![Arxiv:2104.09346V1 [Quant-Ph] 19 Apr 2021 Their Dynamics on a finite Simulating Hardware](http://data.docslib.org/img/c537e48c6e8446edcfd72c4b45f09d2f-1.webp)
UMD-PP-021-02
Toward simulating quantum field theories with controlled phonon-ion dynamics: A hybrid analog-digital approach
1, 2 3 Zohreh Davoudi, ∗ Norbert M. Linke, and Guido Pagano 1Maryland Center for Fundamental Physics and Department of Physics, University of Maryland, College Park, MD 20742, USA. 2Joint Quantum Institute and Department of Physics, University of Maryland, College Park, MD 20742 3Department of Physics and Astronomy, Rice University, 6100 Main Street, Houston, TX 77005, USA. (Dated: April 20, 2021) Quantum field theories are the cornerstones of modern physics, providing relativistic and quan- tum mechanical descriptions of physical systems at the most fundamental level. Simulating real-time dynamics within these theories remains elusive in classical computing. This provides a unique oppor- tunity for quantum simulators, which hold the promise of revolutionizing our simulation capabilities. Trapped-ion systems are successful quantum-simulator platforms for quantum many-body physics and can operate in digital, or gate-based, and analog modes. Inspired by the progress in proposing and realizing quantum simulations of a number of relativistic quantum field theories using trapped- ion systems, and by the hybrid analog-digital proposals for simulating interacting boson-fermion models, we propose hybrid analog-digital quantum simulations of selected quantum field theories, taking recent developments to the next level. On one hand, the semi-digital nature of this pro- posal offers more flexibility in engineering generic model interactions compared with a fully-analog approach. On the other hand, encoding the bosonic fields onto the phonon degrees of freedom of the trapped-ion system allows a more efficient usage of simulator resources, and a more natural implementation of intrinsic quantum operations in such platforms. This opens up new ways for simulating complex dynamics of e.g., Abelian and non-Abelian gauge theories, by combining the benefits of digital and analog schemes.
I. INTRODUCTION question we investigate here is complementary to the lat- ter effort: Given a Hamiltonian formulation, what are the Quantum field theories (QFTs) provide the underlying efficient implementation strategies that maximally take quantum-mechanical descriptions of physical systems, advantage of available hardware capacity and capabil- from relativistic gauge field theories of the Standard ity? We focus on trapped-ion quantum simulators as one Model of particle physics [1,2], to emergent low-energy of the leading quantum-hardware technologies. Nonethe- models in condensed-matter systems [3,4], to effective less, applications of this strategy to other platforms, such field theories in hadronic and nuclear physics [5,6]. as cavity QED and superconducting circuits, can be es- Classical simulation methods have come a long way tablished analogously. to describe phenomena emerging from these underly- QFTs with bosonic fields are extremely common, from ing theories, with notable examples in the realm of lat- scalar field theory descriptions of phase transitions in tice gauge theory (LGT) methods applied in strong- condensed-matter systems or of the inflationary phase interaction physics [7,8]. Nonetheless, there is a need of the early universe and the Higgs mechanism, to gauge for new computational strategies to overcome the limita- field theory descriptions of the fundamental forces of na- tions of the current methods, in order to achieve real-time ture. Due to their infinite-dimensional Hilbert space, simulations of matter, and predictions for equilibrium even on a single space-time point, a truncation must be and out-of-equilibrium phenomena arising from strong- imposed on the various bosonic excitations to contain interaction dynamics. Hamiltonian simulation of phys- arXiv:2104.09346v1 [quant-ph] 19 Apr 2021 their dynamics on a finite simulating hardware. Impor- ical systems, naturally enabled via mapping the prob- tant work has emerged in understanding, qualitatively lem to a quantum simulator or a quantum computer, or quantitatively, the impact of this truncation, from the is an example of such a strategy, with major advances simple quantum harmonic oscillator [75–77] to scalar field reported in recent years in proposing [9–52] and imple- theory [13, 29] and LGTs [47, 50, 74, 78]. While quanti- menting [53–68] simulations of various field theories on ties in the low-energy subspace of the truncated theories the limited existing quantum hardware, and in formu- exhibit exponential convergence to the exact values in lating suitable Hamiltonian descriptions of QFTs and the full theory, as guaranteed by the Shannon-Nyquist LGTs for resource-efficient and robust quantum simu- sampling theorem [29, 76] in the case of scalar field the- lations [9, 12, 13, 21, 29, 36, 42–45, 47–50, 69–74]. The ory, observables in the high-energy sector or those ob- tained from long-time expectation values exhibit a slow convergence to the exact values. In particular, as shown ∗ Also affiliated with the RIKEN Center for Accelerator-based Sci- in Ref. [74] for the case of the SU(2) LGT coupled to ences, Wako, Japan, during the early stages of this work. matter in 1+1 D, for a certain threshold on the gauge- 2
field cutoff quantities enter a scaling region in which the dom in a trapped-ion simulator [14, 113–115]. This idea dependence on the cutoff becomes exponential and sys- has led to concrete gate-based protocols for simulating tematically improvable. However, for high-energy spec- interacting fermion-boson models, such as the Holstein tra and long-time dynamical quantities, such a threshold model of electron-phonon dynamics in condensed-matter is reached for a large value of the cutoff, increasing the physics [115], as well as the first experimental implemen- simulation-resource requirement needed to reach a fixed tation of a one-site boson-fermion dynamics [116] con- accuracy. sidering only a single excitation of the boson. In the The most common encoding of bosonic degrees of free- present work, the ideas in these proposals are taken to dom to qubits for digital quantum simulation is to con- the next level to demonstrate that quantum simulation vert their occupation number to a binary representation, of relativistic QFTs, including LGTs, can benefit from a the qubit-number requirement of which is logarithmic in similar hybrid protocol. the cutoff on the highest excitation of the boson. Most This work introduces a complete and enhanced set of importantly for near-term applications, the number of conventional as well as phonon-based gates, including entangling gates required to implement the dynamics as- but not limited to those introduced in Ref. [115], and sociated with boson-boson or boson-fermion interactions deploys them to construct concrete circuits for the dig- grows polynomially with the system size, as a large num- itized time-evolution operator within a Yukawa theory ber of controlled operations must be introduced on the of scalar-fermion interactions and the U(1) LGT cou- state of the boson register, see e.g., Refs. [39, 75, 77]. pled to matter. The gates include single-spin (qubit) This problem is particularly severe in the case of bosonic operations, and entangling spin-spin, spin-phonon, and field theories as, for example, a scalar field in a finite dis- phonon-phonon operations. In order to ensure efficient cretized space represents quantum harmonic oscillators, implementation of all these gates in one experiment, nor- or bosons, associated with each momentum mode, the mal modes of motion, local modes of motion, or both number of which scales linearly with the volume of the types of modes should be controlled and manipulated. reciprocal lattice. Each harmonic oscillator then requires The simulated theories considered are in 1+1 dimensions the allocation of dedicated qubits and the implementa- (D) but the generalization to higher dimensions, and tion of associated entangling gates. A cost analysis of this in particular challenges involved in the case of higher- kind for a fully-digital implementation is made explicit in dimensional LGTs, are briefly discussed. Numerical ex- this paper and is contrasted with our analog-digital pro- amples are provided to supplement our proposals with tocol, to be introduced next. concrete experimental parameters, and to demonstrate their feasibility within the current technologies. Quali- A trapped-ion quantum simulator [79–81] can operate tative comparisons of the simulation cost within digital in an analog mode, with which dynamics generated by and hybrid analog-digital simulations of the same the- the Ising and Heisenberg spin Hamiltonians can be stud- ories will be provided, along with a discussion of the ied [32, 82–90], or a digital mode, with which universal outlook of a hybrid approach for generalization to more computations expressed in terms of single- and two-qubit complex QFTs, including non-Abelian LGTs. gates can be performed [54, 91–104]. Typically in these systems, specific pairs of internal states encode qubits, which are manipulated via ion-laser or ion-microwave in- II. THE HYBRID ANALOG-DIGITAL teractions. The collective excitations of the motion of BUILDING BLOCKS ions in the trap, i.e., the phonons, are often used as me- diators of the interactions among the qubits, enabling the engineering of effective spin Hamiltonians or in turn The trapped-ion quantum simulator considered in this entangling operations among two or more qubits [81]. work consists of a number of ions confined in a radio- In recent years, promising progress has been reported in frequency Paul trap [117]. The qubit is encoded in two controlling the phonon dynamics in a trapped-ion simu- stable internal levels of the ion, which are separated in 1 lator, leading to experimental demonstrations of phonon energy by an angular frequency ω0. The confining po- hopping, phonon interference, and a phonon quantum tential is sufficiently stronger along the transverse axes, x walk in Paul traps [105–109]. This opens up the pos- and y, of the trap so that the ions form a linear crystal in 2 sibility of taking advantage of phonons, in addition to the axial, z, direction. With appropriate anharmonic the qubits, as dynamical degrees of freedom to encode axial confinement forces, the ions can be arranged in and process information [110, 111]. In fact, there ex- ist quantum-simulation proposals for an Abelian LGT, i.e., the lattice Schwinger model, which take advantage of the phonons in a fully-analog setting to encode the 1 Planck’s constant ~ is set to unity throughout. 2 gauge or fermion degrees of freedom [112], but they are x, y and z should not be confused with the x, y and z indices to be introduced on the quasi-spins of the qubit. While the former not experimentally feasible yet. A more viable option (in san-serif font) correspond to spatial physical coordinates of for near-term applications is to combine the flexibility the trap, hence labeling the Cartesian components of the wave- of gate-based digital simulations with the versatility of vector of the laser beams, the latter (in italic font) correspond both the controllable spin and phonon degrees of free- to the Bloch-sphere axes in the qubit Hilbert space. 3 an equally-spaced configuration for individual laser-beam the axial modes, i.e., those along the z axis of the trap, addressing [118, 119]. The typical spacing between adja- by indices m = 2N + 1, 2N + 2, , 3N. With the in- cent ions is a few micrometers in present-day trapped-ion troduction of two counter-propagating··· laser beams with simulators. wave-vector difference ∆kj, frequency difference (beat- L Given the long-range Coulomb force among the ions note) ∆ωj ωj , and phase difference ∆φj φj, detuned and the common trapping potential applied, the motion from an excited≡ internal level of the ion, the≡ two-photon of the ions can be described in terms of a set of collec- Raman transition among the two states of the qubit can tive normal modes. The displacement of the ions around be induced with a Rabi frequency Ωj at the location of an equilibrium position can then be expressed in terms ion j. The beams are assumed to address the ions indi- of phononic degrees of freedom, whose excitation ener- vidually, hence the subscript j. In an interaction picture gies are quantized in units of the normal-mode frequen- that rotates with the free Hamiltonian in Eq. (1), the cies. While such a normal-mode picture is sufficient for interacting ion-laser Hamiltonian can be written as [84] simulating the dynamics of certain fermion-boson field N theories, switching to a local-mode picture is required 3N iω t iω t Ωj i η (a e− m +a† e m ) H0 = e m=1 m,j m m once the bosonic field exhibits non-trivial dynamics, as is ion-laser 2 × j=1 P the case in the gauge-theory example considered. Local X L modes can be addressed by making the trapping poten- i(ωj ω0)t+iφj + e− − σj + h.c., (2) tial tighter along the transverse directions, or conversely by relaxing the axial confinement to increase the spacing L where the condition ωj ω0 ω0 is assumed. The among the ions, such that the displacement of any given prime on H is to denote| that− this| Hamiltonian is in the ion from its equilibrium position is much smaller than the interaction picture. Later on, we need to adopt a dif- distance among the adjacent ions [120]. The local modes ferent interaction picture rotating with shifted frequen- can be addressed as long as the laser excitation on indi- cies and so it is important to bear in mind the origin vidual ions happens on timescales faster than the sepa- of Eq. (2). Multiple beatnote frequencies, amplitudes, ration of the corresponding normal modes, which defines and phases can be applied simultaneously, hence a sum 3 the timescale for phonon hopping. In such a regime, the over laser parameters can be introduced, a possibility phonon hopping probability among the ion sites is small that we take advantage of later. The Lamb-Dicke pa- and the quanta of motion are localized. The interactions (∆k)2 rameters ηm,j are defined as ηm,j = bm,j, where associated with each of these scenarios, along with the 2Mionωm expanded gate-set for the hybrid analog-digital compu- bm,j are the (normalized) normal-modeq eigenvector com- tation of this work, will be introduced in the following. ponents between ion j and mode m, and Mion denotes the mass of the ion. In the Lamb-Dicke regime in which 2 1/2 ηm,j (am + am† ) 1, transitions in the space of cou- A. Phonons as excitations of normal modes of pledh spin-phononi system take a simpler form, and can be motion realized through quantum gates. ∆kj is assumed to be the same for lasers addressing each ion j, i.e., ∆kj ∆k. The free Hamiltonian of the trapped-ion system in the ≡ The carrier transition corresponds to Hion-laser0 (η0) |O absence of laser-ion interactions is and is obtained by setting ωL ω = 0. With this setting, j − 0 N 3N the Hamiltonian corresponding to ion j becomes ω0 z 1 Hion = σ + ωm(a† am + ). (1) 2 j m 2 Ω j=1 m=1 σ j iφj iφj x iφj iφj y H (φ ) = (e +e− )σ +(e e− )iσ . (3) X X j j 4 j − j Here, σ is a Pauli operator acting on the space of the ions’ σ quasi-spin, i.e., the qubit, with the standard commuta- The superscript on Hj denotes that this Hamiltonian tion relations. am† (am) are the bosonic creation (anni- only acts on the qubit space with the noted σ operators. hilation) operators for the normal modes of motion with The single-spin rotations of arbitrary angle θj, to be de- σ σ the associated frequencies ωm and the commutation re- noted as Rj (θj, φj), can be obtained by applying Hj (φj) σ lations [a , a ] = [a† , a† ] = 0 and [a , a† ] = δ . for time τgate, to be deduced from the relation m m0 m m0 m m0 m,m0 We label the transverse modes of motion along the x-axis x y σ σ σ iθj (cos φj σ sin φj σ ) iH (φ )τ of the trap by indices m = 1, 2, ,N, those along the y R (θ , φ ) e− j − j = e− j j gate . ··· j j j ≡ axis of the trap by indices m = N +1,N +2, , 2N, and (4) ··· For example, rotations around the x and y axes of the Bloch sphere of the qubit j can be realized as 3 One may also consider a micro-trap where the global trapping x potential is replaced by local potentials applied at the position of σ iθj σ R (θj, 0) e− j , (5) each ion [121]. For applications studied in this work, a Paul trap j ≡ with a tight transverse potential suffices to produce the desired σ 3π iθ σy R (θ , ) e− j j , (6) dynamics, and this additional possibility will not be considered. j j 2 ≡ 3
lasers are applied for certain times. The virtual phonons where in the last term, the subscript j on the phonon op- in this process then e↵ectively induce a spin-spin transi- erators is a local index now corresponding to the quanta tion, which is well known as the Molmer-Sorenson tran- of motion at position of the jth ion. sition [].4 The corresponding Hamiltonian between ions For the field theories considered in this work, the sim- j and j0 is ulated theory does not exhibit boson hopping and the phonon hopping in the simulator needs to be suppressed. b b m,j m,j0 x x In a typical trapped-ion system with tens of Ytterbium Hj,j = ⌦j⌦j0 R j j , (12) 0 µ2 !2 0 x m m ions, ! 2⇡ 5 MHz2 and d0 is of the order of a few X ⇠ ⇥ x 3 4 x x 2 microns. Then 10 10 and ! 2⇡ 1 where R ( k) /2Mion is the recoil frequency of the ion kHz, and hence the dynamics⇠ associated with the⇠ phonon⇥ ping probability among the ion sites. The⌘ interactions The superscript on H denotes that this Hamiltonian with Mion being the mass of the ion, and µ !L j !0. hopping can be neglected compared with spin-phonon associated with each of these scenarios,The qubit-phonon along with rotations the only of arbitrary acts on angle the qubit⌘✓ ,tobe space with the noted operators. j,j0 (with typical strength 100 kHz) and spin-spin dynam- expanded gate-set for the hybrid digital-analog compu- The single-qubit rotations of arbitrary angle ✓j,tobede- denoted as Rj,j (✓j,j0 ), can be obtained by applying Hj,j ⇠ 0 0 ics (with typical strength 10 kHz). If further sup- 4 tation of this work, will be introducedfor time in⌧ this, section. to be deducednoted from as theRj relation(✓j, j), can be obtained by applying Hj ( j) ⇠ gate pression of the phonon hopping is desired to improve for time ⌧gate, to be deduced from the relation x x the accuracy of the simulated model. one can either im- i✓ j iH ⌧gate R (✓ ) e j0 = e j,j0 (13) Hamiltonian Model cousins [] A scalar field theory coupled to (non- j,j0 j,j0 x y ⌘ i✓j (cos j sin jpose ) tighteriH ( trappingj )⌧ potential along the transverse di- A. Phonons as excitations of normal modes of R (✓ , ) e j j = e j gate (4) relativistic) fermions is also the e↵ective field theory de- j j j rection or spread out the ions further by reducings.w. the ax- 2 2 x x motion In reality, designing high-fidelity two-qubit⌘ entangling H = F cos (kxj)=F cos ⌘(a + a †) scription of the Yukawa interactions among the nucleons ial potential. Additionally, the nearest-neighborj hopping j j gates of this type requires complexFor example, pulse rotations shaping tech- around the x and y axes of the 2 4 x x 4 x x 2 and pions and enter the quantum many-body description can be actively blocked using techniques demonstrated=2F ( ⌘ + ⌘ )(aj†aj)+2⇥ F ⌘ (aj†⇤aj) + . Consider the free Hamiltonianniques of the to trapped-ion achieve a sys- perfectBloch creation sphere and ofannihilation the qubit j ofcan be realized as e e ··· of nuclei []. It is therefore interesting to investigate the the virtual phonon in the process, which is necessary to in Ref. []. Such possibilities will not be considered fur- (15) tem, x prospects of quantum simulation of such theories, includ- i✓therj j in the current work. Finally we note that the local 3 leave the system in the same phonon-excitationRj state(✓j, 0) be- e , (5) e e e ing the suitability of the hybrid analog-digital approach. N 3N ⌘ mode frequencies can be expressed as the commonHere, k = trapk xˆ where k is the wave-vector of the stand- fore and after the gate operation. 