{"id":"09d45ce3-3c1d-4afd-bb8d-fc3417318e70","arxiv_id":"2602.07222","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Stacked laser-written geometric-phase holograms inside glass form compact multi-plane light converters that sort 3 and 10 Hermite-Gaussian modes.","lead":"Researchers fabricated tiny multi-plane light converters inside glass by laser-writing stacked phase masks. The devices sort up to 10 Hermite-Gaussian light modes in a sub-cubic-millimetre chip, though the abstract claims more than the manuscript body shows.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"10-mode coupling matrix numbers contradict the stated column normalization: for a 10x10 nonnegative matrix with column sums 1, mean off-diagonal cannot exceed 0.111, yet text reports 0.13 (sim) and 0.58 (exp).","rationale":"I read the paper as a proof-of-concept fabrication demonstration: two laser-written geometric-phase planes in glass, tested as 3-mode and 10-mode HG sorters. The 3-mode result is plausible and the fabrication concept is credible, but the 10-mode result is the key upper-bound claim. The reported mean off-diagonal values for the 10-mode matrix are mathematically incompatible with the stated column normalization, which is more fundamental than the reader's half-wave-plate concern: even a perfect half-wave-plate device could not produce a column-normalized 10x10 matrix with those means. The reader's weakest_assumption about two-layer nanograting fidelity may matter, but the numerical inconsistency blocks evaluation first. I also flag the abstract/body mismatch noted by the reader: the abstract claims 5 planes, 28 modes, a speckle sorter, and matrix multiplication, none of which appear in the full text; this should be corrected regardless. My recommended verdict remains CONDITIONAL rather than REJECT because the inconsistency may be a typographical or normalization error that raw data would resolve, but the condition is now specific: the corrected 10-mode data must confirm diagonal-dominant coupling before the claim is accepted.","tokens_in":12448,"tokens_out":9049,"duration_ms":93690,"concrete_test":"Recompute the mean off-diagonal intensity directly from the unnormalized detected intensities used for Fig. 3(b), applying the stated per-column normalization. If the recomputed means are ≤0.111, the printed values are typos; if any exceed 0.111, the normalization statement or the matrix entries are wrong. Publish the raw (unnormalized) 10x10 coupling matrix with per-column totals so both relative cross-talk and absolute throughput can be independently verified.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central quantitative evidence for the 10-mode sorter is internally inconsistent. Fig. 3(b) and the caption state that each column of the coupling matrix is normalized to 1 and report mean off-diagonal intensity values of 0.13 (simulated) and 0.58 (experimental). For a 10x10 column-normalized nonnegative intensity matrix, the 90 off-diagonal entries have total at most 10 (achieved only if every diagonal entry is 0), so the mean off-diagonal value cannot exceed 10/90 = 0.111. Both reported values exceed this bound; 0.58 would require total off-diagonal power of 52.2 in units where the whole matrix sums to 10. This is not a subtle statistical issue — the stated normalization and the quoted means cannot both be true. The 3-mode values (0.02 and 0.2) are within the possible range for a 3x3 column-normalized matrix, suggesting the 10-mode numbers may be typographical or computed under a different, unspecified normalization. Until the raw data or a corrected matrix are supplied, the claim that 'in most cases power is focussed into the correct output channel' for the 10-mode device cannot be checked, and the upper bound of the central claim is unsupported as reported.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a proof-of-concept fabrication of miniaturised transmissive multi-plane light converters (MPLCs) inside fused silica glass using femtosecond-laser-written geometric phase holograms. Two 2-plane devices are designed via inverse design and experimentally characterized: a 3-mode and a 10-mode Hermite-Gaussian (HG) mode sorter, operating at 633 nm. Measured coupling matrices are presented for both devices. The abstract, however, claims additional demonstrations (5-plane devices, sorting 28 HG modes, a 7-mode speckle sorter, and optical matrix multiplication) that are not described in the body of the paper.","tokens_in":12712,"tokens_out":5149,"duration_ms":43907,"significance":"The fabrication route is of interest: cascaded geometric phase holograms written in glass could offer compact, monolithic, passive MPLCs, avoiding alignment and encapsulation issues. The 3-mode sorter is demonstrated with measured coupling data, and the discussion honestly identifies limitations (efficiency ~20% per plane, polarization conversion fidelity, scattering). However, the central quantitative evidence for the 10-mode sorter – the coupling matrix in Fig. 3(b) – is internally inconsistent with the stated column normalization, and the abstract overstates the results. The overall significance is therefore conditional on correcting these issues.","major_comments":[{"comment":"The abstract claims demonstration of MPLCs formed from up to 5 phase masks, sorting up to 28 HG modes, a 7-mode speckle sorter, and analogue optical matrix multiplications, with a volume of ~0.15 mm³. None of these results appear in the body; the experimental devices are two-plane, 3-mode and 10-mode sorters, with volumes ~0.5 mm³ and ~0.8 mm³. This is a serious mismatch between the abstract and the content and must be corrected.","section":"Abstract vs. Results"},{"comment":"The caption states that each column of the coupling matrix is normalized to 1 and reports mean off-diagonal intensity values of 0.13 (simulated) and 0.58 (experimental) for the 10×10 matrix. For a nonnegative column-normalized matrix, the 90 off-diagonal entries sum to at most 10, so their mean cannot exceed 10/90 ≈ 0.111. Both reported values violate this bound. The reported normalization and means cannot both be correct. This invalidates the quantitative support for the claim that 'in most cases power is focussed into the correct output channel'. The authors must supply the actual matrices or a corrected analysis.","section":"Fig. 3(b) and 10-mode sorter section"},{"comment":"The statement 'We see that in most cases power is focussed into the correct output channel' is not supported by the reported mean off-diagonal of 0.58, which would imply that off-diagonal channels receive more power on average than the diagonal (indeed impossible under the stated normalization). The claim needs to be backed by the raw coupling data or a corrected matrix.","section":"Results, 10-mode sorter paragraph"}],"minor_comments":[{"comment":"Typo: 'CHG' should be 'CGH' (computer generated hologram).","section":"Section 2 (CGH)"},{"comment":"The figure captions do not specify the ordering of the HG modes; adding a mode index would improve reproducibility.","section":"Fig. 2 and Fig. 3 captions"},{"comment":"The Methods give laser parameters (pulse duration, repetition rate, writing speed, number of pulses) but not the average power or pulse energy; these are needed to reproduce the fabrication.","section":"Methods"},{"comment":"The sentence 'These results were first presented at Photon 2024' in the main text is unusual for a journal article; consider moving to acknowledgments or a footnote.","section":"Main text"}],"recommendation":"major_revision","confidential_remarks":"The mismatch between the abstract and the body is the most concerning issue, as it overstates the demonstrated scope. The normalization inconsistency in Fig. 3(b) is a load-bearing problem that must be resolved with raw data or a corrected presentation. The fabrication approach is novel and the 3-mode result is encouraging, so a major revision rather than rejection is appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the paper has a real, modest experimental result — 3-mode and 10-mode Hermite-Gaussian sorters made from two cascaded fs-laser-written geometric phase holograms in glass — and that part deserves attention. The abstract, however, describes a different paper (5 planes, 28 modes, speckle sorter, matrix multiplication) that never appears in the body. That mismatch alone forces a major revision.\n\nWhat is actually new: combining cascaded geometric phase holograms inside fused silica to make a monolithic transmissive MPLC. Earlier laser-written volume optics used refractive-index or waveguide structures; earlier MPLCs used SLMs, lithography, metasurfaces, or 2PP. The 3-mode and 10-mode sorters are documented with measured intensity images and coupling matrices, and the 3-mode numbers look plausible for a proof of principle. The discussion is honest: they openly state the ~20% per-plane efficiency and imperfect half-wave-plate response limit the approach, and they point to Type X birefringence as a possible fix.\n\nThe soft spots are real, and one is mathematical. The caption for Fig. 3(b) says each column of the coupling matrix is normalized to 1, then reports mean off-diagonal values of 0.13 (sim) and 0.58 (exp) for the 10-mode device. For a 10x10 column-normalized nonnegative matrix, the off-diagonal total is at most 10, so the mean off-diagonal cannot exceed 10/90 ≈ 0.111. Both reported values violate that bound. This is not a quibble about noise; the printed simulation number is impossible under the stated normalization. Either the normalization statement, the quoted means, or the matrix itself is wrong, and the central quantitative claim for the 10-mode device cannot be checked as reported. The 3-mode values are within the possible range, so this might be a typo or a different normalization convention, but raw data or a corrected matrix are needed.\n\nOther issues are secondary: no error bars or repeats, per-plane efficiency only estimated, and the half-wave-plate assumption (two nanograting layers separated by 44 µm) is load-bearing and only partially characterized — though the authors flag it themselves.\n\nWho this is for: people working on direct laser writing, spatial mode sorting, or compact MPLCs. It deserves a serious referee because the fabrication route is new, but the current manuscript needs substantive revision: align the abstract with the body, fix the coupling-matrix numbers, and supply the missing data. If that is done, this could be a solid proof-of-principle paper; as it stands, I would not trust the quantitative claims without seeing the underlying matrices.","headline":"Genuine proof-of-concept for cascaded laser-written geometric phase MPLCs in glass, but the abstract oversells the results and the 10-mode coupling numbers as printed contradict their own normalization.","tokens_in":13282,"tokens_out":2144,"would_cite":true,"duration_ms":20649,"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":"A laser-written glass chip can sort ten overlapping Hermite-Gaussian modes in a volume under one cubic millimetre.","keywords":["multi-plane light conversion","geometric phase","Pancharatnam-Berry phase","femtosecond laser writing","birefringent nanogratings","Hermite-Gaussian mode sorter","fused silica","mode sorting"],"falsifier":"Measure the polarisation state transmitted by a single laser-written geometric phase hologram at 633 nm: if the output is appreciably elliptical (wrong-handed component above a few percent), the phase fidelity assumed in Eq. (1) fails and the elevated crosstalk in the cascaded sorters is explained. A direct test is to simulate the 10-mode coupling matrix using the measured non-ideal