{"id":"f991389c-3773-4766-9b8c-109305f95fdc","arxiv_id":"2608.09518","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Under identical experimental conditions, photon addition outperforms photon subtraction in multi-mode thermal and sub-Poissonian beams, while photon subtraction outperforms photon addition in multi-mode twin beams; temporal correlations in compound twin beams enable near-ideal photon addition.","lead":"This paper uses \"compound beams\", synthetic multi-mode light fields built from many weak twin-beam detections, to directly compare photon addition and photon subtraction on thermal, sub-Poissonian, and twin-beam states. It finds that addition works better for thermal and sub-Poissonian beams, while subtraction wins for twin beams, and demonstrates a post-selection scheme that approximates ideal photon addition.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'photon addition' analyzed here is an incoherent convolution with an independent heralded source (SM Eqs. 20, 24), not the standard a† operation; the PA-vs-PS comparison may not apply to actual photon addition.","rationale":"The paper's experimental platform is credible: it uses 645×10^6 weak TWB detections, concatenates up to 200 windows, and provides a detailed detection model (Eq. 3) and maximum-likelihood reconstruction. I take the central claim to be that PA is better than PS for thermal/sub-Poissonian beams and worse for TWBs, with the comparisons made under identical experimental conditions. For that claim to hold, the operation called PA must be the operation the literature calls photon addition (a†). The SM reveals otherwise. Eq. (20) constructs PATS by convolving p_th with p_a; Eq. (24) does the same for PATWB. An ideal heralded auxiliary gives a rigid shift p_in(n-1), while a† gives p'_m ∝ m p_{m-1}. These are different distributions with different Fano factors and nonclassicality depths. The difference is not a subtle mode-ordering correction; it is the difference between adding a photon number in an independent mode and applying a creation operator to the target mode. This invalidates the comparison with PS, which is modeled as a standard beam-splitter subtraction in Eq. (23). The temporal-vs-spatial mapping concern raised by the reader is related but not the sharpest version: even if one accepts the mapping, the PA construction remains a convolution. The concrete test—replacing Eq. (24) with the standard PA distribution and recomputing Figs. 3/5—would settle whether the ordering survives. If it does not, the abstract and conclusions overstate the result. Because the underlying data and compound-beam methodology have value and the flaw is in the operation label and its consequences, I recommend a conditional verdict rather than rejection.","tokens_in":17094,"tokens_out":16684,"duration_ms":165573,"concrete_test":"Recompute the PATS and PATWB photon-number distributions from the reconstructed original states using the standard PA formula p'(n_s,n_i) ∝ n_s p(n_s-1,n_i) (for one added photon) rather than the SM Eq. (24) convolution, keeping all experimental parameters fixed. Compare the resulting Fano factors and nonclassicality depths τ with those reported in Figs. 3 and 5; if the ordering PA-over-PS in thermal/sub-Poissonian states or PS-over-PA in twin beams changes, the paper's operation is not the standard photon addition. An analytical check: for a single-mode thermal state with mean \\bar n, standard PA gives F=2\\bar n(1+\\bar n)/(1+2\\bar n), while the paper's shift operation gives F=\\bar n; these differ already at \\bar n=0.5 (0.75 vs 0.5).","verdict_should_be":"CONDITIONAL","load_bearing_attack":"In the SM, Eqs. (20) and (24) define the photon-added states by convolving the original photon-number distribution with the conditional distribution p_a of an auxiliary TWB: p_a_th(n; \\bar c_a)=Σ p_th(n-n') p_a(n') and p_a_TWB(n_s,n_i)=Σ p_TWB(n_s-n'_s,n_i) p_a(n'_s). For an ideal auxiliary source p_a(n')=δ_{n',1}, this gives p_out(n)=p_in(n-1), a rigid shift of the photon-number distribution. The standard photon-addition operation a†ρa on a Fock-diagonal state gives p'_m ∝ m p_{m-1}, which is not a shift. For a single-mode thermal state with mean \\bar n, the shift gives Fano factor F=\\bar n, whereas standard PA gives F=2\\bar n(1+\\bar n)/(1+2\\bar n); the two distributions are unequal for all \\bar n>0. The same mismatch affects the PATWB joint distribution, where the added photon is not correlated with the idler. Thus the operation compared with photon subtraction is not the non-Gaussian operation called photon addition in the literature and in the abstract; it is incoherent addition of a heralded photon number in an independent mode. Since the central claim is an ordering between PA and PS, and PA is not the standard operation, the headline comparison lacks support for actual photon addition. This concern is independent of the temporal-vs-spatial mapping issue: even a spatial compound beam built from independent modes would have the same convolution structure.