{"id":"79b6ca3d-db5f-4fb2-ac3d-7dae8150408f","arxiv_id":"2608.11160","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A systematic family of quantum codes realizes arbitrary fine Z-rotation gates transversally, generalizes code switching to rotated-surface-code geometries, and is piloted in a 45-qubit simulation of a Clifford proxy for the T magic state.","lead":"This paper constructs new quantum error-correcting code families that can implement tiny Z-axis rotations of encoded qubits directly, using the doubling trick extended to color codes and rotated surface codes. It also generalizes code switching, moving quantum information between two codes, and simulates a small 45-qubit demonstration, though the simulated state is a Clifford stand-in rather than a true magic state.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The local-geometry condition underpinning Theorem V.1 is asserted, not proved: the r-orthogonal surface-code families of Section IV require an explicit verification of Definition IV.2 before the code-switching protocol is justified.","rationale":"I focused on the strongest claim: the existence of r-orthogonal codes with rotated-surface-code local geometry and the transversal-CNOT code-switching theorem built on them. The reader's weakest assumption is the same one I would flag: Definition IV.2 is the load-bearing hypothesis of Theorem V.1, and its verification for the constructed families is sketched rather than demonstrated. Theorem III.8 itself appears internally coherent: the recursion S_r(k)=S_r(k-1)+2S_{r-1}(k), the closed-form sum, and the divisibility identity (III.4) make the claimed distance 2k-1 and the 2^r-divisibility of X-stabilizers go through when the roles of Q1 and Q2 in the doubling step are assigned correctly. The proof of Theorem V.1 is likewise sound once Definition IV.2 and the logical-operator representative conditions hold; the only underived assumptions are the stabilizer-space identities for the constructed families. I am not treating the S|+>-as-proxy-for-T|+> simulation as the load-bearing issue: it is a real overstatement of a demonstration, but it does not break the mathematical construction. Similarly, comparisons with previously known code families are external-consensus matters, not internal correctness risks. The paper does provide explicit stabilizer matrices and closed-form parameters for small instances, which makes the proposed computational verification straightforward and decisive for the claimed examples; if those pass, the central concern is a proof gap that can be closed, not a demonstrated failure.","tokens_in":38386,"tokens_out":19095,"duration_ms":167054,"concrete_test":"For the Corollary IV.6 instances [[39,1,3]] (r=3,k=2) and [[177,1,5]] (r=3,k=3), explicitly build the stabilizer generator matrix G from the recursion in Theorem IV.5 and the rotated-surface-code stabilizer matrix S_d with d=2k-1. Then check: (i) every Schur product of up to r rows of G has even weight; (ii) puncturing G to the first d^2 coordinates generates exactly the X-stabilizer code of the [[d^2,1,d]] rotated surface code; (iii) the Z-stabilizers of the larger code with support in those d^2 coordinates coincide with the Z-stabilizer code of the surface code; and (iv) with logical X=(1_{d^2},0,...), the logical operators satisfy property (3) of Theorem V.1. If any check fails, Theorem V.1 does not apply to that family; if all pass, the claimed small-distance instances are real and the remaining issue is proof completeness rather than correctness.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim depends on Theorem V.1, whose hypothesis is that Q1 has the local geometry of Q2 in the sense of Definition IV.2. All downstream applications — the r-orthogonal surface-geometry families, the transversal-CNOT code switching, and the 45-qubit magic-state-preparation simulation — rest on this condition. The proof of Theorem IV.5 says only that 'one can easily verify' the recursive stabilizer matrix G is r-orthogonal and does not explicitly verify Definition IV.2. Theorem IV.7's proof in Appendix D establishes self-orthogonality and distance via a symplectic-basis argument, but it never checks that puncturing the X-stabilizer space of the larger code to the first d^2 coordinates yields the surface-code X-stabilizers, and that the Z-stabilizers supported on those coordinates