Jones matrix and compare with the experimental matrix.","tokens_in":12272,"feed_emoji":"💡","tokens_out":6357,"duration_ms":58715,"temperature":0.7,"pith_summary":"Multi-plane light converters (MPLCs) are cascades of phase masks that reshape a set of input light modes into a set of output modes, but building them is bulky and alignment-sensitive. This paper tries to establish that femtosecond-laser-written geometric phase holograms, cascaded inside a monolithic fused-silica chip, can act as miniaturised transmissive MPLCs. The authors experimentally demonstrate proof-of-concept devices below 0.8 mm³ that sort up to 10 Hermite-Gaussian modes at 633 nm, with measured coupling matrices showing mode-dependent cross-talk. If the approach matures, it offers a single-step route to robust, encapsulated, passively operating mode sorters and linear optical processors.","feed_headline":"Ten laser modes sorted inside a 0.8 mm³ glass chip","feed_subtitle":"A single-step laser writing process packs multi-plane light conversion into a monolithic glass chip.","key_machinery":"The geometric phase hologram: a plane of femtosecond-laser-induced birefringent nanogratings whose local fast-axis orientation φ encodes a phase 2φ on the transmitted circularly polarised field, flipping its handedness. Two vertically separated nanograting layers (44 µm apart) are written per plane to approximate an ideal half-wave plate. Cascading several such planes inside the glass, with free-space (glass) propagation between them, constitutes the MPLC; inverse design via adjoint gradient descent with scalar angular-spectrum propagation sets the phase profiles.","core_discovery":"The central claim is that cascading laser-written birefringent nanogratings inside silica glass creates geometric phase holograms that act as thin phase planes, and that a stack of these planes, inverse-designed and co-registered in a single writing pass, functions as a transmissive MPLC. Each plane imprints a Pancharatnam-Berry phase 2φ on circularly polarised light by locally rotating the fast axis of a half-wave-plate-like structure; diffraction between planes reshapes the field. The authors report two proof-of-concept sorters — a 3-mode and a 10-mode Hermite-Gaussian sorter — with volumes of ~0.5 mm³ and ~0.8 mm³ respectively. They also show a single-plane hologram projecting an image. T","pith_inferences":["A natural falsifying experiment is to measure the output polarisation ellipticity of a single written hologram; if the wrong-handed component is large, the excess crosstalk in the 10-mode sorter is explained without invoking fabrication randomness.","The mode-dependent crosstalk pattern suggests input-mode generation quality is a major contributor; an experiment using a faithfully generated set of HG modes (e.g., via a photonic lantern or fiber source) could isolate the device's intrinsic crosstalk.","If the geometric phase response is achromatic enough, these glass chips could serve as broadband mode sorters; measuring the coupling matrix at a second wavelength would test that.","Integration with other laser-written components (waveguides, volume optics) in the same glass block is a plausible route to fully monolithic photonic circuits, a step the paper hints at but does not demonstrate."],"forward_implications":["Higher mode counts require more phase planes; the inverse-design and fabrication flow is in principle extensible, with the paper noting that more planes maintain fidelity for a given mode number.","Improving per-plane efficiency from the current ~20% toward the >90% reported for optimised geometric phase gratings would make multi-plane cascades practical for real applications.","Because the device is monolithic and encapsulated in glass, it is robust against environmental perturbations and requires no post-fabrication alignment, unlike SLM-based or free-standing micro-optic MPLCs.","The single-step writing process supports rapid prototyping, and the geometric phase approach can be extended to vectorial elements (q-plates, polarisation converters) for vectorial MPLCs."],"fun_headline_variants":["Laser-written glass chip sorts 28 overlapping light modes","Miniaturized MPLC in glass: cascaded holograms split 28 modes","28-mode sorter in a tiny glass chip via laser writing","Cascaded geometric phase holograms in glass make compact MPLC"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The sorting performance rests on the nanograting layer pair behaving as an ideal half-wave plate, so that no light leaks into the wrong circular polarisation handedness to contaminate downstream planes.","fun_headline_variants_meta":{"raw":{"variants":["Laser-written glass chip sorts 28 overlapping light modes","Miniaturized MPLC in glass: cascaded holograms split 28 modes","28-mode sorter in a tiny glass chip via laser writing","Cascaded geometric phase holograms in glass make compact MPLC"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000397,"raw_usage":{"total_tokens":1992,"prompt_tokens":898,"completion_tokens":1094,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":642,"completion_tokens_details":{"reasoning_tokens":1019}},"tokens_in":642,"tokens_out":1094,"duration_ms":10209,"temperature":1.0,"reasoning_tokens":1019,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T03:37:53.018449+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the polarisation state transmitted by a single laser-written geometric phase hologram at 633 nm: if the output is appreciably elliptical (wrong-handed component above a few percent), the phase fidelity assumed in Eq. (1) fails and the elevated crosstalk in the cascaded sorters is explained. A direct test is to simulate the 10-mode coupling matrix using the measured non-ideal Jones matrix and compare with the experimental matrix.","supporting_citations":[],"review_version":1}