Pith. sign in

REVIEW 4 major objections 5 minor 22 references

Trainmon: a framework for reverse engineering potentials in superconducting Qubits

T0 review · 4 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A circuit built from parallel chains of Josephson junctions — the Trainmon — can reverse-engineer the potential wells of other superconducting qubits, reproducing the Quarton potential to 0.02% error and the Fluxonium transition frequencies

desk verdict Potential fitting works nicely, but the fractional-charge Hilbert space undermines the physical claims. read the letter →

arxiv 2509.00819 v1 pith:DNT7MLA3 submitted 2025-08-31 quant-ph

Saeed Hajihosseini , Seyed Iman Mirzaei , Hesam Zandi , Mohsen Akbari This is my paper · ORCID
classification quant-ph
keywords superconductingqubitsJosephsonjunctionarrayspotentialengineeringdiscretecosinetransformFluxoniumQuartonchargebasisdephasing
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes Trainmon, a superconducting circuit made of parallel branches of identical Josephson junctions, whose potential energy is a sum of cosines of fractionally divided node phases — mathematically a discrete cosine transform. The claim is that by numerically fitting the junction energies in a small number of branches, one can reverse-engineer the potential well of a different qubit design, then solve the Trainmon Hamiltonian in a charge basis to reproduce the original qubit's spectrum. For the Quarton qubit, the fitted 124-Trainmon (branches of 1, 2, and 4 junctions) reproduces the potential with a maximum relative error of 0.02%; for the Fluxonium, the reconstructed potential correlates at 0.991 and the first two transition frequencies differ by less than 1%. The paper also estimates the dephasing time of the Fluxonium-like Trainmon (about 2340 μs) against the original Fluxonium (2542 μs). If the approach holds, a single junction-and-capacitor architecture could emulate a variety of carefully engineered qubit potentials without inductors or exotic elements.

What carries the argument

The central object is the Trainmon potential, a cosine series V(φ) = −Σᵢ nᵢ E_J^{nᵢ} cos(φ/nᵢ + …) with period 2π·lcm({nᵢ}) (8π for the 124-Trainmon). The work it does is twofold: numerically, the coefficients E_J^{nᵢ} are fitted with a least-squares optimizer to match a target potential; analytically, each cos(φ/n) term acts in the charge basis as a hopping operator between fractional charge states |k ± 1/n⟩, so the full Hamiltonian becomes a sparse matrix in a fractionally charged Hilbert space (Appendix A). Negative coefficients are realized by flux-biasing loops to shift the cosine phase by π, satisfying the fluxoid quantization condition of each loop.

What would settle it

Fabricate a 124-Trainmon and measure the transition frequencies E₀₁ and E₁₂ at the designed bias. The fractional-charge model predicts an 8π-periodic band structure with quarter-Cooper-pair charge sensitivity; a conventional Cooper-pair circuit should show 2e periodicity. If the measured E₀₁ and E₁₂ deviate from the sub-percent match predicted here, the mapping between the fitted potential and the physical circuit is broken.

Watch

Extended reading notes

Core claim

Trainmon's potential term is a harmonic series in the node phase: V(φ) = −Σᵢ nᵢ E_J^{nᵢ} cos(φ/nᵢ + φ_ext^{nᵢ}/nᵢ), where each branch of nᵢ identical junctions contributes a cosine at a fraction of the phase. Because this sum behaves like a discrete cosine transform, the coefficients E_J^{nᵢ} can be numerically fitted to approximate the potential of a target qubit. Negative fitted coefficients, which are difficult to realize directly, are handled by applying external magnetic fluxes that shift each branch's cosine by π while satisfying fluxoid quantization. Using a 1-, 2-, and 4-junction Trainmon, the paper reconstructs the Quarton potential over [−π, π] with a maximum relative error of 0.02

Load-bearing premise

The reconstruction and the charge-basis solution assume the node phase is compact with period 8π for the 1-2-4 circuit, so the conjugate charge comes in quarter-Cooper-pair steps — a fractional charge quantization stated without physical justification.