coincide with the surface-code Z-stabilizers. These are not cosmetic details: the transversal CNOT of Theorem V.1 requires the full stabilizer-space match, not merely the displayed generator rows, and also requires the logical-operator representatives used in its proof (property (3)). If Definition IV.2 fails for a claimed family, the transversal CNOT is not a valid logical gate between the constructed code and the surface code, and the code-switching protocol collapses. The construction is plausible — the block form of G and logical X=(1,0,0) point the right way — but the missing verification is exactly the load-bearing point. No internal contradiction is apparent; the gap is an omitted proof, not an observed counterexample.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops recursive 'doubling' constructions of CSS codes whose X-stabilizers have high divisibility, producing single-logical-qubit color-code families with closed-form lengths S_r(k) and transversal logical R_Z(π/2^{r-1}) gates. It then adapts the construction to r-orthogonal codes that are claimed to have the local geometry of rotated surface codes, and uses these in a transversal-CNOT code-switching protocol to prepare logical R_Z(θ) states. The final section reports a circuit-level simulation of S|+> preparation in a distance-three rotated surface code using 45 physical qubits.","tokens_in":38694,"tokens_out":8676,"duration_ms":79051,"significance":"If the local-geometry claims are fully proved, the paper would materially extend code switching beyond color codes: it gives systematic closed-form families for arbitrary fine Z-rotations, improves several tabulated parameters, and provides concrete low-footprint candidates. The algebraic recurrences in Lemmas B.1 and B.2, the parameter tables, and the explicit stabilizer matrices for the [[15,1,3]] and [[31,1,3]] codes are valuable, reproducible resources. At present the advertised applications outrun the proofs: the local-geometry condition on which Theorem V.1 rests is asserted rather than demonstrated for the main constructions, and the simulated 'magic state' is actually the Clifford state S|+>, so the headline claims need qualification.","major_comments":[{"comment":"The proof of Theorem IV.5 in Appendix C concludes that 'one can easy verify' the recursive matrix G forms an r-orthogonal quantum code, but it does not verify the local-geometry condition of Definition IV.2 for the constructed family. This condition is the load-bearing hypothesis of the paper's main application: Theorem V.1 requires that restricting the X-stabilizer space of Q1 to the first n2 coordinates gives exactly the X-stabilizer space of Q2, and that the Z-stabilizers of Q1 supported on those coordinates coincide with the Z-stabilizer space of Q2. Displaying generator rows is not sufficient, because the punctured stabilizer subgroup may be larger than the image of the listed generators. If this condition fails for a claimed family, the transversal CNOT in Theorem V.1 is not a valid logical gate between that family and the rotated surface code, so the code-switching protocol collapses. Please supply a complete verification of Definition IV.2 for the r-orthogonal families, including the distance and logical-operator representatives.","section":"§IV, Theorem IV.5 and Appendix C"},{"comment":"The symplectic-basis argument in Appendix D establishes self-orthogonality and lower-bounds the distance, but it never checks the two structural equalities required by Definition IV.2: puncturing the X-stabilizer space of Q to the first d^2 coordinates yields the surface-code X-stabilizer space, and the Z-stabilizers of Q supported on those coordinates coincide with the surface-code Z-stabilizers. These equalities are needed for the code-switching circuit of Theorem V.1, not merely for the parameter count. The proof also does not explicitly show that the logical X operator (1_{d^2},0,...,0) and the logical Z operators of the surface code survive the puncturing/shortening rules. Please provide an explicit verification for the DSPS construction and for the triorthogonal codes built from it.","section":"§IV, Theorem IV.7 and Appendix D"},{"comment":"The statement of Theorem V.1 does not include the assumption, introduced only in item (3) of the proof, that the logical X and Z operators of Q1 and Q2 act transversally on the first n2 qubits. The