Arxiv:1112.1778V1 [Quant-Ph] 8 Dec 2011 Oaie Spsil Ihnterqieet Fquantum of Requirements the Theory

Home , Coherent state, Fock state

arXiv:1112.1778v1 [quant-ph] 8 Dec 2011 oaie spsil ihnterqieet fquantum of requirements the theory. within as possible was hence as and localized relation satis- uncertainty wavefunction was Heisenberg wavepacket this the of its fied the spread of mimicking the Furthermore spread thereby the fixed. oscillator, and evolution, harmonic evolu- period the classical of the from followed tion obtained wavefunction distribution the Gaussian of the centroid oscil- This harmonic a [1]. Gaussian of orig- lator evolution a was the as state describe 1926 to coherent wavepacket Schrödinger in The by introduced classical. inally to close evo- quantum is the lution when particularly dynamics, quantum I.Implementations III. I umr n Conclusions and Summary VI. V ocasclproperties Nonclassical IV. I Formalism II. .Apiain n Implications and Applications V. .Introduction I. oeetsae lya motn oei representing in role important an play states Coherent References Acknowledgments .Mtoa ere ffedm10 freedom of degrees Motional D. .Lechrn ttsadtersproiin 6 superpositions their and states coherent Lie D. .Cmlmnaiy11 Complementarity A. .Prmti mlfiainadpooeeto 9 photodetection and amplification Parametric A. .Chrn ttso h ipeharmonic simple the of states Coherent A. .Cvt unu lcrdnmc 10 electrodynamics quantum Cavity C. .Sproiin fchrn tts5 states coherent of Superpositions C. .Etnlmn 11 Entanglement B. .Nnierotc 9 optics Nonlinear B. .Chrn ttsi unu pis4 optics quantum in states Coherent B. .Bs-isencnests11 condensates Bose-Einstein E. .Etnldchrn tts8 states coherent Entangled E. oscillator nttt o unu nomto cec,Uiest fC of University Science, Information Quantum for Institute lsia falsae fadnmclsystem. dynamical a cohere of the states of all entanglements of are classical they optics because quantum Des intriguing propagating-wave are conditional physics. a mathematical in and produced optics, superse quantum quantum to processing, important are states coherent Entangled erve nage oeetsaersac ic t rti first its since research state coherent entangled review We .INTRODUCTION I. Contents eiwo nage oeetStates Coherent Entangled of Review ar .Sanders C. Barry 13 13 12 12 11 3 9 3 1 ino w oeetsae,wt ag mltd phases amplitude large with superposi- by considered states, the separated are coherent distinct, two life and of and valid tion death macroscopically vs of dead be states being the to of As superposition a [15]. of in alive state be whose to cat, seems defer- the existence in of Schrödinger’s state’, paradox ‘cat to term dis- ence the nearly earned of states coherent Superpositions tinct [14]. Knight Buˇzek and by coherent generalized pho- [6–13]. to states extensions squeezing, environmen- and to as decoherence), robustness tal (such produce and distribution, mode to properties number single how ton their Studies a concerned states, field for 5]. such electromagnetic states [4, the coherent Stoler of of and Yurke superpositions by of detail of was and in type states coherent 3], analyzed certain of [2, a superpositions of Milburn of manifestation by the study evolution super- a Hamiltonian such in was nonlinear of appeared states Evidence first coherent interest. positions widespread of of superpositions rapidly of introduction phenomena. optical quan- quantum and of classical descriptions connecting by tum optics important quantum an play in electric- states role point coherent ideal Thus by detectors. classical detection dipole direct to in of correspond fields eigenstates hence noiseless are de- operator, states optical annihilation coherent the the classical addition of the In evolution in scription. the variables follows canonical variables) classical are canonical values the states (mean of centroid Coherent The states. minimum-uncertainty transition. quantum-to-classical quantum connecting for tool optics. state classical coherent the equa- and the as the oscillator, significant by harmonic of became governed simple superposition a a one of as each tions regarded modes, be classical can electromagnetic many space the free As in systems. field optical anal- concomitant coherent with the of electrodynamics and yses quantum laser juxtapose the to of desire advent the with tation estvt siprati nomn sa owysuch why to as this us fact informing extreme In in to important due is decoherence. sensitivity be environmental could to this sensitivity fundamentally and duce, uepstoso oeetsae aebe reviewed been have states coherent of Superpositions the quasiclassical, as regarded are states coherent As the embodies optics quantum in state coherent The represen- important an as emerged state coherent The uepstoso oeetsae r icl opro- to difficult are states coherent of Superpositions tsae,wihaei es h most the sense a in are which states, nt eto rnils unu information quantum principles, lection ieterihrn rglt hyhave they fragility inherent their pite elzto.Fnaetlytestates the Fundamentally realization. pii s n16 otepresent. the to 1967 in use mplicit lay let 2 N,Canada 1N4, T2N Alberta algary, π ain,i nlgu oti paradox. this to analogous is radians, 2 peculiar states are prevalent in nature. Experimental Sanders in a study concerning production of entangled efforts to create cat states have concentrated on creat- coherent states by using a nonlinear Mach-Zehnder in- ing superpositions of coherent states that have limited terferometer [32, 33]. The nonlinear interferometer com- distinguishability [16]. Such states have been dubbed prises a nonlinear medium in one path of a Mach-Zehnder Schrödinger kittens [17]. interferometer. The nonlinear medium alone could suffice Soon after the introduction of these single-mode super- to produce entangled coherent states [27–30], but the in- positions of coherent states, entangled coherent states (or terferometric set-up has analogies with the homodyne de- superpositions of multimode coherent states) became of tection concept for superpositions of coherent states [4]. widespread interest. These superpositions of multimode Linear optics alone is known to be insufficient to gen- coherent states arose independently in several papers. erate entangled coherent states so a nonlinearity is re- The earliest entangled coherent state appears in Eq. (10) quired [34]. of the 1967 Aharonov and Susskind charge-superselection Soon after the proposals to create entangled coherent analysis that shows charges could appear in superposi- states in two-mode propagating fields, a cavity quantum tion [18]. Entangled coherent states provided a represen- electrodynamics realization was proposed using one atom tation by which superpositions of charge states could be traversing two cavities and post-selecting on atomic mea- understood within the confines of superselection [19]. surement. This scheme was suggested for realizing entan- The next appearance of the entangled coherent state glement between a coherent state in one mode and the appeared in Eq. (11) of Yurke and Stoler’s 1986 seminal vacuum in the other mode [35]. paper on generating superpositions of coherent states [4]. Entanglement of a coherent state with a vacuum state They showed that mixing a superposition of coherent (which is a coherent state of zero amplitude) was a partic- states with a vacuum state at a beam splitter yields ular focus of the Sanders proposal [32]. In this analysis, a an entangled coherent state output. Subjecting one bipartite entangled coherent states was shown to violate of the two output field modes to homodyne detec- a phase-coherence Bell inequality [36] in the few-photon tion [20–22], which corresponds to a quadrature measure- limit [32, 33]. Later the entangled coherent states were ment [4, 5, 23]. also shown to violate a formal Bell inequality in the large The entangled coherent state representation appeared photon number limit [37]. again in considerations of pair coherent states by Agar- Coherent states generated by a Kerr nonlinearity, wal [24–26]: the entangled coherent state representa- within or without an interferometer, obey a conservation tion of the pair coherent state appears in Eq. (2.6) of rule for total photon number, which constrains the phase Agarwal’s 1988 work [25] Although pair coherent states, relationship between the components of the multimode which are a special case of the Barut and Girardello coherent state superposition. Chai [38] introduced entan- SO(2,1) SU(1,1) SL(2,R) coherent state [8], are the ob- gled coherent states as superpositions of two-mode coher- ≃ ≃ jects of study, these pair coherent states are elegantly ent states with equal amplitude but opposite in optical represented as continuous bipartite entangled coherent phase and allowed an arbitrary phase relationship be- states. tween the two components of the bipartite superposition. Entangled coherent states as entities of physical inter- The analysis then focused on the two-mode extension to est in their own right, first arose in a study by Tombesi single-mode even and odd coherent states [39], which are and Mecozzi [27, 28], where they generalize the nonlin- sometimes called ‘even entangled coherent states’ and ear birefringent evolution of Milburn [2, 3] and Yurke and ‘odd entangled coherent states’ [40]. These even and Stoler [4] to multimode coherent states. Rather than em- odd entangled coherent states naturally generally for q- ploy the single-mode nonlinear evolution associated with coherent states [41, 42]. the ideal optical Kerr nonlinearity, they treat the ideal Chai studied statistical properties of even and odd en- Hamiltonian evolution of two orthogonally polarized light tangled coherent states and showed that the joint pho- beams interacting in a nonlinear birefringent medium. ton number distribution of such entangled coherent states After an appropriate evolution time, an initial two-mode vanished for certain values of the photon number sum or coherent state evolves to an entangled coherent state [27– difference. He also evaluated the squeezing properties of 30]. these even and odd entangled coherent states. These even Tombesi and Mecozzi then studied the relevant statis- and odd entangled coherent states have quantum metro- tics of these states, such as photon number distribution logical applications [43, 44]. Such states could be con- and squeezing, as well as robustness to decoherence [28]. structed by multimode parametric amplifiers [40]. En- Agarwal and Puri [31] studied evolution of a two-mode tangled coherent states also have applications to quan- coherent state through an optical Kerr medium and stud- tum information processing [45] ied the entangled coherent