{"id":"6a813182-0004-46ec-8ac1-3caaa1725012","arxiv_id":"1908.01563","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A DMD-based optical potential combined with a magnetic atom chip trap and iterative feedback realizes programmable 1D potentials for ultracold gases.","lead":"Researchers used a digital micro-mirror device to shine shaped laser light onto a 1D gas of rubidium atoms trapped on an atom chip, letting them carve the trap into boxes, barriers, rippled boxes, and V shapes. The technique gives cold atom experiments a flexible new dial for designing quantum systems and studying their dynamics.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Potential fidelity is inferred from the same density used for optimization, so the reported 4-6% residuals do not independently verify the potential claim; an atom-number scaling test would settle it.","rationale":"The paper is a solid experimental methods contribution: it demonstrates DMD-based programmable longitudinal potentials on an atom chip, with an autonomous density-feedback optimization, four different target profiles, and a six-hour stability check. The optical setup is described in detail, and the use of NPSE target densities is appropriate. The reader's verdict of ACCEPT is reasonable. The weakest point is not that the optimization is unstable or the examples are poor, but that the evidence for the potential itself is indirect: density residuals are the optimization target, and the plotted potentials are reconstructed from those same densities through Eq. (2). The update rule also assumes a local monotonic Thomas-Fermi response, and no convergence guarantee is given, but the empirical convergence is demonstrated. The suggested atom-number scaling test provides a clean, non-circular falsification: if the optimized potential is truly the target potential, the density at a different atom number should be predictable without further optimization. Running this check would either confirm the central claim or require it to be qualified as density control rather than verified potential control. Because the demonstrated density control and stability already support the practical utility of the method, the concern does not change the overall ACCEPT verdict; it identifies a worthwhile additional verification.","tokens_in":12357,"tokens_out":8329,"duration_ms":103756,"concrete_test":"After optimizing one of the demonstrated patterns, change the atom number by roughly a factor of two (e.g., by holding the cloud for a controlled loss time or by changing the final evaporation step) and measure the longitudinal density profile without changing the DMD pattern. Compare this measured profile with the NPSE ground state computed for the claimed target potential at the new atom number. Agreement at the same few-percent level would confirm that the optimized potential is indeed the target potential; systematic disagreement would show that the feedback loop matched density only at the calibration atom number and that the actual potential differs from the target.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of arbitrary potential control is verified only through density measurements, and the potential reconstruction is not independent of the optimization target. The convergence metric epsilon_RMS (Eq. 1) and the update rule in Sec. 4 act on the longitudinal density rho(z), while the target density is computed as the NPSE ground state of the desired potential. The 'measured potentials' in Fig. 7 are then reconstructed from the same measured density via the Thomas-Fermi relation (Eq. 2). Consequently, a density match almost automatically produces an apparent potential match, up to the accuracy of the TF approximation. If the actual optical potential deviates from the target in a way that still yields the target density at the calibration atom number, the claim of potential control would be overstated. This could happen if the DMD point-spread function couples neighboring grid points and the greedy local update converges to a density-matching but potential-mismatched pattern, or if local TF/locality breaks down near barriers and box edges. The green 'broadened target potentials' in Fig. 7 are a partial check, but they are not an independent measurement of the potential. The paper would be strengthened by a quantitative cross-check that does not reuse the optimization density.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports an experimental method for shaping the longitudinal potential of a 1D Bose gas on an atom chip by superposing a DMD-shaped, blue-detuned optical dipole potential on the magnetic chip trap. The central technical contribution is an iterative feedback procedure: starting from a measured density profile, the DMD pattern is updated locally so that the measured longitudinal density approaches the NPSE ground-state density of a desired target potential. The authors demonstrate four potentials—a 160 µm box, a double box with a barrier, a box with sinusoidal modulation, and a V-shaped potential—with final density residuals between 4.2% and 5.9%, and they report stability of a box potential for more than six hours. They argue that the method provides flexible and autonomous control of longitudinal confinement for experiments in quantum thermodynamics and quantum simulation.","tokens_in":12561,"tokens_out":5406,"duration_ms":56344,"significance":"If the central claim holds, this is a useful technical advance: it combines the advantages of atom-chip magnetic trapping (fast evaporative cooling, RF dressed-state capabilities) with the flexibility of DMD-shaped optical potentials, and the optimization loop is described in sufficient detail to be