{"id":"3f57c402-b04f-4d82-bd11-93d5449ee44b","arxiv_id":"2502.04866","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Spinless fermions in a tilted 1D lattice show an edge whose width grows as (2t+V)/F and, for strong repulsion, hosts a charge density wave.","lead":"This paper studies spinless fermions in a one-dimensional chain with a linear electric-field potential and nearest-neighbor repulsion. It finds that the boundary region between filled and empty sites has a width that grows with the interaction and shrinks with the field, and that strong repulsion creates a charge-density-wave pattern in that region.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The edge-width definition may exclude the saturated CDW core, so the claimed V-linear scaling is unverified for strong interactions.","rationale":"The reader's weakest assumption targeted Eq. (37), treating the band-edge energies -2t-V and 2t+V as heuristic. Those endpoints are actually exact limits of the chemical potential for this Hamiltonian, so the band-edge formula is not the soft spot. The genuine concern is operational: the threshold-based width xi_2 (Eqs. (21)-(22)) stops counting sites once the CDW order parameter makes occupations exceed 0.98 or drop below 0.02. The paper's Fig. 10 shows F(xi_2-1) vs V only up to V=5, which is likely below saturation for the fields used, so the strong-V behavior is untested. The noninteracting analytics are exact and well validated, and the DMRG data in the tested regime support the linear scaling, so the paper should not be rejected. However, the central claim as stated in the abstract and conclusions is broader than what is verified: the full edge width in the saturated CDW regime has not been shown to scale as (2t+V)/F. The proposed test would settle whether the claim needs qualification. This does not change the existing CONDITIONAL verdict, hence UNCHANGED.","tokens_in":16435,"tokens_out":21624,"duration_ms":223237,"concrete_test":"Using DMRG with the same parameters as Fig. 7 (L=150, F=0.13), compute xi_2 for V=6, 8, 10, and 15, and compare with a threshold-free edge width W_edge = j_last(n_j>10^-4) - j_first(n_j<1-10^-4) and with LCDW obtained from the charge-charge correlator. If F(xi_2-1) deviates from 4+2V while W_edge or LCDW continues to grow linearly with V, the central claim holds only for weakly saturated CDW and must be reformulated. If both xi_2 and the alternative widths follow Eq. (37), the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central interacting claim, stated in the abstract, Eq. (37), Fig. 10, and the Conclusions, is that the width of the intermediate edge region scales as 2(2t+V)/F. The energy endpoints in Eq. (37) are actually exact single-particle/hole addition energies: adding one particle to the empty system costs -2t-V and removing one from the filled system costs +2t+V for the (n-1/2)(n-1/2) interaction, so the chemical-potential range over which the local filling goes from 0 to 1 is exactly 4t+2V. The load-bearing assumption is therefore not the band-edge formula itself but the identification of the numerically measured width with this range. The DMRG width xi_2 (Eqs. (21)-(22)) counts sites with 0.02<n_j<0.98. Inside the CDW plateau for V>2t, occupations approach 1 on one sublattice and 0 on the other; when the CDW order parameter saturates, these sites are excluded from xi_2. Thus for large V, xi_2 measures only the two domain walls, not the full edge including the CDW core, and the linear scaling F(xi_2-1) vs V in Fig. 10 (shown only up to V=5) may saturate even though the CDW core continues to grow as V/F. The separate CDW-size measure LCDW (Fig. 9) is only plotted vs 1/F at V=9, not vs V, so the claimed linearity in V is not established in the strongly saturated regime.