3⇡ y In the following, we focus on a simple case: a scalar field !0 z 1 i✓j j H = + ! (a† a + ) (1) Rj (✓j, ) e frequency, plus additionallasers(6) are local applied corrections. for certain These times. correc- The virtual phonons (0) where in the1 last term, the subscript j on the phonon op- ion j m m m ing wave beam and, ⌘ kx = k x is the theory coupled to one flavor of staggered fermions in 1+1 2 2 2 ⌘ ⌘ 2Mion! j=1 m=1 tions scale similarin to this the process hopping then strength e↵ectively and are inducee hencee a spin-spine transi- erators is a local index nowD corresponding and without the to the possibility quanta of self interaction among X X suppressed by attion, least which a factor is well 10 3 knowncomparedcorresponding as the with Molmer-Sorenson the Lamb-Dickie tran- parameter.of motionq Note at position that the of the jth ion. B. Phonons as excitationsrespectively. of local modes Rotations of motion along the z axis of the Bloch e e the scalar fields. Later, we comment of the applicability Here, is a Pauli operator acting on the space of the Z 4 standing wave is onlye exciting the local modes of mo- sphere, Rj (✓j), defined as trap frequency insition a typical []. The trap corresponding described above. Hamiltonian As a between ions x tionx given the choice of the4 wave-vector made. Ellipsis of this proposal for self-interacting scalar fields and the ion’s quasi-spin, i.e., the qubit, with the standard com- result, !j in Eq.j (14and) canj0 is be replaced with ! to a first z 2 higher-dimensional case, and will point out the requisite Z i✓j in Eq. (15) denotes either higher-orderFor terms the field in ⌘ theoriesthat considered in this work, the sim- mutation relations. am† (am) is the creation (annihila- R (✓ ) e approximation.j (7) In a linear Paul trap the Coulomb energy is compa-j j b b x x 2 1/2 extensions. tion) operator for the normal modes of motion with as- ⌘ will bem,j neglectedm,j0 x x in the regime ⌘ ulated(a + a theory) does1, ornot exhibit boson hopping and the rable to the potential energy associated with the ax- The spin-phonon HamiltoniansH = in⌦ Eq.j⌦j (R8) can be ob- , (12) j j Spin-(normal) Spin-(local) j,j0 0 µ2 !2 j j0 h i ⌧ Consider a one-dimensional spatial lattice with N sites, sociated frequencies !m with the commutationSingle-spin relations gates can be implemented withoutz thetained need for similarly quantum by oper-Two-spin expanding gate the (MS) ion-laserarem interaction phononStanding-wavem non-conserving gate operatorsphonon that hopping rotate withe in the at simulator needs to be suppressed. ial motion of the ions. Namely, the parameter x lattice spacing b, and with periodic boundary conditions † 2 † 3 z2 z phonon gate ⌘ phonon gate leastX the frequency ! and can beIn adiabatically a typical trapped-ion elimi- system with tens of Ytterbium [am,am0 ]=[am† ,am ] = 0 and [am,am ]= m,m0 . We label ation using a classical shift [NorbertHamiltonian to verify]. in Eq. (2) using the local modes of motion. 5 0 e /(0d0Mion! ) & 1, where ! is the trap frequency along 2 2 x e x (PBCs) imposed on the fields. The Hamiltonian of the the transverse modes of motion along the x axis of the The blue and red sideband transitionsThen both correspond the spinwhere toandR the( phononk) /2Mion operatorsisnated the recoil ( asax longand frequency as F ⌘ of/! the ion1[]. ions, ! 2⇡ 5 MHz and d0 is of the order of a few the axial axis of the trap, d0 is the average spacing among j ⌘ j ⌧ ⇠ ⇥ x lattice-regularized3 4 x x Yukawa theory to be simulated with a H and are obtained byx setting ! !with= !Mion being the mass ofSimilar the ion, to and the previousµ !L gates,!0. phonon-phononmicrons. Then rotations 10 10 and ! 2⇡ 1 trap by indices m =1, 2, ,N, thosethe ions, along and theeyisaxis the of electricion-laser charge.0 (⌘ As) a result, the a †) are representingL 0 quantumk operations at location j ⌘ ··· |O j The qubit-phonon rotations of arbitrary angle ✓ ,tobe s.w. ⇠ trapped-ion quantum⇠ simulator⇥ consists of the trap by indices m = N +1,N +2, , 2N, and the and ! , respectively. This setting leads to a simple of arbitrary angle canj,j be0 obtainedkHz, by applying and henceH thefor dynamics associated with the phonon phonons hop over the chain and arek considered as the along the chain, with the associated gate operation iden- e j ··· denoted as R (✓j,j ), can be obtainedaa by applying H axial modes, i.e., those along the z axis of the trap by in- coupled spin-phonon Hamiltonian. In order to achieve j,j0 0 time ⌧gate, to be deduced fromj,j0 thehopping relation can be neglected compared with spin-phonon(I) (II) (III) common normal modes of the system. The same param- tified as HYukawa = H + H + H , (17) a Hamiltonian proportional to j , both sets of bluefor and time ⌧gate, to be deduced from the relation (with typical strength 100 kHz) and spin-spinYukawa dynam-Yukawa Yukawa dices m =2N +1, 2N +2, , 3Neter. With is typically the introduction much smaller for the transverse2 motions x ⇠ ··· z a a a (1) x icsx (with(2) x typicalx 2 strength 10 kHz). If further sup- of two counter-propagating laser beams withR wave-vector(✓j, j),R (✓j) red sidebandR ( operations✓k,j , k,j can) be applied simultaneously x x aa (1) (2) i (aj †aj )+ (aj †aj ) where the purely fermionic Hamiltonian of the ions.j In particular,j if the trapk is,j made extremely Rj (✓j, j)=Rk,j(✓k,j, j) x i✓ j j .R (15)( iHj,j, ⌧gate) e { } { } { } R ak (✓a ,⌘)k,j e⌘j ,✓k,j ✓j0 =je 0 (13) ⇠ { } j,j0 j,jj 0 ⌘ pression of the phonon hopping is desired to improve ping probability among the ion sites.di↵erence The interactions k,frequencydiThe↵erence superscriptbinding ! along! onL, andH oneor phasedenotes both transverse thatwith this a set axes Hamiltonian of ofRaman the ions beams such such that the phases of the ! ⌘! ! s.w. aa ⌘ j iH⇥ j ⌧gate ⇤ N di↵ernece j j, detuned from anx(y) exited internal two sideband beams add up to ⇡. With this setting, the = e the accuracy of the(16) simulated model. one cani either im- associated with each of these scenarios, along with⌘ the only actsthat on the qubit1, space the transverse with the noted phonons operators. are bound to a On the other hand,In inreality, the limit designing of a tight high-fidelity transverse two-qubit di- entangling (I) j level of the ion, the two-phonon RamanFIG. transition⌧ 1. The among building blocksHamiltonian of the analog-digital corresponding circuits to ion j inbecomes a trapped-ion quantum simulator. The gates from left to right are pose tighter trapping potentialHYukawa along= the transverse( j† j+1 di- j†+1 j)+m ( 1) j† j expanded gate-set for the hybrid digital-analog compu- The single-qubitsingle ion, rotations with a small of arbitrary strength angle for hopping✓j,tobede- to the neigh- rection, the spin-spingates Hamiltonian of this type in requires Eq. (12 complex) could pulsebe shaping tech- 2b the two states of the qubit can be induced defined with in Eqs. a Rabi (4), (27), (310), (16x(),y) (14), x and(y) (x(18y)), respectively. An important point to notice regardingrection this or spread gate is out that the ions further byj=1 reducing the ax- tation of this work, will be introduced in this section. noted asboringR (✓ ionj, j that), can goes be as obtainede /(d byM applyingion! )=H ( j)! []. niques to achieve a perfect creation and annihilation of X j 0 j ⌘k,j⌦j derived using the normal modes of motion alongaccording the axial to Eq. (15), the relative size of the coe cients frequency ⌦j at the location of ion j . The beams can ad-/ a i k,j i k,j x ial potential. Additionally, the nearest-neighbor hopping (18) for timeFor⌧gate the, to quantum be deduced simulation from the of relation quantumHk,j ( k,j field)= theories(e a(ork + thee otherak†) lessj .the tight(8) virtual transverse) phonon direction in the process, so to havex whichx is necessaryx x 2 to (1) (2) dress the ions along the chain globally or individually. In 2 of the aj†aj and (aj†aj) terms,can be/ actively,isfixedto blocked using techniques demonstrated that exhibit non-trivial boson-field dynamics, it is neces- su ciently strongleave all-to-all the system spin-spin in the couplings. same phonon-excitation As will state be- describes the hopping term and the mass term of one x y 2 2 in Ref. []. Such possibilities will not be considered fur- A. Phonons as excitations ofan normal interaction modes picture of that rotates with freei✓ Hamiltonianj (cos j j sin j j ) iHj ( j )⌧gate ( 1+⌘ )/⌘ , while the overall strength of these terms Rj (✓jsary, j) torespectively. adopte a local mode Rotations picture.In= ordere along Assuming to the ensurez only(4)axis only one of one of the ofBlochbe the shown, phononeach both modes, beatnote typesforea andk, of afterthe independently, gates, the gate one operation. based the Hamiltonian on local flavor of staggered fermions with mass m . Note that motion in Eq. (1), the interacting ion-laser Hamiltonian⌘ z can be can be tuned arbitrarily by changingtherF in, the as long current as the work. Finally we note that the local the transversesphere, R directionsj (θj), defined of thecouples as trap, to e.g., thex qubit, supports in this lo- setup,modes and of to motions minimize and the the other based on normal modes PBCs impose the identification .Thefree written as [] 2 η Ω condition F ⌘2/!x 1 is not violated.mode The frequencies former con-can be expressed as the common trap N+1 1 For example,cal phonon rotations modes, around the free the Hamiltoniano↵x-resonantand y axes transitions, of theof the ion chain the Raman beams couldσa be set k,j k,j iφk,je e iφk,j y ⌘ of motion, canH be employed( φk,j ) = in the simulation,(e exhibitingak + e− ⌧a†) σ , scalar field Hamiltonian is z k ,jB. Phonons as excitationsdition of local may seem modes too ofk constraining motionj frequency for simulating plus additional an ar- local corrections. These correc- Consider the free Hamiltonian of the trapped-ion sys- Bloch spherecan be of written the qubit as j can besuch realizedz that as theyiθj σ onlyj couple to oneanother set of normal advantage{ modes} { of ofa hybrid} − digital-analog2 setting. In N Rj (θj) e− (7) k tions scale similar to the hopping strength andN are hence ⌦ 3N i!mt i!mt bitrary Hamiltonian containing both phonon terms. As 2 2 2 tem, ping probability among the ion sites.j i The interactions⌘m,j (ame +a† eThe) superscript onmotion.x Hj ≡denotes For example, that this by setting Hamiltonian k = k xˆ, the beams X e ⇧ m m=1 m N 3Ni✓ other words, by suppressing the hopping of the phonons (II3) j ( 'j) ' 2 Hion-laser0 = e j j will be demonstrated in Sec.(9)IV forsuppressed simulating by the at lattice least a factorH 10 compared= b with+ the + ' (19) associated with each of these scenarios,2 along with the !onlyR0j (✓⇥ actsj, 0)z one the qubitonly, couple space with to1 transverse the(5) noted normal operators. modes along the x axis. Yukawa r j j=1 P Hion = +⌘ !m(a† am + )+ along one axis of the trap, the qubits can still be entan- 2 2 2 N 3N can be implementedj withoutm the need for a quantum op- Schwinger model, by arbitrarily andtrap consistently frequency chang-in a typical trap described above.j=1 As a !expanded0 gate-set for the hybrid1 X digital-analog compu- The2 single-qubit3⇡ i rotations✓The y qubit-phonon of arbitrary2 rotations angle ✓j of,tobede- arbitrary angle ✓k,j,tobeIn a linear Paul trap the Coulomb energy is compa- x Xx z i(! !0)t+i j +j=1 m=N+1j j gled with highcan fidelity be implemented by taking advantage directly. of another Note set that Ωj is pro- Hion = + !m(a† am + ) (1) L erationRj ( using✓j, a) classicale , phase shift a on(6) the controller for ing the reference frame in obtainingresult, the interaction!j in Eq. (14 pic-) can be replaced with ! to a first tation ofj this work, willm be introduced in thise section. notedj X+h.c as.R(2)j (✓Xj, jdenoted), can be as obtainedR (✓k,j by, applyingk,j), canH bej ( obtained j) by applyingrable to the potential energy associated with the ax- 2 2 2 ⌘ N k,j of modes atmoted the same to time.Ω to reflect the freedom in the choice of where ⇧j is the conjugate momentum corresponding to j=1 m=1 addressingfor time beam⌧ ,j to. be A deduceda useful gate from in the1 the relationa circuits con- k,j ture, desired ratios of thez coe cientsapproximation. in the simulated X X gate Hk,j ( k,j)x forx timex ⌧gate, to be deducedFor the from simulation the relationial motion of the lattice of the Schwinger ions. Namely, model, the the parameter ! (a †a + ), (14) mode-dependent Rabi frequencies. The spin-phonon ro- the scalar field 'j. In other words, ['j, 'j0 ]=[⇧j, ⇧j0 ]= where the condition !L respectively.!0 !0 structedmust Rotations be assumed. in alongthe following the z axis sections ofj thej j is Bloch the phase gate Sj, e2/(d3M !z2) 1, wheretheory!z is the can trap still frequency be engineered. along⌘ Last but not least, if the x y 2 bosons need to interact0 locally.ion Such& an interaction will Here, is a Pauli operator acting on the space of the| | ⌧ Z j=1 z 0 and ['j, ⇧j0 ]=i j, j0. These fields can be quantized A. PhononsThe as excitations prime on H ofis normal to denotesphere, modes thatR of this(✓whichj), Hamiltonian defined can be as implemented is i✓j (cos j via j sinS j= jR) (π/iH4).j ( j )⌧gatei k,j i k,jtationsx of an arbitrary set ofsimulated angles theoryθk,j , requiresto be de- site-dependent coe cients for j R (✓j, j) e X a j =j e i✓k,j (e a(4)k+e a†) j the axial axis of the trap, d0 is the average spacing among ion’s quasi-spin, i.e., the qubit, with the standard com- j R (✓k,j, k,j) e need to be mappedk to aσa local phonon-phonon interaction.{ } The spin-phonon Hamiltoniansin the standard in Eq. (8 way) can to be obtain ob- a representation in terms in themotion interaction picture. Later on, we need to adopt⌘ k,j noted as R ( θk,j , φk,j ),the can phonon-phonon be obtained couplings, by ap- individual standing-wave The blue and red sidebandz transitions⌘ correspond a toa the ions,k ,j and e is the electric charge. As a result, the mutation relations. a† (am) is the creation (annihila- Z i✓j iH ( Tok,j )⌧ engineer this new type{ } of{ interactions} { in} the trapped- tained similarly by expandingof the the (bosonic) ion-laser harmonic interaction oscillator creation (d† ) and m adi↵erent interaction picture rotating withFor shiftedR example,(✓ ) fre-e rotationsj around the x(7)and= e yL axesk,j of thegate σa(9) beamsσa can replace the global beam in Eq. (15). k Hion-laser0 j (jη) and are obtained by setting ωj ω0ion= ω simulator,k plying asphononsH proposedk ,j( hopφk in,j over Ref.) for the [], time the chainτ ionsgate and, can to are be be considered deduced from as the tion) operator for the normal modes of motion with as- | ⌘ − { } Hamiltonian in Eq. (2) usingannihilation the local modes(dk) operators, of motion. Consider the freequencies Hamiltonian and so itof is the important trapped-ion to4 bear sys- in mindBloch theO sphere origin of the qubit j can be realized as common{ } normal modes of the system. The same param- x sociated frequencies ! with the commutation relations Or aand geometricalωk, respectively.phase gate in case This of spin setting interactions leads propor- to a coupledput near thethe minima relation (or maxima) of a standing wave, Then both the spin j and the phonon operators (a m of Eq. (2). can be implementedz withoutz the needIn particular, for quantum setting oper- laser beam phases to proper angles j tem, tional to − []. x eter is typically much smaller for the transverse motions x N/2 1 † † spin-phonon⌦ Hamiltonian. In orderi✓j toj achieve a Hamil-which induces an AC Stark shift corresponding toiφk,j the iφk,j y [am,am0 ]=[am† ,am ] = 0 and [am,am In]= m, them0 . We Lamb-Dicke label ation using regime a classical in shift which [NorbertRcan to(✓ generate verify]., 0) e transitions, proportional(5) to a + a† σaand a i III.θk,j (e A YUKAWAak+e− a† THEORY:)σ and SCALARaj†) are FIELDS representing quantum1 operations1 at loca- 0 0 j j y k Rk of( thek θ ions., φ In particular,) e− ifk the trap is made extremelyk j i2⇡kj/N i2⇡kj/N the transverse modes of motion alongN the x axis3N of1/ the2 tonian proportional to e.g., σ⌘, beatnotes associated to k ,j k,j k,j COUPLED TO (STAGGERED)tion FERMIONSj along the chain. On' thej = other hand, in the(d limitk† e + dke ) (20) 2 The blue and red sideband transitions correspondj toy { } !⌘m,j (am + a† ) 1,1 transitions in the space a†: 3⇡ binding{ along} { one} or≡σa bothP transverseσa axes of the ions such Nb p2"k 0 z m k i✓j j k= N/2 trap by indices m =1, 2, H,N,= those alongh + the y axis!i (a of† a⌧ + ) (1)blue and red sidebandsR (✓j can, be) appliede , simultaneously(6) iH k ,j ( φk,j )τgate of a tight transverse direction, the spin-spin Hamiltonian ion j m m m Hion-laser0 (⌘) and are obtained by settingj !L !0 = !k x(y) = e− { } { } . (10) X ··· 2of coupled qubit-phonon system2 take|O simpler forms, and 2 ⌘ 4 that 1, the transverse phonons are bound to a the trap by indices m = N +1,N +2j=1, , 2Nm,=1 and the with phases that add up to zero [122]. ⇡With this set- x in Eq. (12) could be derived using theN/2 normal1 modes and !k, respectively. This setting leads to a simplea i✓k,j (ak+ak†) j ⌧ A scalar field theory coupled to fermions describes the canX be··· realizedX through applications of quantum gates. Rk,j(✓k,j, ) e , single(10) ion, with a small strength for hopping to the neigh- i "k i2⇡kj/N i2⇡kj/N axial modes, i.e., those along the z axis of the trap by in- of motion along the axial (or⇧ the= other less tight( trans-d† e d e ) (21) coupled spin-phononting,respectively. the Hamiltonian. Hamiltonian0 Rotations correspondingIn order along to the achieve toz axis ion2 j⌘ ofbecomes the Bloch In particular with a single-mode2coupling addressing,3 ofx( they) opportunely Higgsx(y) bosonx(y) to fermions of the Stan- j k k The carrier transition corresponds to Hion-laser0 (⌘ ) and boring ion that goes as e /(d0Mion! )= ! []. Nb 2 Here, is a Pauli operator acting on the space of the |O Z j x dices m =2N +1, 2N +2, , 3N. With the introduction a Hamiltonian proportionalsphere, R ( to✓j ), defined, both sets as of blue a and ✓k,j (ak ak†) j / dard Model through Yukawa interactions,verse) direction and is respon- so to have su cientlyk= strongN/2 r all-to-all ··· is obtained by setting !L !0 = 0. With this setting,j x Rk,j(✓k,j, 0) e . settingFor laser-beam(11) the quantum phases simulation gives rise of to quantum couplings field to differ-theories X of two counter-propagatingion’s quasi-spin, laser beams i.e., the with qubit, wave-vector with the standard com- η Ω ⌘ spin-spin couplings. As will be shown, both types of the the Hamiltonian correspondingred sideband to ion j operationsbecomesσa can be appliedk,j j simultaneouslyiφk,j iφk,j y sible for dynamical mass generation and the fermion H (φk,j) = (e ak + e−z a†) σ , (8) ent phonon-operatorthat exhibit non-trivial combinations boson-field proportional dynamics, to ita isk neces-+ak† di↵erence k,frequencydimutation↵ relations.erence !am† !