{"id":"e5e0a58d-f2b9-43f0-adf5-287ae28df8f7","arxiv_id":"1908.00793","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Differential loss in two pillars coupled only through a shared elastic plate produces an exceptional point, verified by finite-element simulation and experiment.","lead":"Two aluminum pillars carved on a plate act as coupled oscillators whose only connection is through the plate itself, and adding extra damping to one pillar makes their vibration frequencies merge at a special degeneracy called an exceptional point. This mechanical demonstration points to new vibration-based sensors that are extremely sensitive to small changes in their environment.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The square-root EP signature is not independently established: the eigenfrequencies in Fig. 4 are obtained by fitting Eq. 6, a model that already contains the EP, and the reported best-fit exponent is 0.6, not 0.5, without error bars.","rationale":"The paper's central claim is plausible: the COMSOL simulations and the three-level model provide independent support for environment-mediated coupling and for an EP in the weak-coupling regime, and the experimental system appears carefully constructed with differential loss applied via putty. However, the experimental confirmation is not independent of the theoretical model. The fit function Eq. 6 is derived from the same coupled-oscillator picture whose poles contain an EP, so the observation of a near-EP in the fitted spectra is partly circular. The reported exponent α≈0.6 in Fig. 4(b) deviates from the claimed 0.5, and the paper provides no error bars or raw spectral fits, so the quantitative claim is not secured. The reader's weakest assumption concerned the weak-coupling limit, and the Appendix's own admission after Eq. 8 supports that concern; my focus is on the model-dependence of the eigenvalue extraction, which is related but not identical. The proposed model-independent pole-extraction test would settle whether the square-root singularity is in the data or only in the fitting function. Since the physics may well be correct but the experimental evidence is incomplete, the reader's CONDITIONAL verdict remains appropriate; no change in verdict is needed.","tokens_in":10806,"tokens_out":8501,"duration_ms":87904,"concrete_test":"Re-analyze the raw S21 transmission spectra with a model-independent pole extraction (e.g., vector fitting with a sum of complex Lorentzians plus background) that does not impose the coupled-oscillator form of Eq. 6. Identify the two complex poles near 16.6 kHz for each applied loss Γ, propagate uncertainties, and test whether the poles actually coalesce (equal frequencies and equal residues) at a common Γ and whether the complex splitting Δf follows |Γ−ΓEP|^{1/2} with a slope consistent with 0.5 within error bars. If two independent poles merely cross in real part without residue coalescence, or if the fitted exponent is inconsistent with 1/2, the experimental EP evidence is an artifact of the assumed fitting model.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that the measured transmission data reveal an actual branch-point singularity in the physical response. The paper extracts the eigenfrequencies of Fig. 4(a) by fitting each spectrum to Eq. 6, t(u)=t0+γd A e^{iφ} / [(1-u^2+κ+iuγd)−κ^2/(1-u^2+κ+ϵ+iuΓ)]. This is exactly the two-oscillator coupled-mode model whose poles can coalesce at an EP. The extracted 'eigenfrequencies' are therefore the poles of an assumed model, not directly measured resonances, so observing a near-EP in the plotted dispersion is partly an output of the fitting function rather than an independent test. The paper itself states that after the EP the data cannot significantly resolve the linewidth of the broader resonances, leaving the broken-phase side unconstrained. In addition, the best linear fit to the splitting in Fig. 4(b) is reported as α≈0.6, while the claimed square-root law has α=0.5, and no parameter uncertainties are given. The Appendix also concedes (discussion after Eq. 8) that with unequal positive losses a strict EP survives only in the weak-coupling/linearized limit, a regime the experiment assumes but does not independently verify. The most load-bearing gap is thus model-dependent confirmation, not the physical plausibility of environment-mediated EPs.