{"id":"df0ec4bb-ff7a-4963-b8c0-ff3004840f7b","arxiv_id":"1908.06910","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Any small disorder turns a clean Weyl nodal loop semimetal into a multifractal semimetal, and at a critical disorder there is a transition to a diffusive metal with exponents nu=1.0 and z=1.9.","lead":"A little bit of disorder completely changes what a Weyl nodal loop semimetal looks like: the electrons' wave function becomes concentrated along the nodal line in a special fractal pattern. This numerical study maps the disorder-driven transitions and finds a new type of quantum critical behavior.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'arbitrary small disorder' claim depends on an unverified extrapolation: at W<1.5 the observed Γ~L^{-x} (1Γ that the same spurious exponents are reproduced. That demonstration shows the anomaly is compatible with a resolution artifact, but it does not establish that the thermodynamic limit at arbitrarily small W is the Γ~L^{-1} phase. The scaling collapse in Fig. 5(a) even includes W=1.0 and 1.25, inside the suspected resolution-limited region, and SM S4.1 admits that ν cannot be extracted from ξs because of these resolution problems. Since the 'arbitrary small disorder' part of the headline claim requires the true small-W fixed point to have Γ~L^{-1}, this is the load-bearing assumption. If instead a weak-disorder regime with a different exponent exists, the universal multifractal-semimetal phase would only set in above some finite W, and the statement 'for arbitrary small disorder' would fail even though the finite-disorder phase diagram might survive.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"Gonçalves et al. study a two-band lattice model of a Weyl nodal loop (WNL) semimetal with box-distributed on-site disorder using three complementary numerical methods: kernel polynomial method (KPM) for the density of states, Lanczos exact diagonalization (ED) for low-energy eigenstates, and transfer matrix method (TMM) for the high-disorder localization transition. They report a three-regime phase diagram: for small disorder a 'multifractal semimetal' with zero-energy DOS rho0 = 0 and a momentum-space ground-state wavefunction concentrated around the nodal loop with width Gamma ~ L^{-1} and q-dependent generalized dimension Dk(q) (Dk = 1 for q > 1, Dk = 3 for q < 1); for 2.6 +/- 0.1 < W < 11.0 +/- 0.2 a diffusive ('single-fractal') metal with rho0 > 0; and an Anderson insulator for larger W. The semimetal-to-metal transition is characterized by correlation-length exponent nu = 1.0 +/- 0.2 and dynamical exponent z = 1.9 +/- 0.1 and is claimed to belong to a universality class different from both the 3D Anderson transition and the disordered Weyl semimetal transition. The central physics claim is that an infinitesimal amount of disorder destabilizes the clean WNL, which flows to the multifractal semimetal fixed point.","tokens_in":23524,"tokens_out":23076,"duration_ms":215819,"significance":"If the central claim holds, the Letter reports a genuinely new zero-temperature phase of a disordered nodal-line semimetal and a new semimetal-metal universality class, which would be of substantial interest to the disordered-semimetal community: the clean WNL is argued to be unstable to infinitesimal disorder, in contrast to the finite-disorder threshold of isolated Weyl nodes. The finite-disorder part of the phase diagram is credibly established by mutually consistent analyses (KPM rho'(E) crossings and maxima, ED energy-window crossing, tau_k(q) finite-size scaling, the phi_beta construction, and TMM for the Anderson transition), and the authors are explicit about resolution limits (SM S7) and about quantities they cannot extract (nu from xi_s, SM S4.1). The two-region wavefunction ansatz is validated against direct scaling collapses of the raw |Psi_k|^2 data, and the reported exponents constitute falsifiable predictions.","major_comments":[{"comment":"The claim that an infinitesimal short-range disorder drives the clean WNL into