{"id":"53294e8a-6c5a-4307-8035-d335d3a7fd40","arxiv_id":"2508.11489","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"For liquid crystal reconfigurable surfaces, the optimal phase-shifter length, which trades insertion loss against phase-shift range, depends on the user's position: specular users want a short shifter, off-axis users need a longer one.","lead":"This paper studies liquid crystal surfaces that bounce wireless signals, where shortening the surface's phase shifter cuts signal loss but limits how precisely the surface can steer the beam. The simulation shows the best shifter length depends on user position: users near the mirror-reflection direction do best with a short shifter, off-axis users need a longer one.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Fig. 2 appears inconsistent with the stated loss model: the specular-user power should rise exactly 4.8 dB from omega_max=0 to 2pi with FoM=75 deg/dB, but the plotted rise is ~7-8 dB; the central optimal-length claim rests on these curves.","rationale":"The reader's weakest-assumption analysis targeted the linear loss law Eq. (8) and the constant-FoM assumption. That is a reasonable fragility, but it is not the most load-bearing issue: the qualitative existence of an interior optimum is structurally robust to many monotone loss/range models as long as off-specular users are infeasible at zero range and loss grows with range. What is more damaging is that the paper's own numerical evidence, Fig. 2, appears to disagree with the exact 4.8 dB rise implied by Eqs. (6) and (8) at FoM=75 deg/dB. This is an internal inconsistency, not merely a disagreement with external consensus, and it directly affects the curves from which omega*_max is selected. The reader's point about missing details in Algorithm 1 is also relevant: if the rank-one step is omitted, the computed q_k and P_k are not valid for a physical phase-shift vector. I therefore keep the verdict CONDITIONAL (no change): the qualitative claim is plausible and may survive correction, but the reported simulation evidence needs a reproducibility check and a sensitivity analysis around FoM and the loss model before acceptance.","tokens_in":7988,"tokens_out":7782,"duration_ms":97236,"concrete_test":"Reproduce the y=0 specular-user case with FoM=75 deg/dB and Eq. (8): with all-zero phase shifts, record P_k at omega_max=0 and omega_max=2pi. The dB difference must be 4.8 dB (up to numerical tolerance). If it is not, trace whether a different FoM, a nonlinear/phase-dependent loss, or the missing rank-one recovery before computing q_k via (11) caused the deviation. Then add an explicit rank-one recovery step (e.g., leading-eigenvector or Gaussian randomization) to Algorithm 1 and regenerate Fig. 2; if the curve heights or the argmax omega*_max,k locations shift materially, the conclusions in Figs. 3-4 must be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim, that an interior optimal LC phase-shifter length exists and depends on scenario, is established by the numerical curves in Fig. 2 and the selected optima in Figs. 3-4. Under the Section III model, the reflection amplitude is |Gamma|^2 = 1/|Omega|^2 and Eq. (8) gives |Omega|^2_dB = (2pi/FoM)(l/lr) = omega_max/FoM. With FoM = 75 deg/dB, for a fixed lossless channel the required transmit power in dBm must increase by exactly 4.8 dB from omega_max=0 to 2pi. The y=0 (specular) curve in Fig. 2 appears to increase by about 7-8 dB over this interval. If the plotted curve is accurate, the simulation is using a different FoM, a different loss law, or an artifact of Algorithm 1 -- notably, line 11 computes q_k from (11) without an explicit rank-one recovery from S_k to s_k, so the reported P_k may not correspond to a feasible RIS configuration. Since every omega*_max is chosen by comparing these curves, this quantitative inconsistency must be resolved before the paper's central conclusion is supported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the trade-off between insertion loss and phase-shift range in liquid-crystal-based reconfigurable intelligent surfaces (LC-RIS). The authors model a delay-line LC phase shifter whose maximum phase shift and insertion loss both increase linearly with physical length (Eqs. (6) and (8)). For a narrow-band downlink system with blocked direct links, they formulate a transmit-power minimization problem under a per-user SNR constraint and a phase-range constraint, solve it by alternating optimization between beamforming and RIS phases, and simulate the required power versus ω_max for users at different locations and coverage radii. The central claim is that an interior optimal LC phase-shifter length exists and depends on user location and coverage area, so designing for full 2π range is not always optimal.","tokens_in":8110,"tokens_out":9619,"duration_ms":105955,"significance":"If the numerical results are accurate, the paper provides a useful design insight for LC-RIS hardware: the optimal phase-shifter length depends on the deployment scenario, and a smaller-than-2π range can be beneficial in specular-rich settings. The paper contributes a complete optimization formulation, a numerical algorithm, and parametric studies using a