local-geometry definition alone does not imply this property, and the proof uses it essentially to show that the transversal CNOT acts as a logical CNOT on the direct-sum code. As written, the theorem is not self-contained, and applying it to a pair of codes that satisfies only Definition IV.2 may not be justified. The theorem should either be restated with this hypothesis, or the property should be proved for each construction to which the theorem is applied.","section":"§V, Theorem V.1"},{"comment":"The abstract and Section V advertise a 'fault-tolerant magic state preparation' demonstration and claim that the 45-qubit simulation 'validates the complete fault-tolerant implementation' of the proposed code-switching scheme. However, Section V.B explicitly replaces T|+> with S|+>, which is a Clifford state and not a non-Clifford magic state. The simulation therefore does not demonstrate magic-state preparation or a non-Clifford gate. Please reword these claims so that the Clifford proxy is stated as such in the abstract and conclusion, or provide an actual T|+> simulation for the same protocol.","section":"Abstract and §V.B"}],"minor_comments":[{"comment":"The proof states that the doubling construction yields a code with parameters [[S_r(k−1)+2S_{r−1}(k),1,2k]]; the distance should be 2k−1, not 2k. Please correct this typo.","section":"Appendix B, proof of Theorem III.8"},{"comment":"The proof writes S2(k) = (2k−1)(2k)(2k+1)/3, but the theorem states a length d(d+1)(d+2)/3 − 1. The missing '−1' makes the displayed formula inconsistent with the stated parameters.","section":"Appendix C, proof of Theorem IV.4"},{"comment":"The exponential phase 'e^{iπ/θ}' in Eq. (V.2) appears to be a typo; the expected phase is e^{iθ} (up to a global phase convention). As printed, the expression is dimensionally inconsistent.","section":"§V, Theorem V.1, Eq. (V.2)"},{"comment":"The text says the complete circuit uses 40 data qubits together with four ancilla qubits (three for the [[31,1,3]] code and one for the surface code), while Figure 12 and Figure 13 state that five ancilla qubits are used. This discrepancy in the 45-qubit count should be resolved.","section":"§V.B and Figure 13"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within scope and the main constructions are plausible; the missing local-geometry verification in Section IV is likely repairable, so I recommend major revision rather than rejection. Please ensure the advertised 'magic state preparation' claim is corrected to the Clifford proxy S|+> before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The color-code half of this paper is solid and worth engaging with. The recursive doubling construction gives closed-form families of codes that support logical Z-rotations at arbitrary Clifford-hierarchy levels, and the parameter formulas and divisibility proofs in Appendix B check out. The small-distance instances matching known optimal codes ([7,1,3], [17,1,5], etc.) lend credibility, and the generalized code-switching theorem (Theorem V.1) is a clean formal statement: given two codes satisfying the local-geometry definition, a transversal CNOT does what you want. The 45-qubit simulation is a nice engineering demonstration, and the authors are honest in Section V.B that they prepare S|+>, a Clifford proxy, not T|+>. That honesty is undercut, though, by the abstract and conclusion calling it \"fault-tolerant magic state preparation.\" That is an overstatement and should be fixed.\n\nThe real soft spot is Definition IV.2. The proofs of Theorems IV.5 and IV.7 never actually verify the two rules—puncturing the X-stabilizer space to the first d^2 coordinates yields the surface-code X-stabilizers, and the Z-stabilizers supported on those coordinates match as well. Theorem IV.5 says \"one can easily verify\"; the Appendix C proof shows r-orthogonality and the parameter formula but not the geometry preservation. The symplectic-basis proof in Appendix D establishes self-orthogonality and distance but likewise skips the puncturing/shortening check. The stress-test note is right: this is not a cosmetic gap. Theorem V.1 requires the full stabilizer-space match, and if Definition IV.2 fails for any claimed family, the transversal CNOT is not a valid logical gate and the switching protocol collapses. The block structure makes the condition plausible—I see no contradiction—but it remains an omitted proof at the load-bearing point. The single-shot decoding claim is similarly thin: the paper describes meta-checks that could protect syndromes, but provides no theorem or simulation showing single-shot fault tolerance.