reproduced. The paper is honest in reporting convergence curves, error bars, and a stability measurement, and it explicitly describes the calibration and update rules. However, the verification of the central claim is weakened by a circularity: the optimization acts on the density, and the reported potentials are reconstructed from that same density via the Thomas-Fermi relation. The demonstrated density control is convincing, but the claim of arbitrary potential control would be strengthened by an independent cross-check of the realized potential.","major_comments":[{"comment":"The reported potential fidelity is not independently verified. The optimization observable is the longitudinal density ρ(z), and the \"measured potentials\" in Fig. 7 are reconstructed from the same measured density using the Thomas-Fermi relation, Eq. (2). A density match therefore produces an apparent potential match by construction, up to the accuracy of the TF approximation and the assumed chemical potential. The 4.2–5.9% density residuals do not by themselves establish that the realized V(z) equals the target potential, because a pattern that reproduces the target density at the calibration atom number can in principle produce a different potential at other densities. The green \"broadened target potentials\" in Fig. 7 are a useful consistency check but are not an independent measurement. I recommend adding a cross-check that does not reuse the optimization density: for example, measure the density for the same final DMD pattern at a substantially different atom number and compare with the NPSE ground state of the target potential; a genuine potential match should reproduce the density at both atom numbers. The text should also state explicitly that the plotted potentials are inferred from density, not directly measured.","section":"§5, Eq. (2), Fig. 7"},{"comment":"The local update rule assumes a monotonic, local relationship between switching DMD pixels and changes in the local density—effectively a diagonally dominant Jacobian with negligible coupling between neighboring 1.05 µm grid points. Given that the imaging PSF has an Airy diameter of 4 µm, neighboring grid points are coupled, and near steep barriers and box edges the local Thomas-Fermi relation itself becomes questionable. The paper shows empirical convergence for four examples, but it does not discuss conditions under which the loop could stall or converge to a density-matching pattern that is not the intended potential. I ask for a brief analysis of this limitation or an additional experimental test: rerun the optimization for the same target starting from different initial patterns and compare the final DMD patterns and density residuals, and comment explicitly on the role of the finite PSF in the update procedure.","section":"§4, update rule and Fig. 6"}],"minor_comments":[{"comment":"The word \"arbitrary\" overstates the demonstrated capability: the experiment uses only blue-detuned repulsive light, so the optical potential can only add repulsive corrections on top of the magnetic potential. Please qualify the claim (e.g., \"shaped repulsive potentials\" or \"a wide class of one-dimensional potentials\") in the abstract and conclusion.","section":"Abstract and §6"},{"comment":"The normalization factor in Eq. (1) is written as 1/(j−l), but the sum runs from k=j to k=l, so the number of terms is l−j+1. Please check and correct the notation.","section":"Eq. (1)"},{"comment":"The sentence \"This leads to only one spontaneous scattering per second for an ensemble of 10^4 atoms\" appears inconsistent with the quoted rate of ~2π×0.1 mHz: for 10^4 atoms, this would be on the order of several scattering events per second. Please recheck the numerical estimate.","section":"§3, scattering-rate estimate"},{"comment":"The text says the stability check exceeds six hours, while the horizontal axis of Fig. 10(a) extends to 500 minutes (about 8.3 hours). Please align the stated time span with the plotted data or adjust the axis.","section":"Fig. 10"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid experimental methods demonstration, and the density-control results are credible. My main concern is the circularity between the optimization observable and the potential reconstruction; I would like to see this addressed with an independent cross-check or a clear statement of the inferred nature of the reported potentials before publication. The 'arbitrary' claim should also be tempered given that only repulsive optical potentials are implemented."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a solid experimental methods paper. The authors combine a DMD-generated optical dipole potential with an atom chip magnetic trap, and they demonstrate an autonomous density-feedback optimization routine that produces box, barrier, sinusoidally modulated, and V-shaped longitudinal potentials for 1D Bose gases. The measured density residuals of 4.2–5.9% and the six-hour stability data are credible. The writing is direct, the calibration procedure is described in enough detail to be reproduced, and the citation practice is honest: prior DMD beam shaping work (Gauthier, Zupancic) is properly acknowledged, so the novelty is correctly framed as integration and optimization rather than inventing the technique.\n\nThe soft spot is real but not fatal. The potential itself is never measured independently. Equation (2) reconstructs the potential from the Thomas-Fermi inversion of the same density profile that the optimizer was trying to match, so a density match almost guarantees an apparent potential match. The green broadened target potentials in Fig. 7 are just convolutions of the target, not an independent check. An atom-number scaling test—does the density match at a different atom number if the optical potential is correct?