measured FoM from the authors' own device. However, the existence of the trade-off is largely a consequence of the assumed linear loss model, and the quantitative conclusions rest on simulation curves that currently show a discrepancy with the stated model. The work is therefore a promising co-design study rather than a demonstration of a new fundamental principle.","major_comments":[{"comment":"The specular-user (y=0) curve is inconsistent with the loss model in §III. With FoM=75°/dB, Eq. (8) gives |Ω|²_dB = (2π/FoM)(l/l_r) = ω_max/(75π/180), so the required power for the specular user should increase by exactly 4.8 dB between ω_max=0 and 2π. The plotted rise appears to be about 7–8 dB. Since this curve is the baseline against which off-specular optima are measured, please provide the exact simulation parameters, correct the plot, or explain the discrepancy (e.g., a different FoM, additional loss, or suboptimal phase recovery).","section":"§V.B, Fig. 2"},{"comment":"After solving P5, the algorithm selects [S_k*, ω_max,k*] but never recovers the phase vector s_k from S_k* before computing q_k with Eq. (11). If S_k* is only approximately rank-one, the reported P_k may not correspond to a feasible RIS configuration. Please add an explicit rank-one recovery step (e.g., principal eigenvector) and state how the recovered s_k satisfies the phase-range constraint.","section":"§IV, Algorithm 1"},{"comment":"The convex reformulation cC2 is imported from [14, Lemma 2 and Lemma 3] without stating the lemmas or providing a proof/sketch. This reformulation is central to the convex problem P5 and hence to all numerical results. The paper should either prove the reformulation in an appendix or state the lemmas explicitly. The equation as typeset is also garbled, with missing quantifiers and misaligned branches.","section":"§IV, Eq. (15)"},{"comment":"The headline 'fundamental trade-off' is not an independent discovery: combining Eqs. (6) and (8) gives |Ω|²_dB = ω_max/FoM, so loss in dB is proportional to ω_max by construction. The existence of an interior optimum for off-specular users follows whenever the SNR gain from increased phase range saturates. The paper should present the contribution as a quantitative evaluation of this model-dependent trade-off, not as a new fundamental phenomenon, and should state the linear-loss assumption explicitly as a modeling simplification.","section":"§III and §I"}],"minor_comments":[{"comment":"Please make the units of FoM explicit. The text quotes FoM as 75°/dB, but Eq. (8) uses 2π in the numerator; the equation is only correct if FoM is converted to rad/dB or if ω_max is expressed in degrees consistently.","section":"§III, Eqs. (7)–(8)"},{"comment":"The loop 'for ω_max,k = 0, · · ·, 2π' should state the step size or grid (|W|). The complexity expression includes |W| but the discretization is not specified in the algorithm.","section":"§IV, Algorithm 1 line 3"},{"comment":"The symbol P_k is used both for the coverage area set (possible user locations) and for the transmit power variable. This is confusing; please rename one of them.","section":"§II.A and §IV"},{"comment":"The FoM value used in the simulations is stated only in the text around Fig. 2; the figure captions should include it for reproducibility. Similarly, the path-loss and noise parameters are in the text but not in the captions.","section":"§V.B, Figs. 2 and 4"},{"comment":"The conclusion mentions a 'simplified scenario' but does not list which simplifications are most consequential (LOS-only channels, constant FoM, maximum-loss-at-all-phase-states, no fixed overhead losses). A short limitations paragraph would help readers judge practical applicability.","section":"§VI, Conclusion"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid co-design study, but the authors should be asked to verify the simulation code against Eq. (8) and to reframe the 'fundamental' claim as a model-based design insight. The use of their own measured FoM is a strength, but the lack of uncertainty analysis around that value and the missing rank-one recovery step are concerns. The paper fits the journal's scope if the hardware-communication co-design angle is emphasized."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the useful thing: this paper converts the known LC phase-shifter trade-off (loss grows with length, phase range grows with length) into a system-level design curve. For a specular user, it confirms that a shorter shifter is better; for off-specular users, it shows an interior optimal length. That is a genuinely useful sanity check for hardware folks choosing between different LC-RIS designs, and the simulations make the point clearly under the stated model.\n\nThe math is straightforward and the model is transparent. I don't see any hidden circularity: given Eqs. (6) and (8), the interior optimum is an inevitable consequence, but that's fine—the contribution is the quantification, not the discovery of the trade-off.\n\nNow the soft spots, in order.