state and its statistical prop- Although ‘balanced’, or equally weighted, superposi- erties. They also pointed out that their entangled coher- tions of multimode coherent states are typically studied, ent state is a simultaneous eigenstate of operators that ‘unbalanced’, or unequally weighted, superpositions are are quadratic in the annihilation operators for the two possible. An approximation to ideal unequally weighted modes. superpositions can be generated by a nonlinear evolution The term ‘entangled coherent state’ was introduced by within a double cavity system [46]. The requisite nonlin- 3 ear evolution is actually a special case of the nonlinear the nonclassical coalgebraic structure of the generalized evolution that leads to Titulaer-Glauber generalized co- boson algebra hqi(h(1)) [64]. Squeezed states [65] are herent states [6, 7, 11]. a generalizationU of coherent states as orbits of squeezed Entangled coherent states were initially treated as bi- vacuum states under the Heisenberg-Weyl displacement modal states but later generalized to superpositions of operator. Specifically squeezed states are constructed as multimode coherent states [47–49]. Generalizations to orbits of the squeezed vacuum state whereas coherent multimode systems allows the intricacies of multipsartite states are obtained by the same orbit construction but entanglement to become manifest in entangled coherent with the vacuum replacing the squeezed vacuum state states, for example in Greenberger-Horne-Zeilinger and as the fiducial state [23]. Another way of dealing with W types of states [50, 51] and entangled coherent state squeezing in the context of entangled coherent states to is versions of cluster states [52–54]. to subject them to squeezing [66] or to use squeezed light As entangled coherent states exhibit entanglement, in the homodyne detection process to enhance detector which is a resource for quantum information protocols, performance [28]. entangled coherent states have been studied both as a re- Generalized beyond the coherent state framework is source for quantum information protocols and also as an also also useful. For example coherent states can have input to a quantum information protocol. The degree of photons added to them thereby creating ‘photon-added entanglement embodied by entangled coherent states was coherent states’ [26], which leads naturally to entangled studied in the context of quantum information, where en- photon-added entangled coherent states [67, 68]. The tanglement is considered a resource. “single-mode excited entangled coherent states” also in- By showing that the even bipartite entangled coher- volve adding a photon by acting on an entangled coherent ent state can be obtained by mixing an even coherent state directly with a photon-creation operator [69, 70], state with the vacuum at a beam splitter, van Enk and and this state has value as a cyclic representation of the Hirota [55] establish that the even entangled coherent hw algebra [69]. state has precisely one ebit of entanglement, where one This review provides an overview of research into en- ebit is the amount of entanglement in a maximally entangled coherent states and their generalizations, imple- 1 mentations and applications. This field of research is tangled state of two qubits, or spin- 2 particles. The degree of entanglement in a bipartite entangled coherent quite large so not every paper is cited, but this review state generated by a nonlinear Kerr evolution was sub- strives to be comprehensive in covering all the directions sequently shown to yield an arbitrarily large amount of concerning entangled coherent states. With the recent entanglement over proportionately short times and a lim- successful generation of entangled coherent states [71] ited amount of entanglement over longer times [56]. and their potential importance in quantum information Entangled coherent states have been employed in quan- processing, many new discoveries can be expected in the tum teleportation tasks [57] in two ways: as the state near future. being teleported via continuous-variable quantum teleportation [58, 59] and as the entangled resource state II. FORMALISM employed to effect the teleportation [55, 59–61]. Tele- portation gives operational meaning to the amount of entanglement in an entangled coherent state as telepor- A. Coherent states of the simple harmonic oscillator tation consumes prior shared entanglement to transport quantum information through a classical channel [57]. Entangled coherent states can go beyond entangling Coherent states of the simple harmonic oscillator are harmonic oscillator coherent states. Earlier in this well known since the foundational work of Schrödinger [1] section we mentioned the Barut-Girardello coherent and the ubiquity of coherent states in quantum op- states [8], which can be used to construct pair coherent tics [72–74]. The Hamiltonian for the quantized simple states, and these states are also an example of generaliz- harmonic oscillator in one dimension is ing coherent states beyond the Heisenberg-Weyl algebra. pˆ2 1 Gilmore and Perelomov independently showed another Hˆ = + mω2qˆ2 (1) 2m 2 way of generalizing coherent states based on abstracting the displacement operator to general group operations with ˆ signifying an operator, m the mass of the oscilla- acting on minimum- or maximum-weight states [9, 10, tor, ω the angular frequency, andq ˆ andp ˆ the canonically 12, 13]. The orbit of Gilmore-Perelomov states under conjugate Hermitian operators for position and momen- the general group action form the coherent states for the tum, respectively. These conjugate operators satisfy the given group. For example entangled coherent states can commutator relation [ˆq, pˆ] = i~, and the Hamiltonian be constructed as superpositions of tensor products of 1 ~ spectrum is (n + 2 ) ω for n a nonnegative integer. The two or more generalized Perelomov or Barut-Girardello harmonic oscillator thus has a ground state energy level su(2) [62, 63] and su(1, 1) coherent states as well as en- 1 ~ of 2 ω and all excited energy levels are integer multiples tangled binomial states [62]. of ~ω above the ground state energy level. For the simple In fact entangled coherent states arise naturally from harmonic oscillator, n indicates the number of phonons, 4 with each additional phonon increasing the oscillator’s respectively. The quantityn ¯ = α 2 is the mean phonon mechanical energy by ~ω. number, with the phonon number| | given by a Poisson As the energy levels are equally spaced, it is convenient distribution to introduce the phonon lowering operator 2 iϕ −n¯ n Πn¯ = n α = √ne¯ = e n¯ /n! (8) mω ipˆ aˆ† = qˆ+ (2) D E 2~ √ ~ and r 2m ω D(α) = exp(αaˆ† α∗aˆ) (9) and its conjugate raising operatora ˆ†, which satisfy the − commutator relation the displacement operator. The Poisson distribution has the property that [â, aˆ†]= 11 (3) n¯ = (∆n)2 , (10) corresponding to the Heisenberg-Weyl Lie algebra hw(1) comprising generators a,ˆ aˆ†, 11 with 11 the identity op- and the Mandel Q parameter is [75] erator. { } (∆n)2 n¯ The Fock number states n , for n the number of Q = − , (11) phonons, provide a countable,| orthonormali basis for the n¯ Hilbert space H , with n the number of phonons and n which is unity for the Poisson distribution (8), less than | i an eigenstate of the number operator: unity for sub-Poissonian distributions and greater than unity for super-Poissonian distributions. nˆ aˆ†a,ˆ nˆ n = n n . (4) ≡ | i | i The coherent state can be used for multiple harmonic oscillators. For n simple harmonic oscillators, the joint For the extension to n coupled simple harmonic oscilla- coherent state is tors, each indexed by an integer ℓ, the algebra for the identity 11 and the 2n operators is N α = α = D(α) 0 (12) | i | ii | i † † i=1 aˆ , aˆ ; ℓ =1, 2,...,n , [â , aˆ ′ ]= δ ′ 11. (5) { ℓ ℓ } ℓ ℓ ℓℓ Y with D(α) a product of single-mode displacement opera- The number states are energy eigenstates, hence are tors (9) and 0 the joint ground state of all n oscillators. stationary states of the simple harmonic oscillator. An As discussed| above,i a product coherent state with respect alternative is a Gaussian wavefunction (note that a to a specific set of modes transforms to another product Gaussian in position representation is a Gaussian in coherent state with via a linear mode transformation. the momentum representation so it is sufficient to re- Although this subsection has been concerned with the fer to a Gaussian wavefunction without specifying po- motion of the simple harmonic oscillator, and the en- sition or momentum representation), which satisfies the ergy quanta that separate the equally spaced energy lev- Schrödinger equation and is not stationary. The cen- els have been referred to as phonons, the analysis also troid (mean position and momentum) follows the simple applies to the electromagnetic field. The free-space field harmonic motion expected for the classical simple har- is reducible to modes, and the dynamics of each mode monic oscillator, and the Gaussian remains a minimum- corresponds to a simple harmonic oscillator. The energy uncertainty state (with respect to uncertainty in position quanta are photons, not phonons, as discussed in subsec- and momentum) so the Gaussian wavefunction has desir- tion II B. able properties. This Gaussian, with initial conditions such that the position and momentum uncertainties are stationary, is B. Coherent states in quantum optics known as a coherent state due to Glauber’s association of such a state with coherence in quantum optics. Glauber’s In quantum optics, the electromagnetic field can be oscillating wavepacket is an eigenstate of the annihilation decomposed into modes, and the dynamics of each mode operatora ˆ for a given harmonic oscillator. This pure- in free space is equivalent to the dynamics of the sim- state wavefunction is given in the Fock representation by ple harmonic oscillator, with n the number of photons ∞ (rather than phonons) in the given mode. The canoni- α 2 αn α = D(α) 0 = exp | | n , (6) cally conjugate operatorsq ˆ andp ˆ are referred to as the | i | i − 2 √n!