Arxiv:Quant-Ph/0301099V1 20 Jan 2003 Lmn Uicto Saohratraiet Decoherence to Alternative the Another Entan- Is However, to Fault- Puriﬁcation Glement and [10]

Home , Amplitude amplification

arXiv:quant-ph/0301099v1 20 Jan 2003 lmn uicto saohratraiet decoherence to alternative the another entan- is However, to fault- purification glement and [10]. resort [8] [9], can [7], computation quantum codes one tolerant correction de- error errors, quantum become these than of to ideas avoid more tendency To even its errors to coherent. from due protected information be classical to informa- quantum needs correction: the tion error to quantum connection of in found problem application have important they very devised, another were protocols distillation tum been have that [5] states pure [6]. of experimentally distillation implemented meth- exist the also for there polarization Likewise, ods using [4]. protocols (PBS) of splitters type experimen- beam this feasible for are original proposals There protocol tal its protocol. distillation improving BBPSSW this the to at developing as refer aim to shall We focus the performances. the with dis- standard been protocols a has more with that us provided method The has tillation Bob. [3] and of Alice work com- parties, seminal a two of by realization the states protocol mind mixed munication in of having purification type, the bipartite of in regen- source. interested entanglement to are an we devised Here of been leakages have entanglement that erate methods, those cisely [1], therein). see references review and a [2] (for (algorithmics) or computing coding) communication quantum dense quantum entanglement cryptography, for of (teleportation, whether degree protocols need source prescribed may entanglement a we the sustain that devise to must to up we order it mean, in pump precious to entangle- a method the as some assessed for be Thus, contact to unavoidable environment. ment the external by the the degrading with has of Despite it behaviour entanglement, by odd entanglement. exhibited properties of nice sources the the requires stable mechanics of quantum availability in entanglement of ties nadto oteiiilproefrwihtequan- the which for purpose initial the to addition In pre- are protocols purification or distillation Quantum proper- intriguing the of analysis experimental The ASnmes 36.a 03.67.Lx 03.67.-a, numbers: ren PACS a quantum Grover and distillation the di quantum in between for appears relations amplification that the weight distillati amplification of the notion amplitude of the quantum solution stress We general The appendix. limit. continuum the atclr o urt egttepaedarmo h disti the of large- diagram the phase investigate We wit the obtained points. get fixed s region we bipartite stability qutrits mixed the for initial increase particular, of to family states certain qudits a to applied when esuysvrlpoete fdsilto rtcl opu to protocols distillation of properties several study We itlainPooosfrMxdSae fMliee Qubi Multilevel of States Mixed for Protocols Distillation .INTRODUCTION I. eatmnod ´sc eoiaI nvria Compluten Universidad F´ısica Teórica I, de Departamento h unu eomlzto Group Renormalization Quantum the ..MrńDlaoadM Navascués M. and Mart´ın-Delgado M.A. D ii fqdt rtcl n n naayia ouinin solution analytical an find and protocols qudits of limit hc ie oepwru a fdaigwt errors with dealing of [11]. way communication powerful quantum in more a gives which lsia omncto) hycmrs oa unitary local comprise operators on they and allowed Operations Communication): (Local operations LOCC Classical of as denoted type are qubits The these pairs qubits. sharing parties entangled separated of spatially two are Bob and xoetal ihtelnt fteqatmchannel. quantum the of scale length that the errors us with depolarization exponentially allow distances or [16] long absortion avoiding over [15], by pro- communication Other repeaters quantum stablish quantum [14]. to [13], as fidelity known desired tocols the of to name converges faster the dis- much which by new Amplification(QPA) [12] a Privacy purifica- in Quantum protocol, introduced the was BBPSSW of protocol the focus tillation After the maximally a is mea- protocol. to which is tion respect pair pair, with entangled fidelity pure its mixed entangled of a terms pairs. of in extra purity sured wasting of of entan- expense degree pairs: the entangled The at imperfectly created of is set glement larger accuracy) a certain set of a reduced (within out a pairs creates In entangled BBPSSW maximally method. the of like purification will protocol entanglement they a an difficulties, short, up these set overcome dissipa- to To have to trans- due noise. face channels and also quantum tion will their local in Bob errors op- and producing mission Alice quantum imperfections Futhermore, local from errors. These suffer results will channel. measurement erations classical the a of through communication and ments a enitoue 1] 1]bsdo generalization a on based [19] [18], proposal unitary, [17] introduced nice is very been that proposals another gate has Recently, the CNOT Hermitian. the