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript studies spinless fermions on a finite one-dimensional chain with nearest-neighbor tunneling t, a linear potential F(j-jc), and nearest-neighbor repulsion V. In the noninteracting case, the authors derive exact Bessel-function expressions for the Wannier-Stark eigenstates, the ground-state site occupations, several localization-length measures, and the local density of states, and verify these against DMRG. For V>0, DMRG density profiles show a crossover from a smooth occupation edge to an edge containing a half-filled charge density wave for V>2t. The central quantitative claim is that the width of the intermediate edge region scales as 2(2t+V)/F, with numerical support from data collapses using two different width measures, xi2 and xi4.","tokens_in":16727,"tokens_out":13527,"duration_ms":131646,"significance":"The noninteracting part is a valuable exact benchmark: the Bessel-function representation, the various localization-length definitions, and the LDOS are worked out in detail and match DMRG. The interacting part provides a clear numerical scaling law for the edge width in a tilted fermionic chain and identifies a CDW-ordered central region, which is experimentally relevant for cold-atom implementations. The paper is reasonably transparent about the heuristic nature of Eq. (37), but the presentation does not always separate exact statements from fitted or empirical ones. If the definitional issues below are resolved, this would be a solid contribution to the Wannier-Stark and Stark-many-body-localization literature.","major_comments":[{"comment":"The definition of xi2 is ambiguous in the CDW regime, and this ambiguity is load-bearing for the main claim. Section III.D.2 first says xi2 counts sites with epsilon < n_j < 1 - epsilon, but Eq. (22) defines j_R as the largest site with n_j >= epsilon. These two definitions coincide for the monotone V=0 profile but not for the V>2t profiles in Figs. 7-8, where the 'edge' contains alternating sites with occupations close to 0 and 1. If the implemented definition is the one in Eq. (22), xi2 includes the high-density sites of the saturated CDW core and the linear growth with V in Fig. 10 is consistent with the growth of the CDW core; if the upper threshold 1 - epsilon is also enforced, xi2 measures only the two domain walls and would saturate for strong interactions. Since the central claim that the edge width is linear in V rests on Fig. 10, the authors should state precisely which rule was used and, ideally, show that the same linear scaling is obtained with an explicit width definition that counts the whole CDW-containing edge, including a check at larger V than the V<=5 range shown in Fig. 10.","section":"Sec. IV.B.2, Eq. (22), Fig. 10"},{"comment":"Eq. (37) is presented as an analytical expression for the interacting width, but the argument leading to it has an unstated step. The energies -2t-V and 2t+V are indeed the exact single-particle addition and hole-removal energies of the infinite chain for the interaction term of Eq. (5), so the energy range itself is not in question. What is assumed is that the local density in the many-body ground state changes across this energy range in the same way as in the noninteracting case; no derivation is provided for V different from 0. In addition, the constant c in Eq. (37) is left unspecified, while the figures compare with F(xi-1), which corresponds to c=1. I recommend stating that Eq. (37) is a heuristic estimate validated by DMRG, setting c=1 explicitly, and providing a quantitative test of the endpoint identification, for example by extracting the density endpoints from DMRG and comparing them with the solutions of mu_j = +/- (2t+V) for a few values of V.","section":"Sec. IV.B.1, Eq. (37)"}],"minor_comments":[{"comment":"The two threshold statements in this subsection should be reconciled: the text says sites with epsilon < n_j < 1 - epsilon, but Eq. (22) uses n_j >= epsilon. This matters for the CDW profiles and is related to the first major comment.","section":"Sec. III.D.2"},{"comment":"The captions say the dashed line indicates 'the scaled expression for the analytical expression xi1,int, Eq. (37)', but do not state that it is F(xi1,int - 1) with c=1. Please make the plotted quantity explicit in the captions and in the text.","section":"Sec. IV.B.2, Figs. 10 and 11"},{"comment":"In Eq. (36), the integration limit omega=0 corresponds to the chemical potential being zero, which follows from particle-hole symmetry at half-filling. This should be stated explicitly.","section":"Sec. III.E, Eq. (36)"},{"comment":"The last sentence of Appendix B, 'This does not give a linear xiA', is unclear. Please define xiA and explain how the straight line xitilde_xi = 0.9395x in Fig. 13 was obtained from the series expansion.","section":"Appendix B"},{"comment":"The DMRG section would benefit from reporting a representative truncation error or bond dimension for the largest system sizes; 'around five hundred states' is a useful but incomplete convergence