Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

The simplest 2D quantum walk detects chaoticity

C. Alonso-Lobo Alonso.lobo.C@gmail.com Grupo de Sistemas Complejos, Escuela Técnica Superior de Ingenería Agronómica, Agroambiental y de Biosistemas, Universidad Politécnica de Madrid, Avenida Puerta de Hierro 2-4, 28040 Madrid, Spain    Gabriel G. Carlo g.carlo@conicet.gov.ar CONICET, Comisión Nacional de Energía Atómica, Avenida del Libertador 8250, 1429 Buenos Aires, Argentina    F. Borondo f.borondo@uam.es Departamento de Química, Universidad Autónoma de Madrid, CANTOBLANCO-28049 Madrid, Spain
(January 23, 2025)
Abstract

Quantum walks have been actively studied from many perspectives, mainly from the statistical physics and quantum information points of view. We here determine the influence of basic chaotic features on the walker behavior. We consider an extremely simple model consisting of alternate one-dimensional walks along the two spatial coordinates of bidimensional closed domains (hard wall billiards). The chaotic or regular behavior that the shape of the boundary induces in the deterministic classical equations of motion and that translates into chaotic signatures for the quantized problem also results in sharp differences for the spectral statistics and morphology of the eigenfunctions of the quantum walker. Unexpectedly, two different quantum mechanical problems share the same kind of features related to the corresponding classical dynamics of one of them.

Quantum Walks,Quantum Chaos

I Introduction

Quantum Walks (QWs), originally proposed by Aharonov [1], have become an important and very active area of physics. This interest comes from their efficient way of capturing diffusive properties, which in turn provided many applications in quantum information [2, 3, 4] and quantum optics [5, 6]. In particular, they are suitable to explain and build search algorithms [7], like for example in 2D grids [8]. Moreover, they have been implemented in several experiments [9, 10, 11]. In their simplest form, QWs are quantum counterparts of the well-known classical problem of the random walker on the line, which takes a step to the right or left depending on the outcome of a coin toss. The straightforward quantum-mechanical counterpart can be thought of as a spin particle that moves to the right or left. This displacement depends on the spin state whose evolution is given by the so-called quantum coin, the analogue of the classical coin toss.

Given the relevance of QWs in finite systems applications, studying the effects of boundaries is of the utmost importance [12, 13]. Also, it is interesting to investigate their behavior with respect to the complexity of the landscape in which the walker is moving. For instance, could a minimal QW model reveal quantum signatures of chaos in the same way as the quantized models directly derived from the Hamilton’s classical equations of motion do? Considering the different nature of the dynamics involved, this is a challenging question that we want to address in this work. In fact, QWs are not direct quantizations, in the usual sense, of deterministic classical models.

In this respect, the quantum chaos arena [14] is where the signatures of classical chaos in the quantum realm have been traditionally studied. This area of physics provided many important findings, one of the most celebrated results being the so-called Bohigas-Giannoni-Schmit (BGS) [15] conjecture, which has so far been thoroughly tested. It prescribes a random matrix behavior for quantized chaotic Hamiltonian systems. Notably, this subject has become very active today. In particular, the well known level repulsion derived from BGS, generalized to open systems under the name of the Grobe-Haake-Sommers (GHS) conjecture [16], has been recently challenged in [17]. Another well studied phenomenon like the localization on marginally stable orbits of elsewhere chaotic systems – for example on bouncing ball trajectories – constitute a hallmark of quantum chaos.

Hence, our main point here is to borrow some concepts and tools from quantum chaos in order to discover the consequences that a dynamically chaotic domain has on QWs. For that purpose, we consider a very simple model consisting of a 2D QW constrained by a hard wall. A deterministic classical free particle moving inside will show regular or chaotic dynamics depending on the domain shape. We take the rectangle and the paradigmatic Bunimovich stadium billiards as examples of both behaviors, respectively. We conclude that a streamlined QW model consisting only of a single spin and alternate and independent movements along each coordinate perceives the different nature that these boundaries have at the classical level. This is reflected in the spectral behavior and also by the contrasting shapes of the position probability distributions corresponding to the eigenfunctions of the unitary evolution operator.

This paper is organized as follows: in Sec. II we describe our QW model, in Sec. III we show the results where we focus on the spectral behavior and the morphology of the spatial part of the eigenfunctions of the evolution operator. Finally, in Sec. IV we present our conclusions and outline some possible future developments.

II QW model

Refer to caption
Figure 1: (Color online) Desymmetrized Bunimovich stadium billiard. Some integers (m,n)𝑚𝑛(m,n)( italic_m , italic_n ) specifying the particle position in a grid inside the billiard, and the shape functions used in the evolution operator are shown for reference. See Sec. II for details.

Our QW model is defined as a spin 1/2121/21 / 2 particle moving inside a 2D billiard, whose state at any (discrete) time is

|Ψ(t)=m,nUmn(t)|m,n,u+Dmn(t)|m,n,d,ketΨ𝑡subscript𝑚𝑛superscriptsubscript𝑈𝑚𝑛𝑡ket𝑚𝑛𝑢superscriptsubscript𝐷𝑚𝑛𝑡ket𝑚𝑛𝑑\displaystyle|\Psi(t)\rangle=\sum_{m,n}U_{m}^{n}(t)\,|m,n,u\rangle+D_{m}^{n}(t% )\,|m,n,d\rangle,| roman_Ψ ( italic_t ) ⟩ = ∑ start_POSTSUBSCRIPT italic_m , italic_n end_POSTSUBSCRIPT italic_U start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n end_POSTSUPERSCRIPT ( italic_t ) | italic_m , italic_n , italic_u ⟩ + italic_D start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n end_POSTSUPERSCRIPT ( italic_t ) | italic_m , italic_n , italic_d ⟩ , (1)

where Umnsuperscriptsubscript𝑈𝑚𝑛U_{m}^{n}italic_U start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n end_POSTSUPERSCRIPT and Dmnsuperscriptsubscript𝐷𝑚𝑛D_{m}^{n}italic_D start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n end_POSTSUPERSCRIPT represent the probability amplitudes of the particle located at position (m,n)𝑚𝑛(m,n)( italic_m , italic_n ), with (m,n)𝑚𝑛(m,n)( italic_m , italic_n ) integers defining the position in a grid inside the billiard. Such particle can have either spin up, u𝑢uitalic_u or spin down, d𝑑ditalic_d. The evolution is given by unitary operators that act on the spin space (the coin operator reflecting the coin toss) and on the position space (the walk operator reflecting the step taken by the walker conditional on the spin state). Additionally, there is a spin flip each time the walker reaches the border of the billiard. As an explicit motivation for this model, we can view the previous evolution as that corresponding to the motion of an electron inside a cavity. We can think of this model as consisting of two independent QWs, except for the fact that both share the same spin. The unbounded version is usually called Alternate QW (AQW) [18, 19].

