Arxiv:1801.05704V1 [Quant-Ph] 16 Jan 2018 Coherent State in the Number State Basis Is and Sub-Poissonian Statistics for Suitable Values of the Amplitude

Number state ﬁltered coherent state

Nilakantha Meher∗ and S. Sivakumar† Materials Physics Division, Indira Gandhi Centre For Atomic Research, Homi Bhabha National Institute, Kalpakkam, India-603102

Number state filtering in coherent states leads to sub-Poissonian photon statistics. These states are more suitable for phase estimation when compared with the coherent states. Nonclassicality of these states is quantified in terms of the negativity of the Wigner function and the entanglement potential. Filtering of the vacuum from a coherent state is almost like the photon-addition. However, filtering makes the state more resilient against dissipation than photon-addition. Vacuum state filtered coherent states perform better than the photon-added coherent states for a two-way quantum key distribution protocol. A scheme to generate these states in multi-photon atom-field interaction is presented.

I. INTRODUCTION states saturate the uncertainties in the field quadratures to their minimum values allowed by quantum mechanics. Dynamics of the classical electromagnetic field is de- This, in turn, implies that the phase space distribution of scribed by the Maxwell equations. All the features de- a coherent state is well localized. In the context of a har- scribable within this formalism are classified as the clas- monic oscillator, the coherent state wave packet oscillates sical aspects of the field. However, this description is without dispersion. An alternate definition identifies the inadequate as there are experimental observations such coherent state as a displaced vacuum state [1]. Interest- as the anti-bunching of light, quadrature squeezing, sub- ingly, one of the quasi-probability distributions, namely, Poissonian photon statistics, etc., which cannot be under- the Glauber-Sudarshan P-function for the coherent state stood within the classical description [1]. These observa- is a Dirac delta-function, which is a well defined prob- tions are explainable in the quantum description of the ability density. For this reason these states are called electromagnetic field. In this formulation, the most fun- quasi-classical states. Any other choice of coefficients damental states of the electromagnetic field are the num- makes the resultant state non-classical whose P-function ber states, which correspond to states of definite num- is more singular than a delta function. This signals the ber of photons. It is of interest to know those quantum nonclassical aspect of the quantum state [2, 3]. states of the field which have features similar to the classical field. It turns out that the coherent state, which is a Nonclassicality of quantum states is of importance superposition of all the number states with specific super- in quantum optics [5, 6]. The notion of nonclassicality position coefficients, has that property [2, 3]. This spe- is based on the quasi-probability distributions for the cific choice of the superposition coefficients makes the co- quantum states. Negative value of P-function or Wigner herent state very different from these number states. For function is sufficient to identify a quantum state as non- instance, the coherent states are the eigenstates of the an- classical. Many interesting nonclassical states have been nihilation operatora ˆ, the corresponding complex eigen- defined by suitable modifications of the quasi-classical value α is the coherent state amplitude. These states coherent states; for instance, photon-added coherent form an over-complete basis set so that any state of the states [7], even/odd coherent states [8], truncated electromagnetic field is expressible as a continuous su- coherent states [9, 10], etc. The even and odd coherent perposition. As a consequence, the expectation values of states are defined as the symmetric and anti-symmetric an operator-valued function of the annihilation operator superposition of two coherent states whose complex is the statistical average of the corresponding function amplitudes are of equal magnitude but out of phase of the coherent state amplitude. The expression for the by π [8]. These states exhibit quadrature squeezing arXiv:1801.05704v1 [quant-ph] 16 Jan 2018 coherent state in the number state basis is and sub-Poissonian statistics for suitable values of the amplitude. Agarwal and Tara introduced a new class ∞ n 2 X α |αi = e−|α| /2 √ |ni . (1) of nonclassical states by adding quanta to the coherent state. This states are called photon added coherent state n=0 n! (PACS) [7]. For proper choice of parameters, PACS Correlation functions of any order are factorizable for shows quadrature squeezing and sub-Poissonian photon the coherent states [4]. For large amplitude coherent statistics. Truncated coherent states are obtained by state, the photon number fluctuation is much smaller modifying the expression in Eqn. 1 by removing either than the mean number of photon. Hence large ampli- a finite number of contiguous states starting from the tude coherent states represent the classical field. These vacuum or by removing all the states beyond a specified number state [9, 10]. Another important class of nonclassical states is obtained by number state hole burning ∗ [email protected] or filtering in the coherent states [11]. It is known that † [email protected] any pure state in which the vacuum state |0i is absent 2 is nonclassical [12, 13]. A possible generalization is to NSFS is expressible as consider states in which a number state, not necessarily the vacuum, is absent. More generally, states with 1 1 |ψ(α, m)i = (|αi − Cm |mi) = (I − |mi hm|) |αi , many of the number states not being present in the Nm Nm superposition have been studied to assess their possible (5) use in practical applications such as optical data storage 2 m a superposition of |αi and |mi, where C = e−|α| /2 √α and optical communication [11, 14]. It may be noted m m! that schemes to generate such states are known[15–24]. [25]. The overlap between the states |ψ(α, m)i and |αi is