—would have settled this cleanly. The authors do acknowledge this limitation in passing (\"quantitative comparison is instructive,\" Section 5), which counts in their favor.\n\nI do not see this as a load-bearing flaw. For most applications the density profile is the operational quantity, and the paper's claim of 'arbitrary one-dimensional potentials' is honestly hedged by the demonstrated examples. Still, I would have liked one direct potential probe, or at least a more explicit sentence saying the potential plots are model-dependent.\n\nThis paper will be useful to experimental groups working on atom chips, 1D quantum gases, and quantum thermodynamics. It deserves a serious referee and could be accepted after minor revision, mostly to add the caveat about indirect potential verification. No data or code are shared, which is a small downside in 2020.\n\nAll in all, worth publishing. I would cite it and would bring it to a reading group as a good example of feedback optimization in cold atoms.","headline":"A genuinely useful methods paper that does what it says with care, but the potential claim rests on density inferred from the same measurement used for optimization.","tokens_in":13115,"tokens_out":1755,"would_cite":true,"duration_ms":19729,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["67.85.-d","37.10.Gh"],"model":"deepseek-v4-flash","headline":"A digital micro-mirror device plus a feedback loop writes arbitrary longitudinal potentials—boxes, barriers, and ramps—onto one-dimensional Bose gases on an atom chip.","keywords":["atom chip","one-dimensional Bose gas","digital micro-mirror device","optical dipole potential","feedback optimization","box potential","quantum simulation","Thomas-Fermi density"],"falsifier":"Toggle on a single DMD super-pixel column and image the density: if the induced density dip appears at more than one longitudinal position, or if toggling pixels in one region shifts density in a distant region by a comparable amount, the local update rule on which the optimization rests is empirically false.","tokens_in":12189,"feed_emoji":"🔬","tokens_out":7545,"duration_ms":76130,"temperature":0.7,"pith_summary":"This paper establishes a practical route to arbitrary control of the longitudinal potential of a one-dimensional degenerate Bose gas on an atom chip. The authors add a digital micro-mirror device (DMD) shaped blue-detuned light pattern to the magnetic trap, and an automated feedback loop adjusts the pattern until the measured atomic density matches a target density computed from the desired potential. They demonstrate a 160-micrometre box potential, a box with a central tunable barrier, a box with a sinusoidally modulated bottom, and a linear (V-shaped) confining potential, with density residuals of 4.2 to 5.9 percent over the target region, stable for more than six hours. This matters because homogeneous and precisely shaped one-dimensional systems are the foundation for experiments in quantum thermodynamics, transport, and quantum simulation.","feed_headline":"Feedback-shaped light makes atom-chip traps programmable","feed_subtitle":"A digital-mirror feedback loop stamps box, barrier, and V-shaped potentials onto 1D gases, stable for hours.","key_machinery":"The load-bearing mechanism is the DMD as a binary intensity modulator imaged onto the atoms with spatial averaging and spatial filtering. Because many micro-mirror pixels contribute to each diffraction-limited spot, switching individual pixels in a column changes the local optical potential in small steps, giving quasi-gray-scale one-dimensional potentials. The optimization loop then uses the Thomas–Fermi relation between the one-dimensional density and the total potential, $V(z)=\\mu-\\frac{g}{2\\pi a_\\perp^2}\\frac{\\rho(z)}{\\sqrt{1+2a_s\\rho(z)}}-\\frac{\\hbar\\omega_\\perp}{2}\\left(\\sqrt{1+2a_s\\rho(z)}+\\frac{1}{\\sqrt{1+2a_s\\rho(z)}}\\right)$ (Eq. 2), to translate density deviations into pixel updates: where density is too high, pixels are switched on to raise the potential; where too low, outer pixels are shifted away to lower it. The target density itself comes from the non-polynomial Schrödinger equation for each candidate potential.","core_discovery":"The central claim is that superposing an optical dipole potential, locally shaped by a DMD, on the magnetic confinement of an atom chip yields a robust and versatile method to engineer the longitudinal potential of a one-dimensional Bose gas while preserving the chip's transverse trapping and evaporative cooling. The paper shows that an iterative pattern optimization—comparing the measured one-dimensional density with a target density obtained by solving the one-dimensional non-polynomial Schrödinger equation—converges to target potentials without requiring a precise model of the optical setup. In the demonstrated examples, the final density profiles agree with the target to 4.2–5.9% normalized root-mean-square deviation in the target region, and the achieved patterns stay stable for over six hours. The measured potentials match the target potentials once those targets are broadened by the DMD imaging system.","pith_inferences":["Possible extension: the same density-feedback loop could be ported to other atomic species or to two-dimensional optical potentials, but the Thomas–Fermi locality assumption would degrade and would need stronger spatial filtering or a more global update rule.","Possible extension: because the DMD refresh rate (17–50 kHz) is far above the axial trap frequencies, stroboscopic or time-averaged patterns could