| i n=0 in-phase and out-of-phase quadratures, respectively, with X the phase reference being a local oscillator. with α a complex dimensionless amplitude such that the As the phase of the local oscillator can be varied con- mean position and momentum are given by tinuously, it is convenient in quantum optics to define the quadrature operator as 2~ ~ω q¯ = Re(α), p¯ = Im(α), (7) qˆθ qˆcos θ +ˆp sin θ , (13) rmω r2m ≡ 5 withq ˆ =q ˆ0 andp ˆ =q ˆπ/2. Measurements of a quadra- tations as ture are performed by mixing the signal field with the d2α α ρ α coherent local oscillator field in an optical homodyne de- ρ = P (α),Q(α)= h | | i . (15) π π tection apparatus [20–22], with the local oscillator field Z determining the phase θ. In quantum optics, the field is said to be semiclassical In developing a theory of coherence for optical fields, if P (α) satisfies the axiomatic requirements of a prob- Glauber employed the coherent state [72], but, instead of ability density and is ‘quantum’ otherwise. In contrast the variables being position and momentum of a massive the Q-function is always a probability density. There is particle in a harmonic potential, the canonically conju- a continuum of these so-called quasiprobabilities intro- gate variables are the in-phase and out-of-phase quadra- duced by Cahill and Glauber [86], which are obtained tures of each mode of the field. The two quadrature field through Gaussian convolution of P (α) with a Gaussian, operators are constructed from the raising and lowering with s corresponding to +1 for P , 0 for the Wigner func- operators for the field mode with three-vector label ~k and tion W (q,p), and 1 for Q. The Wigner function is an especially important− case because, although it is not polarization index ε, namelya ˆ† anda ˆ , respectively. ~kε ~kε positive-definite, integrals of the Wigner function yield Properties of the coherent state have been discussed in the marginal distributions for position and momentum. Subsec. II A and are investigated in detail by Klauder The position marginal distribution is thus and Skagerstam [73]. ∞ Coherent states have served as a valuable tool for P (qθ)= W (qθ, qθ+π/2)dqθ. (16) studying quantum optics, primarily because of the con- −∞ Z venience of these states as representations. In addition to Titulaer and Glauber introduced “generalized coher- a coherent state being an eigenstate of the annihilation ent states” as states that are fully coherent with respect operatora ˆ, another critical property is that the product to the coherence functions but are not eigenstates of the coherent state is the unique state that transforms to a annihilation operatora ˆ [6]. These states have Poisson product coherent state under the action of a linear mode number distributions but allow an arbitrary phase rela- coupler such as a beam splitter [76]. For the mode cou- tionship between coefficients in the Fock representation pler transformation given by of the state. For dϕ := dϕ/2π 6 ′ iφ ∞ n/2 aˆ1 cos θ e sin θ aˆ1 −n/¯ 2 n¯ iϑn ′ = −iφ (14) √n¯; ϑ = e e n aˆ2 e sin θ cos θ aˆ2 √ | i n=0 n! − E X 2π ∞ a product coherent state then transforms to a product = dϕ √ne¯ iϕ ei(θn−nϕ), (17) coherent state. This property is particularly important 0 6 Z E n=0 in that a coherent state for the field remains a coherent X state under any linear mode transformation. with the last line corresponding to the representation of Whether coherent states can be considered as ontologi- the generalized coherent state as a superposition of co- cally real has been the subject of vigorous debate, both in herent states on a circle, and this representation on a the context of coherence of atomic Bose-Einstein conden- circle was introduced by Bialynicki-Birula [7]. sates [77–79] and with respect to quantum teleportation Spiridinov [41] showed that these generalized coher- of coherent states [19, 80–83]. The problem essentially ent state are eigenstates of a generalized annihilation concerns the establishment of the phase ϕ of the coher- operator that holds the number operatorn ˆ invariant. ent state either through its creation from a source or by a Physically Spiridonov’s transformation corresponds to a phase-sensitive detection of the state. In practice phase- number-sensitive rotation; optically we can understand locking mechanisms exist that ensure that the phase of this transformation as a generalization of the ideal single- the coherent field is correlated with a reference field, and, mode optical Kerr nonlinearity, which effects a phase treating that field as classical, provides classical meaning shift that is a function of field strength, or, equivalently, to the parameter ϕ. This issue is discussed in more de- photon number for the single-mode field. tail in Subsec. ??, which is concerned with the entangled coherent state representation. C. Superpositions of coherent states Given the indisputable value of the coherent state as a representation, there are two useful ways to represent the density matrix ρ of the single-mode field in terms of co- Whereas the coherent state is regarded as the clos- herent states. One makes use of the Glauber-Sudarshan est quantum optical description to a classical coherent P -representation [74, 84] and the other makes use of the field, superpositions of coherent states exemplify the Husimi distribution, or Q function [85] (not to be con- strangeness of quantum theory. In general any pure state of the field ψ can be written as a superposition of co- fused with the Mandel Q [75]) whose advantages with | i respect to superpositions of coherent states were elabo- herent states according to the expression rated by Milburn [2]. d2α ψ = α ψ α . (18) The density matrix can be expressed in these represen- | i π h | i | i Z 6

As the coherent states form an overcomplete basis, it is reminiscent of the Schrödinger’s cat being described as is not surprising that every state can be expressed as a being in the state ‘live’ + ‘dead’ . superposition of coherent states. The term Schrödinger| i cat| statei has also been applied Interestingly, the overcompleteness of the coherent- to the ‘even’ and ‘odd’ coherent states [39], state basis allows quite different ways to write the superposition. One particularly important case is the su- α = N± ( α α ) (22) | i± | i±|− i perposition of coherent states on the circle, which we have encountered in Subsec. II B in studying the Titulaer- for Glauber generalized coherent states [6]. Other states, exp α 2 exp α 2 can also be expressed in this way. For example, the Fock N+ = | | , N− = | | . (23) number state has the appealing representation [7, 14, 87] 2 cosh α 2 2 sinh α 2 | | | | 1 The even-odd terminologyp refers to thep fact that the pho- n = dϕe−imϕ √meiϕ , (19) | i Π (m) 6 ton number distribution is non-zero only for even pho- n Z ton number in the case of the even coherent state α + | i with Π the Poissonp distribution (8) and d := d/2π. Ex- and is non-zero only for odd photon number in the case 6 of the odd coherent state α . As this state does not pression (19) is a superposition over coherent states with | i− complex amplitude restricted to having modulus √m. have a Poisson number distribution, it cannot be evolved States can also be expressed as superpositions of coher- via a generalized unitary Kerr evolution from the coherent states on lines or other subspaces of the α parameter ent state but is a Titulaer-Glauber generalized coherent space. state [6] for which ϑ(ˆn+211) = ϑ(ˆn). Even and odd coher- Evolution of a coherent state under an ideal optical ent states may arise by a conditional Jaynes-Cummings Kerr nonlinearity evolution [35, 89]. Detection of Schrödinger cat states may be achieved by ϑ(ˆn)= ωnˆ + λω2nˆ2 (20) optical homodyne detection, with the measurement results converging to the marginal distributions for canoni- yields a particular form of generalized coherent state [2, cal position and momentum [4, 5] even in the presence of 3, 88]. Yurke and Stoler [4, 5] showed that a superpo- decoherence [27, 28]. Let us consider the ‘balanced cat’ sition of two coherent states could be obtained under of Eq. (21) with its equally weighted superposition of two this evolution, and its generalization to ϑ(ˆn) nˆk could coherent states π out of phase. If the local oscillator is ∝ be expressed as a finite superposition of coherent states in-phase with either of the coherent states, the marginal with different phases for certain evolution times. In fact distribution is equivalent to that for an incoherent mix- the Titulaer-Glauber generalized coherent [6] can be ex- ture of such coherent states. The marginal distribution pressed as a superposition of a finite number of coherent for the conjugate quadrature exhibits interference fringes states on the circle for ϑn+N = ϑn for some N and for that yield information on how coherent, or pure, the su- all n [7, 11]. perposition state is. Spirodonov [41] identified two other interesting cases of So far we have considered superpositions of single- generalized coherent states: q-deformed coherent states mode coherent states, but a superposition of multimode for which ϑn+N = qϑn and parity coherent states ϑn = coherent states of the type (12) is also allowed. Such a nπ. The parity operator is exp iπnˆ , and the (unnor- state can be written as malized) parity coherent state is{ given} by dµ(α)f(α) α (24) e−iπ/4 α +eiπ/4 α , (21) | i | i |− i Z which is a special case of the superpositions of coher- for which the measure dµ(α) can be over the entire parameter space or over subspaces for which the set α) ent states with equal complex field amplitudes and equal { } phase separations studied by Bialynicki-Birula [7] and is an overcomplete basis. Before proceeding onto stud- Stoler [11]. These superpositions are discussed in more ies of this superposition of multimode coherent states, detail in Subsec. II C. we consider how coherent states and their superpositions As we saw in Subsec. IIB, superpositions of coher- are generalized to Lie groups and algebras other than the ent states on a circle can arise via evolution of a coher- Heisenberg-Weyl group for simple harmonic oscillators. ent state according to a generalized Kerr nonlinearity, yielding an evolution operator exp iϑ(ˆn) . The equally weighted superposition of two coherent{ states} that are π D. Lie coherent states and their superpositions out of phase with each other (21), introduced by Yurke and Stoler [4], has been termed a “Schrödinger cat state”, The term “generalized coherent state” has been used or “cat state” for short, because the coherent state is re- in Subsec. II C to refer to loosening the phase relation garded as being an essentially classical field state, and between elements of the coherent state expressed as a the superposition of two highly distinct coherent states superposition in the Fock basis; we have been careful 7 to refer to these as ‘Titulaer-Glauber generalized coher- the lowering operator and (ii) as orbits of a minimum ent states” [6]. The term “generalized coherent state” uncertainty state, analogous to the orbit of the vacuum has also been applied to establishing coherent states for state under the displacement operator (9). Highest- or general Lie groups. Here we refer to Lie group and al- lowest weight states (states that are annihilated by the gebra generalizations of coherent states as “Lie coher- raising and lowering operators, respectively) are typical ent states”. Where the specific Lie algebra is specified, minimum-uncertainty states for the two groups under the notation for the algebra replaces “Lie”, e.g. “su(2) consideration. coherent state”. For n simple harmonic oscillators, the In 1971, Barut and Girardello [8] introduced “new co- operator algebra is hw(n) given in Eq. (5), and the Lie herent states” for non-compact groups based on crite- coherent state for hw(n) is the multimode product co- rion (ii). They identified the lowering operator(s) and herent state (12). found eigenstates for this operator. For su(1, 1), the low- Whereas hw(N) coherent states are (i) displaced vac- ering operator is Kˆ−, and the Barut-Girardello su(1, 1) uum states (orbits of the vacuum state under the action coherent state is of the displacement operator D(α), (ii) eigenstates of ∞ the lowering operator aˆ, and (iii) minimum-uncertainty η k−1/2 ηn k η = | | kn , (30) states, some sacrifice must be made in defining coher- | iBG I (2 η ) | i s 2k−1 n=0 n!