2 2 In this paper, we study the non classical properties of |hα|ψ(α, m)i| = 1 − |Cm| , (6) number state filtered coherent state. The photon count- ing statistics of these states is discussed in Section II. which is minimum if |α|2 = m. Therefore, maximum Sub-Poissonian statistics is shown to improve phase es- nonclassicality of |ψ(α, m)i is expected if |α|2 = m. timation in comparison to the coherent states. In Sec- tion III, quantification of nonclassicality of these states For a given α, the state |ψ(α, m)i corresponding to is carried out. Entanglement potential and negativity different m are linearly independent. Hence the set ∞ of Wigner function are studied in this context. In Sec- {|ψ(α, m)i}m=0 can be used as a basis. Further, for a tion IV, the possibility of generating these states in the given m interaction of a 3-level atom with a cavity field is dis- Z 1 2 2 cussed. In Section V, features of vacuum state filtered Nm |ψ(α, m)i hψ(α, m)| d α = I − |mi hm| , (7) coherent state are analyzed. Specifically, the robustness π of the state against dissipation is exhibited. This feature which is the completeness relation for the Hilbert space enables these states to perform better than the single less the number state |mi of a harmonic oscillator. photon added coherent state (SPACS) when it comes to It is interesting to note that the even and odd coherent two-way quantum key distribution. Use of these states states are expressible as as a source to generate the even and odd coherent states is discussed in Section VI. s 1 1 − |C |2 √ m |ECSi = 4 (|ψ(α, m)i + |ψ(−α, m)i), 2 1 + |C0| II. NUMBER STATE FILTERING IN (for odd m) (8) COHERENT STATE and Number state filtered state (NSFS), denoted by s 1 1 − |C |2 |ψ(α, m)i, is defined as √ m |OCSi = 4 (|ψ(α, m)i − |ψ(−α, m)i), 2 1 − |C0| 2 ∞ e−|α| /2 X αn |ψ(α, m)i = √ |ni , (2) (for even m). (9) Nm n=0,n6=m n!

A. Photon Statistics where Nm is the normalization constant given by r |α|2m Photon statistics is an experimentally measurable as- N = 1 − e−|α|2 . (3) m m! pect of the electromagnetic field. Probability of detecting k photons in a measurement is |hk|ψ(α, m)i|2, where |ki This definition implies that the state is obtained if the represents the number state with k photons. Using the number state |mi is absent in the superposition defined number state expression given in Eqn. 2, in Eqn. 1. 2 2 e−|α| αk αm P = |hk|ψ(α, m)i|2 = √ − √ δ , (10) The state |ψ(α, m)i does not become |αi for any m. k 2 k,m Nm k! m! However, if m >> |α| or vice-versa, then |ψ(α, m)i ≈ |αi. It is worth noting that the coherent state is a superposi- which gives the average number of photons in the state tion of NSFSs, as 2m 2 Nm † 1 2 −|α| |α| |αi = (|ψ(α, m)i + |ψ(−α, m)i) hnî = ha ai = 2 |α| − me . (11) 2 Nm m! Nk + (|ψ(α, k)i − |ψ(−α, k)i), (4) For |α| >> m or |α| << m 2 2 where m is odd and k is even. hnî ≈ |α| , (12) 3