\n\nThis paper deserves a serious referee. The color-code results and the generalized switching theorem are publishable after the overstatements are corrected. The surface-geometry families need an explicit verification of Definition IV.2 before the code-switching application can be regarded as established. I would send it to review, with instructions to the authors to either prove the geometry preservation or weaken the claims.","headline":"A useful recursive code construction with an over-sold surface-code extension: the local-geometry condition underpinning the main code-switching claim is asserted, not proved, and the simulation prepares a Clifford state, not a magic state.","tokens_in":39230,"tokens_out":2550,"would_cite":true,"duration_ms":25725,"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":"This paper constructs quantum code families whose transversal gates include logical Z-rotations at arbitrary levels of the Clifford hierarchy, and shows these codes switch fault-tolerantly with rotated surface codes.","keywords":["quantum codes","code switching","color codes","surface codes","transversal gates","Clifford hierarchy","self-orthogonal codes","triorthogonal codes"],"falsifier":"Take one of the claimed triorthogonal codes with local surface-code geometry, such as the [[31,1,3]] code, and explicitly list its X- and Z-stabilizer generators; apply Definition IV.2's puncturing rule to the first nine qubits and the shortening rule to the Z-stabilizers supported there, and verify that the resulting check matrices are exactly those of the [[9,1,3]] rotated surface code and that the logical X and Z operators commute in the required way. If the punctured picture gives a different code or a logical algebra of the wrong dimension, the claimed transversal CNOT switch would not implement the desired logical operation.","tokens_in":38182,"feed_emoji":"⚛️","tokens_out":11275,"duration_ms":98028,"temperature":0.7,"pith_summary":"The paper sets out to show that one recursive tool, the doubling construction, can generate quantum error-correcting codes that realize any arbitrarily small logical Z-rotation gate transversally, at any level of the Clifford hierarchy. It claims explicit closed-form families with parameters $[\\![ S_r(k),1,2k-1 ]\\!]$ whose X-stabilizers are all $2^r$-divisible, so that transversal physical $R_Z(\\pi/2^{r-1})$ gates implement the corresponding logical gate. The same framework, seeded with rotated surface codes, produces r-orthogonal codes that keep the local geometry of rotated surface codes, and a transversal CNOT between such a code and the smaller surface code implements fault-tolerant code switching for arbitrary Z-rotation gates. A sympathetic reader would care because this moves code switching beyond the color-code/T-gate setting, offering a concrete path to Clifford+$R_Z(\\theta)$ computation on surface-code architectures. The paper's 45-qubit simulation of $S|+\\rangle$ preparation in a distance-three surface code is offered as evidence that the protocol is already practically testable.","feed_headline":"Doubling yields quantum codes for ever-finer Z-rotation gates","feed_subtitle":"The same construction enables fault-tolerant code switching in rotated surface codes, demonstrated on 45 physical qubits.","key_machinery":"The machinery is the doubling construction, which combines a self-orthogonal seed code $Q_1$ and a smaller orthogonal or triorthogonal code $Q_2$ into a larger code whose X-stabilizer generators have a block structure assembled from $Q_1$'s stabilizer matrix $E_1$, $Q_2$'s stabilizer matrix $E_2$, and one connector row that is all ones on the two copies of $Q_1$ and $Q_2$. The paper shows this operation preserves or improves $2^r$-divisibility, so iterating it lifts the transversal gate one level up the Clifford hierarchy each time, and the length recursion $S_r(k)=2\\sum_{i=1}^k S_{r-1}(i)-1$ evaluates to the closed binomial sums above. For the surface-code version, the seed is a rotated surface code plus a punctured surface code, and Definition IV.2's puncturing and shortening rules are what certify that the larger code has the local geometry of the surface code; that geometry is exactly what makes the transversal CNOT of Theorem V.1 a valid logical operation between the two codes.","core_discovery":"The central discovery is that the doubling construction, previously used to build triorthogonal codes with a logical T gate, iterates to produce $2^r$-divisible quantum codes with a transversal logical $R_Z(\\pi/2^{r-1})$ gate for every $r \\ge 2$. Theorem III.8 gives the color-code family $[\\![ S_r(k),1,2k-1 ]\\!]