statement for a paper whose central results are numerical.","section":"Sec. II, DMRG details"}],"recommendation":"major_revision","confidential_remarks":"The main risk is that the definition of the numerically measured edge width in Fig. 10 was not stated consistently in the text. If the authors confirm that xi2 includes the saturated CDW sites and add a check at larger V, the remaining issues are local and the paper should be publishable. I would not reject the paper on the current evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read. The paper does two things: it works out exact Bessel-function results for the non-interacting Wannier-Stark chain—occupations, localization lengths, LDOS—and it uses DMRG to study the same chain with nearest-neighbor repulsion. The non-interacting part is very good. The analytics match the numerics cleanly, and the finite-chain checks are honest. The interacting part has a genuine new result: for V>2t a CDW forms in the middle of the edge, and the width of the transition region grows when V is increased at fixed F. The data collapses in Figs. 10 and 11 are convincing for the parameter ranges shown.\n\nBut the central quantitative claim—that the edge width scales linearly with V as (2t+V)/F—is not as solid as the abstract suggests. The problem is the definition of the width. In Fig. 10, ξ2 counts sites with 0.02<n_j<0.98. Once the CDW amplitude saturates, the sites in the CDW plateau have occupations very close to 1 and 0, so they fall outside this window. ξ2 then measures only the two domain walls, not the full edge including the CDW core. The figure is shown only up to V=5, which is where the CDW is still unsaturated. For larger V, where the core is fully developed, the linear scaling with V has not been demonstrated. The separate measure LCDW in Fig. 9 is plotted against 1/F at V=9 but not against V, so it doesn't fill the gap. The analytical estimate Eq. (37) gives the full edge width using exact addition energies, so it's likely correct, but the numerical test in Fig. 10 doesn't actually measure that quantity at strong V.\n\nThis is a meaningful soft spot, not a fatal one. The qualitative physics—CDW appears at V>2t and the edge broadens with interaction—is probably right. But the paper should either change the width definition to include the CDW core (e.g., a finite fraction of the site after removal of the CDW pattern) or explicitly limit the claim to the unsaturated regime. The DMRG details are also thin; we get '500 states' and 'less than a dozen sweeps' with no convergence checks or error estimates. That's a minor issue for a 1D problem but should be documented.\n\nWho should read it: anyone working on Wannier-Stark localization, Stark MBL, or confined cold-atom chains. I'd bring it to a reading group. It deserves a serious referee, but the referee should push on the strong-V behavior and the width definition. If the authors can supply a proper measure of the full edge in the CDW regime, this could be a nice PRB paper; as it stands, the abstract overclaims.","headline":"Solid non-interacting analytics and a plausible CDW crossover, but the headline linear-in-V edge width is not established for strong interactions because the width measure drops the saturated CDW core.","tokens_in":17342,"tokens_out":6231,"would_cite":false,"duration_ms":60351,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"For spinless fermions in a one-dimensional Wannier-Stark chain, the width of the partially filled edge between the occupied bulk and the empty vacuum grows as (2t+V)/(eF), and for nearest-neighbor repulsion V > 2t that edge develops into…","keywords":["Wannier-Stark localization","charge density wave","spinless fermions","density matrix renormalization group","Bessel functions","localization length","one-dimensional lattice","cold atoms"],"falsifier":"Calculate (or measure in a cold-atom chain) the width of the intermediate-filling region as a function of F for a fixed V>2t; if the width is not proportional to (2t+V)/F, or if no ...101010... CDW region appears around the chain center for V>2t at small F, the central claim fails.","tokens_in":16194,"feed_emoji":"⚡","tokens_out":6477,"duration_ms":55364,"temperature":0.7,"pith_summary":"This paper studies spinless fermions on a finite one-dimensional lattice with a linear electric potential (Wannier-Stark ladder) and nearest-neighbor repulsion V. It aims to show that the intermediate 