of has not purifi- of One but it extension the an Thus, of on qudits. extensions relies for [22]. propose protocols to [21], cation natural [20], quite quan- through up been etc. flux speed cryptography, could information tum that the channels in communication increase the quite an are qubits of apparent: instead (qudits) states quantum tilevel natpclqatmcmuiaineprmn,Alice experiment, communication quantum typical a In h datgso eln with dealing of advantages The h o qain odsilqbt.In qubits. distill to equations flow the h nrcrinrltosi rsne nan in presented is relations recursion on ae.W n hti spsil ouse to possible is it that find We tates. lto rcs iharc tutr of structure rich a with process llation raiainprocesses. ormalization tlainpooosa poe othe to opposed as protocols stillation U iymliee ui tts(qudits) states qubit multilevel rify grtm ieie einvestigate we Likewise, lgorithm. A e 84.Mdi,Spain. Madrid, 28040. se, ⊗ U B nec ie oa unu measure- quantum local side, each on sand ts D dmninlo mul- or -dimensional 2 of the CNOT gate that is both unitary and Hermitian Distillation Protocol for Qubits and gives a higher convergence. In this paper we make a study of the new purification protocols of [18], [19] when 1. Set up ρ ρ ρ with fidelity F . −→ ⊗ they are applied to mixed bipartite states of qudits that 2. Apply bilateral CNOT gate: U . are not of the Werner form. In this way, we combine BCNOT some of the tools employed by the QPA protocols [12] 3. Alice and Bob measure target qubits. with the advantages of the new methods. This paper is organized as follows: in Sect. II we re- 4. Classical communication of results: retain coinci- view simple distillation protocols for qubits not in Werner dences (0A0B or 1A1B). states and we generalize them for the purification of any ρ′ F ′ > F of the Bell states. In Sect. III we extend the previous pro- 5. Go to step 1) with with fidelity . tocols to deal with multilevel qubits and obtain several The simplicity of this protocol also relies on the fact results like an improvement in the size of the stability fi- that we do not need any depolarization step, as it is the delity basin, analytical formulas for the distillation flows, case when dealing with Werner states [3]. In Fig. 1 we phase diagrams, etc. In Sect. IV we apply the distilla- show a schematic picture of a single aplication of the tion protocols for the purification of non-diagonal mixed purification method. Let us comment on the outcomes states that are more easily realized experimentally. In corresponding to the most relevant steps in this protocol. D Sect. V we study the large- limit of these protocols. After step 1/, the 4-quit mixed state ρ ρ shared by Alice In Sect. VI we present a detailed investigation of the re- and Bob reads as follows ⊗ lationships between quantum distillation protocols and 2 + + + + renormalization methods for quantum lattice Hamiltoni- ρ++ ρ++ = F Φ Φ Φ Φ ans. Section VII is devoted to conclusions. In appendix A ⊗ | ih | + F (1 F )[ Φ+Ψ+ Φ+Ψ+ + Ψ+Φ+ Ψ+Φ+ ] (4) we find the general solution for the distillation recursion − | ih | | ih | relations used in the text in the general case of qudits. + (1 F )2 Ψ+Ψ+ Ψ+Ψ+ − | ih | In step 2/, Alice and Bob apply bilaterally the CNOT II. SIMPLE DISTILLATION PROTOCOLS gate taking their first qubit as source and their second WITH QUBITS qubit as target, i.e., qubits first and third are source qubits while qubits second and fourth are target qubits. Our starting point is the orthonormal basis of Bell To obtain the transformed mixed state we must deter- states formed by the first qubit belonging to Alice and mine the action of the bilateral CNOT gate UBCNOT [3] the second to Bob: on the states of the form ϕAϕB . The results of this computation are shown in Table| I.i With the help of this ± 1 table we find Φ := [ 00 11 ] | i √2 | i ± | i 2 + + + + (1) UBCNOTρ++ ρ++UBCNOT = F Φ Φ Φ Φ ± 1 ⊗ | ih | Ψ := [ 01 10 ] + F (1 F )[ Φ+Ψ+ Φ+Ψ+ + Ψ+Ψ+ Ψ+Ψ+ ] (5) | i √2 | i ± | i − | ih | | ih | + (1 F )2 Ψ+Φ+ Ψ+Φ+ We shall use the word “simple” applied to the distillation − | ih | protocols to denote that the mixed state we shall be deal- After steps 3/ and 4/, Alice and Bob measure their tar- ing with is made up of a combination of one state in the get qubits and retain their source qubit whenever they + − set S := Φ , Φ of Bell states that have coincident find, via classical communication, the same results: ei- {| i | i} + bits in Alice’s and Bob’s qubits, with another state in ther 0A0B or 1A1B. This fact selects the state Φ as the set A := Ψ+ , Ψ− of Bell states that do not have | i {| i | i} the only admissible possibility for the target state. Thus, coincidences. Thus, we have 4 possible combinations to only the first and third terms in the RHS of (5) survive to do this type of entanglement distillation. ′ this process and the resulting 2-qubit state ρ++ is again To begin with, we shall choose the following mixed of the same form as the original starting state (2) in step state in order to set up a simple distillation protocol 1/, but with a higher fidelity F ′ > F . In fact, we get