(aLm,) and isthe phase creation (annihila- k,j Z i✓j j k j gates, one based on local modeswhere ofk motionslabels the and corresponding the other momentum mode pk with a set of Raman beams such− that2RImplementingj the(✓j) phasese of simultaneous the blue(7) and red sidebandsary to adopt a local modemixing picture. in Assuming nature. only Its one non-perturbative of studies using ⌘ tion) operator for the⌘ normal modes of motion with as- ⌘ and ak a†: 2⇡k 2 2 di↵ernece j j, detuned from an exited⌦j internal two sideband beams add up to ⇡. Withtransitions this setting, now detuned the from the normal mode frequen-k lattice-regularized Euclidean field theorybased on simulations normal modes has of2⇡ motion,k/(Nb), can and be"k employed= ( in) + m is the correspond- ⌘ H ( )= (ei j + e i j ) x +(ei j e i j )i y (3) the− transverse directions of the trap, e.g., x, supports lo- Nb ' level of the ion, thesociated two-phonon frequencies Ramanj!m transitionjwith the among commutation Hamiltonian relationsj corresponding can bej implemented1 tor ion j becomesb withoutr the need forr(b) quantum oper- drawn considerable interest as it revealsthe simulation, important exhibiting con- another advantage of a hybrid 4 where φk,j = (φ ciesφ will) = notφ , generate with φ dynamicaldenoting phonons the as longcal as thephonon modes, the free Hamiltoniany of the ion chain ing energy, with m' beingq the bare mass of the scalar † † 2 k,j k,j k,j k,j σa iθk,j (ak+ak†)σj the two states of[ theam,a qubitm0 ]=[ canam† ,a be induced] = 0 and with [am a,a Rabi]= m,m0 . We label ation using a classical− shift [Norbert to verify]. R (θ , 0) e− , (11) digital-analog setting. In other words, by suppressing the m0 m0 can be writtenk,j k as,j nections to the cuto↵ scale of the Standard Model and the transverse modes of motion along the x axis of thered (blue)⌘k,j sideband⌦j laser phase. In order to ensure only ≡ frequency ⌦j at the location of ion j. The beams can ad- a The bluei and j red sidebandi j x transitions correspond to π questions regardingy triviality []. Inhopping the context of the of phonons cos- along one axis of the trap, the Hk,j ( j)= (e ak + e a†) j (8) σa θk,j (ak a†)σj dress the ions alongtrap the by chain indices globallym =1 or, 2, individually.,N, those In along the y axis ofone set of2 the phonon modes,k ak, couples to the qubit in R (θNk,j, ) 3eN − k . (12) qubits can still be entangled with high fidelity by taking Hion-laser0 (⌘) and are obtained by setting !L !0 = !k k!,j0 mology and early1 universe, non-perturbative simulations ··· |O z2 ≡ an interaction picturethe trap that by rotates indices withm free= N Hamiltonian+1,N +2, , 2N, and thethis setup,and ! the, respectively. Raman beams This can setting be set leads such to that a simple the Hion = j + !m(am† am + )+ advantage of another set of5 modes at the same time. ··· In order to ensure onlyk one of the phonon modes, a , 2 are required for2 studies of non-equilibrium and real-time Other boundary conditions can be studied with minimal adjust- in Eq. (1), the interactingaxial modes, ion-laser i.e., those Hamiltonian along the canz axis be of the trap by in-wave-vectorcoupled difference spin-phonon ∆k Hamiltonian.is parallel to Ink one order of the to princi-achieve Implementing simultaneousj=1 m=N+1 blue and red sideband ment. couples to the qubit in this setup, and to minimize the X Xdynamics of this theory and its beyond the Standard written as [] dices m =2N +1, 2N +2, , 3N. With the introductionpal axes of the trap. For example, byj setting ∆k = ∆k xˆ, N a Hamiltonian proportional to x, both sets of blue and transitions detuned from the normal mode frequencies1 of two counter-propagating··· laser beams witho↵ wave-vector-resonant transitions, the Raman beams could be set x x x For the simulation of the lattice Schwinger model, the the beamsred sideband only couple operations to transverse can be applied normal simultaneously modes along will not generate dynamical phonons as! longj (aj† asaj + the) lasers, (14) N such that they only couple to one set of normal modes of 2 bosons need to interact locally. Such an interaction will ⌦di↵erence3N k,frequencydii!mt i!↵merencet ! !L, and phase with a set of Raman beams such that the phases of the j=1 j i ⌘m,j (ame +a† e ) the x axis. If this condition is not met and more than are applied for certain time durations. The spin-phonon H0 = e m=1 m ⌘ motion. For example, by setting k = k xˆ, the beams X need to be mapped to a local phonon-phonon interaction. ion-laser di2↵ernece j j, detuned⇥ from an exited internalone settwo of sideband modes are beams coupled, add up the to ⇡ different. With this frequencies setting, the of j=1 P ⌘ only couple to transverse normal modes along the x axis. couplings in this process then effectively induce a spin- To engineer this new type of interactions in the trapped- X level of the ion, the two-phonon Raman transition amongthe modesHamiltonian can be corresponding addressed with to ion lowerj becomes power to avoid i(! ! )t+i + The qubit-phonon rotations of arbitrary angle ✓k,j,tobe spin interaction, which leads to the well-known Mølmer- the two statese ofL the0 qubitj can+h.c be. induced(2) with a Rabi ion simulator, as proposed in Ref. [], the ions can be j simultaneous a coupling to other sets of modes. By ap- 4 5 denoted as Rk,j(✓k,j, j), can be obtained by applying SørensenOr (MS) a geometrical gate [92 phase–94 gate]. The in case corresponding of spin interactions Hamil- propor- frequency ⌦j at the location of ion j. The beams can ad- a ⌘k,j⌦j i j i j x put near the minima (or maxima) of a standing wave, a a H ( j)= (e ak + e a†) (8) z z H ( j) forplying time ⌧ multiple, to bek,j beatnote deduced from frequencies the relation associatedk j with red tonian betweentional to ions j[].and j is where the conditiondress! theL ions!0 along!0 themust chain be globallyassumed. or individually.k,j In gate 2 ⌦ 0 which induces an AC Stark shift corresponding to the The prime on Hanis interaction to| denote | that⌧ picture this that Hamiltonian rotates with is free Hamiltonianand blue sideband transitions, corresponding to the set L i j i j x a a σσ ωm x x a ω Ini✓k order,jω(e =a tok+e ensure ω ak†for) onlyj k oneiH1k,,j of( thej ),N⌧gate phonon, and modes, given theak, in the interactionin picture. Eq. (1), Later the interacting on, we need ion-laser to adopt HamiltonianR (✓k can,j, bej) j e 0 k = e H = ΩjΩj ηm,jηm,j σ σ , (13) k,j {⌘ }couples − to±{ the qubit} in this∈ { setup,··· and} to minimize the j,j0 0 0 (ωL ω )2 ω2 j j0 adi↵erent interactionwritten picture as [] rotating with shifted fre- flexibility to set the amplitude and phase(9) associated with m j 0 m o↵-resonant transitions, the Raman beams could be set X − − quencies and so it is important to bear in mind the origin N such that they only couple to one set of normal modes of σσ 3N i! t Ini! particular,t setting laser beam phases to proper angles Spin-spin rotations Rj,j (θj,j0 ) of arbitrary angle θj,j0 can of Eq. (2). ⌦j i ⌘ (a e m +a† e m ) 0 m=1 m,j m m motion. For example, by setting k = k xˆ, the beams σσ σσ Hion-laser0 = e † be obtained by applying H for time τ given by the In the Lamb-Dicke regime2 in which can generate⇥ transitions proportional to ak + ak and ak j,j0 gate 1/2 j=1 P only couple to transverse normal modes along the x axis. 2 † 4 To achieve a Hamiltonian of the same type that is instead pro- ⌘m,j (am + am† ) 1, transitionsX in the space ak: The qubit-phonon rotations of arbitrary angle ✓ ,tobe h i ⌧ i(!L !0)t+i j + x k,j of coupled qubit-phonon system take simplere forms, and j +h.c. (2) portional to σj , the a red and blue sideband laser phases can be denoted as Rk,j(✓k,j, j),x can be obtained by applying tuned a to⇡ add up toi✓kπ,j.(a Finally,k+a†) phonon-ion Hamiltonian propor- can be realized through applications of quantum gates. R (✓k ,ja, ) e a k j , (10) tionalk,j Hk,j to( σjz)can for be time obtained⌧gate, in to a be circuit deduced by applying from the appropriate relation 5 Or a geometrical phase gate in case of spin interactions propor- where the condition !L !0 0 !0 must be assumed. 2j ⌘ The carrier transition corresponds to Hion-laser0 (⌘ ) and z z | O| ⌧ single-qubit rotations, as isx demonstrated in later sections. tional to σ σ [91]. The prime on H is to denote| that this Hamiltonian is a ✓k,j (ak ak†) j R (✓ , 0) e . i j i j (11)x a a is obtained by setting !L !0 = 0. With this setting, k,j k ,ja i✓k,j (e ak+e a†) iH ( j )⌧ ⊗ in the interaction picture. Later on, we need to adopt R (✓k,j⌘, j) e k j = e k,j gate the Hamiltonian corresponding to ion j becomes k,j ⌘ adi↵erent interaction picture rotating with shiftedImplementing fre- simultaneous blue and red sideband (9) ⌦j quenciesi i and sox it isi importanti toy bear in mindtransitions the origin now detuned from the normal mode frequen- H ( )= (e j + e j ) +(e j e j )i (3) j j 4 of Eq. (2). j j cies will not generateIn particular, dynamical setting phonons laser as beam long phasesas the to proper angles In the Lamb-Dicke regime in which can generate transitions proportional to ak + ak† and ak 2 1/2 ⌘m,j (am + am† ) 1, transitions in the space ak†: of coupledh qubit-phononi ⌧ system take simpler forms, and ⇡ x a i✓k,j (ak+ak†) j can be realized through applications of quantum gates. Rk,j(✓k,j, ) e , (10) 0 2 ⌘ The carrier transition corresponds to Hion-laser0 (⌘ ) and O x | a ✓k,j (ak ak†) j is obtained by setting !L !0 = 0. With this setting, Rk,j(✓k,j, 0) e . (11) the Hamiltonian corresponding to ion j becomes ⌘ Implementing simultaneous blue and red sideband ⌦j i i x i i y transitions now detuned from the normal mode frequen- H ( )= (e j + e j ) +(e j e j )i (3) j j 4 j j cies will not generate dynamical phonons as long as the 5 relation the accuracy of the simulated model, one can either im- pose tighter trapping potential along the transverse di- x x σσ σσ σσ iθj,j σj σ iH τgate R (θ ) e− 0 j0 = e− j,j0 . (14) rections or spread out the ions further by reducing the ax- j,j0 j,j0 ≡ ial potential. Additionally, the nearest-neighbor hopping In reality, designing high-fidelity two-qubit entangling can be actively blocked using techniques demonstrated gates of this type requires complex pulse shaping tech- in Ref. [107, 128] or by applying local optical tweezers niques to close all the mode trajectories in phase space, to vary the local confinement [129–132]. Another option, which is necessary to eliminate any residual spin-phonon which does not require additional blockade ions, is the coupling, and to leave the system in the same phonon use of a mixed-species ion chain with a large mass ra- state before and after the gate operation, see e.g., tio, in which the modes for the different species separate Refs. [123–127]. by mode participation [133]. In an alternating arrange- ment, phonon hopping will be suppressed by the local- mode frequency difference between neighboring ions of B. Phonons as excitations of local modes of motion different species and the increased distance between the ions of the same species. Finally, we note that the local- mode frequencies can be expressed as the common trap frequency plus additional local corrections. These correc- In a linear Paul trap, the Coulomb energy is compa- tions scale similar to the hopping strength and are hence rable to the potential energy associated with the ax- 3 suppressed by at least a factor of 10− compared with ial motion of the ions. Namely, the parameter βz 2 3 z2 z ≡ the trap frequency in a typical trap described above. As e /(d0Mionω ) & 1, where ω is the trap frequency x x a result, ωj in Eq. (15) can be replaced with ω to a first along the axial axis of the trap, d0 is the average spac- approximation. ing among the ions, and e is the electric charge. As a The spin-phonon Hamiltonians in Eq. (8) can be ob- result, the phonons are shared across the chain as exci- tained similarly by expanding the ion-laser interaction tations of collective normal modes. The same param- Hamiltonian in Eq. (2) using the local modes of motion. eter is typically much smaller for the transverse mo- x Then both the spin σj and the phonon operators (aj and tions of the ions. In particular, if the trap is made x extremely tight along the transverse axes of the trap aj†) are representing quantum operations at location j such that βx(y) 1, the transverse phonons describe along the chain, with the associated gate operation iden- the oscillation of a single ion, with a relatively small cou- tified as pling strength for hopping to the neighboring ions that is σa σa 2 3 x(y) x(y) x(y) Rj (θj, φj) = Rk,j(θk,j, φk,j) x , e /(d M ω ) = β ω [120]. For the quantum ak a ,ηk,j ηj ,θk,j θj ,φk,j φj ∝ 0 ion → j → → → simulation of QFTs that exhibit non-trivial boson-field (16) dynamics, it is necessary to adopt a local-mode picture. Assuming that the transverse directions of the trap sup- where the single-mode addressing limit of Eq. (10) is con- port local phonon modes, the free Hamiltonian of the ion sidered. On the other hand, in the limit of a tight trans- chain can be written as verse direction where the hopping of the phonons along transverse axes of the trap is suppressed, the spin-spin N N ω0 z x x x 1 gate in Eq. (14) can be still derived with high fidelity H = σ + ω (a †a + ) + ion 2 j j j j 2 using the normal modes of motion along the axial di- j=1 j=1 X X rection so to have sufficiently strong all-to-all spin-spin 2N 3N couplings, given that βz βx(y). As will be shown, both y y y 1 1 ω (a †a + ) + ωm(a† am + ), (15) types of gates, one based on local modes of motions and j j j 2 m 2 j=XN+1 m=2XN+1 the other based on normal modes of motion, can be em- ployed in the simulation, exhibiting another advantage of where in the second and third terms, the subscript j on a hybrid analog-digital setting. the phonon operators is a local index corresponding to For the simulation of the lattice Schwinger model, the th the quanta of motion of the j ion. bosons need to interact locally. Such an interaction will For the field theories considered in this work, the sim- need to be mapped to a local phonon-phonon interaction. ulated theory does not exhibit boson hopping and so the To engineer this new type of interaction in the trapped- phonon hopping in the simulator needs to be suppressed. ion simulator, as proposed in Ref. [120], the ions can be In a typical trapped-ion system with tens of Ytterbium placed near the minima (or maxima) of a standing wave, x ions, ω 2π 5 MHz and d0 is of the order of a few which induces an AC Stark shift corresponding to the ∼ × x 3 4 x x microns. Then β 10− 10− and β ω 2π 1 Hamiltonian kHz, and hence the dynamics∼ − associated with the∼ phonon× s.w. 2 2 x x hopping can be neglected compared with spin-phonon Hj = F cos (kxj) = F cos η(aj + aj†) (with typical strength 100 kHz) and spin-spin dynam- 2 4 x x 4 x x 2 ∼ = 2F ( η + η )(aj†aj) + 2F η (aj†aj) + . ics (with typical strength 10 kHz). If further sup- − e e ··· pression of the phonon hopping∼ is desired to improve (17) e e e 6
Here, k = k xˆ where k is the wave-vector of the standing- for quantum simulation of such theories, including the (0) k2 suitability of the hybrid analog-digital approach. In the wave beam and η kx = x is the correspond- e e ≡ e 2Mionω following, we focus on a simple case: a scalar field the- ing Lamb-Dicke parameter. Notee that the standing wave q ory coupled to one flavor of staggered fermions in 1+1 is only exciting the locale modes of motion along one di- e D and without the possibility of self-interactions among rection based on the choice of the wave-vector. Eq. (17) the scalar fields. Later, we comment on the applicability does not consider higher-order terms in η2 that can be 1/2 of this proposal for self-interacting scalar fields and the x x 2 neglected in the regime η (a + a †) 1. It also ne- higher-dimensional case, and will point out the requisite h j j i glects phonon non-conserving operators thate rotate with extensions. x at least the frequency ωe, and can be adiabatically elim- inated as long as F η2/ωx 1 [120]. Analogous to the previous gates, phonon-phonon ro- A. The Yukawa model tations of arbitrary angles can be obtained by applying s.w. aa e Hj for time τgate, to be deduced from the relation Consider a one-dimensional spatial lattice with N sites, lattice spacing b, and with periodic boundary conditions (1) x x (2) x x 2 aa (1) (2) i χ (a †a )+χ (a †a ) 6 R (χ , χ ) e− j j j j (PBCs) imposed on the fields. The Hamiltonian of the j ≡ iHs.w.τ aa lattice-regularized Yukawa theory to be simulated with a = e j gate . (18) − trapped-ion quantum simulator consists of An important point to notice regarding this gate is that H = H(I) + H(II) + H(III) , (19) according to Eq. (17), the relative size of the coefficients Yukawa Yukawa Yukawa Yukawa x x x x 2 (1) (2) of the aj†aj and (aj†aj) terms, χ /χ , is fixed to where the purely-fermionic Hamiltonian ( 1 + η2)/η2, while the overall strength of these terms − N can be tuned arbitrarily by changing F via the standing- (I) i j H = (ψ†ψ ψ† ψ ) + m ( 1) ψ†ψ wave intensity, as long as the condition F η2/ωx 1 is Yukawa 2b j j+1 − j+1 j ψ − j j e e j=1 not violated. The ratio constraint may seem too limiting X (20) for simulating an arbitrary Hamiltonian containinge both phonon terms. However, as shown in Sec.IV, the de- describes the hopping term and the mass term of one sired ratio of the coefficients in the simulation of the lat- flavor of staggered fermions with mass mψ. Note that tice Schwinger model can be engineered by appropriately PBCs impose the identification ψN+1 ψ1. The free choosing the laser frequencies, which amounts to choos- scalar-field Hamiltonian is ≡ ing a suitable interaction picture. Last but not least, if N 2 2 2 the simulated theory requires site-dependent coefficients (II) Πj ( ϕj) mϕ H = b + ∇ + ϕ2 , (21) for the phonon-phonon couplings, individual standing- Yukawa 2 2 2 j j=1 wave beams could replace the global beam in Eq. (17), X at the expense of added experimental complexity. where Πj is the conjugate momentum corresponding to
the scalar field ϕj, i.e. [ϕj, ϕj0 ] = [Πj, Πj0 ] = 0 and 1 [ϕj, Πj0 ] = i(Nb)− δj,j0 . These fields can be quantized in III. A YUKAWA THEORY: SCALAR FIELDS the standard way to obtain a representation in terms of COUPLED TO (STAGGERED) FERMIONS the (bosonic) harmonic-oscillator creation (dk† ) and anni- hilation (dk) operators, A scalar field theory coupled to fermions describes the N/2 1 coupling of the Higgs boson to fermions of the Standard − 1 1 i2πkj/N i2πkj/N Model through Yukawa interactions, and is responsible ϕj = (dk† e− + dke ), √ √2εk Nb k= N/2 for dynamical mass generation and fermion mixing in na- X− ture. Non-perturbative studies of the Yukawa theory us- (22) ing lattice-regularized Euclidean field-theory simulations N/2 1 − i εk i2πkj/N i2πkj/N have drawn considerable interest [134–141] as they reveal Πj = (d† e− dke ), √Nb 2 k − important connections to the cutoff scale of the Standard k= N/2 r X− Model and questions regarding quantum triviality [142]. (23) In the context of cosmology and the early universe, non- perturbative simulations are required for studies of non- where k labels the corresponding momentum mode p k ≡ equilibrium and real-time dynamics of this theory and 2πk 2 2 2πk/(Nb), and εk = ( ) + mϕ is the correspond- its beyond-the-Standard-Model cousins [143]. A scalar Nb ing energy, with m being the bare mass of the scalar field theory coupled to (non-relativistic) fermions is also ϕ q the effective field theory description of the Yukawa in- teractions among the nucleons and pions, and enters the quantum many-body description of nuclei [144–146]. It 6 Other boundary conditions can be studied with minimal adjust- is therefore highly relevant to investigate the prospects ments. 7
j j+1 A Yukawa theory: scalar ...... field coupled to fermions
'j 'j+1
Ions in a linear Paul trap
......
