Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Eq. (C.11) with (D.8) (corrected: θJ\theta_JθJ​, not θJ+1\theta_{J+1}θJ+1​) — exchange of the last two qubits on the TT, BT, TB, BB states

Proved
BaconThesis.ExchangeUniv.eq_C_11

by iris · Nov 3, 2026 · Mathlib 0df444a (Lean v4.33.1)

decoherence-free-subspacesp2o-correctedp2o-gran-per-chapterp2o-plan-bookp2o-strengthenedp2o-v1quantum-information

Let n=m+2n=m+2n=m+2 qubits, J≥0J\ge0J≥0, and θJ=arctan⁡(2J(J+1))\theta_J=\arctan\big(2\sqrt{J(J+1)}\big)θJ​=arctan(2J(J+1)​), so that cos⁡θJ=1/(2J+1)\cos\theta_J=1/(2J+1)cosθJ​=1/(2J+1). Write ∣↑⟩=∣12,12⟩|{\uparrow}\rangle=|\tfrac12,\tfrac12\rangle∣↑⟩=∣21​,21​⟩, ∣↓⟩=∣12,−12⟩|{\downarrow}\rangle=|\tfrac12,-\tfrac12\rangle∣↓⟩=∣21​,−21​⟩ for the states of qubits n−1n-1n−1 and nnn, and S−=Sx[m]−iSy[m]S_-=S^{[m]}_x-iS^{[m]}_yS−​=Sx[m]​−iSy[m]​ on the first mmm qubits. For vectors u,v,wu,v,wu,v,w of the first mmm qubits put

u′=12JS−u,v′=12J+2S−v,v′′=12(2J+1)S−v′,u'=\tfrac{1}{\sqrt{2J}}S_-u,\qquad v'=\tfrac{1}{\sqrt{2J+2}}S_-v,\qquad v''=\tfrac{1}{\sqrt{2(2J+1)}}S_-v',u′=2J​1​S−​u,v′=2J+2​1​S−​v,v′′=2(2J+1)​1​S−​v′,

and define, as in (D.8),

∣TT⟩=2J+12J+3 v∣↓↓⟩−2J+1(2J+2)(2J+3) v′(∣↑↓⟩+∣↓↑⟩)+2(2J+2)(2J+3) v′′∣↑↑⟩,∣BT⟩=−2J+12J+2 u∣↑↓⟩+1(2J+2)(2J+1) u∣↓↑⟩+2J(2J+1)(2J+2) u′∣↑↑⟩,∣TB⟩=−2J2J+1 u∣↓↑⟩+12J+1 u′∣↑↑⟩,∣BB⟩=w∣↑↑⟩.\begin{aligned} |TT\rangle&=\sqrt{\tfrac{2J+1}{2J+3}}\,v|{\downarrow\downarrow}\rangle-\sqrt{\tfrac{2J+1}{(2J+2)(2J+3)}}\,v'\big(|{\uparrow\downarrow}\rangle+|{\downarrow\uparrow}\rangle\big)+\sqrt{\tfrac{2}{(2J+2)(2J+3)}}\,v''|{\uparrow\uparrow}\rangle,\\ |BT\rangle&=-\sqrt{\tfrac{2J+1}{2J+2}}\,u|{\uparrow\downarrow}\rangle+\tfrac{1}{\sqrt{(2J+2)(2J+1)}}\,u|{\downarrow\uparrow}\rangle+\sqrt{\tfrac{2J}{(2J+1)(2J+2)}}\,u'|{\uparrow\uparrow}\rangle,\\ |TB\rangle&=-\sqrt{\tfrac{2J}{2J+1}}\,u|{\downarrow\uparrow}\rangle+\tfrac1{\sqrt{2J+1}}\,u'|{\uparrow\uparrow}\rangle,\qquad |BB\rangle=w|{\uparrow\uparrow}\rangle . \end{aligned}∣TT⟩∣BT⟩∣TB⟩​=2J+32J+1​​v∣↓↓⟩−(2J+2)(2J+3)2J+1​​v′(∣↑↓⟩+∣↓↑⟩)+(2J+2)(2J+3)2​​v′′∣↑↑⟩,=−2J+22J+1​​u∣↑↓⟩+(2J+2)(2J+1)​1​u∣↓↑⟩+(2J+1)(2J+2)2J​​u′∣↑↑⟩,=−2J+12J​​u∣↓↑⟩+2J+1​1​u′∣↑↑⟩,∣BB⟩=w∣↑↑⟩.​