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a non-Hermitian elastodynamic platform in which two identical torsion pillars on an aluminum plate are coupled only through plate (environmental) modes. The authors use a coupled-mode model with one environment mode, COMSOL eigenfrequency simulations with complex shear moduli, and transmission measurements with variable putty-induced differential loss to argue that the system exhibits an exceptional point. The central experimental claim is that the measured frequency splitting scales as Δf ∼ sqrt(1 − Γ/Γ_EP), confirming an EP in the weak-coupling regime. The paper also proposes applications in atomic force microscopy and surface-integrity sensing.","tokens_in":11157,"tokens_out":5844,"duration_ms":63361,"significance":"If the central claim is correct, this is a useful extension of EP physics to mechanical/elastodynamic systems and supports a distinct route to EP control via environmental modes. The strengths are the physically credible geometry, the independent COMSOL calculations, explicit weak-coupling analytic results in the Appendix, and the careful calibration of putty-induced damping. The experimental verification is weakened by the model-dependent extraction of eigenfrequencies and by an exponent mismatch, so the current manuscript does not yet provide a quantitatively decisive confirmation of the square-root singularity.","major_comments":[{"comment":"The experimental eigenfrequencies in Fig. 4(a) are not directly measured quantities; they are the poles of Eq. (6), the transmission function of the same weak-coupling coupled-oscillator model that contains the exceptional point. Fitting each spectrum to this function therefore builds the EP structure into the extracted eigenvalue trajectories, and the good agreement with the COMSOL simulation does not remove the circularity for the experimental confirmation. I recommend an independent check: report the raw transmission data at selected Γ values together with the COMSOL prediction using the measured Γ and no free resonance parameters, or extract resonance positions and linewidths directly from the spectra and compare them with the CMT/COMSOL curves.","section":"Section IV, Fig. 4(a), Eq. (6)"},{"comment":"The claimed square-root law is not actually demonstrated by the fit shown. The text states Δf ∼ sqrt(1 − Γ/Γ_EP), but the best linear fit to the log-log plot is reported as (1 − Γ/Γ_EP)^α with α ≈ 0.6, and no error bars, residuals, or confidence intervals are given for α, Γ, or Δf. Given that α = 0.6 differs from 0.5 by 20%, the data are at best consistent with a near-square-root singularity, not a confirmation of the square-root law. The authors should (i) report the uncertainty in α, (ii) fit the data with a fixed exponent 0.5 and show the residuals, and (iii) mark the broken-phase points that are not constrained because the linewidths are unresolved, as stated in Section IV.","section":"Section IV, Fig. 4(b)"},{"comment":"The Appendix explicitly states that a strict exceptional point exists only when γ0 = 0, and that in the experimental situation with unequal positive losses the EP is restored only in the weak-coupling limit where the frequency dependence of the loss terms is linearized. The experiment's validity therefore depends on the smallness of the neglected terms in Eq. (8), but the paper does not quantify this. I ask for a quantitative check: using the fitted values of κ, γd, and Γ, evaluate the exact eigenvalues of Eq. (8) and compare their coalescence behavior with the CMT prediction of Eq. (5); or estimate the relative magnitude of the neglected γΔ and κ^2 terms. Without this check, one cannot distinguish a true branch-point singularity from an avoided crossing rounded by the neglected terms.","section":"Appendix, after Eq. (8)"}],"minor_comments":[{"comment":"There are typographical errors including 'EXEPTIONAL' in the Section IV heading, 'epansion' near Eq. (2), and 'differential' in the abstract; these should be corrected.","section":"Throughout"},{"comment":"The value of Poisson's ratio used in the COMSOL simulations is said to have been fixed earlier but is never reported; please state it explicitly in Section IV or in the Methods.","section":"Section IV"},{"comment":"Fig. 1(c)–(e) are difficult to read because the panels share axes and the labels overlap; the caption also does not clearly identify which plotted quantities correspond to the real and imaginary eigenfrequency axes in each panel. Please redraw or annotate.","section":"Fig. 1"},{"comment":"The fitting parameters in Eq. (6) (t0, A, φ, ε, γd, κ, Γ) are numerous; at minimum the uncertainties in κ and Γ, which enter the EP position and the exponent α, should be reported.","section":"Eq. (6)"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the journal's applied-physics scope, and the COMSOL evidence is strong. The main risk is the self-consistency of the experimental EP extraction and the unquantified exponent mismatch; I believe these are addressable with additional analysis and data presentation, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nRead the Domínguez-Rocha et al. paper. What is actually new: an EP in a pair of elastic pillars coupled only through the base plate, with differential loss rather than gain, and the claim that a single nearby environmental mode controls the EP. That is a legitimate extension of non-Hermitian physics to elastodynamics, and the three-level CMT model with environmental coupling is a reasonable framework.