˜ −2 |α|2 2 as expected since |ψ(α, m)i ≈ |αi under these limits. respectively. Here Nu = e (1 − γ(N + 1, |α| )/N!) Surprisingly, if |α|2 = m, ˜ −2 |α|2 2 and Nl = e γ(N + 1, |α| )/N! with γ(N, x) = h N−1 j i 2 −x P x hnî = |α| , (13) (N − 1)! 1 − e j=0 j! , the incomplete Gamma function. Note that, the states UTCS and LTCS are same as the average number of photons in the coherent orthogonal to each other. state |αi even though overlap between |ψ(α, m)i and |αi is minimum. Nevertheless, the nonclassical nature of the An interesting aspect of these two states is that they photon distribution can be identified by the Mandel Q- are sub-Poissonian for any α. As |α| increases to values parameter much larger than unity, the photon number distribution of LTCS becomes Poissonian, that is, Q tends to zero. ha†2a2i − ha†ai2 Q = . (14) In the same limit, the Q parameter of UTCS becomes ha†ai -1, that is, sub-Poissonian. If Q = 0, the state is Poissonian which corresponds to a The state |ψ(α, m)i is expressible as a linear combina- coherent state. If Q < 0, the state is sub-Poissonian tion of LTCS and UTCS: which is a nonclassical feature. The state is super- Poissonian if Q > 0. For the state |ψ(α, m)i, |ψ(α, m)i = Cu |α, m − 1iu + Cl |α, mil , (18) 2 e−|α| |α|2m where C = hα, m − 1| ψ(α, m)i and Q = 2m|α|2 − m(m − 1) u u 2 C = hα, m| ψ(α, m)i. m!Nmhnî l l 2m 2 |α| −|α|4 − me−|α| , m! 1 (15) 2 | l

where hnˆi is given in Eqn. 11. The dependence of Q on , |C 0.5 2 |

0.2 0.2 1 (a) (b)

0 0 2 | l

−0.2 −0.2 , |C 0.5 2 | u

Q −0.4 Q −0.4 |C

−0.6 −0.6 0 0 1 2 3 4 5 6 7 8 −0.8 −0.8 |α|

−1 −1 2 2 2 4 6 8 0 2 4 6 8 FIG. 2. |Cu| (continuous) and |Cl| (dashed) are plotted as |α| |α| a function of |α| for (a)m = 9 and (b)m = 25. FIG. 1. The Q parameter as a function of |α| for the state Though the states involved in the superposition are |ψ(α, m)i (continuous), |α, m − 1iu (dashed) and |α, mil (dot- dashed) with (a)m = 9 and (b)m = 25. sub-Poissonian, the resultant state is super-Poissonian for a range of values of α as shown in Fig. 1. A key photon statistics of |ψ(α, m)i, consider expressing it as a observation is that the state |ψ(α, m)i exhibits maxi- linear superposition of a lower truncated coherent state mum super-Poissonian character when UTCS and LTCS (LTCS) and an upper truncated coherent state (UTCS) are equally sub-Poissonian. As α changes, |ψ(α, m)i whose number states expressions are [9, 10] exhibits sub-Poissonian feature over a small region and becomes nearly Poissonian for larger and smaller values ∞ N+1+n of |α|. This feature emerges here in a way that is anal- ˜ X α |α, Ni = Nl |N + 1 + ni , (16) ogous to the appearance of super-Poissonian statistics l p(N + 1 + n)! n=0 when two number states, which are sub-Poissonian, are and superposed. If the superposition coeﬃcients are nearly equal in magnitude, measurement of photon number N n gives results corresponding to the two number states ˜ X α |α, Ni = Nu √ |ni , (17) with nearly equal probability. That is, the measurement u n! n=0 results ﬂuctuate and the statistics is super-Poissonian. 4

The fact that the states |ψ(α, m)i exhibit sub- p1 − |C |2 aˆm |ψ(α, m)i = αm 0 |ψ(α, 0)i , (20) Poissonian character indicates that they could be of p 2 1 − |Cm| use in phase estimation using interferometers. Coherent p states provide a phase resolution of ∆θ = 1/ |α|2| sin θ| in which the vacuum state is absent. This state is similar [1]. There is mild improvement if one uses |ψ(α, m)i in- to the SPACS |α, 1i [7]. The overlap between these two stead of the coherent states. This feature is shown in states is Fig. 3, where the minimum detectable phase is shown as |α|2 a function of m. It is clear that sensitivity is maximum | hα, 1| ψ(α, 0)i|2 = , (21) when |ψ(α, m)i exhibits minimum Q. (1 − e−|α|2 )(1 + |α|2)

which is plotted as a function of |α| in Fig. 4. It clearly indicates that for small or large |α|, the states are nearly 0.5 0.4 the same as the overlap is nearly unity. 0.49 (a) 0.3 0.48 0.2 Q ∆θ 0.47 0.1 1 0.46 0 0.45 −0.1 0.8 2

