{"id":"6d55b561-9719-47e0-93e5-aa0a1f04b0be","arxiv_id":"2501.07062","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Near-field XL-MIMO effective degrees of freedom peak at antenna spacing d = sqrt(lambda*L/sqrt(N)), the same spacing where the array gain to the nearest antenna first becomes zero.","lead":"This paper derives a simple spacing rule for antennas in a near-field XL-MIMO link: effective degrees of freedom are maximized when the antenna spacing equals the square root of wavelength times link distance divided by the fourth root of the antenna count. The rule matters because it gives 6G array designers a fast way to choose antenna spacing and tells them when two popular EDoF estimators stop being reliable.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The closed-form EDoF-maximizing spacing rests on an unproven equivalence between minimizing nearest-neighbor array gain and maximizing EDoF.","rationale":"The reader's weakest_assumption identifies the same unproven link between nearest-neighbor array gain null and EDoF maximization, which is exactly the load-bearing step for the paper's headline formula. I verified that the algebraic derivation of Eq. (21) is correct under the stated Fresnel approximation, so the concern is not about the null calculation itself but about its interpretation as the EDoF-optimal spacing. The numerical evidence in Fig. 5 covers only one array size and one link distance, and the EDoF definition relies on a hand-chosen 99.9% cutoff, so it cannot certify the general claim. A concrete multi-geometry SVD check would settle whether the equivalence holds broadly; if it does, the paper's design rule would be far better supported. Given that the concern is substantive but not a demonstrated error, the conditional verdict is appropriate, with no change needed.","tokens_in":9205,"tokens_out":4533,"duration_ms":41460,"concrete_test":"For at least three geometries (e.g., N = 25×25, 35×35, 45×45; L = 500λ, 2000λ, 4000λ), compute the exact nEDoF(d) by SVD of the full Green's-function channel matrix using the paper's 99.9% energy threshold, and separately compute the exact (non-Fresnel) array gain rho1(d) at the nearest receive antenna when focusing on the center antenna. Sweep d over a fine grid bracketing d_threshold = sqrt(lambda L / sqrt(N)). Check whether argmax_d nEDoF(d) coincides with argmin_d rho1(d) (or the first zero of rho1(d)) to within 5% in d for every configuration. If any configuration shows a mismatch beyond numerical tolerance, the assumed equivalence fails and the claimed optimal spacing is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The derivation of d_threshold in Section IV, Theorem 1 is internally consistent under the stated Fresnel/Taylor approximation, and the null of the approximate nearest-neighbor array gain is correctly identified. However, the paper's central claim that this null spacing maximizes nEDoF relies on the sentence before Theorem 1: 'when the transmit array focuses on one antenna, nEDoF is maximized when the interference imposed on the nearest antenna is minimized.' This is asserted as 'reasonable' but is not proved, and it is not an immediate consequence of the definition of EDoF. EDoF is a global property of the channel matrix's singular-value distribution; minimizing one entry of the focused-field pattern (the nearest-neighbor array gain) does not logically imply that the number of significant singular values is maximized. The numerical verification in Fig. 5 uses one geometry (N = 25×25, L = 4000λ) and the 99.9% energy cutoff, so it does not establish the equivalence across the parameter space where the formula is claimed to hold. A secondary concern is the paraxial approximation in Appendix A, Eq. (28): the Taylor step requires L^2 ≫ max(\\bar x_n^2, \\bar y_m^2), but the paper never quantifies the resulting domain of validity for Eq. (22) in terms of N, L, and d. If the derived threshold itself violates this condition for some plausible N or L, the formula would be outside its own derivation regime.