{"id":"b56e52cf-d989-4373-a253-78fcca2c289d","arxiv_id":"2412.03654","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"At special drive frequencies, periodically driven spin chains show symmetry restoration and the quantum Mpemba effect; driven CFTs on a strip show entanglement asymmetry growing as ln(mT) in the heating phase and as ln(ln mT) on the critical line.","lead":"Periodically driven quantum systems can restore a broken symmetry over time when the drive frequency hits special values, and this can make more-asymmetric states relax faster, the quantum Mpemba effect. For driven conformal field theories on a strip, the paper derives that the symmetry-breaking measure grows logarithmically in the heating phase and doubly logarithmically on the critical line.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"CFT scaling laws hinge on an unproven branch-independence assumption in Eq. 43; if the two sqrt branches do not converge, the ΔS~ln m and ln ln m claims lack support.","rationale":"The reader's weakest_assumption identifies exactly the branch-independence assumption after Eq. 43, and I agree this is the most load-bearing gap in the paper. The XY-chain results are supported by exact numerics, first-order Floquet perturbation theory, and a quasiparticle picture, and the special-frequency condition g1/(hbar omega_D) = p is cleanly derived from Eq. 23. The Rydberg-chain results are small-system ED and would benefit from larger sizes or additional checks, but they are not as central to the paper's advertised analytic novelty as the CFT scaling laws. The CFT section claims 'exact analytic computation' of Delta S_n, and Eq. 45 is the key new formula; yet its derivation depends on an unproven branch-independence assertion. The paper explicitly flags this as an assumption, so the concern is not manufactured. A direct numerical evaluation of the two branches for the square-pulse coefficients would settle whether the assumption holds in the regimes where the scaling laws are claimed. If it fails, the CFT contribution would need revision or qualification; if it holds, the conditional acceptance stands. Thus the reader's CONDITIONAL verdict is appropriate and no adjustment is needed.","tokens_in":26966,"tokens_out":10847,"duration_ms":105097,"concrete_test":"For the square-pulse protocol, evaluate the two branches of B_m = (2i / (sqrt(zeta_m) - sqrt(zeta*_m)))^{2 h} using the real-time coefficients in Eqs. 54-56, with zeta = exp(2 pi i ell / L) and h = c(n - 1/n)/24 + Delta(alpha)/n^2. Work in the heating phase (mu0 > 1) and on the critical line (mu0 = 1) for m = 10, 10^2, 10^3, 10^4. Compute the ratio |B_m^+ / B_m^-| and the phase difference arg B_m^+ - arg B_m^- as a function of alpha. If the modulus ratio does not tend to 1 as m grows, or if the phase difference remains alpha-dependent at large m, then Eq. 44 is not a faithful single-valued reduction and the asymptotics of Eq. 45 (and Eq. 65 on the critical line) are not justified. A secondary check: confirm that Eq. 44's beta_m reproduces the modulus of at least one branch of Eq. 43 for these parameters.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central CFT result, Eq. 45, is derived from Eq. 44, which in turn rests on the statement after Eq. 43 that the two branch choices of (sqrt(zeta_m) - sqrt(zeta*_m))^{-2 h_j} give identical results in the large m limit for the square-pulse protocol. The paper explicitly says 'We shall, in the rest of this work, assume that the latter property holds,' but no derivation is given. This is load-bearing: Eq. 44 is the single-valued expression for beta_m, and Eq. 45's logarithmic and log-log growth in the heating and critical phases follow from the m-dependence of ln beta_m. If the branch choice mattered, the exponent and even the functional form could change. The concern is not merely aesthetic: for the square-pulse protocol the coefficients a_m, b_m, c_m, d_m are complex (Eqs. 54-56), so zeta_m is complex and the two branches of sqrt(zeta_m) differ by a sign, multiplying the correlator by a phase e^{-2 pi i h_j}. Since h_j contains the alpha-dependent term Delta_j(alpha)/n^2, this phase can feed into the alpha-integrals in Eq. 33 and alter the Gaussian evaluation that produces Eq. 45. Thus the assumed branch independence is exactly the condition needed to make the advertised scaling laws rigorous.