Γ(n +2k) ent states for other Lie groups. A basis set of operators | | X for a Lie algebra can be expressed as lowering operators p which satisfies Kˆ− k η BG = η k η BG, for Iν (x) the mod- analogous toa ˆ, the conjugate raising operators, and the ified Bessel function| ofi the first| kind.i Cartan subalgebra, which is a set of mutually commuting For both non-compact and compact groups, coherent elements of the algebra. states can be defined analogously to the displaced vac- A Lie algebra generates a k-parameter Lie group uum state by introducing a minimum-uncertainty state with the dimension of the Cartan subalgebra being k. such as the highest- or lowest-weight state in the rep- Studies of superpositions and entanglement of coherent resentation and considering a generalization of the dis- states have so far focused primarily on one-parameter Lie placement operator. For su(2), the usual minimum- groups (with the exception of one study on su(3) coher- uncertainty state is the highest-weight state j j al- ent states [90]), so we restrict our attention to that case; though the lowest-weight state is used as well. | i in fact we can concentrate on su(2) and su(1, 1). For su(1, 1), the lowest-weight state k 0 is used. For Entanglement of su(2) and su(1, 1) coherent states SU(2), the analogue to the displacement| operatori is the were studied by Wang and Sanders [62]. The correspond- “rotation operator” ing algebras are θ −iϕ ˆ iϕ ˆ ˆ ˆ ˆ ˆ ˆ ˆ R(θ, ϕ)= e J− e J+ , (31) [J+, J−]=2Jz, [Jz, J±]= J± (25) 2 − ± h i and and, for SU(1,1), the analogue is the “squeeze operator” [91] [Kˆ , Kˆ ]= 2Kˆ , [Kˆ , Kˆ ]= Kˆ , (26) + − − z z ± ± ± S(ξ) = exp ξKˆ ξ∗Kˆ . (32) respectively, for su(2) and su(1, 1), with J used for the + − − n o compact SU(2) group and K used for the non-compact The term ‘rotation operator’ applies for SO(3), which is SU(1,1) group. The Cartan subalgebras are Jz for su(2) the rotation group in three-dimensional Euclidean space, ˆ2 and Kz for su(1, 1). The Casimir invariants are J with and the term has been extended to apply to SU(2), which spectrum j(j + 1) for irreducible representation, or ir- is a double covering group of SO(3). The term ‘squeeze reps, indexed by j 0, 1/2, 1, 3/2,... , and Kˆ 2 with operator’ is used here because two-boson realizations of ∈ { } spectrum k(k 1) for irreps indexed by su(1, 1) are − k 1/2, 1/3/2, 2,... . (27) Kˆ =â2 (k =1/4), Kˆ =âˆb (k =3/4), (33) ∈{ } − − Within a given irrep, an orthonormal basis is given by and either of these realizations of su(1, 1) converts the unitary operator (32) to the usual one-mode and two- jm ; m = j, j +1,...,j , (28) mode squeeze operators in quantum optics for k = 1/4 {| i − } and k =3/4, respectively. such that The su(2) coherent states were first introduced as “atomic coherent states” [13]. These states are given by Jˆ jm = j m +1 jm 1 , j j 1 0 (29) −| i − | − i | − − i≡ j; θ, φ = R(θ, φ) j j (34) with kn ; np=0, 1, 2,... for su(2). The su(1, 1) can be | i | i constructed{| i in a similar way.} and form an overcomplete basis of the Hilbert space [9, For non-compact groups, there are two inequivalent 12]. The coherent states are orbits of the minimum- ways to construct coherent states: (i) as eigenstates of uncertainty state under the action of a group element. 8

Perelomov undertook a general analysis of such coherent In the single-particle (equivalently single-mode) case, states for any Lie group, and these Lie coherent states such states are the Titulaer-Glauber coherent states [6]. are known as Perelomov coherent states [10]. As an interesting example of discrete bipartite entan- We refer to the Lie coherent state using the nota- gled coherent states, van Enk studied the discrete “mul- tion ℓ ξ with ℓ the irrep parameter (not required for tidimensional entangled coherent states” [56] the hw| (in) algebra) and ξ the orbit parameter. This notation applies equally to eigenstates of the lowering 1 M−1 Φ = eiφq αe2πiq/M , αe2πiq/M , (38) operator (as for the Barut-Girardello states) and for or- | M i √ M q=0 bits of minimum-uncertainty states (as for the Perelomov X E states). The multipartite Lie coherent state is designated which are generated by an ideal nonlinear Kerr evolu- by ℓ ξ , which is a product of Lie coherent states ℓ ξi all from| i the same irrep. | i tion (20), to characterize the entangling power. He shows So far we have only considered entangled coherent that such states have infinite entanglement for short states where each party in the state has the same times τ infinitesimally short evolution times and finite coherent-state structure. For example the entangled co- entanglement after finite times. Finite discrete superpo- herent state can be a superposition of a tensor prod- sitions can serve as a resource for quantum teleportation. uct of hw(1) coherent states or a superposition of ten- Showed that, for very small losses for multidimensional sor product of su(2) coherent states or so on. On the entangled coherent states, approximately 2.89 ebits are other hand a ‘hybrid’ entangled coherent state could be lost per absorbed photon, which could be useful for cre- constructed as a superposition of tensor products of co- ating entangled coherent states with a fixed amount of herent states with coherent states in the tensor-product entanglement [93]. space corresponding to different types of coherent states. Entangled coherent states overlap conceptually with Such hybrid coherent states have not been studied but the pair coherent state, which was introduced as the joint have been realized experimentally in a limited way: hy- eigenstate of the two-mode, annihilation operatora ˆ1aˆ2 brid hw(1)-su(2) entangled coherent state are realized in and the number difference operatorn ˆ1 nˆ2 [24, 25, 94]. Defining the pair coherent state by ζ,− q , with ζ the cavity quantum electrodynamics experiments as entan- | i gled atom-field states [16, 35, 92]. eigenvalue of the pair annihilation operator and q the eigenvalue of the photon number difference operator. These states exhibit sub-Poissonian statistics, correlated E. Entangled coherent states number fluctuations, squeezing, and violations of photon Cauchy-Schwarz inequalities. Pair coherent states are an example of su(1, 1) coherent A superposition of mutimode or mutipartite coherent states, represented as two-boson realizations. The pair states can be expressed in general as [62] annihilation operator can be expressed according to the algebra (26): the pair annihilation operator is Kˆ− and dµℓ(ξ)fℓ(ξ) ℓξ (35) the photon number difference operator is Kˆ . | i z Z The pair coherent state is an example of an entangled with the state ℓξ the Lie coherent state. For the usual coherent state, which is evident by expressing the pair case of harmonic| i oscillators, corresponding to the alge- coherent state as bra hw(n), the index ℓ is superfluous, and we let ξ be 2π |ζ) α Nqe iϕ −iϕ replaced by to obtain (24). This superposition is not ζ, q = dϕ q ζe ζe (39) | i 6 √ζeiϕ | i| i entangled if there exists any representation for the pure Z0 state such that the state can be expressed as a product p p state over the modes. Otherwise the state is entangled. with q-dependent normalization constant Thus entangled coherent states are a special case of su- q/2 perpositions of multimode coherent states, but a rather ζ Nq = | | , (40) large and especially interesting class of states. Iq(2 ζ ) The entangled state (35), which is expressed as an in- | | p tegral of product coherent states, can be reduced to a for Iq the modified Bessel function of the first kind, as sum if the function fℓ(ξ) can be expressed as a sum of expressed by Gerry and Grobe [95]. delta functions The Schrödinger cat state concept, which corresponds to a superposition of coherent states, was extended to f (ξ)= f (ξ )δ(ξ ξ ). (36) a superposition of pair coherent states by Gerry and ℓ ℓ i − i i Grobe [95]. Specifically the superposition of two pair X coherent states can be expressed as Then ζ, q + eiφ ζ, q ζ,q,φ = | i |− i (41) dµ (ξ)f (ξ) ℓξ = f (ξ ) ℓξ . (37) n 2n ℓ ℓ | i ℓ i | ii | i 2+2N 2 cos φ ∞ (−1) |ζ| i q n=0 n!