Physlib

PhyslibAlpha.Mathematics.LadderSystem.Irreducibility

Irreducibility of the excitation-number sector

i. Overview

`LadderSystem.vacuumSpan L Ω n` is `gl(d)`-irreducible: the only submodules of `vacuumSpan L Ω n` invariant under every `E i j` are `⊥` and `vacuumSpan L Ω n` itself. Completeness in an ambient Hilbert space is a separate analytic question. The proof uses linear algebra over a field of characteristic zero.

ii. Key results

- `LadderSystem.vacuumSpan_eq_of_ne_bot` : a nonzero `E i j`-invariant submodule of `vacuumSpan L Ω n` is all of `vacuumSpan L Ω n`.

iii. Table of contents

  • A. Moving a quantum between modes
  • B. Extracting a single basis vector from an invariant submodule
  • C. Connectivity: one basis vector reaches every other
  • D. Irreducibility

iv. References

A. Moving a quantum between modes

B. Extracting a single basis vector from an invariant submodule

C. Connectivity: one basis vector reaches every other

D. Irreducibility

14 declarations

definition

Moving a quantum from mode jj to mode ii in configuration α\alpha

Given a configuration of occupancy numbers α:{0,,d1}N\alpha: \{0, \dots, d-1\} \to \mathbb{N} (representing the number of quanta in each of dd modes) and indices i,j{0,,d1}i, j \in \{0, \dots, d-1\}, the function `moveOneTo` returns a new configuration α\alpha' representing the state after moving one quantum from mode jj to mode ii. The resulting configuration is defined by: α(c)={α(i)+1if c=iα(j)1if c=j and ciα(c)otherwise \alpha'(c) = \begin{cases} \alpha(i) + 1 & \text{if } c = i \\ \alpha(j) - 1 & \text{if } c = j \text{ and } c \neq i \\ \alpha(c) & \text{otherwise} \end{cases} In the intended case where iji \neq j, this corresponds to incrementing the occupancy of mode ii and decrementing the occupancy of mode jj.

theorem

Permutation equivalence of words under quantum transfer from mode jj to ii

Let dNd \in \mathbb{N} be the number of modes and α:{0,,d1}N\alpha : \{0, \dots, d-1\} \to \mathbb{N} be a configuration representing the number of quanta in each mode. Let countWord(d,α)\text{countWord}(d, \alpha) be the list of indices where each mode cc appears α(c)\alpha(c) times. For any two distinct modes i,j{0,,d1}i, j \in \{0, \dots, d-1\} such that iji \neq j, the list formed by prepending ii to countWord(d,α)\text{countWord}(d, \alpha) after removing one instance of jj is a permutation of the list countWord(d,α)\text{countWord}(d, \alpha'), where α\alpha' is the configuration obtained by moving one quantum from mode jj to mode ii.

theorem

Transfer formula for the action of EijE_{ij} on occupation states

Let VV be a vector space over a field KK with a ladder system LL and a vacuum vector ΩV\Omega \in V. For any occupation number configuration α:{0,,d1}N\alpha : \{0, \dots, d-1\} \to \mathbb{N}, let the corresponding basis vector be denoted by vα=L.word(countWord(d,α),Ω)v_\alpha = L.\text{word}(\text{countWord}(d, \alpha), \Omega). For any two distinct modes i,j{0,,d1}i, j \in \{0, \dots, d-1\}, the action of the operator Eij=aiajE_{ij} = a^\dagger_i a_j on vαv_\alpha is given by: Eijvα=α(j)vα E_{ij} v_\alpha = \alpha(j) \cdot v_{\alpha'} where α\alpha' is the configuration obtained by moving one quantum from mode jj to mode ii (i.e., α(i)=α(i)+1\alpha'(i) = \alpha(i) + 1, α(j)=α(j)1\alpha'(j) = \alpha(j) - 1, and α(c)=α(c)\alpha'(c) = \alpha(c) for ci,jc \neq i, j).

theorem

Invariance under an Endomorphism Implies Invariance under its Polynomials

Let VV be a module over a field KK, and let WVW \subseteq V be a submodule. If WW is invariant under an endomorphism MEndK(V)M \in \text{End}_K(V) (meaning M(w)WM(w) \in W for all wWw \in W), then for any polynomial pK[x]p \in K[x], the submodule WW is also invariant under the endomorphism p(M)p(M) obtained by evaluating pp at MM. That is, for every wWw \in W, p(M)wWp(M)w \in W.

theorem

Nonzero EijE_{ij}-invariant Submodule of vacuumSpan(L,Ω,n)\text{vacuumSpan}(L, \Omega, n) Contains a Basis Vector