$ with $S_r(k)=\\sum_{i=0}^r 2^i \\binom{k+i-2}{i}$, whose X-stabilizer weights are divisible by $2^r$. Theorem IV.5 delivers the companion r-orthogonal family with parameters $[\\![ S_r(k),1,2k-1 ]\\!]$ for $S_r(k)=2^{r+2}\\binom{k+r-1}{r+1}+\\sum_{i=0}^{r-1}2^i\\binom{k+i-2}{i}$, and with the local geometry of rotated surface codes of distance $2k-1$. Theorem V.1 then shows that whenever one code has another code's local geometry, a transversal CNOT performs fault-tolerant code switching, so the magic state $|R_Z(\\theta)\\rangle$ can be teleported from the large code to the surface code. The paper also constructs optimized codes by concatenating rotated surface codes with punctured ones, yielding e.g. triorthogonal [[31,1,3]] and [[113,1,5]] codes, and demonstrates the complete switching protocol for $S|+\\rangle$ in a distance-three surface code using 45 physical qubits in a circuit-level simulation.","pith_inferences":["Editorial inference: if the doubling recursion works for any self-orthogonal seed with a prescribed layout, the same scheme should produce geometry-preserving code families for other topological templates such as toric or hyperbolic codes; the paper only demonstrates color-code and rotated-surface-code seeds.","Editorial inference: the 45-qubit demonstration uses $S|+\\rangle$ as a stand-in for $T|+\\rangle$ and a flag-based post-selection pipeline, so the true end-to-end T-state performance, including the cost of discards and decoding, remains untested; a natural next experiment is the same protocol with the actual T gate at $d=3$ and $d=5$.","Editorial inference: the meta-check argument for single-shot Z-syndrome decoding is developed for the doubled color-code family; transplanting it to the r-orthogonal surface-code-geometry codes would require a decoder that consumes redundant Z-syndromes, which the paper does not simulate.","Editorial inference: if the code switching protocol's resource advantage persists beyond distance three, Clifford+$R_Z(\\theta)$ compilation for algorithms like multi-controlled Toffoli networks could cut non-Clifford depth substantially, matching the synthesis-cost trend the paper cites for small-angle rotations."],"forward_implications":["For every $r \\ge 2$ and odd distance $d = 2k-1$ there exists a code of explicit length $S_r(k)$ whose transversal gate set includes the logical Z-rotation $R_Z(\\pi/2^{r-1})$.","The color-code members of the family use fewer data qubits than the traditional, capped, doubled, and stacked color-code families at the same distance and transversal gate (Table I and Figure 5).","The construction yields triorthogonal codes with the local geometry of rotated surface codes, including [[31,1,3]] and [[113,1,5]], so T-gate switching is possible in surface-code-compatible hardware.","The code switching protocol generalizes to any code pair satisfying the local-geometry condition, giving fault-tolerant S, T, or finer Z-rotation magic states without leaving the rotated surface code layout.","The circuit-level simulation of the 45-qubit distance-three protocol prepares a verified $S|+\\rangle$ state with conditional logical error near $10^{-5}$ at physical error rate $10^{-4}$."],"supporting_citations":[{"why":"Supplies the doubling construction of doubled color codes that is the paper's central recursive tool.","marker":"[40]"},{"why":"Provides the short-length transversal Clifford and T-gate codes whose parameters the new families improve upon.","marker":"[41]"},{"why":"Supplies the triorthogonal-code doubling framework and asymptotically good CSS-T construction that this paper extends.","marker":"[42]"},{"why":"Defines triorthogonal matrices and the transversal-T code construction that motivates the r-orthogonal generalization.","marker":"[17]"},{"why":"Characterizes CSS codes optimal for transversal T, the baseline against which the new families' parameters