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript introduces 'compound beams'—concatenated weak twin-beam detection windows—as a platform to compare photon addition (PA) and photon subtraction (PS) applied to multi-mode thermal states, sub-Poissonian states, and twin beams. The authors present experimental photocount histograms for all nine state groups, with beam intensities spanning two orders of magnitude, and analyze Fano factors, noise-reduction parameters, and nonclassicality depths. They conclude that PA is advantageous for thermal and sub-Poissonian states, while PS outperforms PA for twin beams, and that temporal correlations in compound twin beams enable a nearly ideal PA. The theoretical model is fitted to the original TWB characterization and then used to predict the PA/PS outcome histograms, which is a genuine predictive check. A central caveat is that the implemented 'photon addition' is an incoherent convolution with an auxiliary heralded source, not the standard operator a† applied to the state; this affects the interpretation of the headline comparison.","tokens_in":17491,"tokens_out":10519,"duration_ms":95891,"significance":"The compound-beam technique is versatile, and the dataset is substantial; the model's predictive structure—parameters fixed by the original TWB characterization and then compared with independent PA/PS histograms—is a genuine strength. The nonclassicality-depth analysis goes beyond simple Fano-factor tests and gives quantitative lower bounds. If the operation implemented were standard photon addition, the systematic side-by-side comparison over a wide intensity range would be valuable for quantum-state engineering. However, because the 'photon addition' analyzed is a convolution with an independently heralded photon-number distribution rather than the standard a† operation, the headline comparison and the 'nearly ideal PA' demonstration do not support conclusions about the standard non-Gaussian operation. The paper will be of interest to specialists in photocounting statistics and compound-beam techniques, but the broader quantum-optics claims require substantial revision.","major_comments":[{"comment":"The operation labeled 'photon addition' is defined by the convolution of the original photon-number distribution with the conditional distribution p_a of an auxiliary TWB: p_a_th(n) = Σ p_th(n−n') p_a(n') and p_a_TWB(n_s,n_i) = Σ p_TWB(n_s−n'_s,n_i) p_a(n'_s). For an ideal auxiliary source p_a(n')=δ_{n',1}, this gives p_out(n)=p_in(n−1), a rigid shift of the photon-number distribution. Standard photon addition a†ρa on a Fock-diagonal state gives p'_m ∝ m p_{m−1}, which is not a shift. For a single-mode thermal state with mean \\bar n, the shift yields F = \\bar n, whereas standard PA yields F = 2\\bar n(1+\\bar n)/(1+2\\bar n); these are unequal for all \\bar n>0. Thus the abstract's claims about comparing 'photon addition' with photon subtraction, and about demonstrating 'nearly ideal experimental photon addition', refer to an incoherent convolution operation rather than the standard non-Gaussian operation. Please rename the operation (e.g., 'heralded photon-number convolution') or provide a rigorous argument that this convolution reproduces the physically relevant effects of standard PA in the multimode setting.","section":"SM Sec. III, Eqs. (20) and (24); Abstract"},{"comment":"The headline result that photon subtraction outperforms photon addition in twin beams compares the Fano factor F_s^a of the signal beam after PA with F_i^s of the idler beam after PS. These are different marginal beams with different mean photon numbers: PA increases the signal mean, while PS leaves the idler mean unchanged. The text itself states that the smaller F_i^s is 'attributed to smaller mean photon numbers ⟨n_i⟩_s in PS compared to ⟨n_s⟩_a'. Since the Fano factor is mean-dependent for these multi-mode fields, this comparison does not establish superiority of PS over PA; it may reflect the different means rather than the operation. A controlled comparison—same marginal beam, matched mean photon number, or equal nonclassicality depth—is needed to support the conclusion.","section":"§2, 'Experimental data analysis' and Fig. 2(a,b)"},{"comment":"The histograms f_a_sP and f_s_sP are defined with a normalization by the total sum over both cs and ci, which yields a joint distribution rather than the conditional distribution f(cs|ci) implied by the notation and by the main-text Eq. (4). If this is not a typographical error, the PASPS and PSSPS Fano factors in Fig. 3(d) are not computed from the post-selected states claimed. Please correct the normalization to match the conditional definition and confirm that the data processing used the conditional normalization.","section":"SM Eqs. (9) and (13)"}],"minor_comments":[{"comment":"The reference title contains a typo: 'Suplementary material' should be 'Supplementary material'.","section":"Reference [56]"},{"comment":"The heading 'T able 1' appears with an erroneous space; please correct the