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a near-field XL-MIMO link with point antennas on two parallel UPAs, modeled through the Green's function channel. It defines an effective DoF (nEDoF) as the number of significant singular values using a 99.9% energy threshold, discusses two approximate EDoF estimators, and then derives a closed-form antenna-spacing threshold, d_threshold = sqrt(lambda L / sqrt(N)), from the condition that the array gain at the antenna nearest to the focus point vanishes. The authors claim that this spacing maximizes nEDoF and marks the boundary of validity of the two EDoF estimators. Numerical results for a 25×25 array at L = 4000λ are presented to support the claim.","tokens_in":9448,"tokens_out":3179,"duration_ms":32474,"significance":"If the central claim is correct, the paper offers a simple, parameter-free design rule for antenna spacing in near-field XL-MIMO systems and a crisp characterization of when two popular EDoF estimators apply. The Fresnel/Taylor derivation of the nearest-neighbor array-gain null in Appendix A is algebraically correct under the stated paraxial assumptions, and Eq. (22) is an elegant closed form. The significance is conditional, however, on an unproved equivalence between minimizing the nearest-neighbor array gain and maximizing nEDoF, as well as on an unquantified domain of validity of the paraxial approximation.","major_comments":[{"comment":"The load-bearing statement 'when the transmit array focuses on one antenna, nEDoF is maximized when the interference imposed on the nearest antenna is minimized' is assumed to be 'reasonable' but is not proved. nEDoF is a global property of the singular-value spectrum of the channel matrix, and minimizing one element of the focused-field array gain does not logically imply that the number of significant singular values is maximized. Without a proof or a direct spectral characterization, Theorem 1 establishes only the spacing at which the nearest-neighbor array gain first vanishes, not the spacing that maximizes nEDoF. This gap directly affects the validity of Eq. (22) as the main contribution.","section":"Section IV, paragraph before Theorem 1"},{"comment":"The Taylor/Fresnel approximation leading to Eq. (21) and hence to Eq. (22) requires L^2 to be much larger than the squared aperture coordinates, but the paper never quantifies the resulting domain of validity in terms of N, L, and d. Since d_threshold = sqrt(lambda L / sqrt(N)) implies an aperture size of order N^{1/4} sqrt(lambda L), the condition can be violated for large N or small L. The authors should either state explicit inequalities that guarantee the approximation, or demonstrate numerically that Eq. (22) remains accurate outside the paraxial regime.","section":"Appendix A, Eq. (28)"},{"comment":"The numerical verification uses only a single geometry (N = 25 × 25, L = 4000λ) and a fixed 99.9% energy threshold in the definition of nEDoF. This does not establish that the null of the nearest-neighbor array gain coincides with the nEDoF maximum across the parameter space where Eq. (22) is claimed to hold. In particular, the definition of nEDoF in Eq. (11) depends on an arbitrary threshold, and the location of the nEDoF peak could shift if that threshold changes. A sensitivity study over N, L, and the energy threshold is needed to support the claimed equivalence.","section":"Section V, Fig. 5"}],"minor_comments":[{"comment":"The definition of nEDoF should be stated as the smallest n whose normalized cumulative energy reaches 0.999, rather than an argmin over a function that also includes a constraint; the current notation is imprecise.","section":"Eq. (11)"},{"comment":"The sinc function is used without defining its convention; the paper should state whether sinc(x) = sin(pi x)/(pi x) or sinc(x) = sin(x)/x, since this affects the numerical evaluation of Eq. (21).","section":"Appendix A and Theorem 1"},{"comment":"The units of L in Fig. 2 and the system parameters underlying Fig. 3 are not clearly stated; the reader cannot tell whether L is in meters, wavelengths, or normalized units, and Fig. 3 uses a different array size from Fig. 5 while the threshold changes from 3.2λ to 12.65λ.","section":"Section V, Figs. 2 and 3"},{"comment":"The abstract calls Eq. (22) the 'optimal antenna spacing,' while