Our bounded model in a rectangular billiard is implemented in the following way. We initially apply a SU(2) coin operator of the form

C^1=(cosαsinαsinαcosα),subscript^𝐶1matrix𝛼𝛼𝛼𝛼\hat{C}_{1}=\left(\begin{matrix}\cos{\alpha}&\sin{\alpha}\\ -\sin{\alpha}&\cos{\alpha}\end{matrix}\right),over^ start_ARG italic_C end_ARG start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT = ( start_ARG start_ROW start_CELL roman_cos italic_α end_CELL start_CELL roman_sin italic_α end_CELL end_ROW start_ROW start_CELL - roman_sin italic_α end_CELL start_CELL roman_cos italic_α end_CELL end_ROW end_ARG ) , (2)

acting on the spin space. Then, we proceed with a vertical displacement operator (acting solely in that direction) given by

Wn^^subscript𝑊𝑛\displaystyle\hat{W_{n}}over^ start_ARG italic_W start_POSTSUBSCRIPT italic_n end_POSTSUBSCRIPT end_ARG =\displaystyle== 0nU1|n+1n||UU|superscriptsubscript0subscript𝑛𝑈1tensor-productket𝑛1bra𝑛ket𝑈bra𝑈\displaystyle\sum_{0}^{n_{U}-1}|n+1\rangle\langle n|\otimes|U\rangle\langle U|∑ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT - 1 end_POSTSUPERSCRIPT | italic_n + 1 ⟩ ⟨ italic_n | ⊗ | italic_U ⟩ ⟨ italic_U | (3)
+\displaystyle++ 1nU|n1n||DD|+|0,U0,D|superscriptsubscript1subscript𝑛𝑈tensor-productket𝑛1bra𝑛ket𝐷bra𝐷ket0𝑈bra0𝐷\displaystyle\sum_{1}^{n_{U}}|n-1\rangle\langle n|\otimes|D\rangle\langle D|+|% 0,U\rangle\langle 0,D|∑ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT end_POSTSUPERSCRIPT | italic_n - 1 ⟩ ⟨ italic_n | ⊗ | italic_D ⟩ ⟨ italic_D | + | 0 , italic_U ⟩ ⟨ 0 , italic_D |
+\displaystyle++ |nU,DnU,U|.ketsubscript𝑛𝑈𝐷brasubscript𝑛𝑈𝑈\displaystyle|n_{U},D\rangle\langle n_{U},U|.| italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT , italic_D ⟩ ⟨ italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT , italic_U | .

The last two terms in this expression correspond to the reflection at the boundary (i.e. at 00 and nUsubscript𝑛𝑈n_{U}italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT) by means of a spin flip (see a proof of its unitarity in the Appendix). Next, a second coin operator

C^2=(cosβsinβsinβcosβ)subscript^𝐶2matrix𝛽𝛽𝛽𝛽\hat{C}_{2}=\begin{pmatrix}\cos{\beta}&\sin{\beta}\\ -\sin{\beta}&\cos{\beta}\end{pmatrix}over^ start_ARG italic_C end_ARG start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT = ( start_ARG start_ROW start_CELL roman_cos italic_β end_CELL start_CELL roman_sin italic_β end_CELL end_ROW start_ROW start_CELL - roman_sin italic_β end_CELL start_CELL roman_cos italic_β end_CELL end_ROW end_ARG ) (4)

is applied. Finally, the horizontal displacement with reflections at 00 and mRsubscript𝑚𝑅m_{R}italic_m start_POSTSUBSCRIPT italic_R end_POSTSUBSCRIPT (again, acting only on the corresponding direction) is given by

Wm^^subscript𝑊𝑚\displaystyle\hat{W_{m}}over^ start_ARG italic_W start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT end_ARG =\displaystyle== 0mR1|m+1m||UU|superscriptsubscript0subscript𝑚𝑅1tensor-productket𝑚1bra𝑚ket𝑈bra𝑈\displaystyle\sum_{0}^{m_{R}-1}|m+1\rangle\langle m|\otimes|U\rangle\langle U|∑ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_m start_POSTSUBSCRIPT italic_R end_POSTSUBSCRIPT - 1 end_POSTSUPERSCRIPT | italic_m + 1 ⟩ ⟨ italic_m | ⊗ | italic_U ⟩ ⟨ italic_U | (5)
+\displaystyle++ 1mR|m1m||DD|+|0,U0,D|superscriptsubscript1subscript𝑚𝑅tensor-productket𝑚1bra𝑚ket𝐷bra𝐷ket0𝑈bra0𝐷\displaystyle\sum_{1}^{m_{R}}|m-1\rangle\langle m|\otimes|D\rangle\langle D|+|% 0,U\rangle\langle 0,D|∑ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_m start_POSTSUBSCRIPT italic_R end_POSTSUBSCRIPT end_POSTSUPERSCRIPT | italic_m - 1 ⟩ ⟨ italic_m | ⊗ | italic_D ⟩ ⟨ italic_D | + | 0 , italic_U ⟩ ⟨ 0 , italic_D |
+\displaystyle++ |mR,DmR,U|.ketsubscript𝑚𝑅𝐷brasubscript𝑚𝑅𝑈\displaystyle|m_{R},D\rangle\langle m_{R},U|.| italic_m start_POSTSUBSCRIPT italic_R end_POSTSUBSCRIPT , italic_D ⟩ ⟨ italic_m start_POSTSUBSCRIPT italic_R end_POSTSUBSCRIPT , italic_U | .

This completes the definition of our evolution operator for one time step.

For the Bunimovich stadium case, we just consider the upper right quarter of the billiard in order to avoid unwanted spatial symmetries. The boundary is introduced by modifying Eqs. (3) and (5) as follows