Editorial extensions

If this is right

  • A single Trainmon architecture (junctions and a capacitor only, no inductors) can be tuned to emulate multiple qubit designs by fitting branch Josephson energies and loop fluxes.
  • The first two transition energies of the emulated qubit match the target to sub-percent accuracy, so the Trainmon version can stand in for the target qubit for spectroscopy and low-energy operations.
  • The extra flux loops required for negative coefficients cost only modestly in coherence: 2340 μs dephasing versus 2542 μs for the original Fluxonium under the same noise model.
  • Solving in the charge basis is efficient for junction-capacitor circuits, avoiding an explicit differential-equation solution of the engineered potential.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The cosine-series representation suggests a natural accuracy ladder: larger branch sets with a bigger least common multiple give finer harmonic resolution, so a 124-Trainmon is a proof of concept and adding more branch sizes could tighten the Fluxonium match further.
  • The fractional-charge basis (states |k + 1/n⟩) is a computational device; a physical charge measurement should show Cooper-pair (2e) periodicity, so the model's validity hinges on the effective phase-slip regime E_J^n/E_c ≫ 1 — a testable experimental condition.
  • The framework inverts the normal design flow: choose a potential known to suppress a specific noise channel, then synthesize the junction-branch circuit that realizes it, pointing toward noise-engineered wells such as 0-π-like double wells.
  • The two-loop flux bias gives a natural tuning handle for sweet-spot engineering, but the coherence estimate assumes independent 1/f noise per loop; correlated flux noise between the two loops could change the total dephasing time in either direction.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper introduces 'Trainmon,' a superconducting qubit architecture consisting of parallel branches, each containing n identical Josephson junctions in series, shunted by a common capacitor. The authors assert that in the regime E_J^n/E_c^eff >> 1 the circuit reduces to the 1D Hamiltonian in Eq. (2)/(6), and that the potential term can be used as a Fourier-like basis to reverse-engineer target potentials. Using scipy curve_fit, they fit the Josephson energy coefficients to the Quarton and Fluxonium potentials, report very small potential reconstruction errors, compute transition frequencies in a charge basis, and estimate dephasing times. The central numerical claims are that the 124-Trainmon reproduces the Quarton potential to 0.02% maximum relative error and the Fluxonium potential with correlation 0.991, and that transition frequencies match to within <1% (Table 1).

Significance. The idea of systematically shaping a qubit potential with parallel arrays of Josephson junctions is attractive and, taken purely as a numerical fitting exercise, the potential reconstruction is convincing: the 124-Trainmon matches the Quarton potential to 0.02% and the Fluxonium potential with high correlation. The paper also makes concrete, testable predictions for transition frequencies and dephasing times, and uses standard tools (scQubits, QuTiP). However, the central physical reduction to a 1D Hamiltonian with fractional charge states is not rigorously derived, and the numerical agreement is a consistency check of the potential fit rather than an independent validation. If the circuit-to-1D reduction and the charge quantization were properly derived, the framework could become a useful design tool; as it stands, the results rest on an unvalidated model.