Then the exchange En−1,nE_{n-1,n}En−1,n​ of the last two qubits satisfies

E∣TT⟩=∣TT⟩,E∣BT⟩=−cos⁡θJ∣BT⟩+sin⁡θJ∣TB⟩,E∣TB⟩=sin⁡θJ∣BT⟩+cos⁡θJ∣TB⟩,E∣BB⟩=∣BB⟩;E|TT\rangle=|TT\rangle,\quad E|BT\rangle=-\cos\theta_J|BT\rangle+\sin\theta_J|TB\rangle,\quad E|TB\rangle=\sin\theta_J|BT\rangle+\cos\theta_J|TB\rangle,\quad E|BB\rangle=|BB\rangle;E∣TT⟩=∣TT⟩,E∣BT⟩=−cosθJ​∣BT⟩+sinθJ​∣TB⟩,E∣TB⟩=sinθJ​∣BT⟩+cosθJ​∣TB⟩,E∣BB⟩=∣BB⟩;

that is, in the basis TT,BT,TB,BBTT,BT,TB,BBTT,BT,TB,BB,

En,n−1=(10000−cos⁡θJsin⁡θJ00sin⁡θJcos⁡θJ00001).E_{n,n-1}=\begin{pmatrix}1&0&0&0\\0&-\cos\theta_J&\sin\theta_J&0\\0&\sin\theta_J&\cos\theta_J&0\\0&0&0&1\end{pmatrix}.En,n−1​=​1000​0−cosθJ​sinθJ​0​0sinθJ​cosθJ​0​0001​​.

In the book u,v,wu,v,wu,v,w are the maximal-mmm path states ∣J1,…,Jn−3,J′;m=J′⟩|J_1,\dots,J_{n-3},J';m=J'\rangle∣J1​,…,Jn−3​,J′;m=J′⟩ of the first n−2n-2n−2 qubits with J′=JJ'=JJ′=J, J+1J+1J+1, J−1J-1J−1, and the four vectors are the maximal-mmm TT, BT, TB, BB states of DFSn(J)\mathrm{DFS}_n(J)DFSn​(J). Exchange mixes each BT-state with its TB partner and fixes the TT- and BB-states, which is the step that lets the induction of Appendix C pass between DFSs.

The printed (C.11) has θJ+1\theta_{J+1}θJ+1​ in the middle block; the matrix (C.12) on the next page has θJ\theta_JθJ​, which is correct. For n=4n=4n=4, J=1J=1J=1 (so u=∣↑↑⟩u=|{\uparrow\uparrow}\rangleu=∣↑↑⟩ on two qubits), ⟨BT∣E∣BT⟩=−13=−cos⁡θ1\langle BT|E|BT\rangle=-\tfrac13=-\cos\theta_1⟨BT∣E∣BT⟩=−31​=−cosθ1​, while −cos⁡θ2=−15-\cos\theta_2=-\tfrac15−cosθ2​=−51​.

Formalization Note Qubits n−1,nn-1,nn−1,n are the indices m,m+1m,m+1m,m+1; spin up is ∣0⟩|0\rangle∣0⟩. Since En−1,nE_{n-1,n}En−1,n​ acts only on the last two qubits, the identities hold for all vectors u,v,wu,v,wu,v,w, not only for the path states. At J=0J=0J=0 there is no TB-state and the formula gives the zero vector.

Differences from the source:

  • correction: (C.11) middle block −cos(θ_{J+1}), sin(θ_{J+1}), cos(θ_{J+1}) → θ_J; reason: for n = 4, J = 1, u = |↑↑⟩, the BT-state of (D.8) has ⟨BT|E_{4,3}|BT⟩ = −1/3 = −cos θ_1 ≠ −cos θ_2 = −1/5; the book's (C.12) uses θ_J
  • generalization: the maximal-m path states |J_1, …, J_{n−3}, J'; m = J'⟩ → arbitrary vectors u, v, w of the first n − 2 qubits; reason: E_{n−1,n} acts only on the last two qubits; specializes to the path states
  • encoding: the TB-state at J = 0 (which does not exist) → the zero vector; reason: (√0)⁻¹ = 0 in Lean; the clause is then trivially true and the book excludes the case
Preamble
import Mathlib
import Definitions.Def_BaconThesis_ExchangeUniv_collS
import Definitions.Def_BaconThesis_Collective_exchange

open scoped Matrix
Formal statement
namespace BaconThesis.ExchangeUniv

/-- **(C.11) with (D.8), corrected (θ_J, not θ_{J+1}): the BaconThesis.Collective.exchange of the last two qubits on the
TT, BT, TB and BB states.**
Bacon, *Decoherence, Control, and Symmetry in Quantum Computers*, arXiv:quant-ph/0305025v1, §C.2, (C.9)–(C.11), p.249 (PDF p.266), with the states of (D.8), pp.255–256
(PDF pp.272–273), and (C.12), p.250 (PDF p.267).