+ + + + ′ ′ + + ′ + + ρ++ := F Φ Φ + (1 F ) Ψ Ψ (2) ρ := F Φ Φ + (1 F ) Ψ Ψ (6) | ih | − | ih | ++ | ih | − | ih | Alice and Bob will also need to apply the CNOT gate with the new fidelity being defined as usual F 2 F ′ = (7) UCNOT i j := i i j , i,j =0, 1. (3) F 2 + (1 F )2 | i| i | i| ⊕ i − The distillation protocol can be arranged into a set of This relation defines a recursion scheme for entanglement 5 instructions or steps [3],[12],[14]: purification: starting with say NP pairs of Bell states of fidelity F , after every application of the whole protocol 3

Alice Bob We can answer this question aﬃrmatively by computing the action of the bilateral CNOT gate on the tensor product of these mixed states (8). With a similar analysis F PSfrag replacements F which has led us to Table I [3], we obtain

CNOT CNOT 2 + + + + U ρ − ρ −U = F Φ Φ Φ Φ BCNOT + ⊗ + BCNOT | ih | F F + F (1 F )[ Φ−Ψ− Φ−Ψ− + Ψ−Ψ+ Ψ−Ψ+ ] − | ih | | ih | + (1 F )2 Ψ+Φ− Ψ+Φ− − | ih | Quantum Distillation 2 + − + − U ρ− ρ− U = F Φ Φ Φ Φ BCNOT + ⊗ + BCNOT | ih | + F (1 F )[ Φ−Ψ+ Φ−Ψ+ + Ψ−Ψ− Ψ−Ψ− ] (9) − | ih | | ih | + (1 F )2 Ψ+Φ+ Ψ+Φ+ F ′ > F ′ − | ih | F > F 2 + − + − U ρ−− ρ−−U = F Φ Φ Φ Φ BCNOT ⊗ BCNOT | ih | + F (1 F )[ Φ+Ψ− Φ+Ψ− + Ψ+Ψ− Ψ+Ψ− ] − | ih | | ih | + (1 F )2 Ψ+Φ− Ψ+Φ− FIG. 1: Schematic representation of the distillation protocol − | ih | by Alice and Bob. Originally, two pairs of shared entangled We now realize that if we proceed to measure the target qubits represented by enclosed dots are transformed into a single pair of higher purity (doubly enclosed dots). bits and classical communicate the results, we do not end up with the same type of mixed state as we had started with. That is, the protocol as it stands is not valid since |ϕAi|ϕB i UBCNOT|ϕAi|ϕB i it does not yield invariant mixed states. This problem |Φ+i|Φ+i |Φ+i|Φ+i has a solution provided we introduce an additional step |Φ+i|Ψ+i |Φ+i|Ψ+i prior to the measurement of the target qubits by Alice |Ψ+i|Φ+i |Ψ+i|Ψ+i and Bob. This additional step corresponds to a local + + + + |Ψ i|Ψ i |Ψ i|Φ i unitary operation UA UB performed by Alice and Bob on their source qubits.⊗ The form of this local unitary depends on the mixed state we are distilling. We find TABLE I: This table shows the results of applying the bi- the following results: lateral CNOT gate to certain pairs of Bell states needed to Step 2′. Alice and Bob apply a local unitary transforma- distillation. tion UA UB to their source qubits: ⊗ 1 1 For ρ+−, UA = 2 (1+i)(σx +σy), UB = 2 (1 i)(σx σy). 1 − − NP ′ For ρ−+, UA = UB = (1 + i)(σx + σy). we obtain 2 pairs of higher fidelity F > F . Thus, pu- 2 rification is achieved at the expense of halving the num- For ρ−−, UA = σz, UB = 1. After this extra step, we can guarantee that the result- ber of Bell pairs. The fixed points Fc of the recursion ′ ing 4-qubit mixed state has the appropriate Bell pairs relation (7) are defined as F (Fc) := Fc and they are 1 at the source qubits so as to produce the same original given by Fc = 0, 2 , 1. The fixed points Fc = 0, 1 are 1 state, once steps 3/ and 4/ are performed. Moreover, it stable, while Fc = 2 is unstable. The best way to recast these qualitative properties of the flow equation for the is straightforward to prove that the new fidelity for these fidelities (7) is to draw the corresponding flow diagram 3 protocols is also given by the same recursion relation PSfrag replacementsas shown in Fig. 2. (7) as in the first protocol. Finally, if the constraint that the state of fidelity F must be a Φ± state is relaxed, then there are two addi- 1 tional possible mixed states whose analysis can be carried 0 2 1 F out in a similar fashion. Fc