Wn^^subscript𝑊𝑛\displaystyle\hat{W_{n}}over^ start_ARG italic_W start_POSTSUBSCRIPT italic_n end_POSTSUBSCRIPT end_ARG =\displaystyle== 0f(m)1|n+1n||UU|superscriptsubscript0𝑓𝑚1tensor-productket𝑛1bra𝑛ket𝑈bra𝑈\displaystyle\sum_{0}^{f(m)-1}|n+1\rangle\langle n|\otimes|U\rangle\langle U|∑ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_f ( italic_m ) - 1 end_POSTSUPERSCRIPT | italic_n + 1 ⟩ ⟨ italic_n | ⊗ | italic_U ⟩ ⟨ italic_U | (6)
+\displaystyle++ 1f(m)|n1n||DD|+|0,U0,D|superscriptsubscript1𝑓𝑚tensor-productket𝑛1bra𝑛ket𝐷bra𝐷ket0𝑈bra0𝐷\displaystyle\sum_{1}^{f(m)}|n-1\rangle\langle n|\otimes|D\rangle\langle D|+|0% ,U\rangle\langle 0,D|∑ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_f ( italic_m ) end_POSTSUPERSCRIPT | italic_n - 1 ⟩ ⟨ italic_n | ⊗ | italic_D ⟩ ⟨ italic_D | + | 0 , italic_U ⟩ ⟨ 0 , italic_D |
+\displaystyle++ |f(m),Df(m),U|ket𝑓𝑚𝐷bra𝑓𝑚𝑈\displaystyle|f(m),D\rangle\langle f(m),U|| italic_f ( italic_m ) , italic_D ⟩ ⟨ italic_f ( italic_m ) , italic_U |
Wm^^subscript𝑊𝑚\displaystyle\hat{W_{m}}over^ start_ARG italic_W start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT end_ARG =\displaystyle== 0w(n)1|m+1m||UU|superscriptsubscript0𝑤𝑛1tensor-productket𝑚1bra𝑚ket𝑈bra𝑈\displaystyle\sum_{0}^{w(n)-1}|m+1\rangle\langle m|\otimes|U\rangle\langle U|∑ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_w ( italic_n ) - 1 end_POSTSUPERSCRIPT | italic_m + 1 ⟩ ⟨ italic_m | ⊗ | italic_U ⟩ ⟨ italic_U | (7)
+\displaystyle++ 1w(n)|m1m||DD|+|0,U0,D|superscriptsubscript1𝑤𝑛tensor-productket𝑚1bra𝑚ket𝐷bra𝐷ket0𝑈bra0𝐷\displaystyle\sum_{1}^{w(n)}|m-1\rangle\langle m|\otimes|D\rangle\langle D|+|0% ,U\rangle\langle 0,D|∑ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_w ( italic_n ) end_POSTSUPERSCRIPT | italic_m - 1 ⟩ ⟨ italic_m | ⊗ | italic_D ⟩ ⟨ italic_D | + | 0 , italic_U ⟩ ⟨ 0 , italic_D |
+\displaystyle++ |w(n),Dw(n),U|ket𝑤𝑛𝐷bra𝑤𝑛𝑈\displaystyle|w(n),D\rangle\langle w(n),U|| italic_w ( italic_n ) , italic_D ⟩ ⟨ italic_w ( italic_n ) , italic_U |
Refer to caption
Figure 2: (Color online) Time evolution of a centered Gaussian distribution (see main text for details) at t = 38 (a), 76 (b), 152 (c), 232 (d), for the rectangular billiard. A (mR,nU)=(150,75)subscript𝑚𝑅subscript𝑛𝑈15075(m_{R},n_{U})=(150,75)( italic_m start_POSTSUBSCRIPT italic_R end_POSTSUBSCRIPT , italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT ) = ( 150 , 75 ) position grid, and α=β=π/4𝛼𝛽𝜋4\alpha=\beta=\pi/4italic_α = italic_β = italic_π / 4 have been used in the calculations.
Refer to caption
Figure 3: (Color online) Same as Fig. 2 for the Bunimovich stadium billiard.

Two shape functions, f(m)𝑓𝑚f(m)italic_f ( italic_m ) and w(n)𝑤𝑛w(n)italic_w ( italic_n ), have been introduced in the two previous expressions. Function f(m)𝑓𝑚f(m)italic_f ( italic_m ), which corresponds to the maximum n𝑛nitalic_n at each m𝑚mitalic_m, is given by

f(m)={nUifmmCnU2(mmC)2ifmCm<mR=2mC.𝑓𝑚casessubscript𝑛𝑈if𝑚subscript𝑚𝐶missing-subexpressionsuperscriptsubscript𝑛𝑈2superscript𝑚subscript𝑚𝐶2ifsubscript𝑚𝐶𝑚subscript𝑚𝑅2subscript𝑚𝐶missing-subexpressionf(m)=\left\{\begin{array}[]{ll}n_{U}\hskip 81.09052pt\text{if}\;m\leq m_{C}\\ \sqrt{n_{U}^{2}-(m-m_{C})^{2}}\quad\text{if}\;m_{C}\leq m<m_{R}=2m_{C}.\end{% array}\right.italic_f ( italic_m ) = { start_ARRAY start_ROW start_CELL italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT if italic_m ≤ italic_m start_POSTSUBSCRIPT italic_C end_POSTSUBSCRIPT end_CELL start_CELL end_CELL end_ROW start_ROW start_CELL square-root start_ARG italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - ( italic_m - italic_m start_POSTSUBSCRIPT italic_C end_POSTSUBSCRIPT ) start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG if italic_m start_POSTSUBSCRIPT italic_C end_POSTSUBSCRIPT ≤ italic_m < italic_m start_POSTSUBSCRIPT italic_R end_POSTSUBSCRIPT = 2 italic_m start_POSTSUBSCRIPT italic_C end_POSTSUBSCRIPT . end_CELL start_CELL end_CELL end_ROW end_ARRAY (8)

Here, mCsubscript𝑚𝐶m_{C}italic_m start_POSTSUBSCRIPT italic_C end_POSTSUBSCRIPT is the value m𝑚mitalic_m at which the circular part of the boundary begins, while the lower limit is given by 00, as in the rectangular billiard. Similarly, function w(n)𝑤𝑛w(n)italic_w ( italic_n ) is the right limit of the displacement

w(n)=mC+nU2n2n,𝑤𝑛subscript𝑚𝐶superscriptsubscript𝑛𝑈2superscript𝑛2for-all𝑛w(n)=m_{C}+\sqrt{n_{U}^{2}-n^{2}}\quad\forall\ n,italic_w ( italic_n ) = italic_m start_POSTSUBSCRIPT italic_C end_POSTSUBSCRIPT + square-root start_ARG italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - italic_n start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG ∀ italic_n , (9)

with 00 being the leftmost value, as before. See an illustration in Fig. 1. For comparison purposes, the rectangular billiard is taken as the rectangle in which the Bunimovich stadium is contained, both having a horizontal length two times the vertical one. Remind that the classical dynamics in the rectangular domain is regular.