Collective normal modes are j j+1 used to simulate the dynamic of scalar field and to perform Internal states of the ion spin-spin entangling gates. are used to encode the a ,a , ,a ,a dynamic of fermions. { 1 2 ··· N N+1}
FIG. 2. The degrees of freedom of the lattice-regularized scalar field theory coupled to staggered fermions in 1+1 D (top row) are mapped to those in a linear trapped-ion quantum simulator (bottom row) with individual laser-beam addressing. This scheme involves only normal modes of motion.
field. Inputting Eqs. (22) and (23) in Eq. (21), and us- defined after Eq. (1), the mapping to the transverse nor- ing the commutation relations of the bosonic operators: mal modes along the x direction reads: dk := ak+N/2+1 [dk, dk ] = [d† , d† ] = 0 and [dk, d† ] = δk,k , one arrives and d† := a† for k = N/2, N/2+1,, ,N/2 0 k k0 k0 0 k k+N/2+1 − − ··· − at the well-known Hamiltonian 1. To keep the occupation of these modes unaffected N/2 1 while implementing spin-spin entangling gates, the other (II) − 1 set of transverse modes or the axial modes can be ad- H = ε (d† d + ), (24) Yukawa k k k 2 dressed to implement the MS gates. With the degrees of k= N/2 X− freedom in the simulated theory being mapped to qubit describing the energy of N uncoupled quantum harmonic and phonon degrees of freedom of the trapped-ion simu- oscillators. Finally, the interacting fermion scalar-field lator, what is left to identify are the gate operations that iHYukawat Hamiltonian is implement e− . Given the digital setting of this proposal, the time-evolution operator can be digitized N using the lowest-order Trotter-Suzuki expansion [147]. (III) HYukawa = gb ψj†ϕjψj, (25) Higher-order expansions [148, 149], as well other state-of- j=1 the-art simulation algorithms [150–158], can be similarly X investigated within the current approach, but will not be where field ϕ must be realized as the collection of quan- discussed further in this work. tum harmonic oscillators through Eq. (22). To map the Hamiltonian in Eq. (19) to the Hamiltoni- ans associated with the gates introduced in the previous B. The mapping to an analog-digital circuit section, a re-arrangement of the terms is performed such that HYukawa is now broken down to To map the Hamiltonian in Eq. (19) to the building blocks of the analog-digital trapped-ion simulators in- troduced in Sec.II, one may first transform the stag- 1 N H(I)0 = σxσx , (26) gered fermionic fields straightforwardly to the spin op- Yukawa 4b j j+1 j=1 erators through a Jordan-Wigner transformation: ψi = X z z + N l 2 N N is taken advantage of in the application of spin-phonon (IV )0 g b z 1 H = (Ij + σ ) Yukawa 8N j √ε × gates in the following. It should be emphasized again that r j=1 m=1 m X X experimentally, the change in the interaction picture cor- 2πj N 2πj N i N (m 2 1) i N (m 2 1) responds to detuning each of the spin-phonon operations. (am† e− − − + ame − − ) + Note that the (N + 1)th normal mode is not affected by N 1 the chosen interaction picture, as its dynamics are irrel- ε (a† a + ), (29) m m m 2 evant in simulating the scalar fields. m=1 X With these adjustments, the time evolution of the where the mapping to spin and phonon degrees of free- Yukawa theory for duration δt can now be correctly ob- dom has already been carried out. To generate the free tained considering scalar-field Hamiltonian in Eq. (24) for generic mode- (I)0 (II)0 (III)0 dependent coefficients ε ε , one can re- iH0 δt iH δt iH δt iH δt k m=k+N/2+1 e− Yukawa = e− Yukawa e− Yukawa e− Yukawa sort to a change of interaction-picture→ Hamiltonian [115]. × iH(IV )0 δt 2 The gates obtained so far are in an interaction pic- e− Yukawa + (δt) , (31) ture derived using the free Hamiltonian of the ion sys- O tem in Eq. (1), containing the harmonic-oscillator en- and upon expressing each evolution term as ergy term with mode-independent and experimentally- N fixed coefficients. However, one can introduce the term (I)0 1 iHYukawaδt σσ ε (a a + ) in the interacting Hamiltonian by ap- e− = j,j+1(θj,j+1), (32) m m m† m 2 R j=1 propriately choosing the rotating frame. εm is the prop- P Y erly rescaled εm accounting for the ratio of the model time N e iH(II)0 δt σσ variable and the experimental gate time, see the descrip- e− Yukawa = S S (θ )S†S† , (33) e j j+1Rj,j+1 j,j+1 j j+1 tion after Eq. (35). In particular, setting the free Hamil- j=1 1 Y tonian to Hion m εm(am† am+ ), the Hamiltonian in the − 2 δt interaction picture becomes H + ε (a a + 1 ). with θj,j+1 = , P ion-laser0 m m m† m 2 4b Operationally, this impliese adjusting the detuning of the P N red and blue sideband transitions, in other words tuning (III)0 e iHYukawaδt z L e− = Rj (θj), (34) the laser frequency to ωj ω0 = (ωk εk), which would − ± − 1 j=1 implement the desired mode-dependent term εk(ak†ak + 2 ) Y in the evolution. e 1 j with θj = 2 mψ( 1) δt, and However, this is not the full picture yet, sincee the evo- − lution of the Yukawa theory will be implemented in a N (IV )0 π digital manner. In particular, there will be multiple side- iHYukawaδt σ σa e− = Rj ( , 0)R m ,j( θm,j , φm,j ) 4 { } { } { } band operations (when implementing spin-phonon gates) j=1 during each step of the Trotter evolution, and these effec- Y σ π σ π tively implement the phonon-energy term as well. There- Rj ( , 0)RN+1( , 0) × − 4 4 fore, one must ensure that this term will be induced σa σ π with the correct coefficient. As will be seen below, im- R m ,N+1( θm,N+1 , φm,j )RN+1( , 0), × { } { } { } − 4 plementing the time evolution of the Yukawa theory re- (35) quires introducing an ancilla ion and amounts to a to- tal of N + 1 multi-mode spin-phonon gates per Trotter g2b 2πj N with θm,j = θm,N+1 = δt and φm,j = (m step. This necessitates changing the free Hamiltonian 8Nεm N − 2 − 1 1 to Hion εm(a† am + ), which leads to the 1), and with all valuesq of m in the range 1, ,N . − (N+1) m m 2 σ { ··· } interaction-picture Hamiltonian Rj (θj, φj) is defined in Eq. (4) and its operation re- P mains the same despite the change in the interaction- N e ε σa L i(ω m )t picture Hamiltonian. R ( θm,j , φm,j ) is defined in Ωj i η (a e− m− (N+1) + m ,j H00 = e m m,j m { } { } { } ion-laser 2 e Eq. (10) but must be realized with red and blue sideband j=1 L P L 1 σa X X detunings ωj ω0 = (ωk (N+1) εk) with εkτgate = εkδt, ε i(ω m )t σa − ± − m− (N+1) L L am† e ) i(ωj ω0)t+iφj + where τgate is determined from Eq. (10) with the θm,j e e− − σ + h.c. + j values specified above. It should bee notede that expecta- tion values of observables are invariant under the change 1 N 1 ε (a† a + ), (30) in the interaction picture as long as the time-dependent (N + 1) m m m 2 m=1 states are transformed accordingly. For simple observ- X ables such as fermion and boson occupations, measure- e where L is introduced to denote the possibility of si- ments in the original basis obtain the correct expecta- L multaneous application of N beatnote frequencies ωj tion values as the corresponding operators commute with P 1 ≡ σσ ω0 (ωk εk) for all k 1, ,N with their asso- the transformation [159, 160]. Finally, R (θj,j ), de- ± − N+1 ∈ { ··· } j,j0 0 ciated amplitude ΩL Ω and phase φL φ , which fined in Eq. (14), can operate using transverse normal j ≡ k,j j ≡ k,j e 9 Free fermion terms Ancilla ion Free scalar field and fermion- scalar field interactions FIG. 3. The schematic of the analog-digital quantum circuit associated with the time evolution of a four-site Yukawa theory for a single Trotter step, as expressed in Eqs. (32)-(35). The gate symbols are defined in Fig.1. modes other than those used in spin-phonon gates, or ax- for example through a quartic interaction Hamiltonian: ial modes. Note that an ancilla qubit, labeled as N +1, is (IV ) d 4 introduced in Eq. (35) to effectively implement the inter- HYukawa = b λ ϕj . (36) j actions proportional to Ij in Eq. (29)[115]. This ancilla X qubit is an extra ion prepared in the spin up state and This term represents non-local interactions among quan- can be used in all subsequent Trotter steps. A schematic tum harmonic oscillators in momentum space, requir- of the circuit for a single Trotter step is shown in Fig.3. ing all-to-all phonon-phonon couplings to be engineered We will study the benefits of this hybrid proposal for the in the trapped-ion simulator. While inducing phonon- simulation of Yukawa theory in Sec.VA. phonon coupling among the normal modes is possible by taking advantage of the intrinsic non-linearity of the The algorithm above can be generalized to a Yukawa Coulomb interaction [165] or through mediating the in- theory in higher dimensions. In d spatial dimensions, teraction via virtual spin degrees of freedom, such an im- d the number of sites on the lattice is N , where N is the plementation will be more challenging than the other set number of sites along each Cartesian direction. The num- of gates introduced so far. The use of the local phonon d ber of ions required to fully encode the dynamics is N , modes will not be optimal either, as the phonon hop- plus a single ancilla ion as introduced in the 1+1 D case. ping is suppressed beyond nearest-neighbor sites, and its This is because the number of normal modes of motion strength cannot be made homogeneous across the sites as associated with each of the transverse (or axial) direc- required by Eq. (36). Searching for efficient and feasible d tions will also grow as N (considering the ancilla ion, implementations of the Hamiltonian in Eq. (36) that are d as (N + 1) ), which is more than sufficient to encode more natural to a the trapped-ion simulator will be the d the N momentum modes of the scalar field in the har- subject of future work. d monic oscillator basis. Each ion needs to couple to all N The model considered in this work is still of phe- phonon modes, which polynomially increases the number nomenological interest, for example in simulating lattice of spin-phonon gates required to simulate the fermion effective field theory of nucleons coupled to pions. The scalar-field interaction term. The more significant over- self-interactions of pions matter only at higher orders in head in terms of the entangling operations is caused by the chiral effective field theory of nuclear forces [145, 146] the encoding of the fermionic hopping term, which in the and can be neglected at low energies. Besides the triv- Jordan-Wigner transformation involves implementing a ial incorporation of multiple flavors of (non-relativistic) chain of Pauli operators, with the number of Pauli op- fermions representing spin and isospin components of a erators growing polynomially with N. These operations nucleon, and the (local) self-interaction of fermions rep- can be performed through the known decomposition into resenting nucleons two- and three-body contact interac- spin-spin gates, either in series or in parallel. The paral- tions, the fermion scalar-field interacting term in Eq. (25) lel implementation can be achieved either in one go using must be promoted to a derivative coupling in order to a global MS-operation and shelving the ions that do not represent a pion-nucleon coupling. Such a term can be participate in the coupling using individual beams [161], implemented by representing the derivative of the scalar or by more involved optical pulse-shaping methods as field by a finite difference among the fields at adjacent demonstrated in Refs. [162–164]. sites. This then amounts to multiple implementations of the terms of the type in Eq. (25) that are added with ap- Another important generalization of the Yukawa the- propriate signs. Further, there are three different types ory is to incorporate self-interactions of the scalar fields, of pions, each requiring their own harmonic-oscillator 10 representation (two of which being electrically charged). fermions and antifermions allowed by symmetries, and in- These can either be realized through the three sets of volving varying boson occupation whose average is shown normal modes or by enlarging the ion chain to create in the inset of the plots. The unfilled points represent more normal modes of one type for encoding of all the the same quantities evaluated using a Trotter expansion pionic degrees of freedom. As the building blocks of this of the evolution operator according to the split in the construct are identical to the ones for the simple Yukawa Hamiltonian terms outlined in the previous section. The model above, the explicit circuit will not be discussed coarsest Trotter digitization is chosen such that quanti- further. ties can still be described accurately in the time window specified. The corresponding model parameters, shown in Ta- C. An example with realistic parameters bleI, give rise to gate rotation angles denoted in TableIV of AppendixA1 along with associated experimental gate parameters for the Rσa gate, including gate’s operation To demonstrate the viability of near-term experimen- time, in TableV. The gate time required is shorter (a tal implementations of the dynamics in this model with few tens of microseconds for reasonable experimental pa- a trapped-ion simulator, one can investigate the range of rameters) than that of the spin-spin entangling gate (a the gate parameters required to observe interesting phe- few hundreds of microseconds), which is a great improve- nomena in this model. While only small instances of the ment over fully-digital implementation that would have problem can be studied classically, as is evident from the required multiple entangling spin-spin gates, being an or- examples considered, the quantum simulator with tens of der of magnitude slower in general. We will come back ions can still operate in the same gate-parameter range to gate-number scaling of the analog-digital implemen- and already push the limits of classical capabilities, par- tation compared with the fully-digital implementation in ticularly given the native implementation of boson dy- Sec.V. The spin-spin entangling gates required in simu- namics in the simulator. Later on in Sec.V, we demon- lating the solely fermionic dynamics can be driven with strate the advantage of an analog-digital approach by the rotation angles specified in TableIV using well-known making a qualitative cost comparison with a fully-digital laser pulse-shaping techniques. The trap and laser char- implementation of the same model. acteristics, including transverse normal modes of motion An interesting phenomenon in the Yukawa theory con- and the Lamb-Dicke parameter, are provided in TableVI. sidered is the dynamical generation of mass even if the A final technical detail regarding the experimental im- bare fermion mass is set to zero originally. Such an effec- plementation is that all normal-mode vectors must have tive mass controls the rate of fermion generation through- non-zero components so that all Rabi frequencies Ω are out the evolution and depends on the boson accumulation j finite when applying the Rσa gate. To ensure that each on the lattice. As Fig.4 demonstrates for small lattice ion in the simulation couples to all of the phonon modes sizes N and small boson occupation cutoffs Λ, such a used, the total number of ions must be even. Therefore, non-trivial fermion, anti-fermion, and boson occupation studying an N-site theory of the 1+1 D Yukawa model, evolution can be revealed in the Loschmidt echo, i.e., where the number of staggered sites N is even, generally ψ(0) ψ(t) 2, where ψ(0) is the state with no fermion, requires N + 2 ions: one ancilla, which is operated on no| h antifermion,| i | and no boson. After a quench, this state by the specified gates in Fig.3, and one idle ion that is evolves to a superposition of states with any number of only present to ensure that for any given mode, no ion is stationary along the chain. Model parameters IV. U(1) LATTICE GAUGE THEORY COUPLED b N Λ g mψ mϕ εm δt t TO (STAGGERED) FERMIONS IN 1+1 D √ The Schwinger model, the theory of quantum electrody- 1 2 8 5 2 1 1 {3.297, 1} 0.25 5 namics in 1+1 D, has long served as a testbed for simula- tion methods (both classical and quantum) for strongly- interacting gauge theories. It is qualitatively similar to 1 4 1 5 1 1 {3.297, 1.862, 1, 1.862} 0.125 2.5 quantum chromodynamics, the theory of the strong force in nature, as it exhibits confinement, chiral-symmetry breaking, and a nontrivial θ vacuum [166, 167]. Being TABLE I. The parameters of the Yukawa theory considered a low-dimensional and Abelian theory, it is simpler than in the two examples of this section, each corresponding to its higher-dimensional and non-Abelian counterparts and different N and Λ values as noted. The corresponding trap its lattice-regularized form has been the subject of valu- and gate parameters are tabulated in AppendixA1. t is the able quantum-hardware implementation benchmarks in total evolution time and δt is the duration of the Trotter- recent years [54–57, 65, 66]. Fully-analog proposals for evolution segment. Hamiltonian simulation of the lattice Schwinger model 11 Hamiltonian���� can be written as ��� � N =2, ⇤ =8 ▽▽ ����� ▽▽▽ N =8, ⇤ =2 ▽▽ ▽▽▽ � ▽▽ ▽ ▽▽▽ i ▽ ���� (I) �����(II) ▽ (III▽▽▽) ▽ i ��� d ▽ ▽ ▽ ▽▽ ▽▽▽ � ▽ ▽ HU(1) = HU(1) + Hd U(1) +▽ HU(1) ,▽▽ (37) N ▽ ▽ ����� ▽▽ ▽▽ h ▽ ○ N ▽ ▽▽ 2 2 ▽ ○ h ▽ ▽ | � | ▽ ▽▽ ���� �� ▽ ▽▽▽▽ i ○ ▽ ▽▽ i ○ ▽ ) ��� � ) t t where the gauge-matter interacting���� Hamiltonian, the ( � � � � � � ( ���� � � � � � � | | staggered fermion-mass Hamiltonian, and the electric- ��� ○ ○ (0) ○ ○ (0) field���� Hamiltonian are ○○○ h ○ h ○ ○ ○ | ��� ○ | ○ ○ ○ ���� N ○ ○○ (I) i○○ ○ ○ ○ ○ ○ ○ ○ ○ ○ HU(1) = ○ (ψj†Uj†ψj+1 ○ ψjUjψj†+1), ○○ (38) ��� ○ ○ ○ ○ ○ ○ ○ ○ ○ ���� ○○○ 2b ○ ○ ○○○○ − ○ ○○ ○ j=1 � � � � � � X � � N � � � � (II) j ��� ▽ ����H = m ( 1) ψj†ψj, (39) N =4, ⇤ =1 ▽ ▽▽ U(1) ��� ▽ N =8, ⇤ =4− ���� ▽▽ ▽ j=1 i ▽ ▽ ���� X ▽▽▽▽▽▽▽ d ���� ▽ ▽▽ ��� ▽ ▽ ▽ ▽ N i ▽▽ ▽▽▽▽ ��� d ▽ ▽▽ N ▽ 2 ▽ ▽▽▽ 2 ○ ▽ 2 ���� ▽ h ▽▽ g b ▽ | | ��� ▽ ○(III) 2 N ▽▽ h i i ▽ ▽ ����H =○ E , ���� ▽▽ (40) ) ○ ) U(1) ○ j ▽▽▽ t ��� t 2 ▽▽▽▽▽ ( ��� ( ▽ ○ ○ ��� ��� ��� ��� ��� ��� j=1 �� ▽ X | ○ | ���� � � � � � � ○ (0) ��� (0) ○ respectively.