\n\n1. Fig. 2 looks inconsistent with the stated loss model. With FoM = 75°/dB, the specular-user curve should rise exactly 4.8 dB from ωmax=0 to 2π (since loss in dB = ωmax/FoM). The figure as plotted appears to rise 7–8 dB. If that's accurate, either the simulation used a different FoM, applied a different loss law, or there's a bug—possibly the missing rank-one recovery step in Algorithm 1. Since every ωmax* is selected from these curves, this needs to be resolved. It's not fatal to the qualitative conclusion, but it should be fixed before the paper is used for design numbers.\n\n2. Algorithm 1 omits the rank-one recovery from S_k to s_k before computing q_k with (11). As written, q_k may not correspond to a feasible RIS phase configuration. That's a missing implementation detail; easy to fix, but worth stating.\n\n3. The loss model in Eq. (8) is exactly linear in length at constant FoM. Real LC phase shifters have fixed overhead losses and FoM that varies with tuning state and frequency. The paper offers no sensitivity analysis around FoM or the linearity assumption. The qualitative picture would survive a per-element constant loss, but the optimal lengths would shift. A quick sensitivity check would strengthen the claim.\n\n4. The abstract's 'fundamental trade-off between total transmit power and achievable data rate' overstates what is actually done. The experiment fixes the data rate (SNRthr) and minimizes power. The trade-off is between power and phase-shift range, not data rate. The wording should be corrected.\n\nNone of these are load-bearing flaws in the sense that the qualitative conclusion is wrong. The paper is a modest but useful simulation study. It deserves a serious referee; the Fig. 2 inconsistency should be checked by the referee and fixed by the authors. If the simulation code is available, that would help.","headline":"Useful system-level quantification of the LC phase-range/loss trade-off, but the central figure has a numeric inconsistency that needs checking before design use.","tokens_in":8827,"tokens_out":4981,"would_cite":true,"duration_ms":48021,"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":"The paper claims that the optimal LC-RIS phase-shift range depends on user placement, with specular users best served by near-zero phase range and minimal loss.","keywords":["liquid crystal RIS","reconfigurable intelligent surface","phase-shift range","insertion loss","transmit power minimization","mmWave communications","delay-line phase shifter","beamforming optimization"],"falsifier":"Measure the insertion loss of a fabricated LC delay-line phase shifter at 28 GHz for several lengths and across the full tuning range. If the loss per element includes a fixed overhead (so dB loss versus length has a positive intercept) or if the figure of merit varies with tuning state, recompute the required transmit power versus ωmax; an interior optimum may vanish or move to a different ωmax.","tokens_in":7686,"feed_emoji":"📡","tokens_out":3301,"duration_ms":34953,"temperature":0.7,"pith_summary":"Liquid-crystal (LC) reconfigurable intelligent surfaces built from delay-line phase shifters face a trade-off: longer phase shifters allow a wider phase-shift range but add insertion loss. The paper claims that when minimizing transmit power under a quality-of-service constraint, the required power is non-monotone in the maximum phase-shift range, so a scenario-dependent optimal LC length exists. It shows that a user in the specular reflection direction is best served by zero phase shift and minimal loss, while off-specular users need a larger phase range that grows with their angular offset and with the size of their coverage area. If correct, LC-RIS hardware should be designed for the deployment scenario rather than uniformly for the full 2π range.","feed_headline":"Full phase range isn't always best for LC-RIS","feed_subtitle":"Simulations show the optimal LC shifter length depends on user direction, with specular users best served by near-zero phase range.","key_machinery":"The load-bearing identity is the linear loss–phase-range relation |Ω|²_dB = (2π/FoM)(l/lr) = ωmax/FoM, which makes insertion loss in dB literally proportional to maximum phase-shift range. This is combined with a near-field LOS channel model and an alternating-optimization algorithm: maximum-ratio-transmission beamforming for fixed phases, and a rank-one semidefinite relaxation with a penalty method to set phases for fixed ωmax, followed by a grid search over ωmax.","core_discovery":"The paper models an LC-RIS where the maximum phase-shift range is Δωmax = 2πl/lr and the insertion loss in dB is |Ω|²_dB = (2π/FoM)(l/lr), making loss in dB proportional to the maximum phase-shift range. Optimizing beamforming and RIS phases to minimize transmit power subject to an SNR constraint, the paper finds an interior optimum for off-specular users: the SNR benefit of a wider phase range is eventually outweighed by the insertion-loss penalty. For specular users the optimum is at ωmax = 0. The claimed conclusion is that there exists an LC phase-shifter length achieving a scenario-dependent optimal trade-off, with the optimal ωmax increasing as users move away from the specular directio","pith_inferences":["If the linear-loss model holds, the optimal phase range for a given user distribution could be derived from the angular spread of coverage, suggesting a design rule: choose LC length