major comments (4)
  1. [§2, Eq. (2)/(6)] The reduction of the multi-branch circuit to the 1D Hamiltonian is asserted, not derived. The text states that Eq. (2) is valid for E_J^n/E_c^eff >> 1 and that phase slips allow a quasi-1D potential, but no circuit-level derivation from the full multi-node Lagrangian is given. For a physical circuit with several junction arrays, eliminating internal nodes is nontrivial; the phase drop across each junction in a branch is not simply phi/n unless the internal dynamics are properly integrated out. Since all subsequent results (potential fitting, spectra, coherence times) depend on Eq. (6), this is a load-bearing gap. Please derive Eq. (2)/(6) from the complete circuit or provide a rigorous effective-theory argument with error bounds.
  2. [Appendix A, Eqs. (10)-(11)] The fractional-charge basis |k+1/n> is introduced without physical justification. In a standard superconducting circuit with [phi,n]=i and 2π-periodic phi, the charge eigenvalues are integers (Cooper pairs). For the 124-Trainmon, the paper uses charges spaced by 1/4, i.e., sub-Cooper-pair charge units, and explicitly builds a Hilbert space with states such as |k+1/4>. This is not a harmless convention: it changes the compactification of phi and enlarges the Hilbert space compared to the physical circuit. Since Table 1 and the dephasing times are computed in this basis, this is a load-bearing assumption. The authors must either derive the fractional-charge quantization from the circuit's node-phase periodicities or repeat the calculation in the integer-charge basis of the full multi-node circuit and demonstrate that the same results are obtained.
  3. [§4, Table 1] The agreement in transition energies is a consistency check, not an independent prediction. The coefficients E_J^n are obtained by fitting the Trainmon potential to the target potential via curve_fit; the subsequent spectrum therefore reflects the quality of the potential fit. The paper should clearly state this and strengthen the validation by comparing the full low-energy spectra (not just E_01 and E_12) and by testing whether the fitted coefficients are robust to the fitting range and charge cutoff. Without such tests, the near-perfect agreement in Table 1 does not by itself validate the Trainmon construction as a physical circuit.
  4. [§5, Eq. (8)] The dephasing-time calculation is contingent on the unvalidated 1D Hamiltonian and fractional-charge basis. The reported values T_phi^tot = 2340 μs and T_phi^Fluxonium = 2542 μs are obtained from the dispersion of the fitted model with two independent flux-noise terms, but the text describes the noise as 'spatially correlated' while treating the two loops independently. The noise model and the correlation between loops should be specified precisely. Moreover, until the circuit quantization issue is resolved, these coherence numbers should be presented as conditional results, not as reliable predictions for a physical Trainmon device.
minor comments (5)
  1. [Introduction] There are several typos ('frameworks' in the introduction, 'Qurton' in a figure caption). Eq. (1) uses phi_ext inside the cosine without defining the sign convention relative to Eq. (3).
  2. [Eq. (4) and Figure 2] The notation phi_{N_l}/N_l is confusing. Please define explicitly how the external flux is distributed among the junctions in each branch and how Eq. (5) follows from Eq. (4).
  3. [Appendix A, Eq. (11)] The matrix in Eq. (11) is inconsistent with Eq. (10) for zero external flux and unit Josephson energies, unless additional prefactors are being used. Specify the E_J^n values and the charge-state ordering used to generate the matrix, and state explicitly that the diagonal capacitive terms were omitted.
  4. [Abstract / §2] The claim that the Hamiltonian 'resembles a discrete cosine transform' is not elaborated. Over [-π,π], the functions cos(phi/n) are not orthogonal, and the paper uses curve_fit rather than a transform; the analogy should be made precise or removed.
  5. [Table 1 and Numerical Methods] The paper does not report the fitted E_J^n coefficients for the Quarton case, nor does it justify the charge-space truncation [-1,1] as converged. Including these details would improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: potential coefficients are fitted to target potentials, and the spectral agreement is a forward consistency check, not a fitted prediction.

full rationale

The paper's derivation chain is: (1) define the Trainmon Hamiltonian as 4E_c n^2 − Σ n E_J^n cos(φ/n) (Eq. 2); (2) fit the E_J^n coefficients to target Quarton and Fluxonium potentials using curve_fit; (3) numerically solve the resulting Hamiltonian in the charge basis (Appendix A); and (4) compare the resulting transition energies to the original qubits (Table 1). The fit uses potential values only, not the transition frequencies. The transition energies are then a consequence of the fitted potential through the Schrödinger equation, so the agreement in Table 1 is a consistency check of how well the finite cosine basis approximates the target potential, not a statistical prediction forced by fitting the same observable. No parameter is fit to E_01 or E_12, and the paper does not use any self-citation as load-bearing evidence. The fractional-charge Hilbert space and the compact-phase assumption are physically substantive assumptions that may be incorrect or in need of fuller circuit quantization, but they are not circular: they are inputs to the model, not conclusions that reduce to the model's outputs. The abstract's phrasing 'validate the transition frequencies' is an overstatement, since the frequencies are expected to match when the potential approximation is good, but that is a weakness in evidential strength, not a circularity. Therefore the score is 0.