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the dynamics of the entanglement asymmetry ΔS_n in three periodically driven systems: an integrable XY spin chain, a non-integrable Rydberg (PXP) chain, and a conformal field theory (CFT) on a strip. For the spin chains, the authors identify special drive frequencies at which the leading-order Floquet Hamiltonian has an emergent approximate symmetry, leading to dynamical symmetry restoration (ΔS_2 → 0) and a quantum Mpemba effect. For the driven CFT, the paper derives the scaling laws ΔS_n ~ ln(mT) in the heating phase and ΔS_n ~ ln(ln mT) on the critical line, with small oscillations in the non-heating phase, and provides a phase diagram. The XY chain results are supported by exact numerics, first-order Floquet perturbation theory, and a quasiparticle picture; the Rydberg chain is treated by exact diagonalization for L=24; the CFT results are analytic up to an explicitly stated branch-independence assumption in Eq. (43).","tokens_in":27221,"tokens_out":10527,"duration_ms":91169,"significance":"If correct, the results give a systematic picture of how periodic driving controls entanglement asymmetry, and they identify a concrete mechanism (an emergent approximate symmetry of the Floquet Hamiltonian) that generates both symmetry restoration and the quantum Mpemba effect in driven integrable and non-integrable models. The CFT analysis is the first to provide an analytic prediction for entanglement asymmetry in driven CFTs, with distinct logarithmic and log-log growth laws in the heating and critical phases. The paper also contains a useful cross-check of three independent methods for the XY chain, which is a strength. However, the CFT scaling laws rest on an unproven branch-independence assumption, and the XY-chain Mpemba demonstration is not presented in the standard fixed-Hamiltonian form, so the significance is conditional on these points being resolved.","major_comments":[{"comment":"The branch-independence assumption is load-bearing but unproven. The text states, after Eq. (43), that for the square-pulse protocol both branch choices of (√ζ_m − √ζ*_m)^{−2h_j} give identical results in the large m limit, and then says: 'We shall, in the rest of this work, assume that the latter property holds.' No derivation or numerical check is provided. For the square-pulse protocol the coefficients a_m, b_m, c_m, d_m are complex (Eqs. 54–56), so ζ_m is complex and flipping the sign of √ζ_m multiplies the correlator by e^{−2πi h_j}. Since h_j contains the α-dependent term Δ_j(α)/n^2, this phase is not a constant factor across the α-integrals in Eq. (33) and could affect the Gaussian evaluation that produces Eq. (45). The scaling laws ΔS_n ~ ln m and ΔS_n ~ ln ln m follow from the m-dependence of β_m in Eq. (44), which in turn relies on this branch choice. The authors should either prove the branch independence for the square-pulse protocol or verify numerically that the two branches yield the same result for the integral in Eq. (33) in the large-m limit; without this, the central CFT results are not rigorously established.","section":"§4.2.1, after Eq. (43)"},{"comment":"The demonstration of the quantum Mpemba effect in the driven XY chain compares curves with different g0 and κ0 values (green vs. blue symbols). These differences change not only the initial state but also the Hamiltonian, since g0 appears in the drive protocol and in the Floquet Hamiltonian. The quantum Mpemba effect is normally defined for a fixed Hamiltonian with initial states of different symmetry breaking. The crossing in Fig. 1(a) could therefore reflect the different Floquet dynamics rather than the initial-state effect. The authors should either show the crossing for different initial states at fixed drive parameters (for example, by varying κ0 or using the initial-state parameters of Eq. 15 while