FIG. 2: Flow diagram for the ﬁdelity F of the distillation III. MULTILEVEL EXTENSIONS OF protocol given by the recursion relation (7). DISTILLATION PROTOCOLS

Next, we may wonder whether it is possible to devise In order to generalize the simple distillation protocol distillation protocols for the three possible combinations of the previous section to the case of qudits, we must of Bell states, namely, notice that the two main ingredients in that distillation protocol are: + + − − i) the CNOT gate, ρ − := F Φ Φ + (1 F ) Ψ Ψ + | ih | − | ih | − − + + ii) the Bell states (1). ρ− := F Φ Φ + (1 F ) Ψ Ψ (8) + | ih | − | ih | Regarding the CNOT gate, the extension of this gate − − − − ρ−− := F Φ Φ + (1 F ) Ψ Ψ to deal with qudits is not unique. As has been noted in | ih | − | ih | 4

[18],[19], the CNOT gate for qubits (3) has 3 properties on the state (16) we need a previous result about the that make it special, namely action of the gate UBCNOT on pairs of generalized Bell states (12). After some algebra, we arrive at the following † −1 UCNOT = UCNOT, expression † (10) UCNOT = UCNOT, UBCNOT Ψkj Ψk′ j′ = Ψk⊕k′,j ΨD⊖k′,j⊖j′ (17) i j =0 i = j. | i| i | i| i ⊕ ⇔ This is a fundamental result for it means that the space The extension of the CNOT gate for qudits that satisfies of two-pairs of generalized Bell states is invariant under these 3 properties (10) is given by [18],[19] the action of the generalized bilateral CNOT gate. This is a very nice result that condenses in a single formula UCNOT i j := i i j , i,j =0,...,D 1 (11) | i| i | i| ⊖ i − all the possibilities for the outcome of the action of the where i j := i j, mod D, denotes substraction modulus CNOT gates on Bell states, in particular, the whole table D. This⊖ is the− definition that we shall adopt throughout employed by Bennett et al. in [3] for the case of qubits is this paper, unless otherwise stated. contained in equation (17). This property is essential in As for the higher-dimensional extension of Bell states order to have a closed distillation protocol. Actually, it (1), we shall also take the following generalization would have been enough to have obtained only the source [18],[19] qubits as generalized Bell states. Then, with the help of this property (17) we obtain Ψkj := UCNOT [(UF k ) j ] , k,j =0,...,D 1 (12) the action of U on pairs of ρ states, as follows | i | i ⊗ | i − BCNOT where UF is the quantum Fourier transform (QFT) D−1 D−1 UBCNOTρ ρUBCNOT = qk⊖k′ ,j qk′j′ D−1 ⊗ 1 2πiky k,j=0 k′,j′=0 (18) UF k := e D y , (13) X X | i √ | i ′ ′ ′ ′ D y=0 Ψkj ΨD⊖k ,j⊖j Ψkj ΨD⊖k ,j⊖j X | ih | which reduces to the Hadamard gate when dealing with We see that this state is already of the same form in the qubits (D = 2). As a matter of fact, we can readily check source qubits as the original ρ (16). that for the special case of qubits D = 2 we recover the Step 3/. Alice and Bob measure their target qubits in standard Bell pairs (1) in the following form (18). To see the result of this measurement, let us write the explicit form of the target qubits, namely + + Ψ00 = Φ , Ψ01 = Ψ , | i | i | i | i (14) D−1 − − −2 i ′ Ψ10 = Φ , Ψ11 = Ψ 1 π k z ′ | i | i | i | i ΨD⊖k′,j⊖j′ = e D z z (j j ) . (19) √ | i| ⊖ ⊖ i D z=0 Moreover, using the generalized CNOT gate (11), the X generalized Bell states are given by Therefore, coincidences between Alice’s and Bob’s target D−1 qubits will happen only when the following condition is 1 2πiky Ψkj = e D y y j . (15) satisfied | i √ | i| ⊖ i D y=0 X z = z (j j′) j = j′. (20) With these extensions of the CNOT gate and the Bell ⊖ ⊖ ⇐⇒ states, we can set up a generalization of the simple dis- Step 4/. After their measurement, Alice and Bob com- tillation protocols of Sect. II for qudit states. These municate classically their result so that they retain the protocols have the same 5 steps as before. resulting source Bell pairs only when they have coinci- Step 1/. We shall assume a general diagonal mixed state dences, and discard them otherwise. The resulting net of the form effect of this process is to produce a Kronecker delta ′ D−1 function δjj in the target qubits. More precisely, the resulting unnormalized mixed state is given by ρ := qkj Ψij Ψij , | ih | k,j=0 X (16) D−1 D−1 D−1 ′ ρ qk⊖k′ ,jqk′ j′ δjj′ Ψkj Ψkj . (21) ∼ | ih | 1 =: qkj , k,j=0 k′,j′=0 k,j X X X=0 Therefore, we end up with a diagonal mixed state of the where qkj are normalized probabilities. For non-diagonal same form as the starting one mixed states, we refer to Sect. IV. Then, Alice and Bob share pairs ρ ρ of these states (16). D−1 × ′ ′ Step 2/. Alice and Bob apply bilaterally the generalized ρ = q Ψkj Ψkj . (22) kj | ih | CNOT gate (11). To know the result of this operation k,j X=0 5 with the new probabilities given by 1 F ′ D−1 ′ ′ ′ k′=0 qk⊖k ,j qk j 0.8 qkj = D−1 D−1 . (23) ′ ′ k,jP=0 k′=0 qk⊖k ,jqk j 0.6 This is a generalizedP recursionP relation that includes eq. (7) as a particular instance. 0.4 Step 5/. Alice and Bob start all over again the same process with the initial state now being ρ′ in (22-23). The nice feature of these generalized distillation pro- 0.2 tocols for dealing with qudits is the fact that we havePSfrag replacements at our disposal explicit analytical formulas (23) for the 0 0 0.2 0.4 0.6 0.8 1 evolution (flow) of the different weights (probabilities) F of the generalized mixed states to be purified. As these distillation protocols are too general, it is worthwhile to FIG. 3: Plots of the distilled fidelities F ′ as a function of the consider some particular cases of interest separately. The original fidelity F for several values of the dimension D of the general solution to the distillation recursion relations (23) qudits: D = 2, 4, 10. is presented in the appendix A. We hereby provide the following analysis of some examples: i). Let us investigate the closest generalization of the relation, (27) admits an explicit analytical solution for simple protocols introduced in Sect. II. Thus, let us con- the general term of the series Fk given by sider the following type of initial mixed state k F (2 ) M F = , k 1, F := F. (28) k (2k ) 0 k 1−F ≥ ρ := qi Ψ0i Ψ0i ,M D 1, F (2 ) + (D 1) | ih | ≤ − D−1 i=0 − X (24) h i M From this solution, we inmediately find that the fixed 1 =: qi, qi 0, points F =0, 1 are stable while F = 1 is unstable. ≥ c c D i=0 ′ ′ X In Fig. 3 we plot the function F = F (F ) for several This corresponds to working with the subset of all possi- values of the dimension D. From the analysis of these D−1 curves we inmediately obtain the corresponding flow di- ble generalized Bell states of the form Ψ0i i=0 . Inter- estingly enough, this includes the case{| of thei} state ρ agram that we represent in Fig. 4. We check that for ++ D = 2 we recover the flow diagram corresponding to in (2) for D = 2. The recursion relations (23) forPSfrag this replacements special subset of states takes the following simpler form standard qubits (Fig. 2). q2 q′ = i . (25) 1 i M 2 0 D 1 j=0 qj F Fc For M = 2, i.e., consideringP a mixed state formed of just two Bell states of the form Ψ0i , the protocol has FIG. 4: Flow diagram for the fidelity F of the generalized | i the following recursion relation distillation protocol for qudits given by the recursion relation (27). q2 q′ i , i = 2 2 (26) q + (1 qi) i − We see from Fig. 4 that the stability basin is increased with respect to the case of standard qubits, as in Fig. 2. where here the index i stands for any possible pair of Bell This means that we can start with a mixed state having states of the type Ψ . In other words, we have found 0i a fidelity F with respect to the Bell state Ψ lower a direct D-dimensional| i generalization of the distillation 00 that 1 and we still will succeed in purifiying| thati state protocols for qubits in Sect. II, with q := F . 2 i towards fidelity close to 1. Thus, we have found that it ii). For M = D 1 and taking q := F and q := 0 i is more advantageous to distill a given Bell state Ψ if 1−F ,i = 1,...,D −1 we can find a more advantageous 00 D−1 we prepare the mixed state ρ in (24) in the form | i protocol than the− previous one. In fact, in this case we find that D−1 1 F 2 ρ := F Ψ00 Ψ00 + − Ψ0i Ψ0i , (29) ′ ′ F | ih | D 1 | ih | q0 := F = 2 . (27) i=1 2 (1−F ) − X F + D−1 rather than using just one single of those states The fixed points of this recursion relation are now given 1 by Fc = 0, , 1. Despite being a non-linear recursion ρ := F Ψ Ψ + (1 F ) Ψ i Ψ i , i =0. (30) D | 00ih 00| − | 0 ih 0 | 6 6