Assumptions & free parameters 4 free parameters · 5 assumptions · 1 invented entities

The central claim rests on a chain of assumptions: the 1D reduction of the circuit, uniform phase distribution in each branch, the fractional-charge compact-phase quantization, and fluxoid-based sign inversion. Each is asserted rather than derived, and the fractional-charge basis is especially fragile. The fitted Josephson energies and loop fluxes are free parameters obtained by numerical optimization against the target potential.

free parameters (4)
  • E_J^1 (single-junction branch Josephson energy) = not quoted in text
    Fitted via scipy curve_fit to match the target potential; specific values shown only in figure insets.
  • E_J^2 (two-junction branch per-junction energy) = not quoted
    Same fitting procedure; coefficient of cos(phi/2).
  • E_J^4 (four-junction branch per-junction energy) = not quoted
    Same fitting procedure; coefficient of cos(phi/4).
  • External fluxes Phi_e1 and Phi_e2 (loop biases) = Phi_e1 = Phi_0, Phi_e2 = -2*Phi_0 for Fluxonium reconstruction
    Chosen or fitted to shift cosines by pi to realize negative coefficients; values stated for the Fluxonium case but the selection procedure is not described.
assumptions (5)
  • domain assumption Circuit can be reduced to the 1D Hamiltonian of Eq. 6 when E_J^n/E_c^eff >> 1 (phase-slip regime)
    Method section, after Eq. 2; no derivation from the full multi-node circuit Lagrangian is given.
  • domain assumption The phase drop is uniformly distributed across the junctions in each branch (delta_i = phi/n)
    Equation 4; valid only if intermediate island charging energies are negligible.
  • ad hoc to paper The node phase is compact with period 2*pi*lcm(n_i), allowing fractional charge states |k + 1/n>
    Appendix A; this quantization is non-standard and not derived from the circuit's charge conservation.
  • domain assumption Fluxoid quantization with external fluxes can realize negative cosine coefficients (Eq. 5)
    Method, after Eq. 3; assumes the chosen external fluxes satisfy the constraint without altering the 1D reduction.
  • domain assumption The 1/f flux noise model and dephasing formula Eq. 8 apply to both circuits
    Results, coherence time section; parameters omega_ir, omega_uv, t, A_Phi are set to representative values.
invented entities (1)
  • Fractional Cooper-pair charge states |k + 1/n>
    purpose: Basis for diagonalizing the Trainmon Hamiltonian in charge space
    The paper introduces these states without a physical mechanism for sub-Cooper-pair charge quantization; no experiment or prior result supports them.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Trainmon: a framework for reverse engineering potentials in superconducting Qubits." pith.science (2026). https://pith.science/paper/DNT7MLA3

@misc{pith2026250900819,
  author       = {Pith},
  title        = {Pith review of: Trainmon: a framework for reverse engineering potentials in superconducting Qubits},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DNT7MLA3}},
  note         = {Machine review of arXiv:2509.00819}
}
read the original abstract

A framework named Trainmon is introduced to reverse-engineer a quantum potential well for superconducting qubits. Trainmon consists of parallel branches of identical Josephson junctions. The Hamiltonian for this circuit resembles a discrete cosine transform, which can be applied to mimic various potentials. This framework is applied to well-known qubit potentials such as Quarton and Fluxonium, and their Hamiltonians are extracted and solved to validate the transition frequencies and calculate the coherence times of both the original qubits and their Trainmon-based versions.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 11 canonical work pages

  1. [1]

    Moving beyond the Transmon: Noise-Protected Superconducting Quantum Circuits,

    A. Gyenis, A. Di Paolo, J. Koch, A. Blais, A. A. Houck, and D. I. Schuster, “Moving beyond the Transmon: Noise-Protected Superconducting Quantum Circuits,” PRX Quantum, vol. 2, no. 3, p. 030101, Sep. 2021, doi: 10.1103/PRXQuantum.2.030101

    doi:10.1103/prxquantum.2.030101 2021

  2. [2]