(n+q)! Z X q P 9 which is an eigenstate of the squared pair annihilation B. Nonlinear optics operator with eigenvalue ζ 2. This superposition of pair | | coherent states is also an entangled coherent state, which The earliest proposals for creating entangled coherent is a superposition of two entangled coherent states of the states were expressed in the context of quantum optical type (39). fields interacting via a third-order optical nonlinearity known as a Kerr nonlinearity [98]. The optical Kerr nonlinearity features a refractive index n0 + n2I, which is the sum of a linear refractive index n0 and a second term III. IMPLEMENTATIONS that is proportional to the field strength typically characterized by ‘intensity’ I. The Kerr effect, as a third-order Entangled coherent states are fragile due to their en- optical nonlinearity, is a special case of four-wave mixing. tanglement but are nonetheless implementable if the con- The term ‘cross-Kerr nonlinearity’ is ubiquitous in the ditions are right. Furthermore the states serve as a re- entangled coherent state literature and refers to the phe- source for quantum information processing so they have nomenon that one field experiences a phase shift com- utility and value hence are worth making. Many theoret- ponent that is proportional to the strength of the other ical proposals exist for constructing entangled coherent field. The cross-Kerr effect thus leads to ‘cross-phase states in the laboratory but so far the paramount exper- modulation’, which is specifically the phase shift of one imental demonstration uses a photon-subtraction tech- field due to the intensity of the other. In the quantum nique on two approximate Schrödinger cat states so that analysis, the phase shift depends on the photon number the source of the photon is indeterminate [71, 96]. in the other beam. Of course the phase shift of the beam also depends on its own strength, and this is known as ‘self-phase modulation’. The first proposal for generating entangled coherent A. Parametric amplification and photodetection states was introduced by Yurke and Stoler in 1986 [4, 5] followed by the work of Mecozzi and Tombesi in 1987[27, 28] and others [29, 30]. The entangled coherent state Consider two physically separated states of light of the became the chief object of interest in work by Sanders type α + α . These two fields are each passed through a few years later with a proposed implementation that a separate| i |− beami splitter such that each field loses a small inserted a Kerr nonlinearity into one path of a Mach- fraction of its energy. The extracted part of each field is Zehnder interferometer (often called a ‘nonlinear inter- brought together after using a phase shifter to impose a ferometer’ for short) [32, 33] and has been the subject of φ phase difference between the two beams. The fields are further study [29, 30, 99, 100]. Related to this approach, combined at a beam splitters with a photon counter at if an appropriate superposition of two coherent states is one output port. The effect of this final beam splitter provided, then a beam splitter transformation alone suf- is to ensure that the detected photon is equally likely to fices to produce an entangled coherent state from this have come from either beam. The resultant two-mode resource [101] state conditioned on registering a single photon count is The use of a Kerr nonlinearity to entangle the coherent state with a vacuum state, which is a special case of φ φ i sin α α cos α α bipartite entangled coherent states, was generalized by − 2 | i | i− 2 |− i | i Luis to show how to entangle any state with the vacuum φ φ + cos α α + i sin α α . (42) state [102]. Wang showed how a nonlinearity coupled 2 | i |− i 2 |− i |− i with linear optical elements can be employed to generate general bipartite entangled non-orthogonal states [103]. This is the concept behind the successful experimental Variants of nonlinear interacting propagating field re- creation of a close approximation of this state, and the alizations of entangled coherent states have been stud- success of the process is verified by optical homodyne ied. Slow light in a medium with double electromag- tomography on the resultant state [71]. netically induced transparency could be used to enable The actual experiment involves using a pulsed opti- entangled coherent state generation [104, 105]. Entan- cal parametric amplifier as a source of squeezed vacua. glement could first be prepared in matter qubits then The cat state with small amplitude α, known as a “kit- transferred to fields to make entangled coherent states ten state”, can be prepared by subtracting a single pho- by exploiting a cross-Kerr nonlinearity [106]. Nonlocal ton [17, 97]. Using this principle, the two output modes preparation of a bipartite entangled coherent state, where squeezed vacua of orthogonal polarizations are recom- ‘nonlocal’ means that the two fields being entangled never bined at a polarizing beam splitter. A small fraction of meet or directly interact, could be produced by sending each field goes to the photon counter, which conditions a photon through a Mach-Zehnder interferometer with a the rest of the field going out the other beam splitter port nonlinear Kerr medium in each of its two paths, and sep- into the entangled coherent state (42). The state is then arate coherent states are sent through each of these two tomographically characterized. nonlinear media [107]. The bipartite entangled coherent 10 state is post-selected by detecting from which port the can be used instead. Vibrational degrees of freedom photon leaves: whichever port the photon leaves from for a single trapped ion in two dimensions [126] or of post-selects the nonlocal two-mode field in one of two two trapped ions [49, 127, 128] or for collective modes entangled coherent states. (e.g. center-of-mass or breathing modes) of many trapped Generation of various exotic forms of entangled coher- ions [129, 130], can be transformed into entangled coherent states have been investigated. Greenberger-Horne- ent states. Ion traps can be combined with cavity quan- Zeilinger and W types of entangled coherent states could tum electrodynamics set-ups to make hybrid entangled be produced with propagating fields using linear optics coherent states between electromagnetic and motional and Kerr nonlinearities [50, 51]. Similarly cluster-type degrees of freedom [131, 132]. entangled coherent state can be generated with a nonlin- As an example of creating an entangled coherent state ear medium and a laser driving field [53, 108, 109]. in the vibrational degrees of freedom of a single trapped ion in two dimensions, consider the interaction Hamilto- nian [126] C. Cavity quantum electrodynamics Hˆ = ~χ aˆ†aˆ ˆb†ˆb (ˆσ +ˆσ ) (43) Int − − + − Entangled coherent states can be created in cavity fields rather than in propagating fields, which has the witha ˆ and ˆb the annihilation operators for each of the two advantage of large effective nonlinearities. The nonlin- † vibrational modes, χ a coupling coefficient andσ ˆ+ =σ ˆ− earity in the medium could be a macroscopic optical Kerr the electronic-energy lowering operator. Both vibrational medium or one or more multilevel atoms. For example a modes are initialized in coherent states α and β and multimode entangled coherent state can be prepared by the ion in the ground state g . At time|t ithe combined| i letting a a single atom traverse two or more single-mode (unnormalized) state for the| i atom and the two vibra- cavities, each occupied initially by a coherent state, and tional states is then post-selecting on the atomic state [35, 48, 54, 110– 112]. An unbalanced (i.e. unequally weighted) entangled g αeiχt,βeiχt + αe−iχt,βe−iχt | i coherent state could be produced in a double-cavity sys- iχt iχt −iχt −iχt + e αe ,β e αe ,β e (44) tem [46]. Alternatively just one cavity that supports a | i − multimode field is an alternative to multiple-cavity gen- for e the excited state of the atom. The entangled co- eration of entangled coherent states [112–116]. herent| i state can be created post-selectively by measuring Matter-wave interferometry could assist in preparing the electronic state of the ion. entangled coherent states. If a two-mode cavity can be Ion traps could be used to create multipartite entan- prepared in a pair coherent state, then this state can gled coherent states using entanglement swapping oper- be transformed into an entangled coherent state by the ations [49]. Consider two identical ions with each ini- following procedure. Atoms are sent through a double tially prepared in a superposition of ground and excited slit and then interact with the two-mode cavity field. state and the center-of-mass and breathing modes each Atomic-position detection subsequently post-selects the initially prepared in coherent states with the same am- two-mode field into an entangled coherent state [117]. plitude and phase. Then two distinct Raman beams are Artificial atoms, such as quantum dots [118] or Cooper- directed at the two ions independently. One beam is pair boxes [119], can replace real atoms to produce en- directed at the first ion in order to couple it to the fun- tangled coherent states in cavity quantum electrodynam- damental mode, and the second beam is directed at the ics. The microwave regime could prove to be quite ap- second ion in order to couple it to the breathing mode. propriate for generating entangled coherent states with This selective coupling of ion electron levels to mo- Rydberg atoms in millimeter-wave superconducting cav- tional modes is achieved by choosing judicious Raman ities [120, 121]. parameters. Subsequently Bell-state measurements of As with propagating fields interacting with a Kerr the two-ion electronic states result in the two motional medium, exotic entangled coherent states such as modes ‘collapsing’, or being post-selected, into entangled Greenberger-Horne-Zeilinger, W [122] and cluster-type coherent states. This principle is readily extended to the entangled coherent states, can also be created in cav- multimode entangled coherent state case by extending ities [51, 54, 123–125]. Modified entangled coherent from two to as many ions as desired naturally accom- states, such as the “single-mode excited entangled co- panied by as many vibrational modes. The multi-ion herent states”, could also be created in a cavity quantum electronic state is projected onto a maximally entangled electrodynamics setting [69]. state (generalized Bell measurement) thereby resulting in the vibrational modes being in a multimode (or ‘multipartite’ entangled coherent state) [49]. D. Motional degrees of freedom Other physical realizations that are amenable to creating entangled coherent states in motional degrees of free- Instead of creating entangled coherent states in elec- dom include nano-cantilevers [119] and movable nano- tromagnetic field modes, motional degrees of freedom mirrors [133, 134]. More pointedly, entangled coherent 11 states can in principle be realized in any system that A. Complementarity can be described as harmonic oscillators with appropriate nonlinear coupling and sufficiently low loss and deco- Complementarity in double-slit [143] and two-channel herence. interferometery [144] studies is well understand for single- particle inputs. Single photons exit wholly from either one or the other port of a beam splitter but exhibit strong fringe visibility when the experiment is modified E. Bose-Einstein condensates by replacing the beam splitter by an interferometer [36]. Complementarity may be understood by thinking of the photon’s path state as entangled: a superposition of the Bose-Einstein condensates have an inherently high photon traversing one path (e.g. through one slit or down nonlinearity due to atomic collision terms, and it is possi- one channel of the interferometer) and a vacuum state in ble to prepare two separate Bose-Einstein condensates of the other path and the reverse case. Now consider, in- three-level atoms (corresponding to different electronic stead of a photon in one path and vacuum in the other, states of the same atoms) into