keeping g0, g1, and ω_D fixed), or explicitly justify why the comparison in Fig. 1(a) isolates the Mpemba effect.","section":"§2, Fig. 1(a)"}],"minor_comments":[{"comment":"There are several typos: 'it's initial value' should be 'its initial value' (appears in the abstract and in §5); 'CFTS' should be 'CFTs'; 'Renyi' should be 'Rényi' in most places.","section":"§1, Abstract and body"},{"comment":"The definition of θ_k(T) uses arccos, which is multi-valued; the authors should specify the branch (e.g., principal value) to avoid ambiguity in the quasienergy dispersion.","section":"§2, Eq. (12)"},{"comment":"The Rydberg-chain results are limited to L=24 and subsystem size ℓ=4. The claim of symmetry restoration over 'exponentially large' prethermal timescales would be stronger if the authors showed data for a second system size or provided an estimate of the prethermal time from higher-order perturbation theory, even though the present evidence is plausible.","section":"§3, Fig. 3 and text"},{"comment":"The notation in Eq. (57) '2(L/(aπ)^2)' is ambiguous: it should be written as 2(L/(aπ))^2 or 2L^2/(a^2 π^2) to avoid confusion about the powers of a.","section":"§4.2.3, Eqs. (57)–(65)"},{"comment":"After Wick rotation to real time, the authors state that 'am = d*_m and bm = c*_m' but do not explicitly show that these relations hold for the square-pulse protocol; it would be helpful to state that this follows from unitarity of U(mT,0).","section":"§4.2.1"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the journal's scope and the topic is timely. The main technical gap is the unproven branch-independence assumption in the CFT section, which is central to the advertised scaling laws; this is fixable by a derivation or a direct numerical check. The XY-chain Mpemba claim also needs to be presented in the standard fixed-Hamiltonian form. The work is otherwise well-structured and contains strong cross-checks, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things you should know. First, this is the first paper to push entanglement asymmetry into the Floquet setting, and the spin-chain half of it is genuinely good. The driven XY chain shows dynamical symmetry restoration at special frequencies, and the paper nails the mechanism: at g1/(ℏω_D)=p the leading Floquet Hamiltonian develops an emergent approximate τ_z conservation, and they demonstrate this three independent ways — exact free-fermion numerics, first-order Floquet perturbation theory, and a quasiparticle counting formula. The agreement in Fig. 1 is convincing. The Rydberg chain ED is smaller (L=24, ℓ=4) but shows the same physics in a prethermal window, and that is plausible.\n\nSecond, the CFT part is elegant but more fragile than the abstract implies. The central results are ΔS ~ ln(mT) in the heating phase and ~ ln(ln mT) on the critical line. These follow from Eq. 45, which comes from Eq. 44, which comes from Eq. 43. And Eq. 43 explicitly depends on a branch choice for (√ζ_m − √ζ*_m)^{-2h_j}. The paper says: 'We shall, in the rest of this work, assume that the latter property holds' — i.e., that both branches give the same result at large m for the square-pulse protocol. That is not shown, and it is load-bearing. As the stress-test note observes, a sign flip in the square root introduces a phase e^{-2πi h_j}; since h_j carries the α-dependent term Δ_j(α)/n^2, that phase lands inside the α-integrals that produce the Gaussian evaluation. Without branch-independence, the functional forms are unsupported. The paper does not provide a numerical check of the CFT prediction either, so the log and log-log growths are unverified predictions rather than established results.\n\nAlso minor: 'generic CFT' overstates; the derivation assumes Δ(α) quadratic in α, which covers c=1 free bosons and SU(2)_k WZW, not all CFTs. No code or data is shipped; the free-fermion results are at least reproducible from the formulas, but the CFT appendix would benefit from the missing step.