formatting.","section":"Table 1"},{"comment":"The notation 'F_a^s [F_i^s]' is confusing; please define the superscript/subscript convention clearly (e.g., F_x^y with x = beam and y = operation) and use it consistently in the text and figures.","section":"Fig. 2 caption"},{"comment":"The Introduction asserts a 'one-to-one mapping between the temporal and spatial multiplexing' but does not provide an argument or derivation. Since the physical interpretation of the temporal compound beams rests on this mapping, a brief explanation or a reference to a dedicated derivation would strengthen the paper.","section":"Introduction, compound-beam mapping"},{"comment":"The definition of the success probability p_suc would benefit from an explicit statement of the normalization by the total number of realizations M_m, to avoid ambiguity in the histograms built from repeated use of the same weak-TWB data.","section":"SM Eq. (5)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript relies heavily on the authors' prior work (Refs. [31, 37]) for the compound-beam technique and the theoretical model; the incremental contribution is the side-by-side comparison of PA and PS. The main risk is terminological: the operation called 'photon addition' is not the standard a† operation. If the authors can justify their convolution operation as a legitimate multimode generalization of PA, the paper may be salvageable; otherwise the conclusions must be reframed and the title/abstract adjusted. The dataset and predictive-check structure are solid, so the work is worth a major-revision round rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nYou should know two things about arXiv:2608.09518. First, it is a serious experimental study: 645 million detection windows, a single down-conversion source, compound beams assembled by temporal concatenation, and a detailed Gaussian model that predicts the PA/PS outcome distributions from parameters fitted to the unprocessed TWB. The side-by-side comparison of PA and PS on thermal, sub-Poissonian, and twin-beam states, with up to twenty added or subtracted photons across two orders of magnitude in intensity, is genuinely useful and goes beyond the group's earlier work. Second, the operation they call \"photon addition\" is not the standard a† operation. The SM defines the PA state as a convolution of the original photon-number distribution with a conditional distribution from an auxiliary twin beam (Eqs. 20, 24). For an ideal auxiliary this gives p_out(n)=p_in(n-1), a rigid shift. Standard PA on a Fock-diagonal state gives p'_m ∝ m p_{m-1}, which is not a shift. For a thermal state the two give different Fano factors at every mean. So the abstract's central claim that PA outperforms PS is about an incoherent count addition, not about the non-Gaussian operation the literature calls photon addition. This is not a minor semantics point: the added photon in a true a† operation is correlated with the idler, and the resulting nonclassicality is different. The paper's temporal-to-spatial mapping argument, which the reader flagged, would not fix this.\n\nWhere the paper is solid: the model is genuinely predictive, the nonclassicality-depth analysis is careful, and the ideal-PA post-selection via temporal correlations is a neat trick. The twin-beam comparison is also slightly confounded by comparing the signal Fano factor after PA with the idler Fano factor after PS at different mean photon numbers; the authors note this, but it weakens the \"PS outperforms PA\" claim.\n\nI'd send this out to a serious referee because the experimental platform is valuable and the definitional issue can be repaired by relabeling the operation or by implementing true PA with a standard parametric scheme. As it stands, the headline comparison does not support the stated conclusion. With a revised framing—calling it \"heralded photon-count addition\" and adjusting the claims—the paper could be publishable. I would not cite the PA-vs-PS ordering as evidence about photon addition.","headline":"Serious experimental platform, but the operation it calls 'photon addition' is an incoherent convolution with a heralded auxiliary state, not the standard a† operation, so the headline PA-vs-PS ordering is about a different operation.","tokens_in":17985,"tokens_out":8568,"would_cite":false,"duration_ms":77150,"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":"Using compound beams, the paper finds that photon addition beats photon subtraction for thermal and sub-Poissonian beams, while photon subtraction wins for twin beams, and demonstrates nearly ideal photon addition by post-selection.","keywords":["compound beams","photon addition","photon subtraction","nonclassicality","twin beams","thermal states","sub-Poissonian states","temporal multiplexing"],"falsifier":"Construct the same photon-addition and photon-subtraction states using a genuine spatially multi-mode field detected by a photon-number-resolving camera at matched mean photon numbers and matched added and subtracted counts, then compare