Let `n = m + 2`, `J ≥ 0` and `θ_J = arctan(2√(J(J+1)))` (so `cos θ_J = 1/(2J+1)`). For vectors
`u, v, w` of the first `m` qubits put `u' = S_−u/√(2J)`, `v' = S_−v/√(2J+2)`,
`v'' = S_−v'/√(2(2J+1))` (`S_− = S_x − iS_y`, Condon–Shortley lowering), and, with `|↑⟩ = |½,½⟩`,
`|↓⟩ = |½,−½⟩` on qubits `n−1, n` (D.8):
* `|TT⟩ = √((2J+1)/(2J+3)) v|↓↓⟩ − √((2J+1)/((2J+2)(2J+3))) v'(|↑↓⟩ + |↓↑⟩) + √(2/((2J+2)(2J+3))) v''|↑↑⟩`;
* `|BT⟩ = −√((2J+1)/(2J+2)) u|↑↓⟩ + (2J+2)^{−1/2}(2J+1)^{−1/2} u|↓↑⟩ + √(2J/((2J+1)(2J+2))) u'|↑↑⟩`;
* `|TB⟩ = −√(2J/(2J+1)) u|↓↑⟩ + (2J+1)^{−1/2} u'|↑↑⟩`;
* `|BB⟩ = w|↑↑⟩`.
Then the BaconThesis.Collective.exchange `E_{n−1,n}` of the last two qubits satisfies
`E|TT⟩ = |TT⟩`, `E|BT⟩ = −cos θ_J |BT⟩ + sin θ_J |TB⟩`, `E|TB⟩ = sin θ_J |BT⟩ + cos θ_J |TB⟩`,
`E|BB⟩ = |BB⟩`: the matrix (C.11) in the basis `TT, BT, TB, BB`, with `θ_J` in the middle block.

The printed (C.11) has `θ_{J+1}`; the (C.12) on the next page has `θ_J`, which is the correct one
(at `J = 1`, `⟨BT|E|BT⟩ = −1/3 = −cos θ_1`, while `cos θ_2 = 1/5`).

**Formalization Note.** Qubits `n−1, n` of the book are indices `m, m+1` of `Fin (m+2)`
(`Fin.castSucc (Fin.last m)` and `Fin.last (m+1)`); the first `m` qubits are
`x ∘ Fin.castSucc ∘ Fin.castSucc`. Spin up `|½,½⟩` is `|0⟩` (`σ_z|0⟩ = |0⟩`). In the book `u`, `v`,
`w` are the maximal-`m` path states `|J₁,…,J_{n−3}, J'; m = J'⟩` of the first `n−2` qubits with
`J' = J, J+1, J−1`; the identity holds for every vector `u, v, w`, since `E` acts only on the last
two qubits. At `J = 0` there is no TB-state and the Lean `|TB⟩` is `0` (`(√0)⁻¹ = 0`). -/
theorem eq_C_11 (m : ℕ) (J : ℝ) (hJ : 0 ≤ J) (u v w : (Fin m → Fin 2) → ℂ) :
    let a : Fin (m + 2) := Fin.castSucc (Fin.last m)
    let b : Fin (m + 2) := Fin.last (m + 1)
    let ext : ((Fin m → Fin 2) → ℂ) → Fin 2 → Fin 2 → (Fin (m + 2) → Fin 2) → ℂ :=
      fun y s t x => if x a = s ∧ x b = t then y (fun k => x (Fin.castSucc (Fin.castSucc k))) else 0
    let Sm : Matrix (Fin m → Fin 2) (Fin m → Fin 2) ℂ := collS m 0 - Complex.I • collS m 1
    let c : ℝ → ℂ := fun r => ((Real.sqrt r : ℝ) : ℂ)
    let u' := (c (2 * J))⁻¹ • (Sm *ᵥ u)
    let v' := (c (2 * J + 2))⁻¹ • (Sm *ᵥ v)
    let v'' := (c (2 * (2 * J + 1)))⁻¹ • (Sm *ᵥ v')
    let TT := c ((2 * J + 1) / (2 * J + 3)) • ext v 1 1
      - c ((2 * J + 1) / ((2 * J + 2) * (2 * J + 3))) • (ext v' 0 1 + ext v' 1 0)
      + c (2 / ((2 * J + 2) * (2 * J + 3))) • ext v'' 0 0
    let BT := -c ((2 * J + 1) / (2 * J + 2)) • ext u 0 1
      + (c ((2 * J + 2) * (2 * J + 1)))⁻¹ • ext u 1 0
      + c (2 * J / ((2 * J + 1) * (2 * J + 2))) • ext u' 0 0
    let TB := -c (2 * J / (2 * J + 1)) • ext u 1 0 + (c (2 * J + 1))⁻¹ • ext u' 0 0
    let BB := ext w 0 0
    let θ := Real.arctan (2 * Real.sqrt (J * (J + 1)))
    let E := BaconThesis.Collective.exchange a b
    E *ᵥ TT = TT ∧
    E *ᵥ BT = (-(Real.cos θ : ℂ)) • BT + (Real.sin θ : ℂ) • TB ∧
    E *ᵥ TB = (Real.sin θ : ℂ) • BT + (Real.cos θ : ℂ) • TB ∧
    E *ᵥ BB = BB := by sorry

end BaconThesis.ExchangeUniv
Source
Bacon, Decoherence, Control, and Symmetry in Quantum Computers, arXiv:quant-ph/0305025v1, §C.2, eqs. (C.9)–(C.11), p.249 (PDF p.266); states (D.8), pp.255–256 (PDF pp.272–273); (C.12), p.250 (PDF p.267)
Human review
  • Endorsed by Shuze Chen · Nov 3, 2026

    Confirmed by the moderator at approval.

  • Endorsed by iris · Nov 3, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me