We may wonder how is it likely for Alice and Bob to obtain the same values (coincidences) after measuring the target qudits in the step 3/ of the distillation protocol. Let us denote by AB this probability which will depend on the value F ofP the fidelity. From equations (21) and (27) we find this probability of coincidences to be ′ 1 q0 0.75 1 2 (1 F ) 0.5 (F )= F 2 + − . (31) 0.8 PAB D 1 0.25 − 0 0.6 1 0 The minimum of this probability is at F0 = D and its 0.2 0.4 q1 value is ( 1 ) = 1 . Likewise, (1) = 1. Thus, we PAB D D PAB PSfrag replacements 0.4 find that the probability is lower and upper bounded as 0.6 0.2 1 1 (F ) F for F [ , 1]. q 0.8 D ≤PAB ≤ ∈ D 0 One is also interested in knowing the number of steps 1 0 K(ǫ, F0) needed to achieve a certain final fidelity close to 1, say 1 ǫ, starting from an appropriate initial fi- FIG. 6: The desired fidelity q′ for amplification after a single −1 0 delity F0 > D . This number can be computed from our application of the distillation protocol for qutrits (36) as a analytical solution (28) from the condition function of the previous fidelitites q0, q1.

FK(ǫ,F0) := 1 ǫ. (32) − iii). For qutrits, D = 3, the most general diagonal Thus, we find the following analytical formula for the mixed state with the allowed Bell states taking values on number of steps needed to obtain a certain degree of fi- the set Ψ0i is 1 {| i} delity ǫ as a function of the initial fidelity F0 > D ,i.e., ρ := q0 Ψ00 + q1 Ψ01 + q2 Ψ02 | i | i | i (35) ǫ 1 =: q0 + q1 + q2 ln( (1−ǫ)(D−1) ) K(ǫ, F0)= log . (33) 2 ln( 1−F0 ) Let us assume that the state we want to purify is Ψ00 . & (D−1)F0 !' | i Now, our recursion relation for our fidelity q0 depends on In Fig. 5 we plot the number of iterations (33) for a two variables, namely, q2 12 q′ 0 , 0 = 2 2 2 (36) K q0 + q1 + (1 q0 q1) 10 − − and a similar equation for q1 with q0 q1. In Fig. 6 ↔′ ′ 8 the dependence of the function fidelity q0 = q0(q0, q1) for qutrits is plotted. We observe that it is a monotonous 6 incresing function which gurantees that the initial fidelity