phase-locked coherent we have a coherent state in one path and a vacuum in states and couple them together via a Raman interac- the other. Would complementarity be manifested and tion [135]. This approach could be used to entangle an observable in that case? arbitrary state of one Bose-Einstein condensate with a Rice and Sanders showed that, in principle, a form ground state of the other Bose-Einstein condensate [102]. of complementarity is present, but the notion of a phase Alternatively entangled coherent states could be gener- shifter, which is a simple linear optical element for a pho- ated in Raman-coupled Bose-Einstein condensates [136]. ton, is complicated for a coherent state yet necessary to Nonlocal preparation of distant entangled coherent observe the undularity of the coherent state in the con- states could be possible using electromagnetically in- text of entangled coherent states [100]. Joint photode- duced transparency [137]. In this scheme, two strong cou- tection at the two interferometer output ports [46] can pling laser beams and two entangled probe laser beams reveal anticorrelation of the nonlinear Mach-Zehnder in- prepare two distant Bose-Einstein condensates in electro- terferometer output thereby revealing ‘corpuscularity’ of magnetically induced transparency coherent population the coherent state analogous to the anticorrelation reveal- states then forced to interact. The two Bose-Einstein ing corpuscularity for a single photon [100]. The coherent condensates are initially in a product coherent state while state is thus ‘seen’ to follow one path or another and not the probe lasers are initially entangled. The final prepa- be split. ration step involves performing projective measurements The ideal coherent state phase shifter would corre- upon the two outgoing probe lasers. spond to the unitary transformation exp( iφ α α ) for imposed phase shift φ and could be created− in| ih approxi-| mate form in a highly nonlinear medium with appropriate parameters [145]. The creation of this phase shifter would IV. NONCLASSICAL PROPERTIES enable other types of tests of complementarity such as performing two-coherent-state interferometry, even with large numbers of photons. Two-coherent-state interfer- Entangled coherent states are highly nonclassical ometry is analogous to two-particle quantum interferom- states but are peculiar in that they are expressed as etry but with the single-particle Fock state replaced by an entanglement of the most classically well behaved a coherent state [145]. states we know: coherent states. Thus entangled coherent states are especially intriguing in studies of nonclassicality because the state represents an entanglement of B. Entanglement classically meaningful descriptions of objects. Nonclassicality is studied through a variety of measures The nomenclature “entangled coherent state” demands including squeezing [30, 38, 138], sub-Poissonian pho- quantification of the degree of entanglement of such ton statistics [38], violations of Cauchy-Schwarz inequal- states. There is more than one way to study entangle- ities [38], complementarity between particle-like and ment of such states. One can consider Bell inequalities wave-like features of entangled coherent states [100], vio- or generalizations thereof, perhaps to test local realism lations of a Bell inequality [32, 33, 37, 120, 121, 139, 140] or just to show non-factorizability. Another approach to or Leggett’s inequality [141] for testing local realism, studying entanglement of these states is to recognize that and entanglement properties such as index of correla- unentangled coherent states have Gaussian statistics and tion [62], entanglement of formation [142] and other mea- then use the covariance properties to quantify the de- sures [138]. Nonclassicality of generalized entangled co- gree of entanglement in such states [146]. An alterna- herent states, such as su(2) and su(1,1) states [62, 63] and tive approach considers the entangling power of opera- photon-added entangled coherent states [68] has been tions that produce entangled coherent states [56]. Each studied as well. of these approaches is challenging because the Hilbert 12 spaces are infinite-dimensional and the entanglement is A ‘cat state’ superposition of two coherent states readily between non-orthogonal states [37, 103]. serves as a qubit for quantum logical encoding [169]. Quantifying entanglement can instead by studied in The ‘cat state’ qubit also serves as the logical basis the context of performing a quantum information pro- for performing universal quantum computation [170], and cessing task such as quantum teleportation [57]. Tele- entangled coherent states play an important role in such portation enables a qubit to be sent from one party to quantum information processing [148, 148]. In particu- another through a classical channel by sending instead lar this encoding leads to entangled qubits (ebits) corre- two bits of information and consuming one “ebit”, or en- sponding to entangled coherent states [45]. tangled bit (two maximally entangled qubits) of a prior Entangled coherent states also serve an important role shared entanglement resource. in quantum metrology, which harnesses quantum re- Entanglement can then be quantified by determining sources such as entanglement to surpass the standard how well entangled coherent states serve as the ‘quan- quantum limit (due to partition noise in particle inter- tum channel’ (i.e. the consumable prior shared resource) ferometry, which applies to atomic clocks and displace- for teleporting another state. This other state could be ment measurements inter alia) [91]. Multimode even/odd a qubit corresponding to a superposition of single-mode coherent states are especially amenable for quantum coherent states (‘cat state’) [55, 60, 101, 147–150]. The metrology [43]. Entangled coherent states are known to entangled coherent state can supply an entire ebit of re- outperform other popular two-mode entangled states in source despite being an entanglement of non-orthogonal quantum metrology [171, 172], but perhaps its benefit is states [55, 59, 61]. An alternative approach to studying strongest for digital parameter discrimination [173]. quantum resources considers how well a given a resource The entangled coherent state representation [19] plays serves to teleport all or part of an entangled coherent a key role in resolving fundamental issues concerning su- state [58, 59, 151–157]. perselection of angular momentum [18], charge [18] and Entanglement has been studied for various exotic forms phase [19, 77–83]. Essentially the entangled coherent of entangled coherent states. Both the Greenberger- state representation captures, in a mathematically sim- Horne-Zeilinger type of entangled coherent states [50, ple and conceptually appealing way, how superselection 59, 147, 152, 158] and the W type of entangled coher- can be obviated by adding an extra degree of freedom ent states [50, 158–160] have been studied as well as the and splitting the state to provide a reference frame. cluster type of entangled coherent states [161]. The effect of dissipation and decoherence on entanglement and nonlocality is also a subject of intensive inves- VI. SUMMARY AND CONCLUSIONS tigation for all types of entangled coherent states, including the robustness or fragility of the entanglement [162]. This review provides a comprehensive summary of Characterization of probabilistic teleportation of coher- results concerning entangled coherent states and their ent states via entangled coherent state quantum channel generalizations since the inception of entangled coher- in an open system [142]. Non-Markovian decoherence ent states by Aharonov and Susskind in 1967 to obviate dynamics is important for entangled coherent states, and superselection. Coherent states are appealing for their An, Feng and Zhang obtain an exact master equation mathematical elegance as representations and closeness with and without environmental memory using influence to classical physical states, and entangled coherent states functional theory [163]. build on these elegant representation properties. Strategies to mitigate decoherence of entangled coher- Furthermore entangled coherent states have a richness ent states, for example by squeezing [66], are of practi- due to entanglement between these seemingly classical cal value. Entanglement purification for mixed entangled coherent states. Entangled coherent states have many coherent states is also a promising approach [164, 165]. beautiful nonclassical properties and generalize beyond Park and Jeong compare the dynamics of entangled co- the Heisenberg-Weyl algebra of harmonic oscillators to herent states against entangled photon pair states under the cases of spin, squeezing, pair coherent states and be- decoherence and inefficient detection [166]. They dis- yond. cover that entangled coherent states are more robust as Remarkably entangled coherent states have been cre- quantum channels for teleportation whereas entangled ated and observed experimentally. These exquisitely photon pair states are better with respect to photode- fragile states can be manifested in the laboratory given tection inefficiency. sufficient guile. Until now the one successful experimental realization relies on parametric amplification in two modes and photon subtraction. Other realizations could V. APPLICATIONS AND IMPLICATIONS be possible if large low-loss Kerr nonlinearities are created for propagating or for cavity fields. Ion traps could Entangled coherent states have several applications as also be promising for realizing entangled coherent states discussed earlier in this review. For example entangled between vibrational modes, and nanotechnology could coherent states can serve as a resource for quantum tele- open new vistas for entangling coherent states of motion. portation [55, 59, 61] or for quantum networks [167, 168]. Multipartite entanglement is a vast topic of research, 13 and entangled coherent states play an important role in applications of entangled coherent states for upcoming this area. Various multipartite entangled states such as decades, which hold further revelations and surprises. Greenberger-Horne-Zeilinger, W and cluster states are studied for their rich properties and applications, and each of these states has nontrivial analogues with entangled coherent states. Acknowledgments In summary entangled coherent states have been important from superselection arguments in 1967 to today’s applications to quantum information processing. This re- This project has been supported by AITF, CIFAR, view article can serve as a resource to propel studies and NSERC, MITACS and PIMS.