the text more cautiously describes it as the threshold where nEDoF peaks; the stronger wording should be justified or aligned with the derived result.","section":"Abstract and Section IV"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the core Fresnel derivation is sound, but the central claim that the array-gain null spacing maximizes nEDoF is currently an assumption rather than a demonstrated result. This is fixable either by adding a spectral proof or by a substantially broader numerical study that includes parameter sweeps and threshold sensitivity. I recommend major revision rather than rejection, because the proposed formula is plausible and the technical machinery is otherwise competent."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe genuinely useful thing here is Eq. (22): d_threshold = sqrt(lambda L / sqrt(N)), derived from the first null of the nearest-neighbor array gain under Fresnel/paraxial assumptions. The derivation in Appendix A is correct as far as it goes, and the paper deserves credit for turning a standard sinc-null calculation into a one-line design rule for near-field XL-MIMO. The secondary claim—that this threshold marks where two common EDoF estimators (Eqs. (12) and (13)) break down—is supported by the numerics and is a practical observation for people using those estimators.\n\nWhat is soft is the load-bearing step in Section IV, right before Theorem 1: the authors assume, without proof, that minimizing the array gain at the nearest antenna maximizes nEDoF. They call it \"reasonable,\" which is honest, but it is not a consequence of EDoF being a global property of the singular-value spectrum. The null derivation does not depend on that assumption, so the threshold is not a fitted quantity, but the paper's central interpretation—that this spacing maximizes EDoF—does depend on it. The numerical check is one geometry (25x25, L=4000 lambda) with a 99.9% energy cutoff, so it does not establish the equivalence across the claimed parameter range. A second, more minor soft spot: the paraxial Taylor step in Eq. (28) requires L^2 >> aperture coordinates, but the paper never states the resulting domain of validity for Eq. (22) in terms of N, L, and d. For some combinations the derived threshold may fall outside its own derivation regime.\n\nThat said, I do not see a fatal flaw. The assumption is out in the open, the algebra is coherent, and the estimator-applicability boundary is a real contribution even if the EDoF-maximizing claim needs proof or qualification. The citation pattern is normal for the subfield; the related work on EDoF estimation is cited appropriately.\n\nWho is this for? Researchers designing near-field XL-MIMO arrays who want a quick spacing rule, and people who use the EDoF estimators and need to know their validity range. It is a solid conference/letter-level contribution with a clear, testable formula. I would send it to peer review—it merits referee time, with the main request being that the authors either prove or explicitly weaken the nearest-neighbor-to-EDoF equivalence, and quantify the paraxial regime.\n\nRecommendation: engage, but ask for the equivalence to be addressed before publication.","headline":"A clean closed-form antenna-spacing threshold for near-field XL-MIMO, with the EDoF-maximizing claim resting on an openly stated but unproven equivalence.","tokens_in":10010,"tokens_out":864,"would_cite":true,"duration_ms":10343,"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":"Near-field XL-MIMO's EDoF-maximizing antenna spacing has a closed form, $d_{\\mathrm{threshold}}=\\sqrt{\\lambda L/\\sqrt{N}}$, derived from the first null of nearest-neighbor array gain.","keywords":["near-field XL-MIMO","effective degrees of freedom","antenna spacing","array gain","Green's function","uniform planar array","paraxial approximation","EDoF estimation"],"falsifier":"Simulate or measure a $25\\times25$ UPA at $L=4000\\lambda$ (the paper's parameters) with spacing swept over $8\\lambda$ to $18\\lambda$: compute both the focused array gain at the neighbor antenna and the eigenvalue-based $n_{\\mathrm{EDoF}}$. The paper predicts the gain's first null and the $n_{\\mathrm{EDoF}}$ peak both occur at $d=12.65\\lambda$; if the peak occurs at a different spacing, or no null appears there, the theorem is refuted.","tokens_in":8950,"feed_emoji":"📡","tokens_out":8957,"duration_ms":77148,"temperature":0.7,"pith_summary":"This paper claims to settle a design question for near-field extremely large-scale MIMO (XL-MIMO): what antenna spacing maximizes the effective degrees of freedom (EDoF), the number of significant spatial subchannels, for a given array size and link distance. Using a Green's function channel model, the authors derive an approximate closed form for the threshold spacing at which EDoF peaks. The formula is $d_{\\mathrm{threshold}}=\\sqrt{\\lambda L/\\sqrt{N}}$ for a square $N$-element uniform planar array. A reader should care because EDoF controls channel capacity, and this gives an explicit operating point instead of requiring exhaustive SVD searches. The paper also claims that two existing EDoF estimators are accurate only below this threshold.","feed_headline":"One formula sets the antenna spacing for peak near-field MIMO","feed_subtitle":"Below this spacing the standard EDoF estimators stay accurate; above it, they break.","key_machinery":"The load-bearing object is the nearest-neighbor array-gain formula. Starting from the Green's function between point antennas, the paper approximates the focused beam's response at the antenna adjacent to the focal point, using the paraxial (Fresnel) approximation $\\sqrt{1+x}\\approx 1+x/2$ with $L^2$ much larger than the aperture coordinates. The double sum over the planar array separates into a product of one-dimensional geometric series, producing the sinc-ratio expression. The first zero of that expression—when the numerator sinc vanishes and the denominator does not—defines the claimed threshold. This gain formula connects the geometric parameters ($d$, $N$, $\\lambda$, $L$) to the onset of eigenvalue decorrelation that governs EDoF.","core_discovery":"Stated as Theorem 1, the paper's central discovery is an approximate closed form for the array gain at the antenna adjacent to the focus of a focused transmit array in a near-field XL-MIMO link. For a square uniform planar array with $\\sqrt{N}$ elements per side (so $N_S=N_R=N$ total elements), when the transmitter focuses on one receive antenna, the gain at the neighboring receive antenna is\n$$\n\\rho_1 \\approx N\\,\\frac{\\operatorname{sinc}^2\\!\\bigl(\\sqrt{N}\\,\\frac{$d^{2}$}{\\$\\lambda$ L}\\bigr)}{\\operatorname{sinc}^2\\!\\bigl(\\frac{$d^{2}$}{\\$\\lambda$ L}\\bigr)}.\n$$\nThe first zero of this expression occurs at $\\sqrt{N}\\,d^2/(\\lambda L)=1$, giving $d_{\\mathrm{threshold}}=\\sqrt{\\lambda L/\\sqrt{N}}$. The paper claims that at this spacing the effective degrees of freedom $n_{\\mathrm{EDoF}}$ is maximized, and that below it the two standard estimators $n_{\\mathrm{EDoF}1}$ (fringe counting, Eq. (12)) and $n_{\\mathrm{EDoF}2}$ (eigenvalue ratio, Eq. (13)) are accurate while above it they are not. The same quantity $\\epsilon=\\sqrt{N}d^2/(\\lambda L)$ shows why spacing is a more powerful lever than antenna count: to keep $\\epsilon$ fixed at its peak condition, a fourfold increase in $N$ only halves the required $d$.","pith_inferences":["Inference: the first-zero condition places a null on the adjacent antenna; this could be used deliberately for spatial-multiplexing interference nulling between neighboring receive elements, an engineering use beyond the EDoF-maximization framing.","Inference: for rectangular or unequal transmit/receive arrays, the same separated-sum derivation should yield a pair of thresholds (one per axis) or a dependence on the geometric mean of the array dimensions; testing this would extend Eq. (21) naturally.","Inference: the paper's identification of the EDoF peak with the nearest-neighbor gain null rests on numerical observation; a rigorous proof would need to connect the full eigenvalue spectrum to this gain, and a counterexample with a different eigenvalue ordering would separate the two quantities."],"forward_implications":["Setting $d = \\sqrt{\\lambda L/\\sqrt{N}}$ maximizes the effective degrees of freedom for a given square UPA, so