    Charge-insensitive qubit design derived from the Cooper pair box,

    J. Koch et al., “Charge-insensitive qubit design derived from the Cooper pair box,” Phys. Rev. A, vol. 76, no. 4, p. 042319, Oct. 2007, doi: 10.1103/PhysRevA.76.042319

    doi:10.1103/physreva.76.042319 2007

  3. [3]

    black sheep

    + ⋯ = 4𝐸𝑐𝑛̂2 − ∑ 𝑛∞ 𝑛=1 𝐸𝐽 𝑛 𝑐𝑜𝑠 (𝜙 𝑛) Figure 1- schematic of Trainmon. It is important to note that Equation 2 is valid only under the condition 𝐸𝐽 𝑛/𝐸𝑐eff ≫ 1, where 𝐸𝑐eff is the total effective charging energy determined by the capacitance seen by the junctions [4]. Under this condition, phase slips can occur across the junctions, allowing the use of a...

    work page

  4. [4]

    Fluxonium: An Alternative Qubit Platform for High-Fidelity Operations,

    F. Bao et al., “Fluxonium: An Alternative Qubit Platform for High-Fidelity Operations,” Phys. Rev. Lett., vol. 129, no. 1, p. 010502, Jun. 2022, doi: 10.1103/PhysRevLett.129.010502

    doi:10.1103/physrevlett.129.010502 2022

  5. [5]

    Introduction to quantum electromagnetic circuits,

    U. Vool and M. Devoret, “Introduction to quantum electromagnetic circuits,” Int. J. Circuit Theory Appl., vol. 45, no. 7, pp. 897–934, 2017, doi: 10.1002/cta.2359

    doi:10.1002/cta.2359 2017

  6. [6]

    Millisecond Coherence in a Superconducting Qubit,

    A. Somoroff, Q. Ficheux, R. A. Mencia, H. Xiong, R. Kuzmin, and V. E. Manucharyan, “Millisecond Coherence in a Superconducting Qubit,” Phys. Rev. Lett., vol. 130, no. 26, p. 267001, Jun. 2023, doi: 10.1103/PhysRevLett.130.267001

    doi:10.1103/physrevlett.130.267001 2023

  7. [7]

    The superconducting quasicharge qubit,

    I. V. Pechenezhskiy, R. A. Mencia, L. B. Nguyen, Y.-H. Lin, and V. E. Manucharyan, “The superconducting quasicharge qubit,” Nature, vol. 585, no. 7825, pp. 368–371, Sep. 2020, doi: 10.1038/s41586-020-2687-9

    work page doi:10.1038/s41586-020-2687-9 2020

  8. [8]

    Coherence properties of the 0-π qubit,

    P. Groszkowski et al., “Coherence properties of the 0-π qubit,” New J. Phys., vol. 20, no. 4, p. 043053, Apr. 2018, doi: 10.1088/1367-2630/aab7cd

    work page doi:10.1088/1367-2630/aab7cd 2018

Show all 22 references
  1. [9]

    Magnifying Quantum Phase Fluctuations with Cooper-Pair Pairing,

    W. C. Smith et al., “Magnifying Quantum Phase Fluctuations with Cooper-Pair Pairing,” Phys. Rev. X, vol. 12, no. 2, p. 021002, Apr. 2022, doi: 10.1103/PhysRevX.12.021002

    2022 doi

  2. [10]

    Bifluxon: Fluxon-Parity-Protected Superconducting Qubit,

    K. Kalashnikov et al., “Bifluxon: Fluxon-Parity-Protected Superconducting Qubit,” PRX Quantum, vol. 1, no. 1, p. 010307, Sep. 2020, doi: 10.1103/PRXQuantum.1.010307

    2020 doi

  3. [11]

    Superconducting circuit protected by two-Cooper-pair tunneling,

    W. C. Smith, A. Kou, X. Xiao, U. Vool, and M. H. Devoret, “Superconducting circuit protected by two-Cooper-pair tunneling,” Npj Quantum Inf., vol. 6, no. 1, pp. 1–9, Jan. 2020, doi: 10.1038/s41534-019-0231-2

    2020 doi

  4. [12]