���� The Hilbert space of the theory is char- ○ ○ h ○ h (g) (f) | ○ ○ | acterized by on-site quantum○ ○ numbers○ n and n as- ��� ○ j j ���� ○ ○ ○ sociated with the○○ discrete○ spectrum○ of○ a○ quantum○ ro- ○ ○(g) (g○) (g) (g) ○ ○ ○ ○ ○ ○ ○ ○ ○○ ��� ○ ○ tor���� satisfying○○○ Ej n○j ○ =○○nj○ n○j○○○ and○ U○j n○j = (g) (g|) i | i | i ��� ��� ��� ��� ��� ��� n +� 1 with�n Z� , along� with fermionic� occupa-� | j i j ∈ t t (f) (f) tions eigenstates satisfying ψj nj = δn(f),1 nj 1 | i j | − i (f) (f) (f) FIG. 4. The overlap between an initial state with no † (f) and ψj nj = δn ,0 nj + 1 with nj 0, 1 . The fermion, no antifermion, and no boson, and the corre- | i j | i ∈ { } 2 physical states are those annihilated by the Gauss’s law sponding time-evolved state, |hψ(0)|ψ(t)i| , (in blue) along 1 j operator Gj = Ej Ej 1 ψj†ψj + 1 ( 1) at each with the average number of bosons generated hNdi ≡ − − − 2 − − 1 hψ(t)| PN/2−1 d† d |ψ(t)i, (in red) as a function of time, site. In the following, it is assumed that the simulation N k=−N/2 k k starts in a physical state and hence will remain in the t, using exact (solid curve) and Trotterized evolution (unfilled points) for the Yukawa theory. The corresponding model pa- same sector, as long as the gauge-symmetry violations rameters are given in TableI. The time t is in units of lattice arising from the digitization of the time-evolution oper- spacing b, which is set to one. ator, or from hardware imperfections, remain small. Unfortunately, a quantum rotor representing the gauge link in the Schwinger model does not have the same have been put forward using spin degrees of freedom algebra as the phonon modes of the trapped-ion sim- only [32], and both the spin and phonon degrees of ulator for a direct encoding. A highly-occupied boson freedom [112], but they remain relatively challenging model (HOBM) was proposed in Ref. [112] to resolve for hardware implementation. Here, we propose a hy- this mismatch by considering the following mappings: brid analog-digital simulation of the lattice Schwinger Uj dj/√M and Ej d†dj M for an integer M. → → j − model that brings the simulation proposals involving the This transformation keeps the commutation relation be- phonon degrees of freedom a step closer to experimental tween Ej and Uj intact, but modifies that between the realization. gauge links to [Uj,Uj0 ] = δj,j0 /M. Only in the limit M N, with N being the number of sites on the lattice, the HOBM will recover the lattice Schwinger model, necessitating the simulation involving the bosonic A. The Schwinger model degrees of freedom to initiate in a state with a large number of bosons. Nonetheless, the numerical simu- The well–known Hamiltonian of the lattice Schwinger lations of Ref. [112] demonstrate that for M N, a model, introduced by Kogut and Susskind [166, 167], high level of accuracy is still achieved, particularly∼ in can be written in terms of a one-component staggered space- and time-averaged dynamical observables. Apply- fermion ψj defined on site j (with odd and even sites ing the HOBM mapping as well as the Jordan-Wigner corresponding to matter and anti-matter fields, respec- transformation, the spin-boson Hamiltonian of the lat- tively), the gauge link Uj and its conjugate field, the tice Schwinger model can be written as electric field Ej, both defined on the link emanating from site j, and satisfying the commutation relations (I)0 (II)0 (III)0 HU0 (1) = HU(1) + HU(1) + HU(1) , (41) [Ej,Ej0 ] = [Uj,Uj0 ] = 0 and [Ej,Uj0 ] = Ujδj,j0 . The 12 Ej,Uj E ,U { } { j+1 j+1} Lattice Schwinger model ...... j j+1 Ions in a linear Paul trap j j+1 am ...... Collective normal modes x x aj aj+1 used to perform spin-spin entangling gates. Local transverse modes used Internal states of the ion to encode the dynamic of the are used to encode the gauge fields. dynamic of fermions. FIG. 5. The degrees of freedom of the lattice-regularized Schwinger model in the HOBM (top row) are mapped to those in a linear trapped-ion quantum simulator involving local and normal modes of motion (bottom row). where be implemented by the standing-wave gate introduced in N Eq. (18) (I)0 1 + + H = (σj dj†σi−+1 + σi+1djσj−) (42) iH0 δt U(1) 2b√M The Trotterized evolution operator is e− U(1) = j=1 (I) (II) (III) X iH 0 δt iH 0 δt iH 0 δt 2 N e− U(1) e− U(1) e− U(1) + (δt) , where 1 x x y y O = σ (dj + d†)σ + σ (dj + d†)σ + 8b√M j j j+1 j j j+1 j=1 X N (I) x y y x iH 0 δt σσ π σ π iσj (dj dj†)σj+1 iσj (dj dj†)σj+1 , (43) e− U(1) = S ( )R ( , 0) − − − jRj,j+1 4 j 4 × j=1 N Y (II)0 m j z σa σ π σσ π HU(1) = ( 1) σj + const., (44) R (θ , 0)R ( , 0) ( )S† 2 − j j j − 4 Rj,j+1 − 4 j j=1 X N 2 N π π (III)0 g b 2 2 σσ σ H = (d†d M) (45) Sj Sj+1 j,j+1( )Rj ( , 0) U(1) 2 j j − R 4 4 × j=1 j=1 X Y 2 N σa σ π σσ π 2 g b 2 Rj (θj, 0)Rj ( , 0) j,j+1( )Sj† Sj†+1 = 2M(d†dj) + (d†dj) + const. (46) − 4 R − 4 2 − j j j=1 N X π π S S σσ ( )Rσ( , 0) j j+1Rj,j+1 4 j 4 × j=1 Y σa π σ π σσ π B. The mapping to an analog-digital circuit R (θ , )R ( , 0) ( )S† S† j j 2 j − 4 Rj,j+1 − 4 j+1 j This Hamiltonian can be directly mapped to spin and N π π S2 σσ ( )Rσ( , 0) phonon degrees of freedom in the trapped-ion simulator. j Rj,j+1 4 j 4 × j=1 Noting that the electric-field Hamiltonian in the HOBM, Y (III)0 H , consists of local self-interactions of the bosonic σa π σ π σσ π 2 U(1) R ( θ , )R ( , 0) ( )S† , (47) j − j 2 j − 4 Rj,j+1 − 4 j modes, it is necessary to consider a tight trap in the transverse directions, as introduced in Sec.IIB, to take advantage of the local modes of the motion. In particular, the boson encoding should be realized by the mapping where each square bracket corresponds to the time evo- d ax, and the local self-interaction of the phonons can lution of each of the four terms in Eq. (43) and θ = j → j j 13 Fermion-gauge interactions ··· ··· Gauge-field interactions Fermion mass term FIG. 6. The schematic of the analog-digital quantum circuit associated with the time evolution of a four-site Schwinger model for a single Trotter step, as expressed in Eqs. (47)-(49). The fermion-gauge interaction block must be repeated four times with various phases for the spin-phonon gates and single-spin rotations, see Eq. (47). The gate symbols are defined in Fig.1. 1 x x δt, trapped-ion system by a term proportional to a †a 8b√M j j j in a suitable rotating frame. In this way, the correspond- N ing term in the interaction-picture Hamiltonian,P i.e., the (II)0 iHU(1) δt z e− = Rj (θj0 ), (48) first term in brackets in Eq. (46), can take an arbitrary j=1 coefficient, hence relaxing the condition on the ratio of Y x x x x 2 aj†aj and (aj†aj) terms. This will amount to chang- 1 j with θj0 = 2 ( 1) mδt, and with the gates defined in ing the frequency associated with red and blue sideband − σa Secs.IIA, and transitions involved in the operation of Rj (θj, φj) ac- cordingly, as detailed before for the case of the Yukawa N (III)0 theory. iHU(1) δt aa (1) (2) e− = Rj (χ , χ ), (49) j=1 Y Simulating QED in higher dimensions requires engi- (1) 2 (2) 1 2 with χ = g bMδt and χ = 2 g bδt, and with the neering the magnetic Hamiltonian, that is the sum of gates defined− in Secs.IIB. The schematic of the corre- the product of gauge links along the elementary pla- sponding circuit is shown in Fig.6. Note that the MS quette in each two-dimensional plane in coordinate sys- σσ gate j,j+1 in Eq. (47) can be implemented using the tem. In the HOBM of QED, the gauge links on these normalR modes of the motion along the axial direction, as higher-dimensional lattices can still be mapped to the lo- the phonons involved in its implementation are virtual cal modes of motion in a linear chain of ions. However, and not contributing to the dynamics of gauge fields in engineering the magnetic Hamiltonian requires achiev- the simulated theory. Furthermore, the ratio χ(1)/χ(2) in ing four-body interactions among the phonons pinned to Eq. (49) is fixed in the gate design as mentioned before, four adjacent ions, which will be challenging given the and one must ensure that χ(1)/χ(2) = ( 1 + η2)/η2 = experimental setup explained in this work. This is be- 2M, with the standing-wave Lamb-Dicke− parameter η cause the nearest-neighbor phonon hopping needs to be − 2 defined in Sec.IIB. As typically η 1, the requiremente e simultaneously suppressed to simulate the electric-field M N is guaranteed, but the careful tuning of η in ex- Hamiltonian accurately. Analog and digital simulations e periment is required as the thermodynamice limit of the with quantum simulators are proposed for QED in 2+1 HOBM (N ) is taken. If the necessary variatione in D with pure gauge interactions, using both the HOBM or η is not feasible→ ∞ in a given experimental setup, one can other models [49, 168–170]. Future work will investigate adjust the detuning of the laser beatnotes, as explained the design of an analog-digital simulation approach for ine Sec.III, to effectively shift the free Hamiltonian of the this problem. 14 C.��� An example with realistic� parameters ���� N =2, ⇤ =8 ▽▽ N =8, ⇤ =2 ����� � ▽▽ i ▽ ����� ▽▽▽▽▽▽▽▽▽▽ ▽ i ��� d ▽ ▽ ▽ ���� ○ ▽▽ ▽▽ � ▽ ▽ ○ d ▽▽ ▽ N ▽ ▽ ����� ▽▽ ▽▽ ▽▽ h ▽ ▽ N ▽ ▽ ▽▽ ▽ 2 Similar2 to the case of the Yukawa theory, near-term ▽ ▽▽▽ ▽ h ▽ ▽ | | � ▽ ▽▽ �� ▽ ▽ i i ○ ▽ ▽▽ ���� ▽▽ ) experimental) ��� implementations of the� Schwinger model t t ���� ( with realistic( analog-digital gate parameters� � � will� be� vi-� ○ � � � � � � ���� | | ○ able for small��� to intermediate○ systems and can be scaled ○ ○ (0) (0) ○ ○ ○ up straightforwardly similar to what is expected for dig- ���� ○ ○ ○ ○ ○ ○ ○○ h h ○ ○ ○ | | ○ ital implementations.��� It is therefore useful to○ consider ○ ○ ���� ○ ○ ○ ○ ○ ○ small instances of the problem that can○ be simulated○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ��� ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ classically to obtain the range of expected parameters ���� ○○ in experiment� for which� interesting� phenomena� � in the� � � � � � � model can be observed. Among many interesting fea- ��� ▽ ���� N =4, ⇤ =1 ▽ ▽▽ tures of the Schwinger model are the��� fermion-antifermion▽ N =8, ⇤ =4 ���� ▽▽ ▽ i ▽ ▽ pair creation and the string-breakingd dynamics. These ���� ��� ▽ ▽ ▽ ▽ ���� i ▽▽▽▽▽▽▽▽▽▽ ��� ○ d ▽▽ N ▽ ○ ▽▽ 2 ○ ▽ 2 ���� quantities can be evaluated fromh expectation▽▽ values of ▽▽ | | ��� ▽ N ▽▽▽▽ h i i ▽ ▽ ���� ▽▽▽▽▽ fermionic) or bosonic observables in a state evolved from○ ) ���� ▽▽▽▽▽▽ t ��� t ▽▽▽▽▽▽ ( ��� ( ▽▽ ○ ○ ��� ��� ��� ��� ��� ��� �� a trivial initial state, ψ(0) , such as the vacuum of | | i ○ | ���� ○ � � � � � � the model in the limit g , i.e., a○ state with no ○ (0) ��� → ∞ ○ (0) ○ fermion, no antifermion,○ and M○ bosons. Denoting the ○ ○ ○ ○ h ○ h ���� | | ○ ○ ○ ○ ○ ○ time-evolved��� state at time t as ψ(t) , we simply○ consider ○ | i 2 ○○○ ○○ ○ ○ the Loschmidt echo, i.e.,○ ψ(0) ψ(t) . ψ(t) involves ���� ○ ○ ○ ○ | h | ○ i | | i ○ ○○ ○○ ○ ○ a superposition of○ states○ with any number○ of fermions○ ○ ○○○ ○ ��� ○ ○ and antifermions allowed by symmetries, and involving ���� ○ ��� ��� ��� ��� ��� ��� � � � � � � varying boson occupation whose average is shown in the t t inset of the plots in Fig.7. The coarsest Trotter digitiza- tion is chosen such that quantities can still be described FIG. 7. The overlap between an initial state with no accurately in the time window specified, and the corre- fermion, no antifermion, and M bosons, and the corre- sponding values are marked with unfilled points. Λ de- sponding time-evolved state, |hψ(0)|ψ(t)i|2, (in blue) along notes the cutoff on the excitations of the bosons above with the average number of phonons generated, hNdi ≡ the M bosons in the initial state. It is notable that the 1 PN † N hψ(t)| j=1 dj dj |ψ(t)i, (in red) as a function of time using boson cutoff effects are significant even for such a small exact (solid curve) and Trotterized evolution (unfilled points) lattice, motivating further the need for mapping a larger for the Schwinger model within HOBM. The corresponding bosonic Hilbert space to the quantum simulator, which is model parameters are given in TableII. The time t is in units done naturally in the analog-digital proposal of this work of lattice spacing b, which is set to one. using the local phonon modes in the trap. Since the trap is in the tight-binding regime along the transverse directions, the local-mode frequencies at the location of each ion are equal to a good approximation, simplifying the parameters of the gates applied to each ion. For the phonon-phonon gate, it is ensured that the priately. Furthermore, to generate a specific value of the 2 x x x x x 2 parameter F η /ω is small by choosing η and F appro- ratio of the phonon terms a †a and (a †a ) as required by the HOBM, the red and blue sideband transitions e e in the spin-spin and spin-phonon gates must be imple- Model parameters mented with a frequency shift δωx, corresponding to the necessary modification to the− interaction-picture Hamil- tonian. The shift value, along with other relevant trap b N Λ g m M δt t parameters, are tabulated in TableX of AppendixA2. As noted in TablesVIII andIX, the spin-phonon gate time is similar to what was obtained for the Yukawa the- 2 ory. The phonon-phonon gate time needs to be 100 1 8 0.35 1 10 0.125 5 ∼ 4 microseconds to prevent the need for a large value of the standing wave amplitude. Last but not least, the Lamb- Dicke parameters associated with the Raman beams as TABLE II. The parameters of the Schwinger model within well as the standing-wave beam must satisfy η√M 1 the HOBM considered in the examples of this section. The and η√M 1, respectively. This is because the system corresponding trap and gate parameters are tabulated in Ap- must be prepared in a state with phonon occupation M pendixA2. t is the total evolution time and δt is the duration accordinge to the HOBM. These conditions are satisfied of the Trotter-evolution segment. with the realistic experimental parameters chosen. 15 V. A QUALITATIVE COST ANALYSIS mentum mode is truncated at some cutoff value Λ and is expressed in a binary representation, which is then In this section, our hybrid analog-digital approach will be mapped to log Λ qubit registers. The evolution opera- (II) 1 iH δt i ε (d† d + )δt compared with the fully-digital realization of the Yukawa tor e− Yukawa = e− k k k k 2 can then be real- theory and the Schwinger model. In the long term, when ized by a number of single-qubitP rotations around the z fault-tolerant digital quantum computation will be a real- axis of the qubits’ Bloch sphere (see e.g., Ref. [39] for ity, the non-Clifford gate (T-gate) count determines the a similar example of implementing the time evolution cost of the simulation. In the near-term and particu- of the electric-field Hamiltonian in the lattice Schwinger larly over the next decade, however, quantum compu- model). This step, therefore, can be considered cost-free tations will be limited by the slow and error-prone ap- in both digital and analog-digital implementations. plication of entangling (CNOT) gates. In the analog- Next, the time evolution with the fermion scalar-field digital protocol of this work, while the hardware-specific (III) Hamiltonian HYukawa in Eq. (25) is implemented with a spin-phonon gate is an entangling gate in the combined single ancilla qubit and with (N(N + 1)) spin-phonon Hilbert space of the phonons and qubits, its operation gates, which are considered toO be cost-free compared with is governed by dynamics that occur at (η), where η is the entangling spin-spin gates. In a fully-digital imple- O (III) the Lamb-Dicke parameter introduced in Sec.II. This is iH δt mentation, e− Yukawa can be implemented by writing in contrast to the entangling spin-spin operation that is H(III) as the core of the CNOT implementation in a trapped-ion Yukawa digital computer, which is governed by dynamics that oc- g2b 1 2πkj cur at (η2). Since η 1 in the Lamb-Dicke regime (III) HYukawa = cos ψj†ψj(dk + dk† )+ O 2N εk N in which experiments operate, the spin-spin gates are r j k r X X one or two orders of magnitude slower than single-spin 2 and spin-phonon gates. Phonon-phonon self-interaction g b 1 2πkj sin( )ψj†ψji(dk dk† ). 