{"id":"d2bd1333-ab89-4f94-b6c3-2d2e919fcff7","arxiv_id":"1909.05353","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"The shear in BCDI reconstructions is fixed by computing the real-space sampling basis from the Fourier-space sampling basis, with explicit formulas for typical detector and rocking geometries.","lead":"This paper derives a general coordinate transformation that removes shear distortion from three-dimensional images made by Bragg coherent diffraction imaging (BCDI). It gives experimentalists a geometry-aware recipe to render undistorted nanocrystal shapes, demonstrated on a silicon carbide particle.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Experimental demonstration is circular: the SEM image both disambiguates the twin-image degeneracy and serves as the corroborating reference, so the shear correction is not independently validated.","rationale":"The reader's CONDITIONAL verdict is appropriate. The central formula is derived from the DFT phase relation and matches known prescriptions; I see no internal inconsistency in the derivation. The reader's weakest_assumption, that Eq. (11) requires sufficient sampling and no cyclic aliasing, is a real experimental prerequisite, but it is a condition for any BCDI measurement to succeed, and the coordinate mapping itself is exact for the sampled lattice, so I would not treat it as the primary threat to the central claim. The primary threat is evidential: Section 4 uses the SEM image both to break the twin-image degeneracy and as the reference for corroboration, making the demonstration circular, and the comparison is qualitative. This warrants keeping the verdict CONDITIONAL until an independent, quantitative validation, ideally on simulated data with known ground truth, is provided. My proposed check would settle whether the shear-correction formula is correct in practice.","tokens_in":17383,"tokens_out":25334,"duration_ms":244467,"concrete_test":"Simulate a synthetic BCDI dataset: take a known 3D object, compute its diffraction at the non-orthogonal points B_recip n using the Section 3.3 geometry (e.g., Table 1 parameters), run conventional phase retrieval that assumes an orthogonal array, apply the shear correction r = B_real m with B_real = B_recip^{-T}D, and compare the corrected object to the ground truth with a quantitative metric (normalized mean-square error or Fourier ring correlation). The same simulation should be repeated with noise. If the simulation passes, Eq. (12) is independently confirmed. For the real SiC data, also quantify pillar width, height, and taper angle in the shear-corrected reconstruction and compare with SEM measurements on several particles, quoting uncertainties.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The mathematical derivation of Eq. (12) is self-contained and appears sound: B_real = B_recip^{-T} D follows from requiring the discrete phase factor n^T B_recip^T B_real m to equal n^T D m, which is the condition for the measured non-orthogonal samples to be compatible with a standard DFT. The paper itself correctly notes in Section 2 that this presumes adequate fringe sampling and no cyclic aliasing; that is a normal prerequisite for BCDI phase retrieval, not a shear-specific risk. The load-bearing weakness is instead the experimental demonstration in Section 4. The SEM image is used twice: it is first used to resolve the inherent ψ(r) vs ψ*(-r) degeneracy by choosing the solution that best matches the SEM after shear correction, and it is then cited as corroboration of the corrected morphology. Because the same image selects and validates the solution, the agreement is not independent evidence for Eq. (12). The comparison is also purely visual, against a different but nominally identical particle, with no quantitative metric or error analysis. A systematic error in the shear prescription, such as a wrong 1/N_i scaling or a transposed B_recip mapping, could plausibly survive such a qualitative comparison. This does not call the derivation into question, but it means the paper's claim that the correction is 'demonstrated' rests on an unvalidated, possibly circular, example.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a general derivation of the coordinate transformation needed to correct shear distortion in three-dimensional Bragg coherent diffraction imaging (BCDI) reconstructions. The central result is Eq. (12), which states that the real-space sampling basis B_real is related to the Fourier-space sampling basis B_recip by B_real = B_recip^{-T} D, where D = diag(1/N1, 1/N2, 1/N3). The derivation starts from the continuous Fourier relationship between conjugate bases and adapts it to the discrete Fourier transform, assuming adequate oversampling and no cyclic aliasing. Section 3 develops explicit expressions for the Fourier-space sampling vectors q_i, q_j, q_k for a specific goniometer geometry, and Section 4 applies the correction to a SiC nanocrystal reconstructed from BCDI data, comparing the shear-corrected object with an SEM image.","tokens_in":17644,"tokens_out":6057,"duration_ms":56203,"significance":"If the central derivation is correct, this paper provides a useful, parameter-free unification of previously scattered prescriptions for BCDI shear correction. The derivation of Eq. (12) is