the multifractal semimetal phase, which is the first sentence of the abstract and is repeated in the Discussion ('a clean WNL is unstable to an infinitesimal amount of disorder'), is not directly established by the numerics. For W < 1.5 the authors observe Gamma(W,L) ~ L^{-x} with 1 < x < 2 (e.g., x ≈ 1.4 at W = 0.5, Fig. S13(b)) and attribute this to the finite momentum-space resolution RL, demonstrating in SM S7 that sampling a Lorentzian of width Gamma ~ L^{-1} with RL > Gamma reproduces the spurious exponent. That demonstration shows the anomaly is compatible with a resolution artifact, but it does not exclude a genuine weak-disorder fixed point with an exponent different from unity. The point is load-bearing because the 'for arbitrary small disorder' part of the headline claim requires the true small-W fixed point to have Gamma ~ L^{-1}. The scaling collapse in Fig. 5(a) additionally includes W = 1.0 and 1.25, which lie inside the suspected resolution-limited region, and SM Sec. S4.1 states that nu cannot be extracted from xi_s for this same reason. I recommend either (i) a numerical test at larger L, which should be decisive because the resolution in the kz = 0 plane improves as L^{-2} while Gamma ~ L^{-1}, or (ii) a reformulation of the abstract and Discussion that presents the persistence down to W = 0 as an extrapolation supported by the marginal-relevance argument of Ref. [50] rather than as a directly demonstrated result.","section":"Abstract; Fig. 3(d); SM Sec. S7"},{"comment":"The printed definition of the central quantity Gamma is internally inconsistent. Combining the estimate N ≃ 2πGamma^2 P L^3/(2π)^3 with Ik ≡ Ik(q=2) ≃ 1/N gives Gamma = 2π/sqrt(Ik L^3 P), whereas Eq. (3) prints Gamma = 2π sqrt(Ik L^3 P). With the measured MF scaling Ik ~ L^{-1}, the printed equation would yield Gamma ~ L, contradicting the stated Gamma ~ L^{-1}. Footnote [59] states that the plots instead use Gamma = 1/(Ik L^3), which with Ik ~ L^{-1} scales as L^{-2} and is inconsistent with both Eq. (3) and the claimed L^{-1} behavior. Since Gamma enters the characteristic scales lambda_s and lambda_m and hence the phi_beta construction of W_c' (Sec. S3), the definition must be corrected and the quantity actually plotted in Fig. 3(c) stated unambiguously.","section":"Eq. (3) and footnote [59]"},{"comment":"The defining property of the MF-SM phase is a vanishing zero-energy DOS, rho0 = 0, but the manuscript itself records two limitations: KPM does not converge at E = 0 for small W (Fig. 2(a) and surrounding text, 'prevents a direct determination of rho0 for small W'), and the Discussion leaves open whether rare-region effects produce a finite contribution to rho0 in WNLs, citing Refs. [46,66,67]. The evidence for rho0 = 0 is therefore indirect (growth of rho'(E) with decreasing E, negative concavity of rho(E), and ED scaling of the energy window). I ask the authors to make the status of the rho0 = 0 statement precise: either provide a quantitative bound on rho0(W) within the box-disorder model at small W (for instance from ED with larger Nev or a rare-region argument for box disorder), or state explicitly in the abstract that rho0 = 0 is a numerically inferred statement within the attainable resolution rather than a directly measured zero. As written, the abstract overstates the certainty of this defining property.","section":"Discussion; Fig. 2(a); SM Sec. S2"}],"minor_comments":[{"comment":"The sentence 'As shown in Fig. 1(b) for a typical realization of disorder' appears to cite the phase-diagram schematic; the typical-realization wavefunction plots are in Fig. S10, and the momentum-space width is depicted in Fig. 1(a). Please correct the cross-reference.","section":"MF-SF transition paragraph"},{"comment":"The term 'multifractal' is used for a tau_k(q) that is piecewise linear with two slopes (Dk = 3 for q < 1 and Dk = 1 for q > 1), i.e., a bifractal spectrum. Please either justify the use of 'multifractal' for this piecewise-linear spectrum or state explicitly that the numerically accessible