to cover the widest required steering angle plus margin, and no more.","Since a single LC-RIS length is shared across users in TDMA, deployments with mixed specular and off-specular users might benefit from segmented RIS designs where different regions of the surface have different phase-shift ranges.","A testable extension is to measure required transmit power versus ωmax on a fabricated LC-RIS and verify the interior optimum and its dependence on the figure of merit.","The trade-off likely persists with non-line-of-sight components, but multipath may reduce the need for large phase ranges because diffuse reflections provide incidental coverage, lowering the optimal ωmax."],"forward_implications":["Full 2π phase range is not universally optimal; in specular-rich deployments, a minimal-range LC-RIS with low loss can meet quality-of-service at lower transmit power.","Off-specular users impose a larger optimal ωmax, so user geometry should inform LC phase-shifter length selection.","Larger coverage areas raise both the required transmit power and the optimal ωmax, indicating that wide-coverage LC-RIS design needs longer, lossier shifters.","The optimal ωmax increases with figure of merit for off-specular users, so improvements in LC material quality relax the loss penalty.","In a TDMA system with a shared fixed-length RIS, the optimal length depends on the hardest-to-steer user, not on the average user."],"supporting_citations":[{"why":"Supplies the measured figure of merit (75°/dB) of the DGS-IMSL LC phase shifter used in the simulation setup.","marker":"[12]"},{"why":"Provides the defected delay-line LC-RIS architecture and the phase-range versus insertion-loss trade-off that motivates the paper.","marker":"[3]"},{"why":"Models performance degradation from phase-dependent insertion loss, the baseline the paper extends to LC-specific delay-line loss.","marker":"[6]"},{"why":"Supplies the phase-shift optimization approach (Lemma 2 and Lemma 3 and Algorithm 1) reused to solve the RIS configuration for fixed ωmax.","marker":"[14]"},{"why":"Introduces the energy-efficiency optimization framework for RIS-assisted networks that motivates the power-minimization formulation.","marker":"[4]"},{"why":"Provides the penalty method used to handle the rank-one constraint in the semidefinite relaxation P5.","marker":"[15]"}],"fun_headline_variants":["LC-RIS phase range vs loss: not always full","Optimal LC-RIS phase range depends on user angle","Liquid crystal RIS: loss limits phase range benefit","For specular users, LC-RIS wants zero phase shift","LC-RIS: more phase range, more loss—pick the sweet spot"],"cache_read_input_tokens":2816,"weakest_assumption_plain":"The central result rests on the assumption that insertion loss in dB is exactly proportional to LC phase-shifter length (constant figure of merit), with no fixed overhead loss or state-dependent variation; if that linearity fails, the optima can shift or disappear.","fun_headline_variants_meta":{"raw":{"variants":["LC-RIS phase range vs loss: not always full","Optimal LC-RIS phase range depends on user angle","Liquid crystal RIS: loss limits phase range benefit","For specular users, LC-RIS wants zero phase shift","LC-RIS: more phase range, more loss—pick the sweet spot"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000236,"raw_usage":{"total_tokens":1315,"prompt_tokens":695,"completion_tokens":620,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":439,"completion_tokens_details":{"reasoning_tokens":535}},"tokens_in":439,"tokens_out":620,"duration_ms":6660,"temperature":1.0,"reasoning_tokens":535,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T19:55:51.611187+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the insertion loss of a fabricated LC delay-line phase shifter at 28 GHz for several lengths and across the full tuning range. If the loss per element includes a fixed overhead (so dB loss versus length has a positive intercept) or if the figure of merit varies with tuning state, recompute the required transmit power versus ωmax; an interior optimum may vanish or move to a different ωmax.","supporting_citations":[{"cited_title":"Compact liquid crystal-based defective ground struc- ture phase shifter for reconfigurable intelligent surfaces,","cited_arxiv_id":null,"evidence_quote":"Supplies the measured figure of merit (75°/dB) of the DGS-IMSL LC phase shifter used in the simulation setup."},{"cited_title":"Phase dependent loss analysis for RIS systems,","cited_arxiv_id":null,"evidence_quote":"Models performance degradation from phase-dependent insertion loss, the baseline the paper extends to LC-specific delay-line loss."},{"cited_title":"Reconfigurable intelligent surfaces for energy effi- ciency in wireless communication,","cited_arxiv_id":null,"evidence_quote":"Introduces the energy-efficiency optimization framework for RIS-assisted networks that motivates the power-minimization formulation."},{"cited_title":"Power-efficient resource allocation for multiuser miso systems via intelligent reflecting surfaces,","cited_arxiv_id":null,"evidence_quote":"Provides the penalty method used to handle the rank-one constraint in the semidefinite relaxation P5."}],"review_version":1}