clean, self-contained, and appears mathematically sound, and the paper is careful to note the standard prerequisite of adequate fringe sampling. The main weakness is the experimental demonstration: the SEM image is used both to disambiguate the twin-image degeneracy of phase retrieval and then cited as corroboration of the shear-corrected morphology, making the validation circular. The comparison is also qualitative and against a nominally identical but different particle. Because the theoretical contribution is strong, the circularity is a validation issue rather than a flaw in the derivation, but it prevents the paper from claiming an independent experimental confirmation.","major_comments":[{"comment":"The experimental validation is circular. The caption of Fig. 4 explicitly states that the twin-image degeneracy (psi(r) versus psi*(-r)) was resolved by choosing the solution that best matches the SEM image after the shear correction. The text then uses the agreement with that same SEM image as evidence that the correction is valid. This does not independently test Eq. (12), because the SEM image has already influenced which phase-retrieval solution is presented. The paper should either use an independent criterion for twin selection (for example, known support constraints or a second particle not used for selection) or explicitly reframe the example as illustrative rather than confirmatory. In addition, the comparison is purely visual; a quantitative metric (for instance, measured facet angles or dimensions with uncertainties) would materially strengthen the demonstration.","section":"Section 4 and Fig. 4 caption"},{"comment":"The SEM comparison is made against a different particle from the same batch, not the particle that was imaged by BCDI. The text notes that the SEM image is of the batch of pillars prior to release, and the BCDI particle was one of many nominally identical pillars. This introduces particle-to-particle variability into the comparison, which is not discussed. The paper should state this limitation explicitly and, if possible, compare the shear-corrected reconstruction to an SEM image of the same particle or otherwise quantify the expected variability.","section":"Section 4, Fig. 4"},{"comment":"The summary states that the shear correction was 'demonstrated' and 'corroborated with SEM images.' Given the circularity described above, this overstates the evidence. The paper's lasting contribution is the theoretical derivation and the general computational prescription, not the independent experimental validation. The summary and abstract should be revised to separate the theoretical result from the illustrative example, and the claim of experimental corroboration should be either removed or qualified.","section":"Section 5, Summary"}],"minor_comments":[{"comment":"There are typographical errors in the abstract: 'realiable' should be 'reliable' and 'communitcy' should be 'community.'","section":"Abstract"},{"comment":"The phrase 'orthonormal bases defined in this manner are self-congujate' contains a typo: 'self-congujate' should be 'self-conjugate.'","section":"Section 2, around Eq. (8)"},{"comment":"The phrase 'With the the theoretical' contains a duplicated 'the.'","section":"Section 4, first paragraph"},{"comment":"The derivation assumes the phase relation in Eq. (11) holds exactly for all integer indices, which requires sufficient oversampling and no cyclic aliasing. The paper states this, but it would be helpful to state the oversampling condition quantitatively or to estimate the error when the condition is only approximately satisfied.","section":"Section 2, Eq. (11)"},{"comment":"The derivation of the columns of B_real from Eq. (12) is terse; a brief intermediate step showing the reciprocal-lattice-like relation would improve readability for readers not familiar with the analogous Bravais-lattice conversion.","section":"Section 3.3, Eqs. (14)-(16)"}],"recommendation":"major_revision","confidential_remarks":"This is a theoretical methods paper with a sound central derivation but a circular experimental validation. The editor may consider whether the journal is willing to accept the paper as a theoretical contribution with an illustrative example, in which case the recommended revision is to remove the overclaim of independent corroboration. If the journal requires an experimental demonstration, additional work would be needed to validate Eq. (12) non-circularly. I do not see grounds for rejection, as the mathematical derivation is clear and correct within its stated assumptions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: the derivation in this paper is the real contribution and it holds up. Second: the experimental demonstration is not independent, and the circularity the stress-test flags is real—though it doesn't damage the math.\n\nWhat's actually new: the conjugate-space derivation leading to Eq. (12), B_real = B_recip^{-T} D, and the explicit goniometer-parameterized sampling vectors in Eqs. (28)–(30). The paper is honest that Eqs. (14)–(16) recover earlier beamline-specific prescriptions, so the novelty is the unified derivation, not a new formula. The notation is heavy but consistent, the derivation from continuous to discrete Fourier bases is clean, and the appendix on computing strain on a sheared grid is a practical plus. The pathological δ=0 analysis is also useful.