2N εk N − gates based on standing-wave beams are governed by r j k r 4 X X (η ) dynamics but their strength can be compensated (50) byO laser power, as demonstrated by the examples stud- iede before. Without experimental realizations, the fi- The implementation of these terms follows the procedure delity of this new set of gates cannot be accurately pre- explained in Ref. [39] in the case of realizing the dynam- dicted. Nonetheless, the qualitative expectations for the ics of fermion-gauge interaction term in the Schwinger gate times as described here can provide a rough guid- model. First, Eq. (50) indicates that the operations pro- ance to the relative performance of the simulator in an portional to A (dk + dk† ) and B i(dk dk† ) opera- analog-digital mode compared with a fully-digital mode. tors are only performed≡ if the fermion≡ occupies− a given site, necessitating a controlled operation on the qubit reg- ister of the fermion. The A and B operators are two A. The Yukawa theory near-diagonal matrices whose exponential can be imple- mented using the shift operators, that are realized using Since the implementation of the dynamics associated quantum Fourier transform circuits and single-qubit ro- with the fermionic hopping and mass is the same in both tations in the Fourier space. As a reminder, for a bi- digital and analog-digital circuits, here we focus on im- nary number with log Λ digits, each quantum Fourier plementing dynamics involving the scalar field, namely transform requires ((log Λ)2 +log Λ) CNOT operations. evolution with the free scalar-field Hamiltonian as well Relating A and BOoperators to the shift operators re- as the fermion scalar-field interacting Hamiltonian. quires a periodic wrapping of the matrices, i.e., identi- In the analog-digital protocol, the time evolution of fying the least and most values of harmonic-oscillator (II) occupation. This unphysical modification can subse- the free scalar-field Hamiltonian HYukawa in Eq. (24) is implemented at no cost since it amounts to adjust- quently be removed by application of appropriate log Λ- controlled operations, amounting to (log Λ) additional ing laser beatnote frequencies in the sideband opera- O tions, as explained in Sec.IIIB. In the fully-digital im- CNOT gates [39]. Putting everything together, including plementation, the exact circuit design depends on the the controlled operations required on the fermionic reg- mapping of the scalar fields to qubit registers. The ister, and taking into account all the terms in the lattice iH(III) δt field basis [13], the harmonic-oscillator basis [29], and sums in Eq. (50), the time evolution operator e− Yukawa the single-particle basis [44] are among the representa- can be implemented using (N 2(log Λ)2) CNOT opera- tions considered in literature to perform such a mapping, tions. Therefore, assumingO that spin-phonon gates of the each with their own benefits and disadvantages. Here, hybrid scheme are free compared with spin-spin entan- we give a qualitative CNOT-gate count (or in turn MS gling gates, the digital approach is inferior to the hy- entangling-gate count) in the harmonic-oscillator basis, brid approach. Even if the spin-phonon gate performs as it is the representation closely related to what is con- comparably to the spin-spin (CNOT) gate, the digital sidered in Sec.III. In this basis, the occupation number scheme require ((log Λ)2) more entangling operations of each harmonic oscillator corresponding to each mo- which can be significantO when Λ 1. 16 Yukawa theory Fermion hopping Fermion mass Free scalar fields Fermion scalar-field interaction Analog-digital O (N) O (1) O (1) O (1) Digital O (N) O (1) O (1) O N 2 (log Λ)2 Schwinger model Fermion-gauge interaction Fermion mass Electric-field term Analog-digital O (N) O (1) O(1)a Digital O N (log Λ)2 O (1) O N (log Λ)2 a (N) standing-wave phonon-phonon gates which are expected to have comparable fidelity to the spin-spin gates. O TABLE III. The scaling of the spin-spin gate count per Trotter step as a function of the lattice size N and the cutoff on the boson (phonon) excitations Λ for the Yukawa theory and the Schwinger model (within HOBM) assuming analog-digital and fully-digital implementations. The spin-phonon and phonon-phonon gate counts are not denoted in the table for the analog-digital approach but have been discussed in the text. Such an advantage is at the core of the power of the Yukawa theory). As a result, the total entangling-gate hybrid approach: phonons are represented naturally and count of the digital-circuit implementation of this evolu- as many phonon excitations as permitted in the dynam- tion operator is (N(log Λ)2), where Λ denotes the cutoff ics can be generated without the need to cut their spec- on the boson excitationsO (translating to the electric-field trum off. Of course, an excessive number of phonons excitations in the original U(1) theory). Therefore, the in the system can lead to Kerr cross-coupling [165] and hybrid implementation will save a factor of ((log Λ)2) loss of coherence in the simulator, and therefore a bal- in CNOT count compared with the digital implementa-O ance should be established between accuracy of the sim- tion assuming again that the spin-phonon gate performs ulated theory given a truncated boson spectrum and the at a much higher fidelity than the spin-spin gate. experimental error in the simulator. For this reason, ex- Time evolving the electric-field term, that is Eq. (46), perimental benchmarks are necessary in confirming these requires N standing-wave gates, which can be counted to qualitative theoretical expectations. A summary of the be comparable to (N) entangling gates. The same term entangling-gate count of both schemes for evolving each in the fully-digitalO computation requires either free Z- term in the Hamiltonian of the Yukawa theory is provided rotations (for evolving the dj†dj term similar to what was in TableIII. discussed for the free scalar Hamiltonian in the Yukawa theory) or (N log Λ(log Λ 1)) CNOT gates (for evolv- O 2 − ing the (dj†dj) term). This latter count can be un- B. The Schwinger model derstood by noting that the application of dj†dj phase depends on the occupation of the d excitation itself. Except for the time evolution of the fermion mass term, In summary, the analog-digital implementation requires ((log Λ)2) fewer entangling operations (assuming that both the fermion-gauge field interaction and the electric- O field term (the boson self-interactions in the HOBM) are the standing-wave gate can perform comparably to the implemented differently in the hybrid and fully-digital spin-spin gate). A summary of the entangling-gate count schemes. of both schemes for evolving each term in the Hamilto- The circuit in Fig.6 reveals that the time evolu- nian of the HOBM is provided in TableIII. tion of the interacting fermion-boson field in the HOBM requires (N) spin-spin gates and (N) spin-phonon O O gates, with the latter expected to be not too costly. In VI. CONCLUSION AND OUTLOOK iH(I)0 δt the fully-digital scheme, e− U(1) with the Hamiltonian in Eq. (43) can be implemented following the circuit con- Our work combines the benefits of analog and digi- struction described earlier in the case of fermion scalar- tal trapped-ion simulators to enable both short- and field interactions of the Yukawa theory. The only differ- long-term simulations of quantum field theories involv- ences are that now the bosons are defined locally, and as- ing bosonic fields with sizable Hilbert spaces. Among sociated with each site (link) there is one such boson (as the models considered are the Yukawa theory, i.e., scalar opposed to N bosons associate with all momentum modes fields coupled to fermions in 1+1 D and the lattice in the Yukawa theory), and that the fermions correspond Schwinger model within the highly-occupied bosonic to nearest-neighbor sites (as opposed to a local fermion model. By introducing a set of known and new gates, in- occupation operator in the corresponding term in the cluding spin, spin-spin, spin-phonon, and phonon-phonon 17 gates, the Trotterized evolution in each of these models spins to mediate the phonon interactions in this is expressed as a quantum circuit. Small instances of the context, inspired by the Mølmer-Sørensen scheme, models with realistic experimental parameters for near- in which spin-spin interactions are mediated by the term proof-of-principle demonstrations of non-trivial dy- virtual phonons. Promising experimental demon- namics are identified. Experiments with 20 ions using a strations in circuit-QED have emerged [174] and set of normal or local modes, each occupied with only similar ideas may be applicable to the trapped-ion up to two phonon excitations, already push the limits systems. of what is possible on a classical computer. Most im- portantly, higher cutoffs on the bosonic excitations can To make progress toward the goal of simulating be imposed in the model as the bosons are naturally quantum chromodynamics, one must develop fea- represented by the phonons in the ion trap, leading to sible simulation proposals for non-Abelian gauge an (N log Λ) ( (log Λ)) reduction in the number of theories. The SU(2) LGT coupled to fermions is qubitsO and at leastO an (N 2 log Λ)2 ( (log Λ)2 ) re- a valuable non-Abelian gauge theory that has be- duction in the numberO of the entanglingO spin-spin gates come the focus of intense studies in recent years [18, compared with a fully-digital simulation in the Yukawa 19, 21, 22, 52, 58, 59, 68, 73, 74]. Its prepotential theory (Schwinger model). Here, Λ is the cutoff on the formulation in terms of Schwinger bosons [71] ap- boson-field excitation and N is the number of lattice sites pears the most suitable representation to poten- in the simulated theory. These scalings assume that the tially take advantage of the analog-digital setup spin-phonon gate has a higher fidelity compared with the presented here. Basically, despite the U(1) theory spin-spin gate, while the phonon-phonon gate performs in 1+1 D which has only an approximate represen- with a fidelity comparable to that of a spin-spin gate. tation in terms of bosonic degrees of freedom, the Experimental implementations will test these theoreti- SU(2) LGT in any dimension has an exact map- cal expectations in the coming years and establish the ping to bosons. The complication is that each link analog-digital mode of the simulator as a more powerful on the lattice is a collection of four sets of os- paradigm than either of fully-analog and digital schemes cillators and the physical degrees of freedom are in enabling simulations of interacting fermionic-bosonic those made up of gauge-invariant combinations of field theories of phenomenological interest. these bosonic operators (along with the fermionic A number of promising theoretical and experimental operators). The Hamiltonian matrix elements re- directions may result from extensions of this work to tain the knowledge of the SU(2) algebra and hence other quantum simulators and more complex strongly- contain non-trivial factors [71]. Ongoing progress interacting field theories. These include the following: in developing fully-digital algorithms for this the- ory, nonetheless, will be beneficial in developing The bosonic modes are common in a range of quan- the analog-digital counterparts. Here, complica- tum simulators. A notable example is a circuit- tions associated with the non-local phonon-phonon QED platform in which cavity photons are coupled interactions in the trapped-ion simulator must be to superconducting qubits [171, 172]. The physics dealt with as well. of photon-qubit effective interactions is very sim- ilar to that of phonon-ion interactions in an ion It will be interesting to consider efficient state- trap. For example, cavity photons can be taken preparation strategies for QFTs when bosonic fields advantage of to induce non-local couplings among can be encoded into bosonic registers. Besides the the qubits [173]. One can envision encoding the reduction in the number of qubits and entangling dynamics of scalar or gauge fields in the photon gates required for the preparation routine in an modes and that of fermions in the superconduct- analog-digital mode, one can envision more efficient ing qubits, hence devising photon-based gates as state-preparation protocols as well, such that non- discussed in this work. Nonetheless, the degree of trivial initial states involving entangled boson and control and the feasibility of experimental realiza- fermion modes can be achieved more straightfor- tions must be examined carefully before the poten- wardly. This direction is, however, less developed tial of an analog-digital scheme can be evaluated in and requires additional investigation. such platforms. Last but not least, it is important to investi- A highly desirable capability, that was not required gate whether the experimental imperfections and for the current work given the nature of the mod- non-ideal gate fidelities are more significant in an els studied, is the engineering of non-local phonon- analog-digital compared to a fully-digital setting. phonon couplings. Such a capability would allow This is a rather crucial question when it comes to the simulations of a self-interacting scalar field the- gauge-theory simulations, as local gauge invariance ory, as well as the plaquette (magnetic) interactions imposed on an initial state needs to be efficiently in lattice gauge theories in higher dimensions, in an retained throughout the evolution, or the simulated analog-digital fashion. It will be worth exploring physics will be fundamentally different than the the possibility of taking advantage of the virtual desired gauge theory. Recent ideas in protecting 18 gauge invariance in the simulation [175–178] can be ping QFTs to quantum simulators, and by the investigated in the context of analog-digital simu- DOE Office of Science, Office of Advanced Scientific lations as well. Computing Research (ASCR) Quantum Comput- ing Application Teams program, under fieldwork proposal number ERKJ347, for algorithmic de- velopments for scientific applications of near-term ACKNOWLEDGMENTS quantum hardware. GP and NML acknowledge support by the DOE Office of Science, Office of Nuclear Physics, under Award no. DE-SC0021143, We acknowledge valuable discussions with Moham- for designing hardware-specific simulation proto- mad Hafezi and Alireza Seif. ZD is supported in cols for applications in nuclear physics. GP and part by the U.S. Department of Energy’s Office of NML are further supported by the Army Research Science Early Career Award, under award no. DE- Office (W911NF21P0003) and the Office of Naval SC0020271, for theoretical developments for map- Research (N00014-20-1-2695). [1] Ho-Kim and Pham, Elementary Particles and (2012), arXiv:1111.3633 [quant-ph]. Their Interactions (Springer-Verlag Berlin Hei- [13] Stephen P. Jordan, Keith S. M. Lee, and John delberg, 1998). Preskill, “Quantum Computation of Scattering [2] Matthew D. Schwartz, Quantum Field Theory in Scalar Quantum Field Theories,” Quant. Inf. and the Standard Model (Cambridge University Comput. 14, 1014–1080 (2014), arXiv:1112.4833 Press, 2013). [hep-th]. [3] Hagen Kleinert, Gauge fields in condensed matter, [14] J. Casanova, L. Lamata, I. L. Egusquiza, R. Ger- Vol. 2 (World Scientific). ritsma, C. F. Roos, J. J. Garcia-Ripoll, and [4] Eduardo Fradkin, Field theories of condensed E. Solano, “Quantum Simulation of Quantum matter physics (Cambridge University Press, Field Theories in Trapped Ions,” Phys. Rev. Lett. 2013). 107, 260501 (2011), arXiv:1107.5233 [quant-ph]. [5] David B. Kaplan, “Five lectures on effective field [15] Erez Zohar and Benni Reznik, “Confinement and theory,” (2005) arXiv:nucl-th/0510023. lattice QED electric flux-tubes simulated with [6] H. W. Hammer, S. K¨onig, and U. van Kolck, ultracold atoms,” Phys. Rev. Lett. 107, 275301 “Nuclear effective field theory: status and per- (2011), arXiv:1108.1562 [quant-ph]. spectives,” Rev. Mod. Phys. 92, 025004 (2020), [16] L. Tagliacozzo, A. Celi, A. Zamora, and arXiv:1906.12122 [nucl-th]. M. Lewenstein, “Optical Abelian Lattice Gauge [7] S. Aoki et al. (Flavour Lattice Averaging Group), Theories,” Annals Phys. 330, 160–191 (2013), “FLAG Review 2019: Flavour Lattice Averaging arXiv:1205.0496 [cond-mat.quant-gas]. Group (FLAG),” Eur. Phys. J. C 80, 113 (2020), [17] D. Banerjee, M. Dalmonte, M. Muller, E. Rico, arXiv:1902.08191 [hep-lat]. P. Stebler, U.-J. Wiese, and P. Zoller, “Atomic [8] Zohreh Davoudi, William Detmold, Kostas Quantum Simulation of Dynamical Gauge Fields Orginos, Assumpta Parre˜no,Martin J. Savage, coupled to Fermionic Matter: From String Break- Phiala Shanahan, and Michael L. Wagman, “Nu- ing to Evolution after a Quench,” Phys. Rev. clear matrix elements from lattice QCD for elec- Lett. 109, 175302 (2012), arXiv:1205.6366 [cond- troweak and beyond-Standard-Model processes,” mat.quant-gas]. (2020), arXiv:2008.11160 [hep-lat]. [18] Erez Zohar, J.Ignacio Cirac, and Benni Reznik, [9] Tim Byrnes and Yoshihisa Yamamoto, “Sim- “Cold-Atom Quantum Simulator for SU(2) Yang- ulating lattice gauge theories on a quantum Mills Lattice Gauge Theory,” Phys. Rev. Lett. computer,” Phys. Rev. A 73, 022328 (2006), 110, 125304 (2013), arXiv:1211.2241 [quant-ph]. arXiv:quant-ph/0510027. [19] Erez Zohar, J. Ignacio Cirac, and Benni Reznik, [10] J. Ignacio Cirac, Paolo Maraner, and Jian- “Quantum simulations of gauge theories with ul- nis K. Pachos, “Cold atom simulation of inter- tracold atoms: local gauge invariance from angu- acting relativistic quantum field theories,” Phys. lar momentum conservation,” Phys. Rev. A 88, Rev. Lett. 105, 190403 (2010), arXiv:1006.2975 023617 (2013), arXiv:1303.5040 [quant-ph]. [cond-mat.str-el]. [20] Stephen P. Jordan, Keith S. M. Lee, and [11] L. Lamata, J. Casanova, R. Gerritsma, C. F. John Preskill, “Quantum Algorithms for Roos, J. J. Garcia-Ripoll, and E. Solano, “Rel- Fermionic Quantum Field Theories,” (2014), ativistic quantum mechanics with trapped ions,” arXiv:1404.7115 [hep-th]. New J. Phys. 13, 095003 (2011), arXiv:1106.6222 [21] Erez Zohar and Michele Burrello, “Formulation of [quant-ph]. lattice gauge theories for quantum simulations,” [12] Stephen P. Jordan, Keith S. M. Lee, and Phys. Rev. D 91, 054506 (2015), arXiv:1409.3085 John Preskill, “Quantum Algorithms for Quan- [quant-ph]. tum Field Theories,” Science 336, 1130–1133 19 [22] A. Mezzacapo, E. Rico, C. Sabn, I.L. Egusquiza, [36] Henry Lamm, Scott Lawrence, and Yukari Ya- L. Lamata, and E. Solano, “Non-Abelian SU(2) mauchi (NuQS), “General Methods for Digital Lattice Gauge Theories in Superconducting Cir- Quantum Simulation of Gauge Theories,” Phys. cuits,” Phys. Rev. Lett. 115, 240502 (2015), Rev. D 100, 034518 (2019), arXiv:1903.08807 arXiv:1505.04720 [quant-ph]. [hep-lat]. [23] Erez Zohar, Alessandro Farace, Benni Reznik, [37] Andrei Alexandru, Paulo F. Bedaque, Siddhartha and J. Ignacio Cirac, “Digital quantum simula- Harmalkar, Henry Lamm, Scott Lawrence, and tion of Z2 lattice gauge theories with dynamical Neill C. Warrington (NuQS), “Gluon Field Digi- fermionic matter,” Phys. Rev. Lett. 118, 070501 tization for Quantum Computers,” Phys. Rev. D (2017), arXiv:1607.03656 [quant-ph]. 