4 will flow towards 1, under certain conditions. To find these conditions, we find that the set of fixed points of 2 these recursion relations is given by

1 1 1 1 1 1 PSfrag replacements 0 (q0, q1)c = (0, 0), ( 2 , 0), (0, 2 ), ( 3 , 3 ), ( 2 , 2 ), (1, 0), (0, 1) . 0 0.2 0.4 0.6 0.8 1 { }

F0 We have also found the ﬂow diagram associated to these recursion relations which is now two-dimensional and

FIG. 5: Plot of the number of iterations K(ǫ, F0) (33) to we show it in Fig. 7. From this diagram we see that achieve a final fidelity of F = 0.99 as a function of the initial the purification protocol is successful in arriving to the fidelity F0 and for several values of the dimension D of the maximum fidelity q0 = 1 provided the initial fidelity qudits: D = 2, 4, 10. lies in the stability basin of the fixed point (1, 0) which is given by the trapezoid formed by the set of points 1 1 1 1 1 given value of the final fidelity 1 ǫ that we take as the ( 2 , 2 ), ( 3 , 3 ), ( 2 , 0), (1, 0). fixed value of 0.99, and then we− find how is the dependence on the initial fidelity F0. We see that for a given IV. DISTILLATION OF NON-DIAGONAL admissible value of F0, the lowest number of iterations corresponds to the protocol with the higher value of the MIXED STATES qudits dimension D: So far, we have been investigating the properties of

K(ǫ , F )D1 K(ǫ , F )D2 , for D >D . (34) distillation protocols applied to mixed bipartite state of 0 0 ≥ 0 0 1 2 7

(0, 1) in terms of some physical property taking values on i = 0,...,D 1 and that Alice and Bob have a mechanism to select when− they have ii coincident qudits (or diagonal) from ij non-coincident| i qubits (non-diagonal). q 1 To| proceedi with the distillation of the state ρ in (40), we first must express the states Ψ and Ψ in the basis | di | oi 1 , 1 of the generalized Bell states, with the result PSfrag replacements ( 2 2 ) (0, 1 ) 2 Ψ = Ψ , | di | 00i D− 1 1 (41) Ψo = Ψ0i . | i √D | i 1 i=1 − X Next, Alice and Bob share two pairs of non-diagonal mixed states (0, 0) ( 1 , 0) q (1, 0) 2 0 ρ ρ = F 2 Ψ Ψ Ψ Ψ ⊗ | 00 00ih 00 00| D−1 FIG. 7: Two-dimensional flow diagram associated to the dis- F (1 F ) tillation protocol for qutrits (36). + − [ Ψ Ψ i Ψ Ψ j + Ψ iΨ Ψ j Ψ ] D 1 | 00 0 ih 00 0 | | 0 00ih 0 00| i,j=1 − X 2 D−1 diagonal form such as those in (22), (23). It is apparent (1 F ) + − Ψ0iΨ0k Ψ0j Ψ0l , that once we have a general result for the operation of (D 1)2 | ih | i,j,k,l=1 U − X the BCNOT gate of generalized Bell states (17), we can (42) also deal with non-diagonal mixed states, namely, and they apply bilaterally the CNOT gate to it (17) with D−1 D−1 the result ρ := qkjk′ j′ Ψkj Ψk′j′ , | ih | 2 k,j=0 k′,j′=0 U ρ ρU = F Ψ Ψ Ψ Ψ X X (37) BCNOT ⊗ BCNOT | 00 00ih 00 00| D−1 D−1 D F F −1 ′ ′ ′ ′ (1 ) 1 =: qkjk j , qkjk j 0. + − [ Ψ00Ψ0,⊖i Ψ00Ψ0,⊖j + Ψ0iΨ0i Ψ0j Ψ0j ] ′ ′ ≥ D 1 | ih | | ih | k,j=0 k ,j =0 i,j=1 X X − X D−1 Since this is a too much general state, we prefer to extract (1 F )2 + − Ψ iΨ i⊖k Ψ j Ψ j⊖l . from this class of non-diagonal states one type which we (D 1)2 | 0 0 ih 0 0 | i,j,k,l believe it may have potential applications. − X=1 Let us imagine that Alice and Bob are manipulating (43) bipartite qudit states that are diagonal in the computa- The process of measuring the target qudits and retaining tional basis. More explicitly, the entangled state they the source qudits when upon classical communication Al- want to purify is of the form ice and Bob find coincidences in their measures amounts D−1 to retaining the terms in (43) that have the state Ψ00 1 | i Ψ := ii , (38) in the target qudits. This means that only the first term | di √ | i D i=0 and part of the last term in (43) contribute to the fi- X nal source mixed state, which takes the following form while states which are non-diagonal are considered as without normalization acting as disturbing noise that they want to get rid of. 2 D−1 D−1 Specifically, this noise will be represented by the state ′ 2 (1 F ) ρ F Ψ Ψ + − Ψ i Ψ j ∼ | 00ih 00| (D 1)2 | 0 i  h 0 | D−1 i=1 ! j=1 1 − X X Ψo := ij . (39) 2   | i D(D 1) | i 2 (1 F ) i=6 j=0 = F Ψ Ψ + − Ψ Ψ . − X | 00ih 00| D 1 | oih o| − Then, in order to achievep their goal of purifying states (44) of the diagonal form Ψ with respect to non-diagonal d Upon normalization, we arrive again at a non-diagonal states Ψ , they set up| ai distillation protocol based on | oi mixed state of same form as the one we started with sharing copies of the following mixed state ′ ′ ′ ρ = F Ψd Ψd + (1 F ) Ψo Ψo , but with a new | ′ ih | − | ih | ρ := F Ψ Ψ + (1 F ) Ψ Ψ . (40) fidelity F given by | dih d| − | oih o| 2 ′ F We envisage that this scenario is physically feasible since F = 2 . (45) 2 (1−F ) we can imagine that the computational basis is realized F + D−1 8