and gate realization are assessed.","marker":"[39]"},{"why":"States the Schur-product divisibility conditions for transversal logical phase gates that ground Theorem II.5.","marker":"[59]"},{"why":"Establishes the one-way transversal-CNOT code switching protocol that Theorem V.1 generalizes.","marker":"[5]"},{"why":"Introduces gauge-fixing and code switching between color codes, the setting this paper moves beyond to rotated surface codes.","marker":"[26]"},{"why":"Defines rotated surface codes and their local geometry, which the r-orthogonal families are required to inherit.","marker":"[78]"},{"why":"Provides the circuit-level Monte Carlo sampling engine used for the 45-qubit simulation.","marker":"[86]"}],"fun_headline_variants":["Doubling yields quantum codes for arbitrary Z-rotation gates","Code switching on 45 qubits realizes arbitrary Z-gates","Fault-tolerant magic states via code switching in surface codes","45-qubit demo: fault-tolerant code switching for Z-rotations","Finer Z-rotations and surface-code switching from the doubling trick"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the constructed r-orthogonal codes genuinely possess the local geometry of rotated surface codes required by Definition IV.2, a condition the paper asserts can be 'easily verified' and supports only with a sketched proof; if that geometry fails for a claimed family, the transversal CNOT in Theorem V.1 would not be a valid logical gate between the codes.","fun_headline_variants_meta":{"raw":{"variants":["Doubling yields quantum codes for arbitrary Z-rotation gates","Code switching on 45 qubits realizes arbitrary Z-gates","Fault-tolerant magic states via code switching in surface codes","45-qubit demo: fault-tolerant code switching for Z-rotations","Finer Z-rotations and surface-code switching from the doubling trick"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001343,"raw_usage":{"total_tokens":5586,"prompt_tokens":1200,"completion_tokens":4386,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":816,"completion_tokens_details":{"reasoning_tokens":4298}},"tokens_in":816,"tokens_out":4386,"duration_ms":27481,"temperature":1.0,"reasoning_tokens":4298,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T04:59:01.062922+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take one of the claimed triorthogonal codes with local surface-code geometry, such as the [[31,1,3]] code, and explicitly list its X- and Z-stabilizer generators; apply Definition IV.2's puncturing rule to the first nine qubits and the shortening rule to the Z-stabilizers supported there, and verify that the resulting check matrices are exactly those of the [[9,1,3]] rotated surface code and that the logical X and Z operators commute in the required way. If the punctured picture gives a different code or a logical algebra of the wrong dimension, the claimed transversal CNOT switch would not implement the desired logical operation.","supporting_citations":[{"cited_title":"Measurement-free code-switching pro- tocol for low-overhead quantum computation us- ing permutation-invariant codes.PRX Quantum, 6(4):040341, 2025","cited_arxiv_id":null,"evidence_quote":"Provides the short-length transversal Clifford and T-gate codes whose parameters the new families improve upon."},{"cited_title":"Exact topo- logical quantum order in d= 3 and beyond: Branyons and brane-net condensates.Physi- cal Review B—Condensed Matter and Materials Physics, 75(7):075103, 2007","cited_arxiv_id":null,"evidence_quote":"Supplies the triorthogonal-code doubling framework and asymptotically good CSS-T construction that this paper extends."},{"cited_title":"Magic-state distillation with low overhead.Physical Re- view A—Atomic, Molecular, and Optical Physics, 86(5):052329, 2012","cited_arxiv_id":null,"evidence_quote":"Introduces gauge-fixing and code switching between color codes, the setting this paper moves beyond to rotated surface codes."},{"cited_title":"Quan- tum css duadic and triadic codes: New insights and properties","cited_arxiv_id":null,"evidence_quote":"Defines rotated surface codes and their local geometry, which the r-orthogonal families are required to inherit."},{"cited_title":"Decoding algorithms for surface codes","cited_arxiv_id":null,"evidence_quote":"Provides the circuit-level Monte Carlo sampling engine used for the 45-qubit simulation."}],"review_version":1}