100, 114501 (2019), arXiv:1906.11213 [hep-lat]. [24] Ali Hamed Moosavian and Stephen Jor- [38] Siddhartha Harmalkar, Henry Lamm, and Scott dan, “Faster Quantum Algorithm to simulate Lawrence (NuQS), “Quantum Simulation of Field Fermionic Quantum Field Theory,” Phys. Rev. A Theories Without State Preparation,” (2020), 98, 012332 (2018), arXiv:1711.04006 [quant-ph]. arXiv:2001.11490 [hep-lat]. [25] T.V. Zache, F. Hebenstreit, F. Jendrzejewski, [39] Alexander F. Shaw, Pavel Lougovski, Jesse R. M.K. Oberthaler, J. Berges, and P. Hauke, Stryker, and Nathan Wiebe, “Quantum “Quantum simulation of lattice gauge theories Algorithms for Simulating the Lattice using Wilson fermions,” Sci. Technol. 3, 034010 Schwinger Model,” Quantum 4, 306 (2020), (2018), arXiv:1802.06704 [cond-mat.quant-gas]. arXiv:2002.11146 [quant-ph]. [26] Arpan Bhattacharyya, Arvind Shekar, and [40] Dmitri E. Kharzeev and Yuta Kikuchi, “Real- Aninda Sinha, “Circuit complexity in interact- time chiral dynamics from a digital quantum ing QFTs and RG flows,” JHEP 10, 140 (2018), simulation,” Phys. Rev. Res. 2, 023342 (2020), arXiv:1808.03105 [hep-th]. arXiv:2001.00698 [hep-ph]. [27] Jesse R. Stryker, “Oracles for Gauss’s law on [41] Bipasha Chakraborty, Masazumi Honda, Taku digital quantum computers,” Phys. Rev. A 99, Izubuchi, Yuta Kikuchi, and Akio Tomiya, “Dig- 042301 (2019), arXiv:1812.01617 [quant-ph]. ital Quantum Simulation of the Schwinger Model [28] Indrakshi Raychowdhury and Jesse R. Stryker, with Topological Term via Adiabatic State Prepa- “Solving Gauss’s Law on Digital Quantum ration,” (2020), arXiv:2001.00485 [hep-lat]. Computers with Loop-String-Hadron Digitiza- [42] Junyu Liu and Yuan Xin, “Quantum simulation tion,” Phys. Rev. Res. 2, 033039 (2020), of quantum field theories as quantum chemistry,” arXiv:1812.07554 [hep-lat]. JHEP 12, 011 (2020), arXiv:2004.13234 [hep-th]. [29] Natalie Klco and Martin J. Savage, “Digitization [43] Michael Kreshchuk, William M. Kirby, Gary of scalar fields for quantum computing,” Phys. Goldstein, Hugo Beauchemin, and Peter J. Love, Rev. A 99, 052335 (2019), arXiv:1808.10378 “Quantum Simulation of Quantum Field The- [quant-ph]. ory in the Light-Front Formulation,” (2020), [30] Julian Bender, Erez Zohar, Alessandro Farace, arXiv:2002.04016 [quant-ph]. and J. Ignacio Cirac, “Digital quantum simu- [44] Jo˜aoBarata, Niklas Mueller, Andrey Tarasov, lation of lattice gauge theories in three spatial and Raju Venugopalan, “Single-particle digitiza- dimensions,” New J. Phys. 20, 093001 (2018), tion strategy for quantum computation of a φ4 arXiv:1804.02082 [quant-ph]. scalar field theory,” (2020), arXiv:2012.00020 [31] Natalie Klco and Martin J. Savage, “Minimally- [hep-th]. Entangled State Preparation of Localized Wave- [45] Jan F. Haase, Luca Dellantonio, Alessio Celi, functions on Quantum Computers,” Phys. Rev. A Danny Paulson, Angus Kan, Karl Jansen, and 102, 012612 (2020), arXiv:1904.10440 [quant-ph]. Christine A. Muschik, “A resource efficient ap- [32] Zohreh Davoudi, Mohammad Hafezi, Christopher proach for quantum and classical simulations of Monroe, Guido Pagano, Alireza Seif, and Andrew gauge theories in particle physics,” Quantum 5, Shaw, “Towards analog quantum simulations of 393 (2021), arXiv:2006.14160 [quant-ph]. lattice gauge theories with trapped ions,” Phys. [46] Danny Paulson et al., “Towards simulating 2D ef- Rev. Res. 2, 023015 (2020), arXiv:1908.03210 fects in lattice gauge theories on a quantum com- [quant-ph]. puter,” (2020), arXiv:2008.09252 [quant-ph]. [33] Natalie Klco and Martin J. Savage, “Systemati- [47] Yao Ji, Henry Lamm, and Shuchen Zhu (NuQS), cally Localizable Operators for Quantum Simula- “Gluon Field Digitization via Group Space Deci- tions of Quantum Field Theories,” Phys. Rev. A mation for Quantum Computers,” Phys. Rev. D 102, 012619 (2020), arXiv:1912.03577 [quant-ph]. 102, 114513 (2020), arXiv:2005.14221 [hep-lat]. [34] Henry Lamm, Scott Lawrence, and Yukari Ya- [48] Alexander Buser, Hrant Gharibyan, Masanori mauchi (NuQS), “Parton physics on a quantum Hanada, Masazumi Honda, and Junyu Liu, computer,” Phys. Rev. Res. 2, 013272 (2020), “Quantum simulation of gauge theory via orbifold arXiv:1908.10439 [hep-lat]. lattice,” (2020), arXiv:2011.06576 [hep-th]. [35] Niklas Mueller, Andrey Tarasov, and Raju Venu- [49] Julian Bender and Erez Zohar, “Gauge gopalan, “Deeply inelastic scattering structure redundancy-free formulation of compact QED functions on a hybrid quantum computer,” Phys. with dynamical matter for quantum and classical Rev. D 102, 016007 (2020), arXiv:1908.07051 computations,” Phys. Rev. D 102, 114517 (2020), [hep-th]. arXiv:2008.01349 [quant-ph]. 20 [50] Anthony Ciavarella, Natalie Klco, and Martin J. [cond-mat.quant-gas]. Savage, “A Trailhead for Quantum Simulation of [64] Christian Schweizer, Fabian Grusdt, Moritz Bern- SU(3) Yang-Mills Lattice Gauge Theory in the gruber, Luca Barbiero, Eugene Demler, Nathan Local Multiplet Basis,” (2021), arXiv:2101.10227 Goldman, Immanuel Bloch, and Monika Aidels- [quant-ph]. burger, “Floquet approach to Z(2) lattice gauge [51] M.C. Bauls et al., “Simulating Lattice Gauge theories with ultracold atoms in optical lattices,” Theories within Quantum Technologies,” Eur. Nature Physics 15, 1168–1173 (2019). Phys. J. D 74, 165 (2020), arXiv:1911.00003 [65] Alexander Mil, Torsten V. Zache, Apoorva [quant-ph]. Hegde, Andy Xia, Rohit P. Bhatt, Markus K. [52] Raka Dasgupta and Indrakshi Raychowdhury, Oberthaler, Philipp Hauke, J¨urgen Berges, “Cold Atom Quantum Simulator for String and and Fred Jendrzejewski, “A scalable realiza- Hadron Dynamics in Non-Abelian Lattice Gauge tion of local U(1) gauge invariance in cold Theory,” (2020), arXiv:2009.13969 [hep-lat]. atomic mixtures,” Science 367, 1128–1130 (2020), [53] Rene Gerritsma, Gerhard Kirchmair, Florian arXiv:1909.07641 [cond-mat.quant-gas]. Z¨ahringer, E Solano, R Blatt, and CF Roos, [66] Bing Yang, Hui Sun, Robert Ott, Han-Yi “Quantum simulation of the dirac equation,” Na- Wang, Torsten V. Zache, Jad C. Halimeh, Zhen- ture 463, 68–71 (2010). Sheng Yuan, Philipp Hauke, and Jian-Wei [54] E. A. Martinez et al., “Real-time dynamics of Pan, “Observation of gauge invariance in a 71- lattice gauge theories with a few-qubit quan- site Bose–Hubbard quantum simulator,” Nature tum computer,” Nature 534, 516–519 (2016), 587, 392–396 (2020), arXiv:2003.08945 [cond- arXiv:1605.04570 [quant-ph]. mat.quant-gas]. [55] Christian Kokail et al., “Self-verifying variational [67] Erik Gustafson, Yingyue Zhu, Patrick Dreher, quantum simulation of lattice models,” Nature Norbert M. Linke, and Yannick Meurice, 569, 355–360 (2019), arXiv:1810.03421 [quant- “Real-time quantum calculations of phase shifts ph]. using wave packet time delays,” (2021), [56] N. Klco, E. F. Dumitrescu, A. J. McCaskey, arXiv:2103.06848 [hep-lat]. T. D. Morris, R. C. Pooser, M. Sanz, E. Solano, [68] Sarmed A. Rahman, Randy Lewis, Emanuele P. Lougovski, and M. J. Savage, “Quantum- Mendicelli, and Sarah Powell, “SU(2) lattice classical computation of Schwinger model dynam- gauge theory on a quantum annealer,” (2021), ics using quantum computers,” Phys. Rev. A 98, arXiv:2103.08661 [hep-lat]. 032331 (2018), arXiv:1803.03326 [quant-ph]. [69] Ramesh Anishetty, Manu Mathur, and Indrak- [57] Hsuan-Hao Lu et al., “Simulations of Subatomic shi Raychowdhury, “Prepotential formulation of Many-Body Physics on a Quantum Frequency SU(3) lattice gauge theory,” J. Phys. A 43, Processor,” Phys. Rev. A 100, 012320 (2019), 035403 (2010), arXiv:0909.2394 [hep-lat]. arXiv:1810.03959 [quant-ph]. [70] Manu Mathur, Indrakshi Raychowdhury, and [58] Natalie Klco, Jesse R. Stryker, and Mar- Ramesh Anishetty, “Su (n) irreducible schwinger tin J. Savage, “SU(2) non-Abelian gauge field bosons,” Journal of mathematical physics 51, theory in one dimension on digital quantum 093504 (2010). computers,” Phys. Rev. D 101, 074512 (2020), [71] Indrakshi Raychowdhury and Jesse R. Stryker, arXiv:1908.06935 [quant-ph]. “Loop, string, and hadron dynamics in SU(2) [59] Yasar Atas, Jinglei Zhang, Randy Lewis, Amin Hamiltonian lattice gauge theories,” Phys. Rev. D Jahanpour, Jan F. Haase, and Christine A. 101, 114502 (2020), arXiv:1912.06133 [hep-lat]. Muschik, “SU(2) hadrons on a quantum com- [72] Manu Mathur and T. P. Sreeraj, “Lattice Gauge puter,” (2021), arXiv:2102.08920 [quant-ph]. Theories and Spin Models,” Phys. Rev. D 94, [60] Michael Kreshchuk, Shaoyang Jia, William M. 085029 (2016), arXiv:1604.00315 [hep-lat]. Kirby, Gary Goldstein, James P. Vary, and Pe- [73] Erez Zohar and J. Ignacio Cirac, “Removing Stag- ter J. Love, “Light-Front Field Theory on Current gered Fermionic Matter in U(N) and SU(N) Lat- Quantum Computers,” (2020), arXiv:2009.07885 tice Gauge Theories,” Phys. Rev. D 99, 114511 [quant-ph]. (2019), arXiv:1905.00652 [quant-ph]. [61] Christian W. Bauer, Wibe A. de Jong, Ben- [74] Zohreh Davoudi, Indrakshi Raychowdhury, and jamin Nachman, and Davide Provasoli, “Quan- Andrew Shaw, “Search for Efficient Formulations tum Algorithm for High Energy Physics Simu- for Hamiltonian Simulation of non-Abelian Lat- lations,” Phys. Rev. Lett. 126, 062001 (2021), tice Gauge Theories,” (2020), arXiv:2009.11802 arXiv:1904.03196 [hep-ph]. [hep-lat]. [62] Christian W. Bauer, Marat Freytsis, and Ben- [75] Rolando D Somma, “Quantum simulations of one jamin Nachman, “Simulating collider physics on dimensional quantum systems,” arXiv preprint quantum computers using effective field theories,” arXiv:1503.06319 (2015). (2021), arXiv:2102.05044 [hep-ph]. [76] Alexandru Macridin, Panagiotis Spentzouris, [63] Frederik G¨org, Kilian Sandholzer, Joaqu´ın James Amundson, and Roni Harnik, “Electron- Minguzzi, R´emi Desbuquois, Michael Messer, Phonon Systems on a Universal Quantum Com- and Tilman Esslinger, “Realization of density- puter,” Phys. Rev. Lett. 121, 110504 (2018), dependent Peierls phases to engineer quantized arXiv:1802.07347 [quant-ph]. gauge fields coupled to ultracold matter,” Nature [77] Alexandru Macridin, Panagiotis Spentzouris, Phys. 15, 1161–1167 (2019), arXiv:1812.05895 James Amundson, and Roni Harnik, “Digital 21 quantum computation of fermion-boson interact- pher Monroe, “Observation of prethermalization ing systems,” Phys. Rev. A 98, 042312 (2018), in long-range interacting spin chains,” Science ad- arXiv:1805.09928 [quant-ph]. vances 3, e1700672 (2017). [78] Daniel C. Hackett, Kiel Howe, Ciaran Hughes, [90] Fangli Liu, Rex Lundgren, Paraj Titum, Guido William Jay, Ethan T. Neil, and James N. Pagano, Jiehang Zhang, Christopher Monroe, Simone, “Digitizing Gauge Fields: Lattice and Alexey V. Gorshkov, “Confined quasiparticle Monte Carlo Results for Future Quantum Com- dynamics in long-range interacting quantum spin puters,” Phys. Rev. A 99, 062341 (2019), chains,” Phys. Rev. Lett. 122, 150601 (2019). arXiv:1811.03629 [quant-ph]. [91] Juan I Cirac and Peter Zoller, “Quantum compu- [79] Rainer Blatt and Christian F Roos, “Quantum tations with cold trapped ions,” Physical review simulations with trapped ions,” Nature Physics letters 74, 4091 (1995). 8, 277–284 (2012). [92] Klaus Mølmer and Anders Sørensen, “Multipar- [80] David J Wineland, C Monroe, Wayne M ticle entanglement of hot trapped ions,” Physical Itano, Dietrich Leibfried, Brian E King, and Review Letters 82, 1835 (1999). Dawn M Meekhof, “Experimental issues in co- [93] E Solano, RL de Matos Filho, and N Zagury, herent quantum-state manipulation of trapped “Deterministic bell states and measurement of the atomic ions,” Journal of research of the National motional state of two trapped ions,” Physical Re- Institute of Standards and Technology 103, 259 view A 59, R2539 (1999). (1998). [94] GJ Milburn, S Schneider, and DFV James, [81] C. Monroe, W. C. Campbell, L.-M. Duan, Z.-X. “Ion trap quantum computing with warm ions,” Gong, A. V. Gorshkov, P. W. Hess, R. Islam, Fortschritte der Physik: Progress of Physics 48, K. Kim, N. M. Linke, G. Pagano, P. Richerme, 801–810 (2000). C. Senko, and N. Y. Yao, “Programmable quan- [95] Rainer Blatt and David Wineland, “Entangled tum simulations of spin systems with trapped states of trapped atomic ions,” Nature 453, 1008 ions,” Rev. Mod. Phys. 93, 025001 (2021). (2008). [82] Diego Porras and J Ignacio Cirac, “Effective [96] Ben P Lanyon, Cornelius Hempel, Daniel Nigg, quantum spin systems with trapped ions,” Phys- Markus M¨uller, Rene Gerritsma, F Z¨ahringer, ical review letters 92, 207901 (2004). Philipp Schindler, Julio T Barreiro, Markus Ram- [83] R Islam, EE Edwards, K Kim, S Korenblit, bach, Gerhard Kirchmair, et al., “Universal digi- C Noh, H Carmichael, G-D Lin, L-M Duan, C- tal quantum simulation with trapped ions,” Sci- C Joseph Wang, JK Freericks, et al., “Onset of ence 334, 57–61 (2011). a quantum phase transition with a trapped ion [97] Christopher Monroe and Jungsang Kim, “Scaling quantum simulator,” Nature communications 2, the ion trap quantum processor,” Science 339, 377 (2011). 1164–1169 (2013). [84] Ch Schneider, Diego Porras, and Tobias Schaetz, [98] C Monroe, R Raussendorf, A Ruthven, “Experimental quantum simulations of many- KR Brown, P Maunz, L-M Duan, and J Kim, body physics with trapped ions,” Reports on “Large-scale modular quantum-computer ar- Progress in Physics 75, 024401 (2012). chitecture with atomic memory and photonic [85] Petar Jurcevic, Ben P Lanyon, Philipp Hauke, interconnects,” Physical Review A 89, 022317 Cornelius Hempel, Peter Zoller, Rainer Blatt, (2014). and Christian F Roos, “Quasiparticle engineer- [99] Shantanu Debnath, Norbert M Linke, Caroline ing and entanglement propagation in a quantum Figgatt, Kevin A Landsman, Kevin Wright, and many-body system,” Nature 511, 202 (2014). Christopher Monroe, “Demonstration of a small [86] Philip Richerme, Zhe-Xuan Gong, Aaron Lee, programmable quantum computer with atomic Crystal Senko, Jacob Smith, Michael Foss-Feig, qubits,” Nature 536, 63 (2016). Spyridon Michalakis, Alexey V Gorshkov, and [100] Norbert M Linke, Dmitri Maslov, Martin Roet- Christopher Monroe, “Non-local propagation of teler, Shantanu Debnath, Caroline Figgatt, correlations in quantum systems with long-range Kevin A Landsman, Kenneth Wright, and interactions,” Nature 511, 198 (2014). Christopher Monroe, “Experimental comparison [87] J Zhang, PW Hess, A Kyprianidis, P Becker, of two quantum computing architectures,” Pro- A Lee, J Smith, G Pagano, I-D Potirniche, An- ceedings of the National Academy of Sciences drew C Potter, A Vishwanath, et al., “Observa- 114, 3305–3310 (2017). tion of a discrete time crystal,” Nature 543, 217 [101] C. Figgatt, A. Ostrander, N. M. Linke, K. A. (2017). Landsman, D. Zhu, D. Maslov, and C. Mon- [88] PW Hess, P Becker, HB Kaplan, A Kyprianidis, roe, “Parallel entangling operations on a univer- AC Lee, B Neyenhuis, G Pagano, P Richerme, sal ion-trap quantum computer,” Nature (2019), C Senko, J Smith, et al., “Non-thermalization 10.1038/s41586-019-1427-5. in trapped atomic ion spin chains,” Philosophi- [102] Kevin A Landsman, Caroline Figgatt, Thomas cal Transactions of the Royal Society A: Mathe- Schuster, Norbert M Linke, Beni Yoshida, matical, Physical and Engineering Sciences 375, Norm Y Yao, and Christopher Monroe, “Verified 20170107 (2017). quantum information scrambling,” Nature 567, [89] Brian Neyenhuis, Jiehang Zhang, Paul W Hess, 61 (2019). Jacob Smith, Aaron C Lee, Phil Richerme, Zhe- [103] Yunseong Nam, Jwo-Sy Chen, Neal C Pisenti, Xuan Gong, Alexey V Gorshkov, and Christo- Kenneth Wright, Conor Delaney, Dmitri Maslov, 22 Kenneth R Brown, Stewart Allen, Jason M [117] Wolfgang Paul, “Electromagnetic traps for Amini, Joel Apisdorf, et al., “Ground-state en- charged and neutral particles,” Reviews of mod- ergy estimation of the water molecule on a ern physics 62, 531 (1990). trapped ion quantum computer,” arXiv preprint [118] G-D Lin, S-L Zhu, Rajibul Islam, Kihwan Kim, arXiv:1902.10171 (2019). M-S Chang, Simcha Korenblit, Christopher Mon- [104] K Wright, KM Beck, S Debnath, JM Amini, roe, and L-M Duan, “Large-scale quantum com- Y Nam, N Grzesiak, J-S Chen, NC Pisenti, putation in an anharmonic linear ion trap,” EPL M Chmielewski, C Collins, et al., “Benchmarking (Europhysics Letters) 86, 60004 (2009). an 11-qubit quantum computer,” arXiv preprint [119] G Pagano, PW Hess, HB Kaplan, WL Tan, arXiv:1903.08181 (2019). P Richerme, P Becker, A Kyprianidis, J Zhang, [105] Shinsuke Haze, Yusuke Tateishi, Atsushi E Birckelbaw, MR Hernandez, et al., “Cryogenic Noguchi, Kenji Toyoda, and Shinji Urabe, “Ob- trapped-ion system for large scale quantum sim- servation of phonon hopping in radial vibrational ulation,” Quantum Science and Technology 4, modes of trapped ions,” Phys. Rev. A 85, 031401 014004 (2018). (2012). [120] Diego Porras and J Ignacio Cirac, “Bose-einstein [106] Kenji Toyoda, Ryoto Hiji, Atsushi Noguchi, and condensation and strong-correlation behavior of Shinji Urabe, “Hong–ou–mandel interference of phonons in ion traps,” Physical review letters 93, two phonons in trapped ions,” Nature 527, 74– 263602 (2004). 