\n\nOverall: I would send this to referees. The XY chain results and the emergent-symmetry/Mpemba mechanism are worth publishing. But the CFT section needs either a proof of the branch claim or an explicit numerical test (e.g., a free-fermion lattice approximation of the driven strip), and the abstract should be narrowed accordingly. This is a conditional accept after revision, not a full accept as is.","headline":"First real study of entanglement asymmetry in Floquet systems; the spin-chain results are solid, but the CFT scaling laws rest on an explicitly assumed and unproven branch-independence that needs to be closed or tested.","tokens_in":27751,"tokens_out":2594,"would_cite":true,"duration_ms":25713,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.65.Ud","11.25.Hf"],"model":"deepseek-v4-flash","headline":"Periodically driven quantum systems can restore broken symmetry at special drive frequencies and show a quantum Mpemba effect; in driven conformal field theories the entanglement asymmetry grows logarithmically in the heating phase.","keywords":["entanglement asymmetry","periodically driven quantum systems","Floquet Hamiltonian","emergent symmetry","quantum Mpemba effect","conformal field theory","XY spin chain","Rydberg atom chain"],"falsifier":"A direct test is to compute $\\Delta S_n(mT)$ for a square-pulse-driven CFT on a strip using each branch of $\\sqrt{\\zeta_m}-\\sqrt{\\zeta_m^*}$ and check that their large-$m$ limits coincide up to subleading terms; a branch-dependent answer would invalidate Eq. (45). On the spin-chain side, at $g_1/(\\hbar\\omega_D)=p$ with large $g_1$, one can measure $\\Delta S_2(mT)$ on the prethermal plateau: if it does not decay to zero, or if the initially more asymmetric state does not relax faster than the less asymmetric one, the emergent-symmetry mechanism is refuted.","tokens_in":26717,"feed_emoji":"⚛️","tokens_out":10336,"duration_ms":89523,"temperature":0.7,"pith_summary":"This paper asks how the entanglement asymmetry $\\Delta S_n$ — a measure of how much a state breaks a symmetry, computed from the difference between projected and unprojected R\\'enyi entropies — evolves when a quantum system is driven periodically. The authors show that in driven spin chains, both integrable (XY) and constrained non-integrable (Rydberg/PXP), special drive frequencies make the first-order Floquet Hamiltonian conserve the symmetry charge, so the asymmetry decays to zero and does so faster for initially more asymmetric states: the quantum Mpemba effect. For a driven conformal field theory on a strip, they derive an exact expression and find that long-time behavior is set by the Floquet phase: $\\Delta S_n$ grows as $\\tfrac12\\ln(mT)$ in the heating phase, as $\\tfrac12\\ln(\\ln mT)$ on the critical line, and merely oscillates in the non-heating phase. The paper thus extends symmetry-restoration and Mpemba physics from quenches to periodic drives and gives analytically controlled predictions for a field-theory setting.","feed_headline":"Entanglement asymmetry vanishes at special drive frequencies","feed_subtitle":"The same resonances trigger a quantum Mpemba effect; in driven CFTs, asymmetry grows as ln(mT).","key_machinery":"The argument runs on two mechanisms. For the spin chains the load-bearing piece is first-order Floquet perturbation theory in the large-drive-amplitude regime: because the zeroth-order evolution over one cycle is the identity, the leading effective Hamiltonian is $H_F^{(1)}=\\frac1T\\int_0^T dt\\, U_0^\\dagger(t)(H(t)-H_0)U_0(t)$, and at $g_1/(\\hbar\\omega_D)=p$ (XY) or $\\Delta_1T/(2\\hbar)=p\\pi$ (Rydberg) this object conserves the relevant charge. The asymmetry is then extracted from the fermionic correlation matrix (XY) or from exact diagonalization (Rydberg), with the quasiparticle formula $\\Delta S_2(mT)=-\\int\\frac{d\\alpha}{2\\pi}\\exp[(A_2(\\alpha)+B_2(\\alpha,mT))\\ell]$ reproducing the decay at the resonance. For the CFT, the central object is the composite twist operator with vertex operators at the endpoints of the interval; the drive is encoded as an $SU(1,1)$ M\\\"obius map $\\zeta\\mapsto (a_m\\zeta+b_m)/(c_m\\zeta+d_m)$, and the key identity fixes $\\beta_m=(2L/\\pi)^2|X_m||Y_m|/(\\mathrm{Im}\\sqrt{X_mY_m^*})^2$, whose