Let VV be a vector space over a field KK equipped with a ladder system LL of dd modes and a vacuum vector Ω\Omega. Let vacuumSpan(L,Ω,n)\text{vacuumSpan}(L, \Omega, n) be the nn-th excitation-number sector, defined as the KK-linear span of all vectors formed by applying nn creation operators to Ω\Omega. Suppose WvacuumSpan(L,Ω,n)W \subseteq \text{vacuumSpan}(L, \Omega, n) is a nonzero submodule that is invariant under the action of the gl(d)\mathfrak{gl}(d) generators Eij=aiajE_{ij} = a^\dagger_i a_j for all i,j{0,,d1}i, j \in \{0, \dots, d-1\}. Then WW contains at least one occupation-number basis vector of the form vα=ai1ai2ainΩv_\alpha = a^\dagger_{i_1} a^\dagger_{i_2} \dots a^\dagger_{i_n} \Omega corresponding to a configuration α\alpha where c=0d1α(c)=n\sum_{c=0}^{d-1} \alpha(c) = n.

theorem

Total Occupation Number Equals the Length of the Canonical Word

Let dd be the number of modes and α:{0,,d1}N\alpha: \{0, \dots, d-1\} \to \mathbb{N} be a count function representing the occupancy of each mode. The total occupation number, given by the sum c=0d1α(c)\sum_{c=0}^{d-1} \alpha(c), is equal to the length of the canonical word countWord(d,α)\text{countWord}(d, \alpha) representing that configuration.

theorem

Moving a quantum between modes preserves total occupation number

Let α:{0,,d1}N\alpha: \{0, \dots, d-1\} \to \mathbb{N} be a configuration representing the occupancy numbers of dd modes. For any two distinct modes i,j{0,,d1}i, j \in \{0, \dots, d-1\} such that the source mode jj is not empty (α(j)0\alpha(j) \neq 0), the total occupation number remains invariant when one quantum is moved from mode jj to mode ii: c=0d1(moveOneTo(α,i,j))(c)=c=0d1α(c). \sum_{c=0}^{d-1} (\text{moveOneTo}(\alpha, i, j))(c) = \sum_{c=0}^{d-1} \alpha(c).

theorem

The occupancy of the target mode ii after `moveOneTo` is α(i)+1\alpha(i) + 1

Let α:{0,,d1}N\alpha: \{0, \dots, d-1\} \to \mathbb{N} be a configuration representing the number of quanta in each of dd modes. For any two modes ii and jj, let α=moveOneTo(α,i,j)\alpha' = \text{moveOneTo}(\alpha, i, j) be the configuration obtained by moving a quantum from mode jj to mode ii. The occupancy of the target mode ii in the new configuration is given by α(i)=α(i)+1\alpha'(i) = \alpha(i) + 1.

theorem

Moving a quantum from jj to ii and back is the identity on occupancy configurations

Let α:{0,,d1}N\alpha: \{0, \dots, d-1\} \to \mathbb{N} be a configuration of occupancy numbers representing the number of quanta in each of dd modes. For any two distinct modes i,j{0,,d1}i, j \in \{0, \dots, d-1\} where the occupancy of mode jj is non-zero (α(j)0\alpha(j) \neq 0), moving a quantum from mode jj to mode ii and then immediately moving a quantum from mode ii back to mode jj returns the system to its original configuration α\alpha. That is, moveOneTo(moveOneTo(α,i,j),j,i)=α \text{moveOneTo}(\text{moveOneTo}(\alpha, i, j), j, i) = \alpha

theorem

EijE_{ij}-invariant Submodules are Closed Under Moving a Quantum Between Modes

Let VV be a KK-vector space with a ladder system of dd modes and a vacuum vector Ω\Omega. Let Eij=aiajE_{ij} = a^\dagger_i a_j be the bilinear operators for i,j{0,,d1}i, j \in \{0, \dots, d-1\}. Suppose WVW \subseteq V is a submodule invariant under all EijE_{ij}. Let β:{0,,d1}N\beta: \{0, \dots, d-1\} \to \mathbb{N} be an occupancy configuration, and let β|\beta\rangle denote the state vector produced by applying the corresponding creation operators to Ω\Omega. If βW|\beta\rangle \in W and there exists a mode kk such that the occupancy β(k)>0\beta(k) > 0, then for any mode lkl \neq k, the state β|\beta'\rangle obtained by moving one quantum from mode kk to mode ll is also contained in WW.

theorem

αW    (N,0,,0)W|\alpha\rangle \in W \implies |(N, 0, \dots, 0)\rangle \in W for EijE_{ij}-invariant submodules WW