[1] E. Schrödinger, Naturwissenschaften 14, 664 (1926). [31] G. S. Agarwal and R. R. Puri, Phys. Rev. A 40, 5179 [2] G. J. Milburn, Phys. Rev. A 33, 674 (1985). (1989). [3] G. J. Milburn and C. A. Holmes, Phys. Rev. Lett. 56, [32] B. C. Sanders, Phys. Rev. A 45, 6811 (1992). 2237 (1986). [33] B. C. Sanders, Phys. Rev. A 46, 2966 (1992). [4] B. Yurke and D. Stoler, Phys. Rev. Lett. 57, 13 (1986). [34] B. He and J. A. Bergou, Phys. Rev. A 77, 053818 (2008). [5] B. Yurke and D. Stoler, Phys. Rev. A 35, 4846 (1987). [35] L. Davidovich, A. Maali, M. Brune, J. M. Raimond, and [6] U. M. Titulaer and R. J. Glauber, Phys. Rev. 145, 1041 S. Haroche, Phys. Rev. Lett. 71, 2360 (1993). (1965). [36] P. Grangier, M. J. Potasek, and B. Yurke, Phys. Rev. [7] I. Bialynicki-Birula, Phys. Rev. 173, 1207 (1968). A 38, 3132 (1988). [8] A. O. Barut and L. Girardello, Commun. Math. Phys. [37] A. Mann, B. C. Sanders, and W. J. Munro, Phys. Rev. 21, 41 (1971). A 51, 989 (1995). [9] A. M. Perelomov, Commun. Math. Phys. 26, 222 [38] C.-l. Chai, Phys. Rev. A 46, 7187 (1992). (1972). [39] V. V. Dodonov, I. A. Malkin, and V. I. Man’ko, Physica [10] A. M. Perelomov, Generalized Coherent States and 72, 597 (1974). Their Applications (Springer, Berlin, 1986). [40] V. V. Dodonov, V. I. Man’ko, and D. E. Nikonov, Phys. [11] D. Stoler, Phys. Rev. D 4, 2309 (1971). Rev. A 51, 3328 (1995). [12] R. Gilmore, Ann. Phys. (N.Y.) 74, 391 (1972). [41] V. Spiridonov, Phys. Rev. A 52, 1909 (1995). [13] F. T. Arecchi, E. Courtens, R. Gilmore, and H. Thomas, [42] V. Spiridonov, Phys. Rev. A 53, 2903 (1996). Phys. Rev. A 6, 2211 (1972). [43] N. A. Ansari, L. D. Fiore, M. A. Man’ko, V. I. Man’ko, [14] V. Buˇzek and P. L. Knight, Progress in Optics 34, 1 S. Solimeno, and F. Zaccaria, Phys. Rev. A 49, 2151 (1995). (1994). [15] E. Schrödinger, Naturwissenschaften 23, 812 (1935). [44] N. A. Ansari and V. I. Man’ko, Phys. Rev. A 50, 1942 [16] M. Brune, E. Hagley, J. Dreyer, X. Maitre, A. Maali, (1994). C. Wunderlich, J. M. Raimond, and S. Haroche, Phys. [45] W. J. Munro, G. J. Milburn, and B. C. Sanders, Phys. Rev. Lett. 77, 4887 (1996). Rev. A 62, 052108 (2001). [17] A. Ourjoumtsev, R. Tualle-Brouri, J. Laurat, and [46] B. Wielinga and B. C. Sanders, J. Mod. Opt. 40, 1923 P. Grangier, Science 312, 83 (2006). (1993). [18] Y. Aharonov and L. Susskind, Phys. Rev. 155, 1428 [47] I. Jex, P. Törmä, and S. Stenholm, J. Mod. Opt. 42, (1967). 1377 (1995). [19] B. C. Sanders, S. D. Bartlett, T. Rudolph, and P. L. [48] S.-B. Zheng, Quant. Semiclass. Opt. B: J. European Knight, Phys. Rev. A 68, 042329 (2003). Opt. Soc. B 10, 691 (1998). [20] H. P. Yuen and J. H. Shapiro, IEEE Trans. Inf. Theory [49] X. Wang and X. Sanders, Phys. Rev. A 65, 012303 26, 78 (1980). (2001). [21] H. P. Yuen and V. W. S. Chan, Opt. Lett. 8, 177 (1983). [50] H. Jeong and N. B. An, Phys. Rev. A 74, 022104 (2006). [22] T. Tyc and B. C. Sanders, J. Phys. A: Math. Gen. 37, [51] H.-M. Li, H.-C. Yuan, and H.-Y. Fan, Int. J. Theor. 7341 (2004). Phys. 48, 2849 (2009). [23] B. C. Sanders, K. S. Lee, and M. S. Kim, Phys. Rev. A [52] P. P. Munhoz, F. L. Semião, and Vidiello, Phys. Lett. 52, 735 (1995). A 372, 3580 (2008). [24] G. S. Agarwal, Phys. Rev. Lett. 57, 827 (1986). [53] W.-F. Wang, X.-Y. Sun, and X.-B. Luo, Chin. Phys. [25] G. S. Agarwal, J. Opt. Soc. Am. B 5, 1940 (1988). Lett. 25, 839 (2008). [26] G. S. Agarwal and K. Tara, Phys. Rev. A 43, 492 (1991). [54] E. M. Becerra-Castro, W. B. Cardoso, A. T. Avelar, and [27] A. Mecozzi and P. Tombesi, Phys. Rev. Lett. 58, 1055 B. Baseia, J. Phys. B: At. Mol. Opt. Phys. 41, 085505 (1987). (2008). [28] P. Tombesi and A. Mecozzi, J. Opt. Soc. Am. B 4, 1700 [55] S. J. van Enk and O. Hirota, Phys. Rev. A 64, 022313 (1987). (2001). [29] B. C. Sanders and D. A. Rice, Opt. Quant. Electronics [56] S. J. van Enk, Phys. Rev. Lett. 91, 017902 (2003). 31, 525 (1999). [57] C. H. Bennett, G. Brassard, C. Crépeau, R. Jozsa, [30] B. C. Sanders and D. A. Rice, Phys. Rev. A 61, 013805 A. Peres, and W. K. Wootters, Phys. Rev. Lett. 70, (2000). 1895 (1993). 14

[58] T. J. Johnson, S. D. Bartlett, and B. C. Sanders, Phys. Brouri, and P. Grangier, J. Opt. Soc. Am B 27, A137 Rev. A 66, 042326 (2002). (2010). [59] X. Wang, Phys. Rev. A 64, 022302 (2001). [97] A. Ourjoumtsev, A. Dantan, R. Tualle-Brouri, and [60] J. Janszky, A. Gabris, M. Koniorczyk, A. Vukics, and P. Grangier, Phys. Rev. Lett. 98, 030502 (2007). J. K. Asbóth, J. Opt. B: Quantum Semicl. Opt. 4, S213 [98] P. Weinberger, Phil. Mag. Lett. 88, 897 (2008). (2002). [99] C. C. Gerry, Phys. Rev. A 59, 4095 (1999). [61] H. Jeong, M. S. Kim, and J. Lee, Phys. Rev. A 64, [100] D. A. Rice and B. C. Sanders, Quant. Semiclassical 052308 (2001). Opt.: J. European Opt. Soc. B 10, L41 (1998). [62] X. Wang, B. C. Sanders, and S.-h. Pan, J. Phys. A: [101] J. K. Asbóth, P. Adám,´ M. Koniorczyk, and J. Janszky, Math. Gen. 33, 7451 (2000). Eur. Phys. J. D 30, 403 (2004). [63] C. C. Gerry, A. Benmoussa, E. E. Hach, and J. Albert, [102] A. Luis, Phys. Rev. A 65, 034102 (2002). Phys. Rev. A 79, 022111 (2009). [103] X. Wang, J. Phys. A: Math. Gen. 35, 165 (2002). [64] N. Aizawa, R. Chakrabarti, and J. Segar, J. Phys. A: [104] M. Paternostro, M. S. Kim, and B. S. Ham, Phys. Rev. Math. Gen. 38, 9007 (2005). A 67, 023811 (2003). [65] R. Loudon and P. L. Knight, J. Mod. Opt. 34, 709 [105] Y. Guo and L.-M. Kuang, Chin. Phys. Lett. 22, 595 (1987). (2005). [66] C.-W. Lee and H. Jeong, Phys. Rev. A 80, 052105 [106] J. Lee, M. Paternostro, C. Ogden, Y. W. Cheong, (2009). S. Bose, and M. S. Kim, New J. Phys. 8, 23 (2006). [67] M. X. Tao, H. Lu, and W. L. She, Acta Physica Sinica [107] C. C. Gerry and R. Grobe, Phys. Rev. A 75, 034303 51, 1996 (2002). (2007). [68] J.-S. Zhang and J.-B. Xu, Physica Scripta 79, 025008 [108] N. B. An, K. Kim, and J. Kim, Quant. Inf. Commun. (2009). 11, 124 (2011). [69] L. Xu and L.-M. Kuang, J. Phys. A: Math. Gen. 39, [109] D.-G. He, Y.-L. Dong, J.-X. Fang, L. Qian, and J. Hu, L191 (2006). Comm. Theor. Phys. (Beijing, China) 55, 426 (2011). [70] Z.-Z. Ren, R. Jing, and X.-Z. Zhang, Chin. Phys. Lett. [110] K.-H. Song, W.-J. Zang, and G.-C. Guo, Eur. Phys. J. 25, 3562 (2008). D 19, 267 (2002). [71] A. Ourjoumtsev, F. Ferreyrol, R. Tualle-Brouri, and [111] K.-H. Song, W.-J. Zang, and G.