spectrum is bifractal.","section":"Fig. 3(a)"},{"comment":"The sentence stating that the critical exponents differ from those of a disordered Weyl semimetal is too strong: the quoted nu = 1.0 +/- 0.2 is fully consistent with the Weyl value nu ≈ 1, so the case for a distinct universality class rests on z = 1.9 +/- 0.1 versus z ≈ 1.5 and on the multifractal-to-single-fractal character of the transition, not on nu. Please rephrase accordingly.","section":"Discussion (universality class)"},{"comment":"The two KPM estimates (Wc = 2.61 +/- 0.01 from the maxima and Wc = 2.74 +/- 0.02 from the crossings) differ by more than their quoted statistical errors; the 'least squares' combination yielding 2.64 +/- 0.05 is not described, and the ED value 2.68 +/- 0.03 is quoted without stating how its error was obtained. Please specify the combination procedure or quote a systematic error covering the spread.","section":"Wc determination (Fig. 2(b))"},{"comment":"The two-region form |Psi_k|^2 ~ L^{-1} (k' < Gamma) and L^{-3} k'^{-2} (k' > Gamma) is a phenomenological parametrization inferred from the same data used to compute tau_k(q); the resulting tau_k(q) agrees with the scaling collapses in Fig. 3(d) and Fig. S11, but this is a consistency check rather than an independent derivation, and the text should say so explicitly.","section":"Fig. 3(d) and Eq. (2)"}],"recommendation":"major_revision","confidential_remarks":"To the editor: my assessment is that the finite-disorder phase diagram (W ≳ 1.5) is solid and the paper is unusually honest about its limitations (SM S7 and S4.1). The revision hinges on the small-W extrapolation: a decisive numerical test is feasible (larger L at fixed W < 1.5, since the kz = 0 resolution improves as L^{-2}), and the authors could alternatively qualify the abstract's 'arbitrary small disorder' claim. The Eq. (3)/footnote [59] definitional inconsistency should also be resolved. I see no grounds for rejection; the paper fits the scope of the journal, and the conditional verdict expressed in the prior reader's report is consistent with mine."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper. First, the central phase diagram—zero-energy DOS vanishing up to a finite disorder Wc≈2.6, then a diffusive metal, then an Anderson transition—is well supported by several independent numerical methods. Second, the stronger claim that any infinitesimal disorder drives the clean nodal loop semimetal into a multifractal semimetal is plausible but not directly shown; it rests on an extrapolation across a resolution-limited region.\n\nWhat is genuinely new: the multifractal semimetal phase and its transition to a single-fractal metal are not in the cited literature. Previous work only had perturbative hints that disorder is relevant. The authors use KPM for the DOS, exact diagonalization for energy-level scaling, wave-function scaling collapses, and TMM for the high-disorder Anderson transition. They get consistent estimates of Wc around 2.6 from several of these, and they honestly document where resolution limits bite. The two-region model of the momentum-space wave function (a Lorentzian-like core of width Γ plus a power-law tail) is simple and explains the q-dependent multifractal exponents.\n\nThe soft spot is the infinitesimal-disorder claim. For W<1.5 the authors observe Γ ~ L^{-x} with 11$ in that resolvable regime, or if the zero-energy density of states $\\rho_0$ is found to be nonzero at arbitrarily small disorder, the claimed multifractal semimetal for arbitrary small disorder is ruled out.","tokens_in":22950,"feed_emoji":"🌀","tokens_out":8619,"duration_ms":81313,"temperature":0.7,"pith_summary":"This paper is trying to establish that a clean Weyl nodal loop semimetal—a three-dimensional band structure whose valence and conduction bands touch along a one-dimensional loop—is unstable to arbitrarily small short-range disorder. In the disordered ground state, the density of states at zero energy still vanishes, but the momentum-space wavefunction concentrates around the