In Fig. 2 we show some examples of the probability distribution at different times in the rectangular billiard case. We have taken a grid of size mR=150subscript𝑚𝑅150m_{R}=150italic_m start_POSTSUBSCRIPT italic_R end_POSTSUBSCRIPT = 150 and nU=75subscript𝑛𝑈75n_{U}=75italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT = 75 and a normalized spatial Gaussian initial distribution located at the center of this domain, with amplitudes proportional to exp((m75)2(n38)2)superscript𝑚752superscript𝑛382\exp{(-(m-75)^{2}-(n-38)^{2})}roman_exp ( - ( italic_m - 75 ) start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT - ( italic_n - 38 ) start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT ) with 73m7773𝑚7773\leq m\leq 7773 ≤ italic_m ≤ 77 and 36n4036𝑛4036\leq n\leq 4036 ≤ italic_n ≤ 40 and zero elsewhere; for spin up the amplitude is 1/2121/\sqrt{2}1 / square-root start_ARG 2 end_ARG and for spin down is i/2𝑖2i/\sqrt{2}italic_i / square-root start_ARG 2 end_ARG. We have selected four evolution times that approximately correspond to a meaningful part of the evolved probability being reflected by the wall. In Fig. 3 we do the same as in the previous case, but now considering the Bunimovich stadium. It is clear that before the bounce on the circular part of the Bunimovich stadium both probability distributions coincide (a slight difference can be seen at (mC,nU)subscript𝑚𝐶subscript𝑛𝑈(m_{C},n_{U})( italic_m start_POSTSUBSCRIPT italic_C end_POSTSUBSCRIPT , italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT )), as expected. But after the main reflection on the circle at t=76𝑡76t=76italic_t = 76 they become markedly different, already showing regularity and strong symmetry in the rectangular billiard and the lack of them in the stadium. In fact, the probability peak at the center of the rectangular domain is gradually destroyed in the Bunimovich case for later times as can be checked by comparing the lower panels of both Figures.

Refer to caption
Figure 4: (Color online) Unfolded level spacing distributions P(s)𝑃𝑠P(s)italic_P ( italic_s ) vs. s𝑠sitalic_s for the Bunimovich stadium billiard with the following coin angles: α=β=π/8𝛼𝛽𝜋8\alpha=\beta=\pi/8italic_α = italic_β = italic_π / 8 (a), α=β=π/7𝛼𝛽𝜋7\alpha=\beta=\pi/7italic_α = italic_β = italic_π / 7 (b), α=π/8𝛼𝜋8\alpha=\pi/8italic_α = italic_π / 8 β=π/9𝛽𝜋9\beta=\pi/9italic_β = italic_π / 9 (c), and α=π/4𝛼𝜋4\alpha=\pi/4italic_α = italic_π / 4 β=π/6𝛽𝜋6\beta=\pi/6italic_β = italic_π / 6 (d). PWsubscript𝑃WP_{\rm W}italic_P start_POSTSUBSCRIPT roman_W end_POSTSUBSCRIPT and best fitting PBsubscript𝑃BP_{\rm B}italic_P start_POSTSUBSCRIPT roman_B end_POSTSUBSCRIPT (with the parameters and errors reported in Table 1) distributions, see Eqs. (10) and (12), are shown as red (gray) and green (light gray) solid lines, respectively.

III Chaotic signatures: spectral behavior and morphology of the eigenfunctions

Refer to caption
Figure 5: (Color online) Unfolded level spacing distributions P(s)𝑃𝑠P(s)italic_P ( italic_s ) vs. s𝑠sitalic_s for the rectangular billiard with the following coin angles: α=β=π/8𝛼𝛽𝜋8\alpha=\beta=\pi/8italic_α = italic_β = italic_π / 8 (a), α=β=π/7𝛼𝛽𝜋7\alpha=\beta=\pi/7italic_α = italic_β = italic_π / 7 (b), α=π/8𝛼𝜋8\alpha=\pi/8italic_α = italic_π / 8 β=π/9𝛽𝜋9\beta=\pi/9italic_β = italic_π / 9 (c), and α=π/4𝛼𝜋4\alpha=\pi/4italic_α = italic_π / 4 β=π/6𝛽𝜋6\beta=\pi/6italic_β = italic_π / 6 (d). PPsubscript𝑃PP_{\rm P}italic_P start_POSTSUBSCRIPT roman_P end_POSTSUBSCRIPT distribution (errors reported in Table 2) , see Eq. (11), is shown by means of red (gray) solid lines.

An unbounded one-dimensional QW is translationally invariant and as a consequence a simple Dispersion Relation (DR) can be obtained. A similar situation arises for the unbounded AQW in two dimensions [20], leading to a generalized DR which is a straightforward extension of the 1D solution. For a cylinder, a compact domain with periodic boundary conditions in one direction, the same effectively happens, and a DR has been derived for it [21]. On the other hand, general billiards consist of a completely bounded domain with reflective boundaries, which in the context of QWs we have implemented by means of spin flips.

III.1 Spectral statistics

Since in our case there is no analytical DR we proceed, in what follows, with numerical explorations of the behavior of our system by means of the spectral statistics and the morphology of the corresponding eigenfunctions.

We directly diagonalize the evolution operator for one time step, i.e. Wm^C^2Wn^C^1^subscript𝑊𝑚subscript^𝐶2^subscript𝑊𝑛subscript^𝐶1\hat{W_{m}}\hat{C}_{2}\hat{W_{n}}\hat{C}_{1}over^ start_ARG italic_W start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT end_ARG over^ start_ARG italic_C end_ARG start_POSTSUBSCRIPT 2 end_POSTSUBSCRIPT over^ start_ARG italic_W start_POSTSUBSCRIPT italic_n end_POSTSUBSCRIPT end_ARG over^ start_ARG italic_C end_ARG start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT, in the spin-position basis, and study its (unfolded) spectrum for both the Bunimovich and rectangular stadium billiards. To speed up the computations, we take here a grid of size (mR,nU)=(50,25)subscript𝑚𝑅subscript𝑛𝑈5025(m_{R},n_{U})=(50,25)( italic_m start_POSTSUBSCRIPT italic_R end_POSTSUBSCRIPT , italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT ) = ( 50 , 25 ) (notice that this grid defines a smaller position basis for the Bunimovich stadium case). It is also important to underline that desymmetrizing the Bunimovich stadium is a customary procedure in order to avoid unwanted symmetries, and then reveal the Wigner surmise (or other statistics in different systems) for the eigenvalues associated to the corresponding Hamiltonian. Failing to do so would allow the overlapping of uncorrelated spectra corresponding to different symmetry classes. Keeping our line of reasoning, we proceed in this way.