-C. Guo, Eur. Phys. J. P. Grangier, Nature Phys. 5, 189 (2009). D 20, 305 (2002). [72] R. J. Glauber, Phys. Rev. Lett. 10, 84 (1963). [112] M. S. Kim and G. S. Agarwal, Phys. Rev. A 59, 3044 [73] J. R. Klauder and B. S. Skagerstam, Coherent States (1999). (World Scientific, Singapore, 1985). [113] Q. X. Mu, Y.-H. Ma, and L. Zhou, J. Phys. B: At. Mol. [74] E. C. G. Sudarshan, Phys. Rev. Lett. 10, 277 (1963). Opt. Phys. 40, 3241 (2007). [75] L. Mandel, Opt. Lett. 4, 205 (1979). [114] E. Solano, G. S. Agarwal, and H. Walther, Phys. Rev. [76] Y. Aharonov, D. Falkoff, E. Lerner, and H. Pendleton, Lett. 90, 027903 (2003). Ann Phys. (N.Y.) 39, 498 (1966). [115] J.-H. Guo, Commun. Theor. Phys. (Beijing, China) 41, [77] K. M. lmer, Phys. Rev. A 55, 3195 (1997). 37 (2004). [78] K. M. lmer, Phys. Rev. A 58, 4247 (1998). [116] G.-C. Guo and S. B. Zheng, Opt. Comm. 133, 142 [79] J. Gea-Banacloche, Phys. Rev. A 58, 4244 (1998). (1997). [80] T. Rudolph and B. C. Sanders, Phys. Rev. Lett. 87, [117] S. B. Zheng and G.-C. Guo, Opt. Comm. 137, 308 077903 (2001). (1997). [81] H. M. Wiseman, J. Opt. B: Quantum Semiclass. Opt. [118] X. Wang, M. Feng, and B. C. Sanders, Phys. Rev. A 6, S849 (2004). 67, 022302 (2003). [82] J. A. Smolin (2004), quant-ph/0407009. [119] S. Bose and G. S. Agarwal, New J. Phys. 8, 34 (2006). [83] S. J. van Enk and C. A. Fuchs, Phys. Rev. Lett. 88, [120] P. Milman, A. Auffeves, F. Yamaguchi, M. Brune, J. M. 027902 (2002). Raimond, and S. Haroche, Eur. Phys. J. D 32, 233 [84] R. J. Glauber, Phys. Rev. 130, 2529 (1963). (2005). [85] K. Husimi, Proc. Phys.-Math. Soc. Japan 22, 264 [121] S. Haroche, M. Brune, and J. M. Raimond, J. Mod. (1940). Opt. 54, 2101 (2007). [86] K. E. Cahill and R. J. Glauber, Phys. Rev. 177 (1969). [122] M.-f. Chen and S.-s. Ma, Acta Physica Sinica (2007). [87] V. Buˇzek and P. L. Knight, Opt. Comm. 81, 331 (1991). [123] L. Tang, J. Opt. B: Quant. Semicl. Opt. 42, 085502 [88] D. J. Daniel and G. J. Milburn, Phys. Rev. A 39, 4628 (2009). (1989). [124] H. Wei and F. Wang, J. Phys. B: At. Mol. Opt. Phys. [89] J. Gea-Banacloche, Phys. Rev. Lett. 65, 3385 (1990). 43, 208001 (2010). [90] K. Nemoto and B. C. Sanders, J. Phys. A: Math. Gen. [125] L. Tang, J. Phys. B: At. Mol. Opt. Phys. 43, 208002 34, 2051 (2001). (2010). [91] B. Yurke, S. L. McCall, and J. R. Klauder, Phys. Rev. [126] C. C. Gerry, Phys. Rev. A 55, 2478 (1997). A 33, 4033 (1986). [127] L.-H. Lin, L.-Z. Wang, and Y.-K. Jiang, Comm. Theor. [92] L. Davidovich, M. Brune, J. M. Raimond, and Phys. (Beijing, China) 40, 207 (2003). S. Haroche, Phys. Rev. A 53, 1295 (1996). [128] S.-B. Zheng, J. Opt. B: Quantum Semiclass. Opt. 3, 333 [93] S. J. van Enk, Phys. Rev. A 72, 022308 (2005). (2001). [94] C. T. Lee, J. Opt. Soc. Am B 15, 1187 (1998). [129] E. Solano, R. L. de Matos, and N. Zagury, J. Opt. B: [95] C. C. Gerry and R. Grobe, Phys. Rev. A 51, 1698 Quant. Semicl. Opt. 4, S324 (2002). (1995). [130] S.-B. Zheng, Commun. Theor. Phys. (Beijing, China) [96] N. Sangouard, C. Simon, N. Gisin, J. Laurat, R. Tualle- 15

43, 523 (2005). Rev. A 75, 044305 (2007). [131] X. Zou, K. Pahlke, and W. Mathis, Phys. Rev. A 65, [153] J.-Q. Liao and L.-M. Kuang, Chin. Phys. 15, 2246 064303 (2002). (2006). [132] J.-Q. Liao, Y. Guo, H.-S. Zeng, and L.-M. Kuang, J. [154] Y. Sun, Z.-X. Man, and Y.-J. Xia, Chin. Phys. Lett. 26, Phys. B: At. Mol. Opt. Phys. 39, 4709 (2006). 020306 (2009). [133] S. Bose, K. Jacobs, and P. L. Knight, Phys. Rev. A 56, [155] J.-Q. Liao and L.-M. Kuang, J. Phys. B: At. Mol. Opt. 4175 (1997). Phys. 40, 1183 (2007). [134] S.-B. Zheng, Aust. J. Phys. 53, 793 (2000). [156] H. N. Phien and N. B. An, Phys. Lett. A 372, 2825 [135] D. Jaksch, S. A. Gardiner, K. Schulze, J. I. Cirac, and (2008). P. Zoller, Phys. Rev. Lett. 86, 4733 (2001). [157] M. K. Mishra and H. Prakash, J. Phys. B: At. Mol. Opt. [136] L.-M. Kuang, A.-H. Zeng, and Z.-H. Kuang, Chin. Phys. Phys. 43, 185501 (2010). Lett. 21, 999 (2004). [158] N. B. An, Phys. Rev. A 69, 022315 (2004). [137] L.-M. Kuang, Z.-B. Chen, and J.-W. Pan, Phys. Rev. A [159] Y. Guo and L.-M. Kuang, J. Phys. B: At. Mol. Opt. 76, 052324 (2009). Phys. 40, 3309 (2007). [138] B.-L. Zhang and D.-Y. Gao, J. At. Mol. Sci. 2, 155 [160] Y. Guo and L.-M. Kuang, Chin. Opt. Lett. 6, 303 (2011). (2008). [139] C. C. Gerry, J. Mimih, and A. Benmoussa, Phys. Rev. [161] N. B. An and J. Kim, Phys. Rev. A 80, 042316 (2009). A 80, 022111 (2009). [162] D. Wilson, H. Jeong, and M. S. Kim, J. Mod. Opt. 49, [140] A. Gilchrist, P. Deuar, and M. D. Reid, Phys. Rev. A 851 (2002). 60, 4259 (1999). [163] J.-H. An, M. Feng, and W.-M. Zhang, Quant. Inf. Com- [141] C.-W. Lee, M. Paternostro, and H. Jeong, Phys. Rev. mun. 9, 0317 (2009). A 83, 022102 (2011). [164] H. Jeong and M. S. Kim, Quant. Inf. Commun. 2, 208 [142] S.-B. Li and J.-B. Xu, Phys. Lett. A 309, 321 (2003). (2001). [143] W. K. Wootters and W. H. Zurek, Phys. Rev. D 19, 473 [165] J. Clausen, L. Knöll, and D.-G. Welsch, Phys. Rev. A (1979). 66, 062303 (2002). [144] B. C. Sanders and G. J. Milburn, Phys. Rev. A 39, 694 [166] K. Park and H. Jeong, Phys. Rev. A 82, 062325 (2010). (1989). [167] P. van Loock, N. Lütkenhaus, W. J. Munro, and [145] D. A. Rice, G. Jaeger, and B. C. Sanders, Phys. Rev. A K. Nemoto, Phys. Rev. A 78, 062319 (2008). 62, 012101 (2001). [168] A. El Allati, Y. Hassouni, and N. Metwally, Phys. Scr. [146] V. V. Dodonov, A. S. M. de Castro, and S. S. Mizrahi, 83, 065002 (2011). Phys. Lett. A 296, 73 (2002). [169] P. T. Cochrane, G. J. Milburn, and W. J. Munro, Phys. [147] N. B. An, Phys. Rev. A 68, 022321 (2003). Rev. A 59, 2631 (1999). [148] H. Jeong and M. S. Kim, Phys. Rev. A 65, 042305 [170] M. C. de Oliveira and W. J. Munro, Phys. Rev. A 61, (2002). 042309 (2000). [149] J. Janszky, J. K. Asbóth, A. Gábris, A. Vukics, M. Ko- [171] J. Joo, W. J. Munro, and T. P. Spiller, Phys. Rev. Lett. niorczyk, and T. Kobayashi, Fortschritte Phys. - Prog. 107, 083601 (2011). Phys. 51, 157 (2003). [172] J. Joo, W. J. Munro, and T. P. Spiller, Phys. Rev. Lett. [150] P. P. Munhoz, J. A. Roversi, A. Vidiella-Barranco, and 107, 219902 (2011). F. L. Semião, Phys. Rev. A 81, 042305 (2010). [173] H. Hirota, K. Kato, and D. Murakami, Effective- [151] X.-H. Cai, J.-R. Guo, J.-J. Nie, and J. Jin-Ping, Chin. ness of entangled coherent state in quantum metrology, Phys. 15, 488 (2006). arXiv:1108.1517 (2011). [152] H. Prakash, N. Chandra, R. Prakash, and Shivani, Phys.