\n\nThe COMSOL study is the strongest piece. It uses a real geometry, varies the indent to steer an environmental mode, imposes balanced gain/loss in the shear modulus, and shows eigenfrequency and eigenvector coalescence. That is independent evidence for an EP. The experimental design is also thoughtful: putty as controllable loss, clamp calibration of Γ, and Eq. (6) fit with an explicit detuning check ϵ for mass loading. The putty viscoelastic modeling is a nice side contribution.\n\nWhere it gets soft: the experimental eigenfrequencies in Fig. 4(a) are not measured directly. They are extracted by fitting each transmission spectrum to Eq. (6), which is already the coupled-oscillator model whose poles carry the EP. So observing near-EP behavior in the extracted data is partly circular. The log-log plot in Fig. 4(b) bolsters the claim a bit, but the best fit gives α≈0.6 rather than the expected 0.5, with no error bars, and the paper concedes that after the EP the broader resonance linewidth cannot be resolved, leaving the broken-phase side essentially unconstrained. The appendix also admits that strict EPs with unequal positive losses exist only in the linearized weak-coupling limit; the experiment assumes that limit (κ≈45 Hz at ω0≈16.3 kHz, which is about 3×10^-3, so the assumption is plausible but not independently validated). The three-level environmental-mode model is qualitative, as the authors say.\n\nAll that said, I don't think the central physics is wrong. The COMSOL result, the weak-coupling analysis, and the rough agreement in Fig. 4(a) together make the existence of an environment-mediated EP likely. What is missing is a measurement that doesn't rely on the same model to extract the very quantity being tested, plus honest uncertainties and a fit to the full spectrum rather than a few eigenfrequency points.\n\nThis paper deserves a serious referee. The right outcome is a major-revision request: reanalyze the experimental data with a model-independent resonance extraction (e.g., phase-rotation or lineshape fitting on raw spectra), report uncertainties, and present the broken-phase linewidths or explain why they cannot be measured. I would not cite it as a definitive experimental demonstration yet, but I would file it as a useful platform paper.","headline":"Environment-mediated EPs in elastodynamics is a plausible and genuinely new platform, but the experimental square-root confirmation is partly model-dependent; the COMSOL simulation carries the load.","tokens_in":11682,"tokens_out":2432,"would_cite":false,"duration_ms":23908,"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":"Two identical aluminum pillars, coupled only through their base plate, reach an exceptional point when one is subjected to purely dissipative differential loss, with the eigenfrequency splitting following the square-root law near the…","keywords":["exceptional points","elastodynamics","non-Hermitian physics","PT symmetry","coupled-mode theory","mechanical resonators","differential loss","environmental mode coupling"],"falsifier":"Measure both the real and imaginary parts of the two torsional eigenfrequencies with fine steps of the loss parameter through the would-be exceptional point: the claim requires the two real parts to become exactly equal and the two imaginary parts to become exactly equal at a single value of $\\Gamma$, with the splitting scaling as $|\\Gamma-\\Gamma_\\mathrm{EP}|^{1/2}$ in the neighborhood. If instead the real parts show an avoided crossing or the imaginary parts retain a finite difference at the closest approach, the singularity is not present. This test is achievable with the paper's own transmission-fitting method, which currently loses linewidth resolution just beyond the exceptional point.","tokens_in":10590,"feed_emoji":"⚙️","tokens_out":10267,"duration_ms":94197,"temperature":0.7,"pith_summary":"This paper demonstrates an exceptional point in a purely mechanical system: two identical aluminum pillars carved from a single block and coupled only through their common base plate. By applying a specially prepared putty layer to one pillar, the authors add damping without adding mass, creating a differential loss that tunes the two torsional resonances toward