0 5 10 15 20 25 30 35 40 | m 〉

,0) 0.6

0.38 0.2 α (

(b) ψ

,1| 0.4 α

〈 | Q

∆θ 0.36 0 0.2

αm+1 aˆm+1 |ψ(α, m)i = |αi , (22) p 2 B. Subtraction of photons from NSFS 1 − |Cm| the coherent state. In short, subtraction of m+1 photons Photon-addition to a coherent state generates a non- from |ψ(α, m)i generates the coherent state. This has to classical state [7]. Like photon addition, photon subtrac- be compared with the action ofa ˆ on a number state |mi. tion is a process which may introduce nonclassicality in After m successive photon subtractions, |mi becomes the many states. However, photon subtraction leaves the co- vacuum state |0i and one more subtraction annihilates herent state unaltered as it is an eigenstate ofa ˆ. Action the state. It is to be noted that while subtraction of ofa ˆ on other states generates photon-subtracted states m+1 photons from |ψ(α, m)i generates a coherent state, [26]. One way of realizing photon subtraction experi- addition of photons to a coherent state does not generate mentally is by using beam splitters. Though |ψ(α, m)i |ψ(α, m)i, a consequence of the non-commutativity of a is necessarily nonclassical, the eﬀect of photon subtrac- and a†. tion is of interest. The resultant state on subtracting a 5

From Eqn. 22, it is seen that |ψ(α, m)i satisfies the III. NONCLASSICALITY OF NSFS eigenvalue equation Phase space representation of quantum states provides h m+1i m+1 additional insights. For instance, classical aspects of (I − Pˆm)a |ψ(α, m)i = α |ψ(α, m)i , (23) quantum states are readily obtained by such represen- tations. In quantum optics, the most used phase space where Pˆm = |mi hm|. distributions are the P -, Q- and Wigner-functions. Neg- ative values of Glauber-Sudarshan P -function is a sufficient condition for qualifying the state as non-classical. In x-representation, the wavefunction for a coherent 2 state is a Gaussian. Probability density for x in |ψ(α, m)i The condition h(∆ˆn) i < hnî requires the P -function 2 to be negative. Hence, the amplitude squeezing is a is |ψα,m(x)| , which is shown in Fig. 5. For comparison, the corresponding probability densities for coherent state non-classical effect [1]. A state with amplitude squeez- |αi of amplitude α = 2 and the number state |mi with ing possesses sub-Poissonian photon statistics. The sub- m = 4 is also shown. From the figure it is clear that, the Poissonian character of |ψ(α, m)i has been discussed in probability distribution is not a Gaussian. In fact, the the Section II. distribution is oscillatory.

A. Wigner function 0.8

In this section, we discuss the nonclassicality of 0.6 |ψ(α, m)i in terms of its Wigner function. Wigner function, though not singular for any state, can become neg- 0.4 ative in some regions. The negativity is an indicator of P(x) quantumness [27]. Wigner function for a state |φi is,

Considering the time evolution of |ψ(α, m)i under U = † e−iωta a, the evolved state is

On ﬁltering the number state |mi from the coherent state |αi, the resultant state has maximum deviation from |αi if |α|2 ≈ m, which is the condition for max- FIG. 6. Wigner function for the states |ψ(2, 5)i (left) and imum nonclassicality of the NSFS. On the other hand, |ψ(3, 9)i (right) showing negative regions. NSFS exhibits maximum super-Poissonian photon statistics under the same condition. In fact, a state with super- For the state |ψ(α, m)i, Poissonian photon statistics does not refer to the nonclassicality of the state. In order to clarify the condition for ∞ 2 X ∗ 2 maximum nonclassicality of NSFS, we study the negativ- W (β) = (−1)n e−iIm(βα )hα − β|ni − C∗ D (β)t . π m mn ity of Wigner function and entanglement potential in the n=0 next section. (26) 6

Here Wigner functions for |ψ(α, m)i with α = 2, m = 5 and α = 3, m = 9 are shown in Fig. 6. Both the states r n! 2 are nonclassical as their respective Wigner functions are D (β) = e−|β| /2βm−nL(m−n)(|β|2), m > n mn m! n negative in some regions of phase space. Interestingly, (27) the number of peaks equals m + 1. To quantify the r nonclassicality, the volume under negative portion of the m! −|β|2/2 ∗ m−n (n−m) 2 Dmn(β) = e (−β ) Ln (|β| ), m ≤ n Wigner function is shown in Fig. 7. It is readily inferred n! from the ﬁgure that the negativity is maximum when (28) |α|2 ≈ m. This is consistent with the result that the

(k) overlap of |ψ(α, m)i with the classical state |αi reaches where Ln is the associated Laguerre polynomial. its minimum when |α|2 = m.