growth rate determines $\\Delta S_n\\simeq\\tfrac12\\ln(c_0\\ln\\beta_m/n^2)$ through Eq. (45).","core_discovery":"The central claim is that periodic driving can generate an approximate emergent symmetry in the Floquet Hamiltonian — a symmetry possessed by neither the undriven Hamiltonian nor the initial state — and that this symmetry dictates how entanglement asymmetry $\\Delta S_n$ evolves. For the square-pulse-driven XY chain, when $g_1/(\\hbar\\omega_D)=p$ with integer $p$, the leading Floquet term $H_F^{(1)}$ commutes with $\\tau_z$, so $\\Delta S_2(mT)\\to 0$ at late stroboscopic times; higher-order terms break this symmetry only at order $1/(\\hbar\\omega_D)$, so the restoration survives over long prethermal timescales, and initial states with more symmetry breaking relax faster, the quantum Mpemba effect. Exact diagonalization of the driven Rydberg (PXP) chain shows the same behavior at $\\Delta_1 T/(2\\hbar)=p\\pi$, again traced to $[H_F^{(1)},\\sigma^z_j]=0$. For a driven CFT on a strip, the paper derives an exact expression for $\\Delta S_n(mT)$ from a composite twist operator with vertex insertions at the interval endpoints; the stroboscopic drive acts as an $SU(1,1)$ M\\\"obius transformation, and the asymmetry reads $\\Delta S_n\\simeq \\tfrac12\\ln\\gamma_m$ with $\\gamma_m=c_0\\ln\\beta_m/n^2$. In the heating phase $\\beta_m$ grows exponentially in $mT$, giving $\\Delta S_n\\sim \\ln(mT)$; on the critical line $\\beta_m\\sim m^4$, giving $\\Delta S_n\\sim \\ln(\\ln mT)$; in the non-heating phase the coefficients remain oscillatory and $\\Delta S_n$ performs small-amplitude oscillations about its initial value.","pith_inferences":["Editorial inference: the logarithmic-versus-log-log crossover of $\\Delta S_n$ in the driven CFT gives an experimentally accessible marker of the critical line, since symmetry-resolved R\\'enyi entropies can be measured in cold-atom and trapped-ion simulators.","Editorial inference: the resonance condition $g_1/(\\hbar\\omega_D)=p$ is not special to the XY model; any integrable chain whose Floquet evolution has an $SU(2)$ structure should show the same emergent symmetry and Mpemba effect at integer drive ratios.","Editorial inference: because the emergent symmetry is broken by higher-order terms, the spin-chain predictions describe a prethermal transient; at sufficiently late times $\\Delta S_2$ should eventually rise again, and the rise time as a function of $1/\\omega_D$ is a testable prediction the paper leaves implicit.","Editorial inference: the non-heating phase's oscillating asymmetry suggests periodic driving can also suppress symmetry restoration, offering a Floquet-engineering knob to freeze or release symmetry breaking on demand."],"forward_implications":["At drive frequencies where the first-order Floquet Hamiltonian conserves the charge, driven XY and Rydberg chains show dynamical symmetry restoration, with $\\Delta S_2\\to 0$ over prethermal timescales.","At those same frequencies the relaxation is Mpemba-like: a state with a larger initial symmetry breaking decays to the symmetric state faster than one with smaller breaking.","Away from the special frequencies the spin-chain asymmetry does not vanish; for the XY chain it saturates to a finite value, and for the Rydberg chain it shows Rabi-frequency oscillations about a finite value.","For driven CFTs, the behavior of $\\Delta S_n$ distinguishes the Floquet phases: logarithmic growth in the heating phase, $\\ln\\ln mT$ growth on the critical line, and persistent small oscillations in the non-heating phase, so the asymmetry is a phase diagnostic.","The equilibrium strip result $\\Delta S_n\\simeq\\tfrac12\\ln\\ln(\\ell/a)$ is recovered in the undriven limit and agrees with the boundary-CFT scaling found for intervals on a semi-infinite line."],"supporting_citations":[{"why":"Defines entanglement asymmetry as the difference between projected and unprojected R\\'enyi entropies and supplies the correlation-matrix method used throughout.","marker":"[11]"},{"why":"Gives the XY-chain quench analysis and quasiparticle picture that the driven-chain