nodal line with a width that shrinks as $1/L$, and its statistics are multifractal rather than plain ballistic or diffusive. The resulting multifractal semimetal persists up to a critical disorder $W_c = 2.6 \\pm 0.1$, where it becomes a compressible diffusive metal in a universality class characterized by $\\nu = 1.0 \\pm 0.2$ and $z = 1.9 \\pm 0.1$. A separate Anderson metal-insulator transition appears at larger disorder, $W_c^l \\approx 11$. The broader point is that the semimetallic phase in dirty samples is fundamentally different from the clean nodal semimetal.","feed_headline":"Tiny disorder makes Weyl nodal-loop semimetals multifractal","feed_subtitle":"Density of states still vanishes, but the wavefunction hugs the nodal line until a new transition at Wc ≈ 2.6.","key_machinery":"The central object is the generalized momentum-space inverse participation ratio, $I_k(q) = \\left(\\sum_{k,\\alpha}|\\Psi_{k,\\alpha}|^2\\right)^{-1}\\sum_{k,\\alpha}|\\Psi_{k,\\alpha}|^{2q} \\propto L^{-\\tau_k(q)}$, with $\\tau_k(q)=D_k(q)(q-1)$. Its q-dependence is the diagnostic: a constant $D_k$ means a single fractal, a q-dependent $D_k$ means multifractality. The wavefunction's width around the nodal loop is defined as $\\Gamma = 2\\pi\\sqrt{I_k L^3}/P$, with $I_k \\equiv I_k(2)$ and $P$ the loop perimeter; the phase distinction is carried by whether $\\Gamma \\sim L^{-1}$ (multifractal semimetal) or $\\Gamma \\sim L^0$ (metal). The critical-point analysis is carried by the scaling forms of the density of states, $\\rho(E) \\sim \\delta^{\\nu(d-z)} F_\\gamma(\\delta^{-\\nu z}|E|)$ and $\\rho(E) \\sim |E|^{d/z - 1}$ at $W = W_c$, whose collapse fixes $\\nu$ and $z$.","core_discovery":"The central claim is that a clean Weyl nodal loop semimetal flows to a strong-coupling fixed point under any infinitesimal short-range disorder: a phase the authors call the multifractal semimetal. In this phase the zero-energy density of states obeys $\\rho_0 = 0$, while the momentum-space ground-state wavefunction is concentrated along the nodal loop with width $\\Gamma \\sim L^{-1}$ and exhibits a q-dependent generalized dimension $D_k(q)$—equal to 3 for $q<1$ and 1 for $q>1$—which is the signature of multifractality. At the same critical disorder $W_c = 2.6 \\pm 0.1$ where $\\rho_0$ becomes nonzero, the wavefunction becomes single-fractal with $D_k(q)=3$, i.e. a standard three-dimensional diffusive metal. The transition is characterized by correlation-length exponent $\\nu = 1.0 \\pm 0.2$ and dynamical exponent $z = 1.9 \\pm 0.1$, obtained from finite-size scaling of the width $\\Gamma$ and from scaling collapse of the density of states; these differ from the 3D Anderson transition and from the disordered Weyl-semimetal transition. At considerably higher disorder, $W_c^l = 11.0 \\pm 0.2$, the metal localizes through an Anderson transition.","pith_inferences":["A testable extension follows from the sharp momentum-space structure: in cold-atom or photonic realizations of nodal-loop Hamiltonians, momentum-resolved images of the ground state should show a bright ring around the loop whose thickness shrinks as $1/L$ for weak disorder and saturates for strong disorder.","If the small-$W$ exponent $x>1$ is not a resolution artifact, the phase diagram would acquire an additional crossover or a distinct low-disorder fixed point; checking this requires studying $\\Gamma$ at system sizes where the $k_z$-plane resolution is finer than $\\Gamma$.","The multifractal semimetal may have anomalous dynamical signatures, such as non-diffusive spreading of wavepackets or a non-standard conductivity exponent, because the momentum-space wavefunction is quasi-one-dimensional around the loop even though the system is three-dimensional in real space."],"forward_implications":["If the central claim is right, clean nodal-loop semimetals are not the appropriate zero-disorder starting point for describing real dirty samples: even tiny disorder produces the multifractal semimetal, so transport and spectral properties computed from clean