In Fig. 4 we show the spectral statistics, where the eigenphases (i.e. the phases of the complex unimodular eigenvalues) of the evolution operator have been considered to construct the histograms P(s)𝑃𝑠P(s)italic_P ( italic_s ) of the unfolded level spacings s𝑠sitalic_s. After ordering the eigenphases between 00 and 2π2𝜋2\pi2 italic_π, the unfolding is performed by simply dividing the distances between nearest neighbors by the mean (arithmetic) distance. In the different panels, we show the results for four different pairs of α𝛼\alphaitalic_α and β𝛽\betaitalic_β values; in Figs. 4(a) and (b) they are equal (symmetrical coins), while in (c) and (d) they are different (asymmetrical coins, see caption for details). Together with the histograms, we plot the Wigner surmise [22] as the red (gray) solid lines given by

PW(s)=π2sexp(πs24),subscript𝑃W𝑠𝜋2𝑠𝜋superscript𝑠24P_{{\rm W}}(s)=\frac{\pi}{2}\ s\ \exp{\left(\frac{-\pi\ s^{2}}{4}\right)},italic_P start_POSTSUBSCRIPT roman_W end_POSTSUBSCRIPT ( italic_s ) = divide start_ARG italic_π end_ARG start_ARG 2 end_ARG italic_s roman_exp ( divide start_ARG - italic_π italic_s start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG start_ARG 4 end_ARG ) , (10)

which describes the spectral behavior of chaotic Hamiltonian evolution and in principle is not directly applicable to our QW scenario. In the other end of level statistics, for regular systems the spectral spacings satisfy the Poisson distribution given by

PP(s)=exp(s).subscript𝑃P𝑠𝑠P_{{\rm P}}(s)=\exp{(-s)}.italic_P start_POSTSUBSCRIPT roman_P end_POSTSUBSCRIPT ( italic_s ) = roman_exp ( - italic_s ) . (11)

Given that the QW dynamics is different from the one corresponding to the quantum Hamiltonian of a free particle inside a billiard, deviations from PW(s)subscript𝑃W𝑠P_{{\rm W}}(s)italic_P start_POSTSUBSCRIPT roman_W end_POSTSUBSCRIPT ( italic_s ) could be expected. Hence, we have also fitted the data (minimizing the root mean square error) with a very simple function that interpolates between Poisson and the Wigner surmise, the Brody distribution [23], whose expression is

PB(s)=asδexp(bsδ+1),subscript𝑃B𝑠𝑎superscript𝑠𝛿𝑏superscript𝑠𝛿1P_{{\rm B}}(s)=a\ s^{\delta}\ \exp{(-b\ s^{\delta+1})},italic_P start_POSTSUBSCRIPT roman_B end_POSTSUBSCRIPT ( italic_s ) = italic_a italic_s start_POSTSUPERSCRIPT italic_δ end_POSTSUPERSCRIPT roman_exp ( - italic_b italic_s start_POSTSUPERSCRIPT italic_δ + 1 end_POSTSUPERSCRIPT ) , (12)

where a=(δ+1)b𝑎𝛿1𝑏a=(\delta+1)bitalic_a = ( italic_δ + 1 ) italic_b, b=Γ((δ+2)/(δ+1))δ+1𝑏Γsuperscript𝛿2𝛿1𝛿1b={\Gamma((\delta+2)/(\delta+1))}^{\delta+1}italic_b = roman_Γ ( ( italic_δ + 2 ) / ( italic_δ + 1 ) ) start_POSTSUPERSCRIPT italic_δ + 1 end_POSTSUPERSCRIPT, and δ𝛿\deltaitalic_δ is a parameter. The corresponding results are displayed by means of green (light gray) solid lines in Fig. 4 and the fitting values are shown in Table 1. We notice a general behavior that approximately follows the Wigner surmise since the best fitting Brody curve is closer to it (δ0.550.83𝛿0.55similar-to0.83\delta\in 0.55\sim 0.83italic_δ ∈ 0.55 ∼ 0.83 and the error slightly lower in general) than Poisson (δ=0𝛿0\delta=0italic_δ = 0). It is interesting to see that, besides the deviations from the exact Wigner curve, in all situations selected but one there is a high bar corresponding to the smallest level spacings, which abruptly falls in the next bin. This does not happen for α=π/4𝛼𝜋4\alpha=\pi/4italic_α = italic_π / 4 and β=π/6𝛽𝜋6\beta=\pi/6italic_β = italic_π / 6 (Fig. 4(d)), where level repulsion is indeed present, but the Wigner surmise performs worse. Symmetries due to the coin operator shape could exist. These symmetries though, are not enough to significantly uncorrelate the spectrum and lead to Poisson statistics.

Table 1: Fitting results and spectral distributions errors for the QW in the Bunimovich stadium billiard
Coins Error
α𝛼\alphaitalic_α β𝛽\betaitalic_β δ𝛿\deltaitalic_δ PBsubscript𝑃BP_{\rm B}italic_P start_POSTSUBSCRIPT roman_B end_POSTSUBSCRIPT PWsubscript𝑃WP_{\rm W}italic_P start_POSTSUBSCRIPT roman_W end_POSTSUBSCRIPT
π/8𝜋8\pi/8italic_π / 8 π/8𝜋8\pi/8italic_π / 8 0.730 0.104 0.111
π/7𝜋7\pi/7italic_π / 7 π/7𝜋7\pi/7italic_π / 7 0.718 0.104 0.113
π/8𝜋8\pi/8italic_π / 8 π/9𝜋9\pi/9italic_π / 9 0.833 0.088 0.092
π/4𝜋4\pi/4italic_π / 4 π/6𝜋6\pi/6italic_π / 6 0.552 0.093 0.126

In the case of the rectangular billiard, whose associated Hamiltonian is regular, we have evaluated the error of the Poisson distribution (see Fig. 5 and Table 2).

Table 2: Poisson distribution errors for the QW in the rectangular billiard
Coins Error
α𝛼\alphaitalic_α β𝛽\betaitalic_β PPsubscript𝑃PP_{\rm P}italic_P start_POSTSUBSCRIPT roman_P end_POSTSUBSCRIPT PPsuperscriptsubscript𝑃PP_{\rm P}^{\prime}italic_P start_POSTSUBSCRIPT roman_P end_POSTSUBSCRIPT start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT
π/8𝜋8\pi/8italic_π / 8 π/8𝜋8\pi/8italic_π / 8 0.113 0.038
π/7𝜋7\pi/7italic_π / 7 π/7𝜋7\pi/7italic_π / 7 0.106 0.037
π/8𝜋8\pi/8italic_π / 8 π/9𝜋9\pi/9italic_π / 9 0.113 0.045
π/4𝜋4\pi/4italic_π / 4 π/6𝜋6\pi/6italic_π / 6 0.152 0.079

It becomes clear that the behavior is not strictly Poissonian, but it closely resembles it. In particular, when not considering the first and very large bars of the histograms the error of the Poisson distribution (we call it PPsuperscriptsubscript𝑃PP_{{\rm P}}^{\prime}italic_P start_POSTSUBSCRIPT roman_P end_POSTSUBSCRIPT start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT in Table 2) is much lower. Hence, though the QW dynamics is clearly different from that of a quantum free particle moving inside a hard wall billiard, the statistical behavior of the spectra in these two cases is similar and allows to distinguish between them. Can we also identify other coincidences of this kind at the eigenfunction level?