synthesize effective Floquet potentials on the same chip; fast switching is mentioned in the paper but no Floquet experiment is reported.","Possible consequence: pushing the measurement noise down, for example with higher-resolution imaging, should drive the converged density residuals below the 4–6% level, since the paper's own termination criterion is set by the standard error of the mean density."],"forward_implications":["Homogeneous box-like one-dimensional gases remove density inhomogeneity that otherwise masks intrinsic many-body dynamics, enabling cleaner studies of recurrence, thermalization, and transport.","Barriers inside a box allow longitudinal splitting and tunable tunneling between homogeneous segments, extending the double-well physics already available via radio-frequency dressed states.","Because DMD patterns can be changed faster than the axial dynamics, the same setup can implement time-dependent potentials and quenches, for example exciting a chosen mode by imprinting its density modulation.","The autonomous optimization means no detailed optical model is needed: any target potential can be reached from a black screen or from a previous pattern, limited mainly by imaging noise and resolution."],"supporting_citations":[{"why":"Establishes that the optical dipole potential is proportional to light intensity with sign set by detuning, the basis for the added potential.","marker":"[11]"},{"why":"Demonstrates direct DMD imaging for configurable microscopic optical potentials, the technique this paper adapts and extends to atom-chip one-dimensional traps.","marker":"[12]"},{"why":"Provides the atom-chip magnetic trapping platform whose longitudinal confinement is to be shaped.","marker":"[19]"},{"why":"Shows radio-frequency dressed-state potentials on the chip, the manipulation capability the new method preserves.","marker":"[22]"},{"why":"Documents potential roughness near lithographically fabricated atom chips, the imperfection the optical potential corrects.","marker":"[23]"},{"why":"Supplies the absorption imaging technique used to measure the one-dimensional density profiles in the feedback loop.","marker":"[32]"},{"why":"Gives the non-polynomial Schrödinger equation used to compute target densities from target potentials.","marker":"[38]"}],"fun_headline_variants":["Digital mirror crafts arbitrary 1D potentials on atom chips","Feedback-shaped light makes atom-chip traps fully programmable","DMD shapes box, barrier, and V traps for 1D Bose gases","Stable arbitrary potentials for atom chips via DMD-light shaping"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The update rule assumes that turning DMD pixels on or off changes the potential locally and monotonically lowers or raises the local density, per the Thomas–Fermi relation, with negligible coupling to neighbouring positions.","fun_headline_variants_meta":{"raw":{"variants":["Digital mirror crafts arbitrary 1D potentials on atom chips","Feedback-shaped light makes atom-chip traps fully programmable","DMD shapes box, barrier, and V traps for 1D Bose gases","Stable arbitrary potentials for atom chips via DMD-light shaping"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000267,"raw_usage":{"total_tokens":1557,"prompt_tokens":828,"completion_tokens":729,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":444,"completion_tokens_details":{"reasoning_tokens":658}},"tokens_in":444,"tokens_out":729,"duration_ms":8007,"temperature":1.0,"reasoning_tokens":658,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:09:27.580268+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Toggle on a single DMD super-pixel column and image the density: if the induced density dip appears at more than one longitudinal position, or if toggling pixels in one region shifts density in a distant region by a comparable amount, the local update rule on which the optimization rests is empirically false.","supporting_citations":[{"cited_title":"Opticaldipoletrapsforneutralatoms,","cited_arxiv_id":null,"evidence_quote":"Establishes that the optical dipole potential is proportional to light intensity with sign set by detuning, the basis for the added potential."},{"cited_title":"Direct imaging of a digital-micromirror device for configurable microscopic optical potentials,","cited_arxiv_id":null,"evidence_quote":"Demonstrates direct DMD imaging for configurable microscopic optical potentials, the technique this paper adapts and extends to atom-chip one-dimensional traps."},{"cited_title":"Controlling cold atoms using nanofabricated surfaces: Atom chips,","cited_arxiv_id":null,"evidence_quote":"Provides the atom-chip magnetic trapping platform whose longitudinal confinement is to be shaped."},{"cited_title":"Radiofrequency-dressed-state potentials for neutral atoms,","cited_arxiv_id":null,"evidence_quote":"Shows radio-frequency dressed-state potentials on the chip, the manipulation capability the new method preserves."},{"cited_title":"Potential roughness near lithographically fabricated atom chips,","cited_arxiv_id":null,"evidence_quote":"Documents potential roughness near lithographically fabricated atom chips, the imperfection the optical potential corrects."},{"cited_title":"Absorption imaging of ultracold atoms on atom chips,","cited_arxiv_id":null,"evidence_quote":"Supplies the absorption imaging technique used to measure the one-dimensional density profiles in the feedback loop."},{"cited_title":"Effective wave equations for the dynamics of cigar-shaped and disk-shaped bose condensates,","cited_arxiv_id":null,"evidence_quote":"Gives the non-polynomial Schrödinger equation used to compute target densities from target potentials."}],"review_version":1}