    Engineering Framework for Optimizing Superconducting Qubit Designs,

    F. Yan et al., “Engineering Framework for Optimizing Superconducting Qubit Designs,” Jun. 07, 2020, arXiv: arXiv:2006.04130. doi: 10.48550/arXiv.2006.04130

    2020 doi arXiv

  5. [13]

    Double-Fourier engineering of Josephson energy-phase relationships applied to diodes,

    A. M. Bozkurt, J. Brookman, V. Fatemi, and A. R. Akhmerov, “Double-Fourier engineering of Josephson energy-phase relationships applied to diodes,” SciPost Phys., vol. 15, no. 5, p. 204, Nov. 2023, doi: 10.21468/SciPostPhys.15.5.204

    2023 doi

  6. [14]

    Experimental Proof of Magnetic Flux Quantization in a Superconducting Ring,

    R. Doll and M. Näbauer, “Experimental Proof of Magnetic Flux Quantization in a Superconducting Ring,” Phys. Rev. Lett., vol. 7, no. 2, pp. 51–52, Jul. 1961, doi: 10.1103/PhysRevLett.7.51

    1961 doi

  7. [15]

    Experimental Evidence for Quantized Flux in Superconducting Cylinders,

    B. S. Deaver and W. M. Fairbank, “Experimental Evidence for Quantized Flux in Superconducting Cylinders,” Phys. Rev. Lett., vol. 7, no. 2, pp. 43–46, Jul. 1961, doi: 10.1103/PhysRevLett.7.43

    1961 doi

  8. [16]

    Symmetries and Collective Excitations in Large Superconducting Circuits,

    D. G. Ferguson, A. A. Houck, and J. Koch, “Symmetries and Collective Excitations in Large Superconducting Circuits,” Phys. Rev. X, vol. 3, no. 1, p. 011003, Jan. 2013, doi: 10.1103/PhysRevX.3.011003

    2013 doi

  9. [17]

    Universal Fast-Flux Control of a Coherent, Low-Frequency Qubit,

    H. Zhang et al., “Universal Fast-Flux Control of a Coherent, Low-Frequency Qubit,” Phys. Rev. X, vol. 11, no. 1, p. 011010, Jan. 2021, doi: 10.1103/PhysRevX.11.011010

    2021 doi

  10. [18]

    Scqubits: a Python package for superconducting qubits,

    P. Groszkowski and J. Koch, “Scqubits: a Python package for superconducting qubits,” Quantum, vol. 5, p. 583, Nov. 2021, doi: 10.22331/q-2021-11-17-583

    2021 doi

  11. [19]

    Computer-aided quantization and numerical analysis of superconducting circuits,

    S. P. Chitta, T. Zhao, Z. Huang, I. Mondragon-Shem, and J. Koch, “Computer-aided quantization and numerical analysis of superconducting circuits,” New J. Phys., vol. 24, no. 10, p. 103020, Nov. 2022, doi: 10.1088/1367-2630/ac94f2

    2022 doi

  12. [20]

    Decoherence in a superconducting quantum bit circuit,

    G. Ithier et al., “Decoherence in a superconducting quantum bit circuit,” Phys. Rev. B, vol. 72, no. 13, p. 134519, Oct. 2005, doi: 10.1103/PhysRevB.72.134519

    2005 doi

  13. [21]

    Noise and Decoherence in Quantum Two-Level Systems,

    A. Shnirman, Y. Makhlin, and G. Schön, “Noise and Decoherence in Quantum Two-Level Systems,” Phys. Scr., vol. 2002, no. T102, p. 147, Jan. 2002, doi: 10.1238/Physica.Topical.102a00147

    2002 doi

  14. [22]

    Dephasing of Solid-State Qubits at Optimal Points,

    Y. Makhlin and A. Shnirman, “Dephasing of Solid-State Qubits at Optimal Points,” Phys. Rev. Lett., vol. 92, no. 17, p. 178301, Apr. 2004, doi: 10.1103/PhysRevLett.92.178301. Appendix A One of the main advantages of Trainmon is solving the Hamiltonian in charge space, the Traim...

    2004 doi

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.