77 (2015). [121] Juan Ignacio Cirac and Peter Zoller, “A scalable [107] S Debnath, NM Linke, S-T Wang, C Figgatt, quantum computer with ions in an array of mi- KA Landsman, L-M Duan, and C Monroe, “Ob- crotraps,” Nature 404, 579–581 (2000). servation of hopping and blockade of bosons in a [122] C. Monroe, W. C. Campbell, L. M. Duan, Z. X. trapped ion spin chain,” Physical review letters Gong, A. V. Gorshkov, P. Hess, R. Islam, K. Kim, 120, 073001 (2018). N. Linke, G. Pagano, P. Richerme, C. Senko, [108] Ryutaro Ohira, Takashi Mukaiyama, and Kenji and N. Y. Yao, “Programmable quantum simula- Toyoda, “Phonon-number-resolving detection of tions of spin systems with trapped ions,” (2019), multiple local phonon modes in trapped ions,” arXiv:1912.07845 [quant-ph]. Phys. Rev. A 100, 060301 (2019). [123] S.-L. Zhu, C. Monroe, and L.-M. Duan, [109] Masaya Tamura, Takashi Mukaiyama, and “Trapped ion quantum computation with trans- Kenji Toyoda, “Quantum walks of a phonon verse phonon modes,” Phys. Rev. Lett. 97, in trapped ions,” Physical Review Letters 124, 050505 (2006). 200501 (2020). [124] Christian F Roos, “Ion trap quantum gates with [110] C. Fl¨uhmann, T. L. Nguyen, M. Marinelli, amplitude-modulated laser beams,” New Journal V. Negnevitsky, K. Mehta, and J. P. Home, “En- of Physics 10, 013002 (2008). coding a qubit in a trapped-ion mechanical oscil- [125] Todd J. Green and Michael J. Biercuk, “Phase- lator,” Nature 566, 513–517 (2019). modulated decoupling and error suppression in [111] Wentao Chen, Jaren Gan, Jing-Ning Zhang, qubit-oscillator systems,” Phys. Rev. Lett. 114, Dzmitry Matuskevich, and Kihwan Kim, 120502 (2015). “Quantum computation and simulation with vi- [126] Pak Hong Leung, Kevin A. Landsman, Caro- brational modes of trapped ions,” (2021), line Figgatt, Norbert M. Linke, Christopher Mon- arXiv:2103.14299. roe, and Kenneth R. Brown, “Robust 2-qubit [112] Dayou Yang, Gouri Shankar Giri, Michael Jo- gates in a linear ion crystal using a frequency- hanning, Christof Wunderlich, Peter Zoller, and modulated driving force,” Phys. Rev. Lett. 120, Philipp Hauke, “Analog quantum simulation of 020501 (2018). (1+1)-dimensional lattice qed with trapped ions,” [127] Reinhold Bl¨umel,Nikodem Grzesiak, Nhung H. Physical Review A 94, 052321 (2016). Nguyen, Alaina M. Green, Ming Li, Andrii [113] J. Casanova, A. Mezzacapo, L. Lamata, and Maksymov, Norbert M. Linke, and Yunseong E. Solano, “Quantum Simulation of Interacting Nam, “Efficient, stabilized two-qubit gates on Fermion Lattice Models in Trapped Ions,” Phys. a trapped-ion quantum computer,” (2021), Rev. Lett. 108, 190502 (2012), arXiv:1110.3730 arXiv:2101.07887. [quant-ph]. [128] Ryutaro Ohira, Shota Kume, Kyoichi Takayama, [114] L. Lamata, A. Mezzacapo, J. Casanova, and Silpa Muralidharan, Hiroki Takahashi, and Kenji E. Solano, “Efficient quantum simulation of Toyoda, “Blockade of phonon hopping in trapped fermionic and bosonic models in trapped ions,” ions in the presence of multiple local phonons,” EPJ Quant. Technol. 1, 9 (2014), arXiv:1312.2849 Phys. Rev. A 103, 012612 (2021). [quant-ph]. [129] Tobias Olsacher, Lukas Postler, Philipp [115] A Mezzacapo, J Casanova, L Lamata, and Schindler, Thomas Monz, Peter Zoller, and E Solano, “Digital quantum simulation of the hol- Lukas M Sieberer, “Scalable and parallel tweezer stein model in trapped ions,” Physical review let- gates for quantum computing with long ion ters 109, 200501 (2012). strings,” PRX Quantum 1, 020316 (2020). [116] Xiang Zhang et al., “Experimental quantum sim- [130] Yu-Ching Shen and Guin-Dar Lin, “Scalable ulation of fermion-antifermion scattering via bo- quantum computing stabilised by optical tweez- son exchange in a trapped ion,” Nature Commun. ers on an ion crystal,” New Journal of Physics 9, 195 (2018), arXiv:1611.00099 [quant-ph]. 22, 053032 (2020). 23 [131] Juan Diego Arias Espinoza, Matteo Mazzanti, 1–75 (2011), arXiv:1105.2919 [nucl-th]. Katya Fouka, Rima X Sch¨ussler, Zhenlin Wu, [146] David B. Kaplan, Martin J. Savage, and Mark B. Philippe Corboz, Rene Gerritsma, and Argha- Wise, “Two nucleon systems from effective field van Safavi Naini, “Engineering spin-spin interac- theory,” Nucl. Phys. B 534, 329–355 (1998), tions with optical tweezers in trapped ions,” arXiv arXiv:nucl-th/9802075. preprint arXiv:2103.10425 (2021). [147] Masuo Suzuki, “General theory of fractal path in- [132] Yi Hong Teoh, Manas Sajjan, Zewen Sun, tegrals with applications to many-body theories Fereshteh Rajabi, and Rajibul Islam, “Manipu- and statistical physics,” Journal of Mathematical lating phonons of a trapped-ion system using op- Physics 32, 400–407 (1991). tical tweezers,” arXiv preprint arXiv:2103.03273 [148] Nathan Wiebe, Dominic Berry, Peter Høyer, and (2021). Barry C Sanders, “Higher order decompositions [133] Ksenia Sosnova, Allison Carter, and Christo- of ordered operator exponentials,” Journal of pher Monroe, “Character of motional modes for Physics A: Mathematical and Theoretical 43, entanglement and sympathetic cooling of mixed- 065203 (2010). species trapped-ion chains,” Physical Review A [149] Andrew M Childs, Yuan Su, Minh C Tran, 103, 012610 (2021). Nathan Wiebe, and Shuchen Zhu, “A theory [134] Z. Fodor, J. Hein, K. Jansen, A. Jaster, and of trotter error,” arXiv preprint arXiv:1912.08854 I. Montvay, “Simulating the electroweak phase (2019). transition in the SU(2) Higgs model,” Nucl. Phys. [150] Dominic W Berry, Graeme Ahokas, Richard B 439, 147–186 (1995), arXiv:hep-lat/9409017. Cleve, and Barry C Sanders, “Efficient quantum [135] P. Gerhold and K. Jansen, “The Phase struc- algorithms for simulating sparse hamiltonians,” ture of a chirally invariant lattice Higgs-Yukawa Communications in Mathematical Physics 270, model for small and for large values of the 359–371 (2007). Yukawa coupling constant,” JHEP 09, 041 [151] Dominic W Berry and Andrew M Childs, “Black- (2007), arXiv:0705.2539 [hep-lat]. box hamiltonian simulation and unitary im- [136] P. Gerhold and K. Jansen, “Lower Higgs bo- plementation,” arXiv preprint arXiv:0910.4157 son mass bounds from a chirally invariant lat- (2009). tice Higgs-Yukawa model with overlap fermions,” [152] Andrew M Childs and Nathan Wiebe, “Hamilto- JHEP 07, 025 (2009), arXiv:0902.4135 [hep-lat]. nian simulation using linear combinations of uni- [137] John Bulava, Philipp Gerhold, Karl Jansen, Jim tary operations,” arXiv preprint arXiv:1202.5822 Kallarackal, Bastian Knippschild, C. J. David (2012). Lin, Kei-Ichi Nagai, Attila Nagy, and [153] Dominic W Berry, Andrew M Childs, Richard Kenji Ogawa, “Higgs-Yukawa model in chirally- Cleve, Robin Kothari, and Rolando D Somma, invariant lattice field theory,” Adv. High Energy “Exponential improvement in precision for simu- Phys. 2013, 875612 (2013), arXiv:1210.1798 [hep- lating sparse hamiltonians,” in Proceedings of the lat]. forty-sixth annual ACM symposium on Theory of [138] David Y. J. Chu, Karl Jansen, Bastian Knipp- computing (2014) pp. 283–292. schild, and C. J. David Lin, “Higgs-Yukawa [154] Dominic W Berry, Andrew M Childs, and Robin model on the lattice,” EPJ Web Conf. 175, 08017 Kothari, “Hamiltonian simulation with nearly op- (2018), arXiv:1710.09737 [hep-lat]. timal dependence on all parameters,” in 2015 [139] Karl Jansen, Hubert Simma, and Rainer Som- IEEE 56th Annual Symposium on Foundations of mer, “Tests of the Standard Model and Lattice Computer Science (IEEE, 2015) pp. 792–809. Simulations,” NIC series 47, 41–48 (2014). [155] Dominic W Berry, Andrew M Childs, Richard [140] N. Brambilla et al., “QCD and Strongly Coupled Cleve, Robin Kothari, and Rolando D Somma, Gauge Theories: Challenges and Perspectives,” “Simulating hamiltonian dynamics with a trun- Eur. Phys. J. C 74, 2981 (2014), arXiv:1404.3723 cated taylor series,” Physical review letters 114, [hep-ph]. 090502 (2015). [141] Oscar Akerlund, Quantum field theory on the lat- [156] Guang Hao Low and Nathan Wiebe, “Hamilto- tice: approximate and exact methods, with appli- nian simulation in the interaction picture,” arXiv cations beyond the standard model, Ph.D. thesis, preprint arXiv:1805.00675 (2018). Zurich, ETH (2016). [157] Guang Hao Low and Isaac L Chuang, “Hamilto- [142] David J. E. Callaway, “Triviality Pursuit: Can nian simulation by qubitization,” Quantum 3, 163 Elementary Scalar Particles Exist?” Phys. Rept. (2019). 167, 241 (1988). [158] Amir Kalev and Itay Hen, “Simulating hamilto- [143] Daniel Boyanovsky, Hector J. De Vega, Richard nian dynamics with an off-diagonal series expan- Holman, and Matthew R. Martin, “Nonequi- sion,” arXiv preprint arXiv:2006.02539 (2020). librium large N Yukawa dynamics: Marching [159] Dingshun Lv, Shuoming An, Zhenyu Liu, Jing- through the Landau pole,” Phys. Rev. D 65, Ning Zhang, Julen S. Pedernales, Lucas Lamata, 045007 (2002), arXiv:hep-ph/0108113. Enrique Solano, and Kihwan Kim, “Quantum [144] Steven Weinberg, “Effective chiral Lagrangians Simulation of the Quantum Rabi Model in a for nucleon - pion interactions and nuclear forces,” Trapped Ion,” Phys. Rev. X 8, 021027 (2018), Nucl. Phys. B 363, 3–18 (1991). arXiv:1711.00582 [quant-ph]. [145] R. Machleidt and D. R. Entem, “Chiral effective [160] JS Pedernales, I Lizuain, S Felicetti, G Romero, field theory and nuclear forces,” Phys. Rept. 503, L Lamata, and E Solano, “Quantum rabi model 24 with trapped ions,” Scientific reports 5, 1–7 [173] Guanyu Zhu, Yi˘git Suba¸sı, James D Whit- (2015). field, and Mohammad Hafezi, “Hardware- [161] Julio T Barreiro, Markus M¨uller, Philipp efficient fermionic simulation with a cavity– Schindler, Daniel Nigg, Thomas Monz, Michael qed system,” npj Quantum Information 4, 1–10 Chwalla, Markus Hennrich, Christian F Roos, Pe- (2018). ter Zoller, and Rainer Blatt, “An open-system [174] Yvonne Y Gao, Brian J Lester, Yaxing Zhang, quantum simulator with trapped ions,” Nature Chen Wang, Serge Rosenblum, Luigi Frun- 470, 486–491 (2011). zio, Liang Jiang, SM Girvin, and Robert J [162] Caroline Figgatt, Aaron Ostrander, Norbert M Schoelkopf, “Programmable interference between Linke, Kevin A Landsman, Daiwei Zhu, Dmitri two microwave quantum memories,” Physical Re- Maslov, and Christopher Monroe, “Parallel en- view X 8, 021073 (2018). tangling operations on a universal ion-trap quan- [175] Minh C. Tran, Yuan Su, Daniel Carney, and Ja- tum computer,” Nature 572, 368–372 (2019). cob M. Taylor, “Faster Digital Quantum Simula- [163] Yao Lu, Shuaining Zhang, Kuan Zhang, Wen- tion by Symmetry Protection,” P. R. X. Quan- tao Chen, Yangchao Shen, Jialiang Zhang, Jing- tum. 2, 010323 (2021), arXiv:2006.16248 [quant- Ning Zhang, and Kihwan Kim, “Global entan- ph]. gling gates on arbitrary ion qubits,” Nature 572, [176] Henry Lamm, Scott Lawrence, and Yukari 363–367 (2019). Yamauchi (NuQS), “Suppressing Coherent [164] Nikodem Grzesiak, Reinhold Bl¨umel, Kenneth Gauge Drift in Quantum Simulations,” (2020), Wright, Kristin M Beck, Neal C Pisenti, Ming arXiv:2005.12688 [quant-ph]. Li, Vandiver Chaplin, Jason M Amini, Shantanu [177] Jad C. Halimeh, Haifeng Lang, Julius Milden- Debnath, Jwo-Sy Chen, et al., “Efficient arbitrary berger, Zhang Jiang, and Philipp Hauke, simultaneously entangling gates on a trapped-ion “Gauge-Symmetry Protection Using Single-Body quantum computer,” Nature communications 11, Terms,” (2020), arXiv:2007.00668 [quant-ph]. 1–6 (2020). [178] Valentin Kasper, Torsten V Zache, Fred Jen- [165] C. Marquet, F. Schmidt-Kaler, and D.F.V. drzejewski, Maciej Lewenstein, and Erez Zo- James, “Phonon–phonon interactions due to non- har, “Non-abelian gauge invariance from dynami- linear effects in a linear ion trap,” Applied Physics cal decoupling,” arXiv preprint arXiv:2012.08620 B 76, 199–208 (2003). (2020). [166] John Kogut and Leonard Susskind, “Hamiltonian formulation of wilson’s lattice gauge theories,” Physical Review D 11, 395 (1975). [167] Tom Banks, Leonard Susskind, and John Kogut, Appendix A: Details of the numerical examples “Strong-coupling calculations of lattice gauge the- ories:(1+ 1)-dimensional exercises,” Physical Re- view D 13, 1043 (1976). This appendix supplements the numerical examples [168] Erez Zohar, J. Ignacio Cirac, and Benni provided in Secs.IIIC andIVC of the main text. Reznik, “Simulating (2+1)-Dimensional Lattice QED with Dynamical Matter Using Ultracold Atoms,” Phys. Rev. Lett. 110, 055302 (2013), 1. Yukawa theory arXiv:1208.4299 [quant-ph]. [169] Robert Ott, Torsten V. Zache, Fred Jendrze- jewski, and J¨urgenBerges, “Scalable cold-atom This section contains the numerical values for the quantum simulator for two-dimensional QED,” gate rotation angles (TableIV) and gate param- (2020), arXiv:2012.10432 [cond-mat.quant-gas]. [170] Alessio Celi, Benoˆıt Vermersch, Oscar Viyuela, eters (TableV) required in implementing the ex- Hannes Pichler, Mikhail D. Lukin, and Peter amples studied in Sec.IIIC. The relevant trap and Zoller, “Emerging Two-Dimensional Gauge Theo- laser parameters for a realistic near-term experi- ries in Rydberg Configurable Arrays,” Phys. Rev. ment are also provided (TableVI). X 10, 021057 (2020), arXiv:1907.03311 [quant- ph]. [171] Alexandre Blais, Jay Gambetta, Andreas Wall- raff, David I Schuster, Steven M Girvin, Michel H 2. Schwinger model Devoret, and Robert J Schoelkopf, “Quantum- information processing with circuit quantum elec- This section contains the numerical values for the trodynamics,” Physical Review A 75, 032329 (2007). gate rotation angles (TableVII) and gate param- [172] Sebastian Schmidt and Jens Koch, “Circuit qed eters (TablesVIII andIX) required in implement- lattices: towards quantum simulation with su- ing the examples studied in Sec.IVC. The relevant perconducting circuits,” Annalen der Physik 525, trap and laser parameters for a realistic near-term 395–412 (2013). experiment are also provided (TableX). 25 Gate angles Nion θj,j+1 θj θm,j = θm,N+1 φm,j 2+1 {0.0625, 0.0625} {−0.125, 0.125} {{0.344, 0.344, 0.344} , {0.625, 0.625, 0.625}} {{−π, −2π}, {0, 0}} {{0.061, 0.061, 0.061, 0.061}} , {{−π, −2π, −3π, −4π, −5π} , 4+1 {0.0312, 0.0312, 0.0312, 0.0312} {−0.0625, 0.0625, −0.0625, 0.0625} {0.081, 0.081, 0.081, 0.081} , {−π/2, −π, −3π/2, −2π, −5π/2} , {0.110, 0.110, 0.110, 0.110} , {0, 0, 0, 0, 0} , {π/2, π, 3π/2, 2π, 5π/2}} {0.081, 0.081, 0.081, 0.081}} TABLE IV. The gate angles associated with the circuit implementation of the Trotterized dynamics in the Yukawa theory as specified in Eqs. (32)-(35). The values are associated with the model parameters given in TableI of the main text. σa Implementing R{m},j ({θm,j } , {φm,j }) Ωm,j σa Nion ηm,j 2π [kHz] τgate [ms] 2 + 1 (m = 1, 3, j = 1, 2, 3) {{−0.039, −0.039, −0.039} , {{98.6, 98.6, 98.6} , 20 × 10−3 {0.028, −0.057, 0.028}} {−248.4, 124.2, −248.4}} {{−0.028, −0.028, −0.028, −0.028, −0.028} , {{69.7, 69.7, 69.7, 69.7, 69.7} , 4 + 2 (m = 1, ··· , 4, j = 1, ··· , 5) {−0.042, −0.023, −0.008, 0.008, 0.023} , {61.8, 109.5, 336.3, −336.3, −109.5} , 10 × 10−3 {0.038, −0.009, −0.029, −0.029, −0.009} , {−91.7, 380.6, 120.8, 120.8, 380.6} , {0.025, −0.038, −0.0196, 0.0196, 0.038}} {−102.3, 67.4, 131.8, −131.8, −67.4}} TABLE V. The spin-phonon gate parameters associated with the gate rotation angles specified in TableIV for the examples studied in Sec.IIIC. Normal-mode spectrum and interaction-picture shift 1 1 εm Nion η 2π ωm [kHz] 2π N+1 [kHz] e 2+1 0.068 {4000, 3938.3, 3850.2}{2.2, 0, 0.7} 4+2 0.068 {4000, 3938.3, 3849.4, 3735.5, 3596.4, 3430.7}{1.3, 0.7, 0.4, 0.7, 0, 0} TABLE VI. The characteristics of the ion trap and the lasers relevant for the examples studied in Sec.IIIC and the associated gate parameters in TableV. 26 Gate angles 0 (1) (2) θj θj χ χ 0.008 for all j {−0.0625, 0.0625, −0.0625, 0.0625, −0.0625, 0.0625, −0.0625, 0.0625} −0.153 0.008 TABLE VII. The gate rotation angles associated with the circuit implementation of the Trotterized dynamics in the Yukawa theory as specified in Eqs. (47)-(49). The values are associated with the model parameters given in TableII of the main text. σa Implementing Rj (θj , φ) σa Ωj /2π [kHz] τgate [ms] −43.9 for all j 10−3 TABLE VIII. The spin-phonon gate parameters associated with the gate rotation angles specified in TableVII for the examples studied in Sec.IVC. The required φ angles are al- ready noted in the relevant circuit expression in Eq. (47). aa (1) (2) Implementing Rj (χ , χ ) √ 2 x aa F /2π [kHz] ηe ηe MF ηe /ω τgate [ms] 1949.6 0.05 0.16 0.001 50 × 10−3 TABLE IX. The phonon-phonon gate parameters associated with the gate rotation angles specified in TableVII for the examples studied in Sec.IVC. Local-mode properties and interaction-picture shift √ 1 x 1 x Nion 2π ω [kHz] ηj ≡ η η M 2π δω [kHz] 8 6000 0.056 0.18 9.2 TABLE X. The characteristics of the ion trap and lasers rele- vant for the examples studied in Sec.IVC and the associated gate parameters in TablesVIII andIX.