computation generalizes, including the quantum Mpemba effect.","marker":"[12]"},{"why":"Provides the exact Floquet Hamiltonian for the square-pulse-driven XY chain in momentum space used for the correlation-matrix computation.","marker":"[58]"},{"why":"Derives the first-order Floquet perturbation theory in the large-amplitude regime that yields the emergent conserved $H_F^{(1)}$ for the XY chain.","marker":"[52]"},{"why":"Supplies the perturbative Floquet Hamiltonian and prethermal evolution method for the driven Rydberg chain.","marker":"[69]"},{"why":"Provides the higher-order Floquet corrections and prethermal timescale for the Rydberg chain used in the exact-diagonalization comparison.","marker":"[70]"},{"why":"Introduces the composite twist-operator formalism with vertex insertions for symmetry-resolved entanglement that underlies the CFT calculation.","marker":"[21]"},{"why":"Supplies the replica/OPE and strip-to-UHP mapping machinery used to evaluate the projected replica partition function on the strip.","marker":"[95]"},{"why":"Establishes the two-step conformal transformation and the heating/non-heating/critical phase classification of driven CFTs used in Section 4.","marker":"[71]"},{"why":"Gives the boundary-CFT result $\\Delta S_n\\sim\\tfrac12\\ln\\ln(\\ell/a)$ that the equilibrium strip computation matches.","marker":"[24]"}],"fun_headline_variants":["Periodic driving creates symmetry, erases entanglement asymmetry","Quantum Mpemba effect emerges from Floquet symmetry restoration","Entanglement asymmetry vanishes at resonant drive frequencies","Driven CFT: asymmetry grows as ln time in heating phase","Symmetry emerges under drive, triggering Mpemba effect"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The CFT derivation assumes that, for the square-pulse protocol, both branches of the multivalued factor $(\\sqrt{\\zeta_m}-\\sqrt{\\zeta_m^*})^{-2h_j}$ give the same answer in the large-$m$ limit; if the branch choice mattered, the single-valued expression for $\\beta_m$ and the resulting $\\ln(mT)$ and $\\ln(\\ln mT)$ laws would not be justified.","fun_headline_variants_meta":{"raw":{"variants":["Periodic driving creates symmetry, erases entanglement asymmetry","Quantum Mpemba effect emerges from Floquet symmetry restoration","Entanglement asymmetry vanishes at resonant drive frequencies","Driven CFT: asymmetry grows as ln time in heating phase","Symmetry emerges under drive, triggering Mpemba effect"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000672,"raw_usage":{"total_tokens":3202,"prompt_tokens":1228,"completion_tokens":1974,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":844,"completion_tokens_details":{"reasoning_tokens":1904}},"tokens_in":844,"tokens_out":1974,"duration_ms":13918,"temperature":1.0,"reasoning_tokens":1904,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T22:15:52.423595+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct test is to compute $\\Delta S_n(mT)$ for a square-pulse-driven CFT on a strip using each branch of $\\sqrt{\\zeta_m}-\\sqrt{\\zeta_m^*}$ and check that their large-$m$ limits coincide up to subleading terms; a branch-dependent answer would invalidate Eq. (45). On the spin-chain side, at $g_1/(\\hbar\\omega_D)=p$ with large $g_1$, one can measure $\\Delta S_2(mT)$ on the prethermal plateau: if it does not decay to zero, or if the initially more asymmetric state does not relax faster than the less asymmetric one, the emergent-symmetry mechanism is refuted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the exact Floquet Hamiltonian for the square-pulse-driven XY chain in momentum space used for the correlation-matrix computation."},{"cited_title":"Banerjee and K","cited_arxiv_id":null,"evidence_quote":"Derives the first-order Floquet perturbation theory in the large-amplitude regime that yields the emergent conserved $H_F^{(1)}$ for the XY chain."},{"cited_title":"Calabrese and J","cited_arxiv_id":null,"evidence_quote":"Supplies the replica/OPE and strip-to-UHP mapping machinery used to evaluate the projected replica partition function on the strip."}],"review_version":1}