In a ladder system LL on a KK-vector space VV with d>0d > 0 modes and vacuum state Ω\Omega, let WVW \subseteq V be a submodule invariant under the gl(d)\mathfrak{gl}(d) generators Eij=aiajE_{ij} = a^\dagger_i a_j for all i,j{0,,d1}i, j \in \{0, \dots, d-1\}. If WW contains an occupation-number state α|\alpha\rangle (the state formed by applying creation operators to Ω\Omega according to the occupancy configuration α:{0,,d1}N\alpha: \{0, \dots, d-1\} \to \mathbb{N}), then WW must also contain the "hub" state (N,0,,0)|(N, 0, \dots, 0)\rangle, where N=c=0d1α(c)N = \sum_{c=0}^{d-1} \alpha(c) is the total number of excitations and all excitations are concentrated in mode 00.

theorem

Hub state W    \in W \implies all states of same excitation number W\in W for EijE_{ij}-invariant WW

Let VV be a vector space over a field KK with a ladder system LL of d>0d > 0 modes and a vacuum vector ΩV\Omega \in V. Let WVW \subseteq V be a submodule invariant under the action of the bilinear operators Eij=aiajE_{ij} = a^\dagger_i a_j for all i,j{0,,d1}i, j \in \{0, \dots, d-1\}. If WW contains the "hub" state with total excitation number nn, where all excitations are in the 00-th mode: (a0)nΩW, (a^\dagger_0)^n \Omega \in W, then WW contains every state with the same total excitation number nn. That is, for any occupation number distribution α:{0,,d1}N\alpha: \{0, \dots, d-1\} \to \mathbb{N} such that c=0d1α(c)=n\sum_{c=0}^{d-1} \alpha(c) = n, the corresponding state belongs to WW: (c=0d1(ac)α(c))ΩW. \left( \prod_{c=0}^{d-1} (a^\dagger_c)^{\alpha(c)} \right) \Omega \in W.

theorem

Connectivity of Occupation States with Equal Total Excitation Number (αc=αc)(\sum \alpha_c = \sum \alpha'_c)

Let LL be a ladder system on a KK-vector space VV with d>0d > 0 modes and vacuum vector Ω\Omega. Let WW be a submodule of VV that is invariant under the action of the bilinear operators Eij=aiajE_{ij} = a^\dagger_i a_j for all i,j{0,,d1}i, j \in \{0, \dots, d-1\}. Suppose α\alpha and α\alpha' are two occupation count functions from {0,,d1}\{0, \dots, d-1\} to N\mathbb{N} such that they have the same total excitation number: c=0d1α(c)=c=0d1α(c) \sum_{c=0}^{d-1} \alpha(c) = \sum_{c=0}^{d-1} \alpha'(c) If the state wα=(a0)α(0)(a1)α(1)(ad1)α(d1)Ωw_\alpha = (a_0^\dagger)^{\alpha(0)} (a_1^\dagger)^{\alpha(1)} \dots (a_{d-1}^\dagger)^{\alpha(d-1)} \Omega is contained in WW, then the state wα=(a0)α(0)(a1)α(1)(ad1)α(d1)Ωw_{\alpha'} = (a_0^\dagger)^{\alpha'(0)} (a_1^\dagger)^{\alpha'(1)} \dots (a_{d-1}^\dagger)^{\alpha'(d-1)} \Omega is also contained in WW.

theorem

gl(d)\mathfrak{gl}(d)-Irreducibility of the nn-th Excitation-Number Sector

Let VV be a vector space over a field KK with a ladder system LL of d>0d > 0 modes and a vacuum vector ΩV\Omega \in V. Let vacuumSpan(L,Ω,n)\text{vacuumSpan}(L, \Omega, n) denote the nn-th excitation-number sector, defined as the KK-linear submodule of VV spanned by all vectors of the form ai1ai2ainΩa^\dagger_{i_1} a^\dagger_{i_2} \dots a^\dagger_{i_n} \Omega. Suppose WW is a submodule of vacuumSpan(L,Ω,n)\text{vacuumSpan}(L, \Omega, n) that is invariant under the action of the bilinear operators Eij=aiajE_{ij} = a^\dagger_i a_j for all i,j{0,,d1}i, j \in \{0, \dots, d-1\}. If WW is not the zero submodule (W{0}W \neq \{0\}), then WW must be the entire excitation-number sector: W=vacuumSpan(L,Ω,n) W = \text{vacuumSpan}(L, \Omega, n) This implies that the nn-th excitation-number sector is irreducible under the action of the gl(d)\mathfrak{gl}(d) generators EijE_{ij}.