states need revision.","The semimetal-to-metal transition at $W_c \\approx 2.6$ belongs to a new universality class, so scaling predictions based on the 3D Anderson transition or on the disordered Weyl-semimetal transition do not apply near that critical point.","The finite $\\rho_0$ in the metallic phase and its vanishing in the multifractal semimetal imply that the compressibility and low-frequency response change sharply at $W_c$, measurable in principle by transport and thermodynamic probes.","At $W_c^l \\approx 11$, the system undergoes a conventional Anderson metal-insulator transition, so any experimental signature of localization at strong disorder should be governed by the orthogonal-class Anderson universality class."],"supporting_citations":[{"why":"Supplies the kernel polynomial method (KPM) used to compute the density of states $\\rho(E)$ for system sizes up to $L=10^3$.","marker":"[52, 53]"},{"why":"Introduces the generalized momentum-space inverse participation ratio $I_k(q)$ from which the multifractal exponents $\\tau_k(q)$ and wavefunction width $\\Gamma$ are extracted.","marker":"[54, 55]"},{"why":"Provides the scaling form of the density of states near a semimetal-metal transition used to extract $\\nu$ and $z$ and to perform the data collapse.","marker":"[60]"},{"why":"Assesses rare-region effects for disordered nodal-point semimetals; the paper relies on [47] to argue such effects do not destroy the finite-disorder semimetallic phase, supporting $\\rho_0=0$ in the multifractal semimetal.","marker":"[46, 47]"},{"why":"Gives the critical exponents of the 3D Anderson transition for all symmetry classes, used to show that the new transition belongs to a different universality class.","marker":"[61–64]"},{"why":"Transfer-matrix method used to locate the higher-disorder Anderson metal-insulator transition at $W_c^l \\approx 11$.","marker":"[69–71]"}],"fun_headline_variants":["Infinitesimal disorder makes Weyl loops multifractal","Weyl loops go multifractal at any disorder","Disorder: Weyl nodal loops become multifractal","Multifractal semimetal from infinitesimal disorder","New phase: Weyl loops multifractal at any disorder"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the apparent scaling $\\Gamma \\sim L^{-x}$ with $1 < x < 2$ seen for $W \\lesssim 1.5$ is only a finite-resolution artifact and would become $\\Gamma \\sim L^{-1}$ at larger system sizes; if that extrapolation is wrong, the 'arbitrary small disorder' part of the central claim collapses.","fun_headline_variants_meta":{"raw":{"variants":["Infinitesimal disorder makes Weyl loops multifractal","Weyl loops go multifractal at any disorder","Disorder: Weyl nodal loops become multifractal","Multifractal semimetal from infinitesimal disorder","New phase: Weyl loops multifractal at any disorder"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000448,"raw_usage":{"total_tokens":2268,"prompt_tokens":961,"completion_tokens":1307,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":577,"completion_tokens_details":{"reasoning_tokens":1224}},"tokens_in":577,"tokens_out":1307,"duration_ms":10160,"temperature":1.0,"reasoning_tokens":1224,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:32:00.048720+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Examine the ground-state wavefunction width $\\Gamma$ for $W < 1.5$ at system sizes large enough that the average spacing of momentum points near the loop is well below $\\Gamma$ (in particular resolving the $k_z$ direction). If the scaling exponent of $\\Gamma$ versus $L$ remains $>1$ in that resolvable regime, or if the zero-energy density of states $\\rho_0$ is found to be nonzero at arbitrarily small disorder, the claimed multifractal semimetal for arbitrary small disorder is ruled out.","supporting_citations":[{"cited_title":"Kobayashi, T","cited_arxiv_id":null,"evidence_quote":"Provides the scaling form of the density of states near a semimetal-metal transition used to extract $\\nu$ and $z$ and to perform the data collapse."}],"review_version":1}