'edge' region, where the occupation falls from 1 to 0, has a width proportional to (2t+V)/(eF), growing linearly with the interaction and inversely with the field, and that for V > 2t this region becomes a half-filled charge density wave. For the non-interacting case, the paper provides exact Bessel-function expressions for the density profile, the localization lengths, and the local density of states, and confirms them against density matrix renormalization group numerics. The relevance is that cold-atom chains with a tunable potential slope could directly test this predicted scaling and the emergence of ordered CDW domains inside a confined system.","feed_headline":"Tilted-chain edge widens as (2t+V)/F, then orders into a CDW","feed_subtitle":"Cold-atom experiments can test the predicted linear scaling of the wall width and the V>2t charge density wave.","key_machinery":"The central object is the Wannier-Stark eigenfunction expansion b†_m = sum_j J_{j-m}(2t/(eF)) c†_j, which diagonalizes the non-interacting Hamiltonian and yields the occupation sum n_j(x) = sum_{m<=0} $J^{2}$_{m-j}(x) with x=2t/(eF). This Bessel-function machinery provides exact expressions for the density profile, the localization lengths (ξ1 = 4/F+1, ξ2 via the intermediate-filling window, ξ3 = 2√2 ξ̃ + 1 with ξ̃² = 2(t/F)², and the many-particle length ξ4), and the local density of states, all inversely proportional to F. For finite V, the same bandwidth argument that gives the V=0 edge (a site is partially filled when its potential lies between the effective band edges) is extended to -2t - V < μ_j < 2t + V, yielding the interacting width ξ_{1,int} = 2(2t+V)/(eF) + c. DMRG simulations then confirm that the numerically measured edge width ξ2 and many-particle length ξ4 collapse onto this linear-in-V, linear-in-1/F prediction, and that for V > 2t a CDW region with n_j n_{j+1} = 0 appears in the middle of the chain.","core_discovery":"The paper establishes that in the ground state of the interacting Wannier-Stark chain, the domain wall between the fully occupied bulk and the empty vacuum is characterized by a single length scale: the width of the intermediate-filling region is 2(2t+V)/(eF) plus a constant of order one. In the non-interacting limit V=0, this reduces to 4t/(eF)+1, and the site occupations n_j = sum_{m<=0} J_{j-m}^2(2t/eF) of the infinite chain reproduce the finite-chain DMRG results exactly when the system is longer than the localization length. When V exceeds 2t, the middle of the chain, lying within the charge gap, orders into a ...101010... charge density wave whose spatial extent grows with V and shrinks with F, while remaining surrounded by the same linear edge. The local density of states evolves from a Wannier-Stark ladder of equally spaced peaks at V=0 into a broadened structure as V grows, consistent with the increased localization length.","pith_inferences":["The paper's bandwidth argument treats the interaction as a rigid shift of the band edges; a natural next step would be to test whether quantum fluctuations around the CDW boundary modify the effective edge position by a V-dependent constant or by logarithmic corrections at fixed F.","The same Bessel-function framework could be extended to spinful fermions or to next-nearest-neighbor interactions, where the phase diagram may contain additional CDW periodicities; the paper predicts these appear only at the uniform half-filling CDW for the spinless model.","The predicted linear scaling suggests a practical method to calibrate interaction strength in optical-lattice experiments: measuring the slope of the edge width versus inverse tilt gives (2t+V) directly, independent of the microscopic model details beyond nearest-neighbor terms."],"forward_implications":["The width of the intermediate-filling edge in interacting Wannier-Stark chains is quantitatively predicted by a simple one-particle band-edge formula, so cold-atom measurements of the density profile can directly extract both t and V from the slope of width versus 1/F.","For V > 2t, a tunable electric field controls the spatial extent of a half-filled CDW domain inside the chain, offering a clean experimental knob to create or destroy ordered regions in a disorder-free system.","The Bessel-function expressions for the non-interacting occupation profile and LDOS are exact for infinite chains and remain accurate for finite chains as long as the localization