III.2 Eigenfunctions morphology

To answer this we turn to analyze the morphology of the eigenfunctions. For that purpose we use the Participation Ratio (PR), a commonly used measure of localization in quantum chaos. The PR is calculated by first normalizing the eigenfunctions |ΦketΦ|\Phi\rangle| roman_Φ ⟩ and then evaluating PR=(m,n|UΦmn|4+|DΦmn|4)1PRsuperscriptsubscript𝑚𝑛superscriptsuperscriptsubscriptsubscript𝑈Φ𝑚𝑛4superscriptsuperscriptsubscriptsubscript𝐷Φ𝑚𝑛41{\rm PR}=(\sum_{m,n}|{U_{\Phi}}_{m}^{n}|^{4}+|{D_{\Phi}}_{m}^{n}|^{4})^{-1}roman_PR = ( ∑ start_POSTSUBSCRIPT italic_m , italic_n end_POSTSUBSCRIPT | italic_U start_POSTSUBSCRIPT roman_Φ end_POSTSUBSCRIPT start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n end_POSTSUPERSCRIPT | start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT + | italic_D start_POSTSUBSCRIPT roman_Φ end_POSTSUBSCRIPT start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n end_POSTSUPERSCRIPT | start_POSTSUPERSCRIPT 4 end_POSTSUPERSCRIPT ) start_POSTSUPERSCRIPT - 1 end_POSTSUPERSCRIPT (UΦmnsuperscriptsubscriptsubscript𝑈Φ𝑚𝑛{U_{\Phi}}_{m}^{n}italic_U start_POSTSUBSCRIPT roman_Φ end_POSTSUBSCRIPT start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n end_POSTSUPERSCRIPT and DΦmnsuperscriptsubscriptsubscript𝐷Φ𝑚𝑛{D_{\Phi}}_{m}^{n}italic_D start_POSTSUBSCRIPT roman_Φ end_POSTSUBSCRIPT start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_n end_POSTSUPERSCRIPT are the components of |ΦketΦ|\Phi\rangle| roman_Φ ⟩). For example, in the rectangle the values of the PR range from 1111 to 2mRnU2subscript𝑚𝑅subscript𝑛𝑈2m_{R}n_{U}2 italic_m start_POSTSUBSCRIPT italic_R end_POSTSUBSCRIPT italic_n start_POSTSUBSCRIPT italic_U end_POSTSUBSCRIPT (2×50×25250252\times 50\times 252 × 50 × 25), roughly indicating how many position basis elements participate into a given eigenfunction (besides the factor 2222 coming from the spin part of the Hilbert space). As a consequence, larger PR values mean less localized states in this basis. In Fig. 6 we show the histogram corresponding to P(PR)𝑃PRP(\rm{PR})italic_P ( roman_PR ) vs. PR.

Refer to caption
Figure 6: (Color online) P(PR)𝑃PRP(\rm{PR})italic_P ( roman_PR ) vs. PR for both the Bunimovich stadium and the rectangular billiards. Dashed lines stand for the Bunimovich stadium histogram and solid ones for the rectangle case. Blue (black) lines correspond to α=π/4𝛼𝜋4\alpha=\pi\ /4italic_α = italic_π / 4 and β=π/3𝛽𝜋3\beta=\pi\ /3italic_β = italic_π / 3, while red (gray) lines to α=β=π/4𝛼𝛽𝜋4\alpha=\beta=\pi\ /4italic_α = italic_β = italic_π / 4.

It can be directly observed that localization on around 1000100010001000 basis elements is the approximate typical behavior for the Bunimovich stadium while on around 1250125012501250 and 1400140014001400 for the rectangle, which is only in part due to the different effective dimensions of the position basis in the two cases. This represents about half of the bases sizes (again, notice the spin space dimension). But not only the Bunimovich stadium billiard is more localized than the rectangle on average, the distribution is biased towards lower PR values in a meaningful way (PR approximately [400,800]absent400800\in[400,800]∈ [ 400 , 800 ]). In what follows, by analyzing some representative cases we are going to deepen into this features.

We display some examples of eigenfunctions that help to understand the behavior of localization. In Fig. 7 we show two of the most delocalized and localized eigenfunctions of the Bunimovich stadium (upper row) and of the rectangular billiard (lower row), both for the symmetrical coins case. we notice that the delocalized states (left) are different with the Bunimovich’s one looking more irregular (or chaotic in the quantum chaos sense) and the rectangle’s one looking markedly more regular. For both maximally localized states the analogy with the Hamiltonian system becomes difficult and the different nature of the dynamics manifests itself. As a matter of fact, these extremely peaked eigenfunctions are not usually found in quantum billiards and deserve further study in the future.

Refer to caption
Figure 7: (Color online) Upper row: eigenfunctions of the Bunimovich stadium with PR=1200.021200.021200.021200.02 (a) and PR=57.1357.1357.1357.13 (b). Lower row: same for the rectangular billiard with PR=1557.571557.571557.571557.57 (c) and PR=104.01104.01104.01104.01 (d). In all cases α=β=π/4𝛼𝛽𝜋4\alpha=\beta=\pi\ /4italic_α = italic_β = italic_π / 4.

We have also considered the case of asymmetrical coins in Fig. 8. In this case both systems apparently behave in an irregular fashion, but if one takes a closer look the only thing that changes from the previous Figure is the tilting provided by the asymmetrical coins (it amounts to biasing one of the 1D translations with respect to the other). Thus, the differences between regular and irregular behavior due to the shape of the boundaries in the corresponding Hamiltonian dynamics can still be detected.

Refer to caption
Figure 8: (Color online) Upper row: eigenfunctions of the Bunimovich stadium with PR=1145.021145.021145.021145.02 (a) and PR=36.4136.4136.4136.41 (b). Lower row: same for the rectangular billiard with PR=1636.231636.231636.231636.23 (c) and PR=148.11148.11148.11148.11 (d). In all cases α=π/4𝛼𝜋4\alpha=\pi\ /4italic_α = italic_π / 4 and β=π/3𝛽𝜋3\beta=\pi\ /3italic_β = italic_π / 3.

More interestingly, in the Bunimovich stadium billiard we have found localization on structures similar to those found in the corresponding Hamiltonian system, and which could be responsible for the greater localization found in this case (notice their typical PR values). In fact, there are bouncing ball states that closely resemble the ones ubiquitous in the quantum chaos literature [24, 25]. We show them by means of Fig. 9 for the symmetrical coins case and of Fig. 10 for the asymmetrical one. In the traditional quantum chaotic model, this family of eigenfunctions are localized on marginally stable orbits that form a continuous family. In the QW scenario this is remarkable, having in mind that we expect a purely diffusive behavior. Moreover, the typical sequence of horizontal excitations is also present in the QW and the only difference comes at the time of considering the asymmetrical coins which just tilt the bouncing balls.