0.12 B. Beam splitter transformation 0.1 Another way to quantify nonclassicality is the entan- 0.08 glement at output of a beam splitter. If one of the input 0.06 states is a coherent state, it requires a nonclassical state at the second port to generate entanglement at the out- Negativity 0.04 put [29]. The amount of entanglement is a measure of 0.02 the nonclassicality of the input state to the second port. The unitary transformation eﬀected by a 50 : 50 beam 0 splitter is −4 −3 −2 −1 0 1 2 3 4 α " √1 √i # U = 2 2 . (29) FIG. 7. Negativity of Wigner function as a function of α for √i √1 2 2 |ψ(α, m)i corresponds to m = 2 (dot-dashed),4 (dashed) and 9(continuous). The input bipartite state |ψ(α, m)i |βi transforms to

" ∞ r ∞ ˜s 1 X 2 α˜ X ˜ 2 β |ψ i = U |ψ(α, m)i |βi = e−|α˜| /2 √ e−|β| /2 √ |r, si AB N m r=0 r! s=0 s!  ∞ n m n r 0 0 X 2 β X X (q + q )!(m + n − q − q )! −C e−|β| /2 √ m  m!n! n=0 n! q q0 0 0 !# 1 n+q−q i m−q+q × √ √ |q + q0, m + n − q − q0i , (30) 2 2

√ √ whereα ˜ = (α + iβ)/ 2 and β˜ = (iα + β)/ 2. Here |βi values of m with α = β = 3. It is clear from the figure is a coherent state. that the state at the output ports of the beam splitter is entangled, which indicates that the input state is non- Entanglement in the output of the beam splitter is classical. For large m, |ψ(α, m)i ≈ |αi and hence the quantified by the linear entropy of one of the reduced entanglement of the output state is zero. Beam splitter density operators of the bipartite output. It is defined as generates maximally entangled state if |α|2 = m for a [30], given α and independent of β. This is consistent with the observation made earlier in Section III that the non- 2 LA(|ψABi) = 1 − T r(ρA), (31) classicality of |ψ(α, m)i is expected to be maximum if |α|2 = m. where ρA = T rB(|ψABi hψAB|). Here A and B are refer to the two output ports. For pure bipartite states, linear entropy is a good measure of entanglement. Non-zero values of LA imply entanglement in |ψABi. Fig. 8 shows the entanglement in the output state |ψABi for various 7

state at time t is 0.25 1 √ √ imφ |ψ(t1)i = cos( m!gt1) |0, fi − ie sin( m!gt1) |m, gi . 0.2 (33)

The atomic transition |ei → |fi comes in to the picture 0.15 when the atom is driven by an external laser with Rabi A

L frequency . The corresponding Hamiltonian is 0.1 1 H = ω σ + ω a†a + g(σ a + σ a†) 2 2~ fe z ~ c ~ + − 0.05 −iωdt iωdt + ~(e σ− + e σ+), (34)

† † 2 IV. GENERATION OF NSFS Heff = ~χ[|fi hf| + (a a + λ(a + a) + λ )σz], (35)

In this section, a scheme to generate |ψ(α, m)i is where χ = g/(ωfe − ωc) and λ = /g. If the state of presented. This process involves driving a system of the atom is |gi, then the cavity field is unaffected. If an atom and a cavity by two consecutive Hamiltonian the atomic state is |ei or |fi, the cavity evolution is evolutions. First, a multi-photon interaction drives the non-trivial. Cavity driven is conditional depending on atom-cavity system. The resultant state is modified by the atomic state. driving the atomic system with a classical field. Details of the processes are indicated here. Under the evolution governed by Heff , |ψ(t1)i evolves to

V. NSFS FOR QUANTUM INFORMATION PROCESSING FIG. 9. Ladder type 3-level atom with energy levels |fi , |ei and |gi are labeled in decreasing order of their energies. The elementary unit for quantum computation and information processing is a qubit. Any two orthogonal atom-cavity is |0, fi, where the cavity ﬁeld is in its vac- quantum states of a system form a qubit. In the context uum and the atom is in the state |fi. The time evolved of cavities, the vacuum state and the single photon 8

state form a qubit. Various quantum protocols such as 1 quantum teleportation [33, 34], quantum state transfer

[35], entanglement generation[36], etc. have been imple- 0.8 mented in this system. SPACS has also been suggested for implementing various quantum information protocols 0.6 [37, 38]. Performance of any quantum information

processing system is aﬀected by external factors such Fidelity 0.4 as dissipation, decoherence, etc. It has always been of interest to use qubits which are resilient against such 0.2 processes.