\n\nThe soft spot is Section 4. The SEM image is used twice: first to pick the phase-retrieval solution that best matches the SEM after shear correction (resolving the ψ vs ψ* twin-image degeneracy), then that same SEM is cited as corroboration of the corrected morphology. That's circular as a validation, and the comparison is visual, against a different but nominally identical particle, with no error metric. A systematic error in the shear prescription could plausibly survive. This is exactly the concern in the stress-test note, and I agree with it.\n\nBut the load-bearing result is fine. Eq. (12) follows from the discrete phase condition with no fitted parameters, and the underlying assumptions—adequate fringe oversampling, no cyclic aliasing—are standard BCDI prerequisites. So the theory stands; only the word 'demonstrated' is too strong. Labeling the example as illustrative, or adding an independent quantitative check, would fix it.\n\nWho this is for: BCDI and Bragg ptychography practitioners, and anyone building reconstruction software for new beamlines. It'd be a reasonable reading-group paper, though the discussion should focus on what counts as validation.\n\nDecision: yes, send it to peer review. A careful referee will ask for the validation caveat to be handled, but the general derivation deserves to be in the literature. I'd cite it if I worked in this area.","headline":"The math holds and the paper is honest about its novelty, but the SEM-based validation is circular and needs to be fixed before publication.","tokens_in":18221,"tokens_out":2950,"would_cite":true,"duration_ms":26985,"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":"Bragg coherent diffraction images carry a predictable shear that one linear-algebra identity removes.","keywords":["Bragg coherent diffraction imaging","phase retrieval","shear correction","coordinate transformation","scattering geometry","Bragg ptychography","conjugate spaces","nanocrystal strain"],"falsifier":"Simulate a faceted nanocrystal, sample its noiseless Fourier transform on the non-orthogonal grid defined by the paper's $\\mathbf{q}_i, \\mathbf{q}_j, \\mathbf{q}_k$, reconstruct by phase retrieval, apply Eq. (12), and compare rendered facet angles and edge lengths to the ground truth; any systematic discrepancy beyond pixel or angular resolution would show the conjugacy relation is incomplete.","tokens_in":17209,"feed_emoji":"🔬","tokens_out":9544,"duration_ms":81039,"temperature":0.7,"pith_summary":"This paper establishes that the three-dimensional shear seen in Bragg coherent diffraction imaging (BCDI) reconstructions is a predictable geometric consequence of how Fourier space is sampled, not a flaw of phase retrieval, and can be removed by a general post-processing transformation. The argument reduces to a matrix identity linking the Fourier-space sampling basis to the real-space sampling basis of the reconstructed object, so the goniometer settings alone determine how to render an undistorted image. This matters because BCDI encodes a single component of the lattice strain tensor, a frame-dependent quantity, so displaying the recovered complex field on the wrong grid can corrupt both morphology and strain readings. The paper generalizes earlier facility-specific prescriptions into a frame-agnostic relation and demonstrates it on a silicon carbide nanoparticle checked against scanning electron microscopy.","feed_headline":"Sheared Bragg images fixed by one matrix identity","feed_subtitle":"Detector geometry alone predicts the distortion-free grid for rendering a 3D crystal.","key_machinery":"The load-bearing object is the identity $B_{\\mathrm{real}} = B_{\\mathrm{recip}}^{-T} D$ (Eq. 12), the discrete analogue of the continuous reciprocal-basis relation $B_q = B_r^{-T}$. $B_{\\mathrm{recip}}$ is the matrix whose columns are the non-orthogonal Fourier-space step vectors $\\mathbf{q}_i, \\mathbf{q}_j, \\mathbf{q}_k$ fixed by pixel size, object-detector distance, wavelength, Bragg angles, and rocking increment; $D = \\mathrm{diag}(1/N_1, 1/N_2, 1/N_3)$ accounts for the array size. The identity converts the easily measured Fourier-space geometry into the correct real-space grid and reveals the shear as unavoidable: in Bragg geometry $\\mathbf{q}_i \\perp \\mathbf{q}_j$, while $\\mathbf{q}_k$ cannot be perpendicular to both. The paper also supplies explicit laboratory-frame expressions for $\\mathbf{q}_i, \\mathbf{q}_j, \\mathbf{q}_k$ from two detector rotations and the rocking step, together with a Rodrigues rotation formula so the same computation can be adapted to other goniometer arrangements.","core_discovery":"On its own terms, this paper establishes that the sheared appearance of a phase-retrieved BCDI object is not an arbitrary artifact but the exact discrete consequence of sampling Fourier space on a non-orthogonal grid. The central identity is Eq. (12): $B_{\\mathrm{real}} = B_{\\mathrm{recip}}^{-T} D$, with $D=\\mathrm{diag}(1/N_1, 1/N_2, 1/N_3)$, which follows from requiring the discrete phase $q^T r = il/N_1 + jm/N_2 + kn/N_3$ to hold on every array index. Here $B_{\\mathrm{recip}} = [\\mathbf{q}_i\\ \\mathbf{q}_j\\ \\mathbf{q}_k]$ lists the Fourier-space step vectors fixed by detector pixel