coalescence. Near the critical loss the eigenfrequency splitting follows the square-root law $\\Delta f \\sim \\sqrt{1-\\Gamma/\\Gamma_\\mathrm{EP}}$, the definitive signature of an exceptional point. The result matters because it transfers exceptional-point physics from photonics to elastodynamics, where the coupling itself is mediated by 'environmental' plate modes, opening the door to new mechanical sensors for surface integrity and a differential atomic force microscope.","feed_headline":"Two aluminum pillars reach an exceptional point when one is made lossy","feed_subtitle":"Frequency splitting between coupled torsional modes obeys the square-root law of a non-Hermitian degeneracy.","key_machinery":"The central object is the two-mode coupled-mode theory Hamiltonian in the weak-coupling limit (Eq. 5 of the paper), which maps the unequal-loss mechanical dimer onto an effective $2\\times 2$ matrix with the same square-root singularity structure as a parity-time-symmetric dimer. The pivotal linearization is $1-u^2 \\approx -2\\Delta$ with $u=1+\\Delta$, and the Appendix derives the exceptional point position $\\Gamma_\\mathrm{EP}=2(\\sqrt{1+2\\kappa}-1)+\\gamma_d$ and the splitting $\\Delta\\omega/\\omega_0 = \\tfrac{2}{3}\\kappa^{3/2}$. A three-level coupled-mode model (Eq. 4), adding one explicit environmental level with detuning $\\nu$ and couplings $\\lambda_1,\\lambda_2$, shows how a nearby plate mode qualitatively modifies the two-level exceptional-point behavior. On the experimental side, a viscoelastic boundary-layer model of the putty layer gives the key constitutive relation that damping can be added without mass loading, which is what keeps the loss contrast tunable while preserving the resonator frequencies.","core_discovery":"Using a pair of aluminum torsion pillars sharing a base plate, the paper shows that an exceptional point — a non-Hermitian degeneracy where two eigenfrequencies and their eigenvectors coalesce — can be induced by purely dissipative differential losses. The inter-pillar coupling is not a direct spring but is mediated by the elastic modes of the base plate, and the loss is introduced as a boundary layer of putty that adds damping with negligible mass loading. Experiment and COMSOL simulation both show that, as the imposed loss $\\Gamma$ increases, the real parts of the two torsional eigenfrequencies merge and the splitting obeys $\\Delta f \\sim \\sqrt{1-\\Gamma/\\Gamma_\\mathrm{EP}}$ near the exceptional point, with the imaginary parts becoming equal in the exact phase. A coupled-mode theory in the weak-coupling limit and a three-level version including one explicit environmental mode reproduce the observed behavior, and the authors note that a strict exceptional point for unequal positive losses exists only in this weak-coupling approximation, which their device enforces with a small coupling splitting ($\\kappa \\approx 45$ Hz at $\\omega_0 \\approx 2\\pi\\times 16.29$ kHz).","pith_inferences":["A testable extension the authors do not pursue: sweeping both loss terms (or loss and coupling) along a closed loop around the EP should produce a chiral mode swap, turning this pair of pillars into a mechanical analog of topological mode transfer.","An experiment the paper does not run: place a phononic band-gap lattice in the base plate so that no environmental mode lies near the pillar frequency; the model predicts the EP would then disappear or move, which would confirm the causal role of the environmental mode.","A scaling question left open: the putty boundary-layer mechanism is specific to the 16 kHz scale, and extending the same idea to MEMS resonators would require a material whose boundary-layer dynamics remains loss-dominated without mass loading at much higher frequencies.","A practical consequence the authors do not state: if the square-root law holds, the same measurement curve of $\\Delta f$ versus $\\Gamma$ contains both the coupling strength $\\kappa$ and the intrinsic damping $\\gamma_d$, so the EP calibration could be used as a non-invasive way to extract these mechanical parameters from a single transmission measurement."],"forward_implications":["Near the exceptional point, a small change in loss contrast $\\Gamma$ produces a large, square-root response in the eigenfrequency splitting, so the structure acts as a sensitive transducer for dissipation and contact.","Because the coupling is mediated by plate modes, modifying the plate (for example, adding an indent) changes the exceptional-point characteristics without touching the pillars, giving a new control knob labeled 'environmental mode control.'","In the weak-coupling limit the unequal-loss mechanical dimer is mathematically equivalent