system designers have an explicit operating point instead of running SVD searches over spacings.","Below $d_{\\mathrm{threshold}}$, the two EDoF estimators in Eqs. (12) and (13) agree with the SVD-based $n_{\\mathrm{EDoF}}$; above it they diverge, so the threshold doubles as the validity boundary for those estimators.","The capacity computed with $n_{\\mathrm{EDoF}}$ closely matches the full $n_{\\mathrm{DoF}}$ capacity, meaning the threshold spacing is a capacity-relevant design target, not just an eigenvalue curiosity.","Because $\\epsilon=\\sqrt{N}d^2/(\\lambda L)$ must approach 1 for peak EDoF, the required spacing shrinks only as $N^{-1/4}$, quantifying the diminishing returns of adding antennas compared with increasing spacing."],"supporting_citations":[{"why":"Supplies the Green's-function channel model and the intensity-fringe EDoF estimator $n_{\\mathrm{EDoF}1}$ (Eq. (12)) whose paraxial validity the paper analyzes.","marker":"[5]"},{"why":"Supplies the eigenvalue-ratio EDoF estimator $n_{\\mathrm{EDoF}2}$ (Eq. (13)) and the claim that its significant singular values are nearly constant in paraxial conditions.","marker":"[9]"},{"why":"Gives the channel-capacity formula in Eq. (9) that ties EDoF to rate, motivating the optimal-spacing analysis.","marker":"[11]"},{"why":"Supports the radiative near-field approximation that amplitude variations across the array are negligible compared with phase variations, used in deriving Eq. (16) and the sinc-ratio gain in Theorem 1.","marker":"[12]"}],"fun_headline_variants":["Near-field MIMO: one spacing rule peaks effective DoF","Antenna spacing sweet spot found for near-field XL-MIMO","Optimal spacing formula doubles as EDoF estimator breakpoint","Peak near-field MIMO capacity: spacing beats antenna count"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the unproved assertion, stated just before Theorem 1, that when the transmit array focuses on one antenna, $n_{\\mathrm{EDoF}}$ is maximized exactly when the array gain on the nearest antenna is minimized; if this link fails, the spacing formula does not necessarily maximize EDoF.","fun_headline_variants_meta":{"raw":{"variants":["Near-field MIMO: one spacing rule peaks effective DoF","Antenna spacing sweet spot found for near-field XL-MIMO","Optimal spacing formula doubles as EDoF estimator breakpoint","Peak near-field MIMO capacity: spacing beats antenna count"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000128,"raw_usage":{"total_tokens":1140,"prompt_tokens":986,"completion_tokens":154,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":602,"completion_tokens_details":{"reasoning_tokens":83}},"tokens_in":602,"tokens_out":154,"duration_ms":2469,"temperature":1.0,"reasoning_tokens":83,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:49:53.900677+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate or measure a $25\\times25$ UPA at $L=4000\\lambda$ (the paper's parameters) with spacing swept over $8\\lambda$ to $18\\lambda$: compute both the focused array gain at the neighbor antenna and the eigenvalue-based $n_{\\mathrm{EDoF}}$. The paper predicts the gain's first null and the $n_{\\mathrm{EDoF}}$ peak both occur at $d=12.65\\lambda$; if the peak occurs at a different spacing, or no null appears there, the theorem is refuted.","supporting_citations":[{"cited_title":"Waves, modes, communications, and opti cs: a tutorial,","cited_arxiv_id":null,"evidence_quote":"Supplies the Green's-function channel model and the intensity-fringe EDoF estimator $n_{\\mathrm{EDoF}1}$ (Eq. (12)) whose paraxial validity the paper analyzes."},{"cited_title":"Ele c- tromagnetic effective degree of freedom of an MIMO system in free space,","cited_arxiv_id":null,"evidence_quote":"Supplies the eigenvalue-ratio EDoF estimator $n_{\\mathrm{EDoF}2}$ (Eq. (13)) and the claim that its significant singular values are nearly constant in paraxial conditions."},{"cited_title":"On limits of wireless comm unications in a fading environment when using multiple antennas,","cited_arxiv_id":null,"evidence_quote":"Gives the channel-capacity formula in Eq. (9) that ties EDoF to rate, motivating the optimal-spacing analysis."}],"review_version":1}