length stays below the system size, providing a benchmark standard for numerical methods.","The linear scaling of the many-particle localization length with V/F implies that Stark many-body localization in this model is governed by the same single length scale as the non-interacting WS ladder, with interaction only renormalizing the prefactor."],"supporting_citations":[{"why":"Provides the Bessel-function diagonalization of the non-interacting Wannier-Stark chain and the analytic eigenfunctions used throughout the paper.","marker":"[20]"},{"why":"Introduces the bandwidth argument for the edge width and the DMRG approach that this paper extends to spinless fermions.","marker":"[4]"},{"why":"Uses a similar band-edge estimate for interacting tilted chains, the basis for the (2t+V) shift in Eq. (37).","marker":"[29]"},{"why":"Establishes that the uniform spinless chain has a half-filled CDW only for V >= 2t, used to explain the CDW region's density.","marker":"[40]"},{"why":"Supplies the Bessel-function identities used to derive localization lengths and occupation sums.","marker":"[34]"},{"why":"Provides the integral representation of J^2 used to obtain the continuous-j occupation formula Eq. (17).","marker":"[35]"},{"why":"Gives the LDOS form for the Wannier-Stark chain that Eq. (35) reproduces.","marker":"[37]"}],"fun_headline_variants":["Tilted chain edge width ~ (2t+V)/F, CDW when V>2t","From Wannier-Stark to CDW: edge width scales as (2t+V)/F","Edge wall width ~ (2t+V)/F; strong V orders it into CDW","Tilted chain: edge width 2(2t+V)/F, CDW for V>2t","Cold atoms can test edge width (2t+V)/F and CDW for V>2t"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The edge width formula assumes the interacting ground state can be described by independent one-particle energies shifted by V, so the boundaries are where the local potential equals -2t - V and 2t + V; if many-body corrections shift these boundaries in a more complicated way with F, the linear scaling would fail.","fun_headline_variants_meta":{"raw":{"variants":["Tilted chain edge width ~ (2t+V)/F, CDW when V>2t","From Wannier-Stark to CDW: edge width scales as (2t+V)/F","Edge wall width ~ (2t+V)/F; strong V orders it into CDW","Tilted chain: edge width 2(2t+V)/F, CDW for V>2t","Cold atoms can test edge width (2t+V)/F and CDW for V>2t"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001149,"raw_usage":{"total_tokens":4802,"prompt_tokens":1024,"completion_tokens":3778,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":640,"completion_tokens_details":{"reasoning_tokens":3650}},"tokens_in":640,"tokens_out":3778,"duration_ms":25267,"temperature":1.0,"reasoning_tokens":3650,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T21:08:49.111412+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Calculate (or measure in a cold-atom chain) the width of the intermediate-filling region as a function of F for a fixed V>2t; if the width is not proportional to (2t+V)/F, or if no ...101010... CDW region appears around the chain center for V>2t at small F, the central claim fails.","supporting_citations":[{"cited_title":"Waschke, H","cited_arxiv_id":null,"evidence_quote":"Provides the Bessel-function diagonalization of the non-interacting Wannier-Stark chain and the analytic eigenfunctions used throughout the paper."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the bandwidth argument for the edge width and the DMRG approach that this paper extends to spinless fermions."},{"cited_title":"Fukuyama, R.A","cited_arxiv_id":null,"evidence_quote":"Uses a similar band-edge estimate for interacting tilted chains, the basis for the (2t+V) shift in Eq. (37)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes that the uniform spinless chain has a half-filled CDW only for V >= 2t, used to explain the CDW region's density."},{"cited_title":"Merlin, Phonon Bloch Oscillations, Phys","cited_arxiv_id":null,"evidence_quote":"Supplies the Bessel-function identities used to derive localization lengths and occupation sums."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the integral representation of J^2 used to obtain the continuous-j occupation formula Eq. (17)."},{"cited_title":"Morong et al., Observation of Stark many-body lo- calization without disorder, Nature 599, 393 (2021)","cited_arxiv_id":null,"evidence_quote":"Gives the LDOS form for the Wannier-Stark chain that Eq. (35) reproduces."}],"review_version":1}