to a PT-symmetric dimer, so established PT-symmetric sensing and transport designs can, in principle, be ported to elastodynamic structures.","The three-level model indicates that a single nearby environmental mode can qualitatively shift or distort the exceptional point, meaning the base-plate mode spectrum can be used to engineer EP behavior in mechanical devices."],"supporting_citations":[{"why":"A review of non-Hermitian physics and parity-time symmetry that situates the exceptional-point concepts used here.","marker":"[2]"},{"why":"The finite-element simulation package (COMSOL Multiphysics) used for all numerical eigenfrequency calculations of the pillar/plate structure.","marker":"[16]"},{"why":"The perturbation-theory monograph that defines eigenvalue/eigenvector coalescence, the foundational notion of an exceptional point.","marker":"[22]"},{"why":"A standard physics review of exceptional points and their square-root topology, establishing the framework the paper follows.","marker":"[23]"},{"why":"The paper that introduced parity-time-symmetric Hamiltonians, whose exact/broken phases the mechanical dimer reproduces in the weak-coupling limit.","marker":"[24]"},{"why":"The appendix of a standard acoustics reference used to model the viscoelastic putty boundary layer, establishing that damping is added without mass loading.","marker":"[29]"}],"fun_headline_variants":["Loss-induced exceptional point in aluminum pillars","Damping differential merges torsional modes at exceptional point","Exceptional point from base-plate coupling and putty loss","Square-root law verifies exceptional point in elastic pillars"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire demonstration relies on the pillar-pillar coupling being weak enough that the unequal-loss system can be linearized into a strict exceptional point; if the coupling is not kept small (here about 45 Hz compared with 16.29 kHz), no true coalescence of eigenvalues occurs and the square-root splitting law is lost.","fun_headline_variants_meta":{"raw":{"variants":["Loss-induced exceptional point in aluminum pillars","Damping differential merges torsional modes at exceptional point","Exceptional point from base-plate coupling and putty loss","Square-root law verifies exceptional point in elastic pillars"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000562,"raw_usage":{"total_tokens":2654,"prompt_tokens":917,"completion_tokens":1737,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":533,"completion_tokens_details":{"reasoning_tokens":1674}},"tokens_in":533,"tokens_out":1737,"duration_ms":13420,"temperature":1.0,"reasoning_tokens":1674,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:32:27.081056+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure both the real and imaginary parts of the two torsional eigenfrequencies with fine steps of the loss parameter through the would-be exceptional point: the claim requires the two real parts to become exactly equal and the two imaginary parts to become exactly equal at a single value of $\\Gamma$, with the splitting scaling as $|\\Gamma-\\Gamma_\\mathrm{EP}|^{1/2}$ in the neighborhood. If instead the real parts show an avoided crossing or the imaginary parts retain a finite difference at the closest approach, the singularity is not present. This test is achievable with the paper's own transmission-fitting method, which currently loses linewidth resolution just beyond the exceptional point.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"A review of non-Hermitian physics and parity-time symmetry that situates the exceptional-point concepts used here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The finite-element simulation package (COMSOL Multiphysics) used for all numerical eigenfrequency calculations of the pillar/plate structure."},{"cited_title":"Kato, Perturbation theory for linear operators (Springer-Verlag, Berlin, 1966), p.p","cited_arxiv_id":null,"evidence_quote":"The perturbation-theory monograph that defines eigenvalue/eigenvector coalescence, the foundational notion of an exceptional point."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"A standard physics review of exceptional points and their square-root topology, establishing the framework the paper follows."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The paper that introduced parity-time-symmetric Hamiltonians, whose exact/broken phases the mechanical dimer reproduces in the weak-coupling limit."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The appendix of a standard acoustics reference used to model the viscoelastic putty boundary layer, establishing that damping is added without mass loading."}],"review_version":1}