation (3)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the specific one-dimensional inversion algorithm used to extract the cluster-size distribution $f_n(s)$."},{"cited_title":"S´ anchez, H","cited_arxiv_id":null,"evidence_quote":"Provides the earlier ballistic $t^3$ propagation result for the double-quantum Hamiltonian in adamantane against which the 97% cluster front is compared."}],"review_version":1}