0 0 1 2 3 4 5 6 7 One way of assessing the robustness of the qubit evolv- t ing under a dissipative channel, is to calculate the overlap between its initial and final states. We consider four FIG. 10. Fidelities F|ψ(α,0)i(continuous), F|αi(dashed), F|α,1i families of qubits using cavity field. One of the states of (dot-dashed) and F|1i(dotted) as a function of time. All curve the qubit is the vacuum state. The other state can be corresponds to |α| = 1. a single photon state |1i, a coherent state |αi, a SPACS |α, 1i or a NSFS |ψ(α, 0)i. Of these choices, the one with |αi is not ideal qubit as the states |0i and |αi are not in the presence of a third party E who eavesdrops with orthogonal. A phenomenological model of damping in a beam splitter. In order to quantify the information a quantum channel is considered to study the effect of gathered by E, the respective probabilities of detecting dissipation on these families of qubits [1]. The fidelity, at least one photon by B and E are calculated. If |0i and which is the overlap of evolved state and the initial state, SPACS are used, then these probabilities are is −|α|2 cos2 θ e 2 2 4 2 PB(n ≥ 1) = 1 − sin θ + |α| sin θ , (40) F|φi = |hφ(0)|φ(t)i| . (38) 1 + |α|2

† where |φ(t)i ∝ e−γa at/2 |φ(0)i. The amplitude damping and rate is γ. The respective fidelities for the states men- 2 2 e−|α| sin θ tioned above are P (n ≥ 1) = 1 − cos2 θ + |α|2 cos4 θ , E 1 + |α|2 −γt F|1i = e , (41) 2 −|α−αd| F|αi = e , where θ is an angular parameter that characterize 2 1 2 1 2 2 the beam splitter. It is clear that for small θ and α, − 2 |α−αd| − 2 (|α| −|αd| ) e − e information loss is less. F|ψ(α,0)i = 2 2 , (1 − e−|α| )(1 − e−|αd| ) 1 2 1 2 2 2 2 2 The state |ψ(α, 0)i is more robust against dissipative 2 |α−α˜d| 2 (|α| −|α˜d| ) −|α−αd| −(|α| −|αd| ) F|α,1i = e e e e losses in a quantum channel. This may endow |ψ(α, 0)i as ∗ 2 (1 + α αd) a better choice than SPACS. The respective probabilities × 2 ∗ , (39) (1 + |α| )(1 + α α˜d) of detecting at least one photon by B and E when |0i and |ψ(α, 0)i are the input to the beam splitter are −γt/2 −γt where αd = αe andα ˜d = αe . The suffix 2 2 2 indicates the state whose fidelity is being studied. e−|α| cos θ − e−|α| P˜ (n ≥ 1) = 1 − , (42) B 1 − e−|α|2 Fidelities for these various quantum states are shown in Fig. 10, assuming γ = 0.01. It is to be noted that the and 2 2 2 state |ψ(α, 0)i is robust against damping in a quantum e−|α| sin θ − e−|α| P˜ (n ≥ 1) = 1 − . (43) channel as its fidelity decays slower than the other states. E 1 − e−|α|2 This makes the state suitable for quantum information processing. In order to compare the advantage of |ψ(α, 0)i over SPACS, the ratios of detection probabilities are consid- Secure communication protocols such as BB84 are be- ered. These are defined as ing carried out using the coherent states. The drawback P˜ (n ≥ 1) in using the coherent states is that the eavesdropper can R = E , (44) |ψ(α,0)i ˜ extract some photons, which may lead to loss of informa- PB(n ≥ 1) tion. In order to prevent photon loss, Mundarain et. al and have shown that SPACS can be used for two-way quan- PE(n ≥ 1) tum key distribution [39]. Their protocol guarantees se- R|α,1i = . (45) cure key distribution between two parties A and B even PB(n ≥ 1) 9

These two ratios are shown in Fig. 11. The ratio R|ψ(α,0)i It is easy to recognize that if m is even and the atom is is always less than R|α,1i indicates that the diversion of detected in the state |ei, the ﬁeld state collapses to the photons by eavesdropper E via beam splitter transfor- odd coherent state. If m is odd and the atom is detected mation is less probable if one uses the state |ψ(α, 0)i in |gi, the ﬁeld state is projected on to the even coherent instead of SPACS. This shows |ψ(α, 0)i performs better state. than SPACS.