the Fano factors and noise-reduction parameters with the temporally concatenated compound-beam results; a systematic divergence would show that the temporal-to-spatial mapping fails for those states.","tokens_in":16909,"feed_emoji":"🔬","tokens_out":9488,"duration_ms":82302,"temperature":0.7,"pith_summary":"Compound beams—multi-mode fields assembled by concatenating many weak twin-beam detection windows from a single down-conversion source—let the authors compare photon addition and photon subtraction under identical experimental conditions. Previous comparisons were rare because different quantum operations usually need different sources, so results were not on the same footing. Using these beams, the paper finds that photon addition outperforms photon subtraction for multi-mode thermal and sub-Poissonian beams: addition induces nonclassicality in thermal states and enhances it in sub-Poissonian states, whereas subtraction cannot make a thermal state nonclassical. In twin beams the reverse holds: photon subtraction produces stronger marginal nonclassicality than photon addition, although it pays for this by disturbing the signal-idler quantum correlations more. The authors also exploit temporal photon-pair correlations in compound twin beams to demonstrate nearly ideal photon addition, which produces more nonclassical states than the realistic operation.","feed_headline":"Photon addition beats subtraction—except in twin beams","feed_subtitle":"Testing both operations on the same light source tells engineers which one best creates nonclassical states.","key_machinery":"The central object is the compound beam: a multi-mode field built from simple experimental blocks, here weak twin-beam detection windows concatenated in time. Photon addition is realized by taking additional detection windows from the same two channels, forming an auxiliary twin beam, and post-selecting on a fixed idler photocount number; the conditioned signal part is then added to the original beam. Photon subtraction is realized virtually by splitting the $M_w$ signal windows into $M_s$ and $M_w-M_s$ groups, mimicking a beam splitter with transmissivity $T_s=1-M_s/M_w$. The statistics are modeled by Mandel-Rice photon-pair and noise components, inverted from photocount histograms by maximum-likelihood reconstruction, and quantified by the Fano factor $F$, the noise-reduction parameter $R$, and the nonclassicality depth, defined as the amount of thermal noise needed to conceal the field's nonclassical features.","core_discovery":"On the paper's own terms, the central claim is that compound beams provide a common experimental platform on which quantum operations can be compared directly, and that this platform changes the practical answer to the question 'which operation is better?'. For thermal states, photon addition creates nonclassicality (Fano factor $F<1$) while photon subtraction only lowers the intensity and cannot cross the quantum–classical border; for sub-Poissonian states, addition and subtraction both preserve the existing nonclassicality, with ideal photon addition clearly enhancing it. For twin beams, photon subtraction gives lower Fano factors in the marginal beams than photon addition, but at the cost of a larger increase in the noise-reduction parameter $R$; at high twin-beam intensity the two operations become indistinguishable because adding or subtracting a few photons barely changes a bright beam. The paper further claims nearly ideal photon addition by post-selecting on temporal photon-pair correlations in compound twin beams, with ideal addition of one, two, and three photons producing stronger nonclassicality than the realistic operation.","pith_inferences":["If the temporal–spatial mapping holds, the same compound-beam construction could benchmark other non-Gaussian operations—photon catalysis, noiseless amplification, or general post-selection filters—on identical footing, giving protocol designers a direct way to choose operations for a given task.","The ideal photon-addition result suggests that improved photon-number-resolving detectors could make heralded addition and subtraction approach ideal nonclassicality limits, potentially strengthening entanglement distillation and cat-state generation.","Since compound beams can be concatenated almost arbitrarily, the method could also serve as a predictive tool: it can estimate the properties of high-intensity multi-mode states before the corresponding real source is built."],"forward_implications":["Photon addition is the operation of choice for generating nonclassicality from thermal light and for pushing sub-Poissonian light further from the classical border.","Photon subtraction is the operation of choice for making the marginal beams of twin beams nonclassical, though it weakens the signal-idler correlations more than addition does.","The larger the number of added or subtracted photons (up to 20), the stronger the induced nonclassicality, but the weaker the remaining quantum correlations; when the beam is much brighter