pitch, sample-detector distance, wavelength, Bragg angle, and rocking step, so once those are known the correct real-space sampling basis $B_{\\mathrm{real}}$ is determined. Rendering the retrieved array on this sheared basis yields the undistorted scatterer, and the paper verifies the prescription on a SiC nanoparticle by matching the corrected rendering to SEM images.","pith_inferences":["The identity is the finite-array analogue of converting a unit cell to its reciprocal cell; if the Fourier aperture is marginal, an aperture-weighted version of Eq. (12) would also predict small shape errors at the object boundary, a testable refinement.","A sensitivity analysis of Eq. (12) would quantify which experimental parameter dominates the distortion, something the paper does not do; this could guide the design of new BCDI beamlines.","Since the derivation only assumes a fixed Fourier sampling matrix, the shear correction should transfer unchanged to Bragg ptychography and non-standard rocking trajectories; a synthetic-data test with an arbitrary rocking axis would confirm the transfer.","If phase retrieval enforces orthogonal-grid real-space support, the post-processing shear correction can only be exact away from constraint boundaries, which is a reason to prefer Part II's in-algorithm approach in constrained reconstructions."],"forward_implications":["After phase retrieval, applying Eq. (12) with the measured goniometer parameters converts any BCDI reconstruction into a physically accurate, shear-corrected 3D image without modifying the reconstruction algorithm.","The corrected grid fixes the scatterer's orientation in the laboratory frame, allowing morphological features to be compared directly with scanning electron micrographs.","The linear dependence of the columns of $B_{\\mathrm{recip}}$ serves as a degeneracy test: a rocking axis in the scattering plane produces parallel sampling vectors and cannot yield a 3D BCDI dataset.","The strain component encoded in BCDI can be computed on the sheared grid by finite differences of the phase along the columns of $B_{\\mathrm{real}}$, avoiding interpolation to an orthogonal grid.","The same conjugacy relation provides the foundation for Part II's in-algorithm modification of the discrete Fourier transform, enabling reconstruction directly on an orthogonal grid with physical constraints."],"supporting_citations":[{"why":"Stated the continuous conjugate-basis relation $B_q = B_r^{-T}$ that this paper generalizes to the discrete DFT case.","marker":"Pateras (2015)"},{"why":"Provided an earlier working-rule post-processing recipe for a specific BCDI geometry that Eq. (12) unifies.","marker":"Pfeifer (2005)"},{"why":"Supplied a prior real-space coordinate inversion prescription whose form Eqs. (14)-(16) reproduce up to the finite Fourier-volume factor.","marker":"Berenguer et al. (2013)"},{"why":"Gave a concurrent prescription for mapping data between sample and detector conjugated spaces that the paper identifies as equivalent.","marker":"Yang et al. (2019)"},{"why":"Computes for standard goniometers the Fourier-space points spanned by $B_{\\mathrm{recip}}$, which supplies the input sampling basis used in Eq. (12).","marker":"Kriegner et al. (2013)"},{"why":"Treats the phase contributions of real- and Fourier-space origin offsets, justifying the paper's zero-offset assumption in the measured intensity.","marker":"Vartanyants & Robinson (2001)"}],"fun_headline_variants":["One matrix identity corrects BCDI shear","Post-retrieval fix for sheared Bragg images","Shear-free BCDI via discrete Fourier mapping","Detector geometry sets the real-space grid"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole result rests on the assumption that the measured diffraction pattern is a well-sampled, un-aliased digital Fourier transform of the object; if the fringes are too coarse or the detector window clips them, the computed real-space grid is unreliable no matter how the shear is corrected.","fun_headline_variants_meta":{"raw":{"variants":["One matrix identity corrects BCDI shear","Post-retrieval fix for sheared Bragg images","Shear-free BCDI via discrete Fourier mapping","Detector geometry sets the real-space grid"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00095,"raw_usage":{"total_tokens":4119,"prompt_tokens":1078,"completion_tokens":3041,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":694,"completion_tokens_details":{"reasoning_tokens":2990}},"tokens_in":694,"tokens_out":3041,"duration_ms":21097,"temperature":1.0,"reasoning_tokens":2990,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:07:38.952889+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate a faceted nanocrystal, sample its noiseless Fourier transform on the non-orthogonal grid defined by the paper's $\\mathbf{q}_i, \\mathbf{q}_j, \\mathbf{q}_k$, reconstruct by phase retrieval, apply Eq. (12), and compare rendered facet angles and edge lengths to the ground truth; any systematic discrepancy beyond pixel or angular resolution would show the conjugacy relation is incomplete.","supporting_citations":[{"cited_title":"& Chamard, V","cited_arxiv_id":null,"evidence_quote":"Supplied a prior real-space coordinate inversion prescription whose form Eqs. (14)-(16) reproduce up to the finite Fourier-volume factor."}],"review_version":1}