1 VII. SUMMARY 0.8 Coherent states are special as they exhibit minimum 0.6 uncertainty in the quadratures, Poissonian statistics, factorizability of coherence functions, etc. All these Ratio 0.4 features which make these states the most classical among the quantum states arise due to the specific 0.2 choice of superposition coefficients. Number state filtering from a coherent state leads to a nonclassical 0 0 1 2 3 4 5 6 7 8 state. These states are realizable in the multi-photon |α| interaction between a three-level atom and a single mode cavity field in a coherent state. The number state FIG. 11. Ratios R|ψ(α,0)i (continuous line) and R|α,1i |mi that has the maximum weightage in a coherent (dashed line) as a function of |α|. Here cos2 θ = 0.9. state |αi corresponds to |α|2 ≈ m. This is the condition for maximal nonclassicality for a number state filtered coherent state. The overlap between the coherent state and NSFS is minimum when this condition is satisfied. VI. GENERATION OF SCHRODINGER¨ CAT The negativity of the Wigner function is maximum under STATE FROM NSFS these conditions. Also, the entanglement potential of these states is highest when |α|2 ≈ m. If m deviates from Generation of the even and odd coherent states from |α|2, the resultant state is still nonclassical as shown |ψ(α, m)i is discussed. It is possible to realize the afore- by the sub-Poissonian. This property enables these mentioned states, for instance, from coherent states in states to perform slightly better than the coherent states cavities [40], mixing squeezed vacuum and vacuum in if used for phase measurements in an interferometric beam splitter [41–43], etc. The procedure for the gen- setup. Surprisingly, NSFS is super-Poissonian when eration of the odd/even coherent states from |ψ(α, m)i |α|2 ≈ m. Emergence of this super-Poissonian statistics is similar to that for the generation of these states from is understood based on the fact that NSFS is a super- the coherent state [40]. The system consists of a Ξ type position of suitably truncated coherent states which are three-level atom as described in Section IV and a cavity sub-Poissonian. The super-Poissonian statistics emerges with resonance frequency ωc. Consider the initial state in an analogous manner as in suitable superposition of two number states which are also sub-Poissonian. 1 |Ψ(0)i = √ (|ei + |gi)|ψ(α, m)i. (46) 2 Further, vacuum state filtered coherent states are re- A dispersive atom-field coupling is assumed so that the silient against the effects of dissipation than the photon- effective Hamiltonian is added coherent states. This makes NSFS more suitable † for quantum information processing. For instance, a ro- Heff = −~χa a |ei he| . (47) bust qubit can be realized using the vacuum and NSFS as the computational basis states. The vacuum state Time evolved state under Heff is filtered coherent state performs better than the single |Ψ(t)i = e−iHeff t/~ |Ψ(0)i , photon-added coherent state in two-way quantum key distribution where eavesdropping is carried out using a 1 iχt = √ (|ei |ψ(e α, m)i + |gi |α, m)i), (48) beam splitter. 2

A Ramsey√ pulse is applied that√ transforms |ei → (|ei + |gi)/ 2 and |gi → (|gi − |ei)/ 2. The evolved state at time t = π/χ is ACKNOWLEDGEMENT 1 |Ψ(π/χ)i = [|ei (|ψ(−α, m)i − |ψ(α, m)i) One of the authors (NM) acknowledges the research 2 fellowship from the Department of Atomic Energy, Gov- + |gi (|ψ(−α, m)i + |ψ(α, m)i)]. (49) ernment of India. 10