than the number of photons operated on, the effect becomes negligible.","Post-selection based on temporal photon-pair correlations can simulate an ideal photon-number-resolving detector, and the resulting ideal photon-added states are more nonclassical than realistically added states."],"supporting_citations":[{"why":"Supplies the compound-beam construction method by temporal multiplexing without requiring genuine photon-number-resolving detection.","marker":"[31]"},{"why":"Supplies the theoretical model and prior experimental treatment of photon addition and subtraction in multi-mode fields that this paper extends to a direct side-by-side comparison.","marker":"[37]"},{"why":"Supplies the Mandel-Rice component model for the multi-mode twin-beam photon-number distribution used in Eq. (4).","marker":"[60]"},{"why":"Supplies the maximum-likelihood reconstruction method used to convert photocount histograms into photon-number distributions.","marker":"[58]"},{"why":"Supports the temporal-to-spatial mapping by realizing the virtual beam-splitter operation with real spatially multi-mode fields.","marker":"[32]"},{"why":"Supplies earlier experimental characterization of multimode photon-subtracted twin beams that is here compared against photon addition.","marker":"[36]"},{"why":"Supplies the nonclassicality depth, the quantifier used to compare the strengths of the nonclassical states created by addition and subtraction.","marker":"[63]"},{"why":"Supplies the nonclassicality witnesses derived from the Fano factor and noise-reduction parameter used to estimate lower bounds on nonclassicality depth.","marker":"[64]"}],"fun_headline_variants":["Photon addition wins for thermal, subtraction for twin beams","For thermal beams, photon addition wins; for twin, subtraction does","Compound beams: direct test settles photon addition vs subtraction","Near-ideal photon addition from post-selected twin pairs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that concatenating many short, weak twin-beam measurements produces a compound beam whose statistics faithfully represent a real multi-mode quantum state that exists simultaneously; the paper acknowledges that these compound beams do not exist in real time and depends on a claimed one-to-one mapping between temporal and spatial multiplexing.","fun_headline_variants_meta":{"raw":{"variants":["Photon addition wins for thermal, subtraction for twin beams","For thermal beams, photon addition wins; for twin, subtraction does","Compound beams: direct test settles photon addition vs subtraction","Near-ideal photon addition from post-selected twin pairs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002134,"raw_usage":{"total_tokens":8273,"prompt_tokens":929,"completion_tokens":7344,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":545,"completion_tokens_details":{"reasoning_tokens":7276}},"tokens_in":545,"tokens_out":7344,"duration_ms":44968,"temperature":1.0,"reasoning_tokens":7276,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T15:45:27.796715+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct the same photon-addition and photon-subtraction states using a genuine spatially multi-mode field detected by a photon-number-resolving camera at matched mean photon numbers and matched added and subtracted counts, then compare the Fano factors and noise-reduction parameters with the temporally concatenated compound-beam results; a systematic divergence would show that the temporal-to-spatial mapping fails for those states.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the compound-beam construction method by temporal multiplexing without requiring genuine photon-number-resolving detection."},{"cited_title":"Lett.\\/ 49 4521--4524","cited_arxiv_id":null,"evidence_quote":"Supplies the theoretical model and prior experimental treatment of photon addition and subtraction in multi-mode fields that this paper extends to a direct side-by-side comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Mandel-Rice component model for the multi-mode twin-beam photon-number distribution used in Eq. (4)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the maximum-likelihood reconstruction method used to convert photocount histograms into photon-number distributions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supports the temporal-to-spatial mapping by realizing the virtual beam-splitter operation with real spatially multi-mode fields."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies earlier experimental characterization of multimode photon-subtracted twin beams that is here compared against photon addition."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the nonclassicality depth, the quantifier used to compare the strengths of the nonclassical states created by addition and subtraction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the nonclassicality witnesses derived from the Fano factor and noise-reduction parameter used to estimate lower bounds on nonclassicality depth."}],"review_version":1}