Refer to caption
Figure 9: (Color online) Bouncing ball eigenfunctions of the Bunimovich stadium. Upper row: PR=684.98684.98684.98684.98 (a) and PR=797.18797.18797.18797.18 (b). Lower row: PR=679.09679.09679.09679.09 (c) and PR=780.15780.15780.15780.15 (d). In all cases α=β=π/4𝛼𝛽𝜋4\alpha=\beta=\pi\ /4italic_α = italic_β = italic_π / 4.
Refer to caption
Figure 10: (Color online) Bouncing ball eigenfunctions of the Bunimovich stadium. Upper row: PR=574.42574.42574.42574.42 (a) and PR=594.73594.73594.73594.73 (b). Lower row: PR=715.89715.89715.89715.89 (c) and PR=678.73678.73678.73678.73 (d). In all cases α=π/4𝛼𝜋4\alpha=\pi\ /4italic_α = italic_π / 4 and β=π/3𝛽𝜋3\beta=\pi\ /3italic_β = italic_π / 3.

IV Conclusions

We have studied the properties of a QW inside compact bidimensional domains with different boundary shapes, characterized by a regular and irregular behavior of their corresponding classical and quantum Hamiltonian dynamics. The simple diffusive evolution given by the QW is able to detect these two regimes which in principle have only been associated to the classical behavior of a free particle inside these cavities and its corresponding quantization. This allows us to conclude that the simplest generalization of the QW on the line to 2D billiards constitutes a new paradigmatic example of quantum chaos.

A deeper theoretical explanation of this behavior is a promising avenue of research. One of the most interesting questions arises from the fact that though bouncing ball states have been found, no evidence of whispering gallery ones were obtained. This naturally indicates that the different properties of diffusive systems like QWs come into play at the time to detect chaoticity. The diffusion directions compared to the domain shape and symmetries are relevant, the details of their influence will be studied in the future. On the other hand, the underlying positions grid automatically associated with discrete QWs could induce different behaviors when compared to the usual billiard models. In fact, a hopping Hamiltonian on lattice billiards has been investigated in [26] showing the same spectral properties as the continuum counterparts, though in the open scenario new lattice scars were found. Is there something similar in the QWs case?

Also, we will consider more complex situations like having two particles inside the domain and other coins that would serve as different models for electrons inside a cavity [27]. Finally, we will investigate if our chaotic QW can improve 2D grid searches (as an alternative to the nonlinear QW model studied in [8], for instance).

Acknowledgment

This research has been partially supported by the Spanish Ministry of Science, Innovation and Universities, Gobierno de España under Contract No. 2021-122711NB-C21. Support from CONICET is gratefully acknowledged.

*

Appendix A Unitarity of the evolution operator

It is worth showing the unitarity of the reflection mechanism at the billiard boundaries. Taking the most general vertical displacement:

Wn^^subscript𝑊𝑛\displaystyle\hat{W_{n}}over^ start_ARG italic_W start_POSTSUBSCRIPT italic_n end_POSTSUBSCRIPT end_ARG =\displaystyle== 0f(m)1|n+1,Un,U|superscriptsubscript0𝑓𝑚1ket𝑛1𝑈bra𝑛𝑈\displaystyle\sum_{0}^{f(m)-1}|n+1,U\rangle\langle n,U|∑ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_f ( italic_m ) - 1 end_POSTSUPERSCRIPT | italic_n + 1 , italic_U ⟩ ⟨ italic_n , italic_U | (13)
+\displaystyle++ 1f(m)|n1,Dn,D|+|0,U0,D|superscriptsubscript1𝑓𝑚ket𝑛1𝐷bra𝑛𝐷ket0𝑈bra0𝐷\displaystyle\sum_{1}^{f(m)}|n-1,D\rangle\langle n,D|+|0,U\rangle\langle 0,D|∑ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_f ( italic_m ) end_POSTSUPERSCRIPT | italic_n - 1 , italic_D ⟩ ⟨ italic_n , italic_D | + | 0 , italic_U ⟩ ⟨ 0 , italic_D |
+\displaystyle++ |f(m),Df(m),U|ket𝑓𝑚𝐷bra𝑓𝑚𝑈\displaystyle|f(m),D\rangle\langle f(m),U|| italic_f ( italic_m ) , italic_D ⟩ ⟨ italic_f ( italic_m ) , italic_U |

and,

W^nsuperscriptsubscript^𝑊𝑛\displaystyle\hat{W}_{n}^{\dagger}over^ start_ARG italic_W end_ARG start_POSTSUBSCRIPT italic_n end_POSTSUBSCRIPT start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT =\displaystyle== 0f(m)1|n,Un+1,U|superscriptsubscript0𝑓𝑚1ket𝑛𝑈bra𝑛1𝑈\displaystyle\sum_{0}^{f(m)-1}|n,U\rangle\langle n+1,U|∑ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_f ( italic_m ) - 1 end_POSTSUPERSCRIPT | italic_n , italic_U ⟩ ⟨ italic_n + 1 , italic_U | (14)
+\displaystyle++ 1f(m)|n,Dn1,D|+|0,D0,U|superscriptsubscript1𝑓𝑚ket𝑛𝐷bra𝑛1𝐷ket0𝐷bra0𝑈\displaystyle\sum_{1}^{f(m)}|n,D\rangle\langle n-1,D|+|0,D\rangle\langle 0,U|∑ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_f ( italic_m ) end_POSTSUPERSCRIPT | italic_n , italic_D ⟩ ⟨ italic_n - 1 , italic_D | + | 0 , italic_D ⟩ ⟨ 0 , italic_U |
+\displaystyle++ |f(m),Uf(m),D|.ket𝑓𝑚𝑈bra𝑓𝑚𝐷\displaystyle|f(m),U\rangle\langle f(m),D|.| italic_f ( italic_m ) , italic_U ⟩ ⟨ italic_f ( italic_m ) , italic_D | .

Hence,

W^nW^nsuperscriptsubscript^𝑊𝑛subscript^𝑊𝑛\displaystyle\hat{W}_{n}^{\dagger}\hat{W}_{n}over^ start_ARG italic_W end_ARG start_POSTSUBSCRIPT italic_n end_POSTSUBSCRIPT start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT over^ start_ARG italic_W end_ARG start_POSTSUBSCRIPT italic_n end_POSTSUBSCRIPT =\displaystyle== 0f(m)1|n,Un,U|superscriptsubscript0𝑓𝑚1ket𝑛𝑈bra𝑛𝑈\displaystyle\sum_{0}^{f(m)-1}|n,U\rangle\langle n,U|∑ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_f ( italic_m ) - 1 end_POSTSUPERSCRIPT | italic_n , italic_U ⟩ ⟨ italic_n , italic_U | (15)
+\displaystyle++ 1f(m)|n,Dn,D|+|0,D0,D|superscriptsubscript1𝑓𝑚ket𝑛𝐷bra𝑛𝐷ket0𝐷bra0𝐷\displaystyle\sum_{1}^{f(m)}|n,D\rangle\langle n,D|+|0,D\rangle\langle 0,D|∑ start_POSTSUBSCRIPT 1 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_f ( italic_m ) end_POSTSUPERSCRIPT | italic_n , italic_D ⟩ ⟨ italic_n , italic_D | + | 0 , italic_D ⟩ ⟨ 0 , italic_D |
+\displaystyle++ |f(m),Uf(m),U|,ket𝑓𝑚𝑈bra𝑓𝑚𝑈\displaystyle|f(m),U\rangle\langle f(m),U|,| italic_f ( italic_m ) , italic_U ⟩ ⟨ italic_f ( italic_m ) , italic_U | ,