[1] C. Gerry and P. Knight, Introductory Quantum Optics, [25] B. Baseia, S. C. G. Granja, and G. C. Marques, Physica edited by Cambridge (Cambridge (England): Cambridge Scripta 55, 719 (1997). UP, 2005.). [26] H. Takahashi, K. Wakui, S. Suzuki, M. Takeoka, [2] R. J. Glauber, Phys. Rev. 131, 2766 (1963). K. Hayasaka, A. Furusawa, and M. Sasaki, Phys. Rev. [3] E. C. G. Sudarshan, Phys. Rev. Lett. 10, 277 (1963). Lett. 101, 233605 (2008). [4] R. J. Glauber, Phys. Rev. 130, 2529 (1963). [27] E. Wigner, Phys. Rev. 40, 749 (1932). [5] L. Mandel, Physica Scripta 1986, 34 (1986). [28] R. F. Bishop and A. Vourdas, Phys. Rev. A 50, 4488 [6] A. Pathak and A. Ghatak, Journal of Electro- (1994). magnetic Waves and Applications 32, 229 (2018), [29] M. S. Kim, W. Son, V. Buˇzek, and P. L. Knight, Phys. https://doi.org/10.1080/09205071.2017.1398109. Rev. A 65, 032323 (2002). [7] G. S. Agarwal and K. Tara, Phys. Rev. A 43, 492 (1991). [30] T.-C. Wei, K. Nemoto, P. M. Goldbart, P. G. Kwiat, [8] V. Dodonov, I. Malkin, and V. Man’ko, Physica 72, 597 W. J. Munro, and F. Verstraete, Phys. Rev. A 67, (1974). 022110 (2003). [9] A. Miranowicz, K. Pia¸tek, and R. Tana´s,Phys. Rev. A [31] M. S. Zubairy and J. J. Yeh, Phys. Rev. A 21, 1624 50, 3423 (1994). (1980). [10] S. Sivakumar, International Journal of Theoretical [32] C. C. Gerry, Phys. Rev. A 65, 063801 (2002). Physics 53, 1697 (2014). [33] E. Lombardi, F. Sciarrino, S. Popescu, and F. De Mar- [11] B. Baseia, M. Moussa, and V. Bagnato, Physics Letters tini, Phys. Rev. Lett. 88, 070402 (2002). A 240, 277 (1998). [34] K.-H. Song, D.-H. Lu, W.-J. Zhang, and G.-C. Guo, [12] C. T. Lee, Phys. Rev. A 52, 3374 (1995). Optics Communications 207, 219 (2002). [13] A. F. de Lima and B. Baseia, Phys. Rev. A 54, 4589 [35] N. Meher, S. Sivakumar, and P. K. Panigrahi, Scientific (1996). Reports 7, 9251 (2017). [14] B. Baseia and J. M. C. Malbouisson, Chinese Physics [36] S. M. Tan, D. F. Walls, and M. J. Collett, Phys. Rev. Letters 18, 1467 (2001). Lett. 66, 252 (1991). [15] K. Sanaka, K. J. Resch, and A. Zeilinger, Phys. Rev. [37] P. V. P. Pinheiro and R. V. Ramos, Quantum Informa- Lett. 96, 083601 (2006). tion Processing 12, 537 (2013). [16] J. Malbouisson and B. Baseia, Physics Letters A 290, [38] D. Wang, M. Li, F. Zhu, Z.-Q. Yin, W. Chen, Z.-F. Han, 234 (2001). G.-C. Guo, and Q. Wang, Phys. Rev. A 90, 062315 [17] A. Avelar and B. Baseia, Optics Communications 239, (2014). 281 (2004). [39] M. Miranda and D. Mundarain, Quantum Information [18] A. T. Avelar and B. Baseia, Phys. Rev. A 72, 025801 Processing 16, 298 (2017). (2005). [40] M. Brune, E. Hagley, J. Dreyer, X. Maˆıtre, A. Maali, [19] C. Gerry and A. Benmoussa, Physics Letters A 303, 30 C. Wunderlich, J. M. Raimond, and S. Haroche, Phys. (2002). Rev. Lett. 77, 4887 (1996). [20] N. Zagury and A. F. R. de Toledo Piza, Phys. Rev. A [41] M. Wolinsky and H. J. Carmichael, Phys. Rev. Lett. 60, 50, 2908 (1994). 1836 (1988). [21] H. Ghosh and C. C. Gerry, J. Opt. Soc. Am. B 14, 2782 [42] M. Dakna, T. Anhut, T. Opatrný,L. Knöll, and D.-G. (1997). Welsch, Phys. Rev. A 55, 3184 (1997). [22] B. M. Escher, A. T. Avelar, T. M. da Rocha Filho, and [43] S. Glancy and H. M. de Vasconcelos, J. Opt. Soc. Am. B B. Baseia, Phys. Rev. A 70, 025801 (2004). 25, 712 (2008). [23] M. Cooper, E. Slade, M. Karpiski, and B. J. Smith, New Journal of Physics 17, 033041 (2015). [24] C.-L. Zou, L. Jiang, X.-B. Zou, and G.-C. Guo, Phys. Rev. A 94, 013841 (2016).