or equivalently

W^nW^nsuperscriptsubscript^𝑊𝑛subscript^𝑊𝑛\displaystyle\hat{W}_{n}^{\dagger}\hat{W}_{n}over^ start_ARG italic_W end_ARG start_POSTSUBSCRIPT italic_n end_POSTSUBSCRIPT start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT over^ start_ARG italic_W end_ARG start_POSTSUBSCRIPT italic_n end_POSTSUBSCRIPT =\displaystyle== 0f(m)|n,Un,U|+0f(m)|n,Dn,D|superscriptsubscript0𝑓𝑚ket𝑛𝑈quantum-operator-product𝑛𝑈superscriptsubscript0𝑓𝑚𝑛𝐷bra𝑛𝐷\displaystyle\sum_{0}^{f(m)}|n,U\rangle\langle n,U|+\sum_{0}^{f(m)}|n,D\rangle% \langle n,D|∑ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_f ( italic_m ) end_POSTSUPERSCRIPT | italic_n , italic_U ⟩ ⟨ italic_n , italic_U | + ∑ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_f ( italic_m ) end_POSTSUPERSCRIPT | italic_n , italic_D ⟩ ⟨ italic_n , italic_D | (16)
=\displaystyle== 0f(m)|nn|(|UU|+|DD|)superscriptsubscript0𝑓𝑚tensor-productket𝑛bra𝑛ket𝑈bra𝑈ket𝐷bra𝐷\displaystyle\sum_{0}^{f(m)}|n\rangle\langle n|\otimes(|U\rangle\langle U|+|D% \rangle\langle D|)∑ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_f ( italic_m ) end_POSTSUPERSCRIPT | italic_n ⟩ ⟨ italic_n | ⊗ ( | italic_U ⟩ ⟨ italic_U | + | italic_D ⟩ ⟨ italic_D | )

which is the identity. The same happens for the most general horizontal displacement, completing the proof.

References

  • [1] Y. Aharonov, L. Davidovich, and N. Zagury, Quantum random walks, Phys. Rev. A 48, 1687 (1993).
  • [2] F. Nosrati et al., Quantum Information Spreading in a Disordered Quantum Walk, J. Opt. Soc. Am. B 38, 2570 (2021).
  • [3] A. P. Hines and P. C. E. Stamp, Quantum walks, quantum gates, and quantum computers, Phys. Rev. A 75, 062321 (2007).
  • [4] A. Ambianis, Quantum search algorithms, ACM SIGACT News 35, 22 (2004).
  • [5] P. M. Preiss et al., Strongly correlated quantum walks in optical lattices, Science 347, 1229 (2015).
  • [6] G. Maquiné Batalha, A. Volta, W.T. Strunz, and M. Galiceanu, Quantum transport on honeycomb networks, Sci. Rep. 12, 6896 (2022).
  • [7] R. Portugal, Quantum walks and search algorithms (Springer Nature, Switzerland, 2018).
  • [8] G. Di Molfetta and B. Herzog, Searching via nonlinear quantum walk on the 2D-grid, Algorithms 13, 305 (2020).
  • [9] Hao Tang et al., Experimental twodimensional quantum walk on a photonic chip, Sci. Adv. 4, 3174 (2018).
  • [10] Q.P. Su et al., Experimental demonstration of quantum walks with initial superposition states, npj Quantum Inf. 5, 40 (2019).
  • [11] Z.Q. Jiao et al., Two-dimensional quantum walks of correlated photons, Optica 8, 1129 (2021).
  • [12] J. Rodrigues, N. Paunkovic, and P. Mateus, A simulator for discrete quantum wals on lattices, Int. J. Mod. Phys. C 28, 1750055 (2017).
  • [13] E. Bach, S. Coppersmith, M.P. Goldschen, R. Joynt, and J. Watrous, One dimensional Quantum Walks with absorbing boundaries, JCSS 69, 562 (2004).
  • [14] F. Haake, Quantum Signatures of Chaos, Springer-Verlag, Berlin, (2004).
  • [15] O. Bohigas, M.J. Giannoni, and C. Schmit, Characterization of chaotic quantum spectra and universality of level fluctuation laws, Phys. Rev. Lett. 52, 1 (1984).
  • [16] R. Grobe and F. Haake, Universality of cubic level repulsion for dissipative quantum chaos, Phys. Rev. Lett. 62, 2893 (1989).
  • [17] D. Villaseñor, L.F. Santos, and P. Barberis-Blostein, Breakdown of the quantum distinction of regular and chaotic classical dynamics in dissipative systems, Phys. Rev. Lett. 133, 240404 (2024).
  • [18] C. Di Franco, M. Mc Gettrick, and Th. Busch, Mimicking the probability distribution of a two-dimensional Grover walk with a single-qubit coin, Phys. Rev. Lett. 106, 080502 (2011).
  • [19] L.A. Bru, M. Hinarejos, F. Silva, G.J. de Valcárcel, and E Roldán, Electric quantum walks in two dimensions, Phys. Rev. A 93, 032333 (2016).
  • [20] E. Roldán, C. Di Franco, F. Silva, and G.J. de Valcárcel, N-dimensional alternate coined quantum walks from a dispersion-relation perspective, Phys. Rev. A 87, 022336 (2013).
  • [21] L.A. Bru, G.J. de Valcárcel, G. Di Molfetta, A. Pérez, E. Roldán, and F. Silva, Quantum walk on a cylinder, Phys. Rev. A 94, 032328 (2016).
  • [22] M.L. Mehta, Random matrices, Academic Press, San Diego, (1991).
  • [23] T. Prosen and M. Robnik, Energy level statistics in the transition region between integrability and chaos, J. Phys. A: Math. Gen. 26, 2371 (1993).
  • [24] E.G. Vergini and G.G. Carlo, Semiclassical quantization with short periodic orbits, J. Phys. A: Math. Gen. 33, 4717 (2000).
  • [25] S. Selinummi, J. Keski-Rahkonen, F. Chalangari, and E. Räsänen, Formation, prevalence, and stability of bouncing-ball quantum scars Phys. Rev. B 110, 235420 (2024).
  • [26] V. Fernández-Hurtado, J. Mur-Petit, J.J. García-Ripoll and R.A. Molina, Lattice scars: surviving in an open discrete billiard, New J. Phys. 16, 035005 (2014).
  • [27] A. Melnikov and L. Fedichkin, Quantum walks of interacting fermions on a cycle graph, Sci Rep 6, 34226 (2016).