Let us notice that this is precisely the same recursion k relation that we found in Sect. III in a diﬀerent context q( )(x) (27). 8

V. CONTINUUM LIMIT OF QUDIT 6 PROTOCOLS 4 For the general case represented by the recursion relations (25) we can also find the general solution for the 2 k-th iteration q(k),i = 0,...,D 1 starting from their i − (0) D−1 (0) PSfrag replacements initial values qi satisfying i=0 qi := 1. We find the 0 following solution 0 0.2 0.4 0.6 0.8 1 P x 2k (0) qi q(k) = . (46) FIG. 8: Evolution of the probability density under the itera- i 2k Mh i(0) tive application of the distillation protocol in the continuum j=0 qj limit (49). Starting with a parabolic distribution k = 0, we h i show the resulting profiles after steps k = 1, 2,..., 6. Let us assume that theP maximum initial value is M := max q(0) and it is p times degenerate. Then, using { i } the general solution (46) we can inmeditely find the fixed density q(x)(k) after k steps of the distillation protocol. points after the evolution with the recursion relations. This is given by We find 2k (k) (0) q(0)(x) lim qi =0, if qi < M, q(k)(x)= . (49) k→∞ 2k (47) 1 q(0)(y) dy (k) 1 (0) 0 lim qi = , if qi = M. k→∞ p This is a closed analyticalR equation that provides us with From the analysis of these distillation protocols and the the evolution of the probability density for any initial way they operate we arrive at the conclusion that they probability profile q(0)(x). In Fig. 8 we plot this evolution resemble a sort of amplitude amplification quite similar for an initial distribution of a parabolic form q(0)(x) = to what happens in the Grover algorithm where there 6(x x2). We see how as we increase the step k of the − exists what is called quantum amplitude amplification. distillation, the new distributions get peaked around the 1 However, there is an important distintion between both highest value of the initial distribution, which is x = 2 in procedures: in the distillation method, the maximum am- this particular case. This behaviour illustrates the idea plification is attained asymptotically, while in Grover al- of the weight amplification and is in agreement with the gorithm it is achieved periodically. The reason for this results (47) for the fixed points of the flow equations. difference relies on the fact that the distillation process is not unitary (since we make measurements and discard states), while Grover is unitary. Thus, we propose to VI. QUANTUM DISTILLATION AND refer to the distillation protocol as weight amplification, QUANTUM RENORMALIZATION since it is certain probability weights of the intial mixed states, and not amplitudes, what are being amplified. It is interesting to notice the analogy between the re- When D is very large, we can approximate the prob- cursive distillation process represented by the equation (0) (7) and Fig. 2 and the truncation process in the Renor- ability weights qi taking values on the discrete set 0, 1,...,D 1 i, by a density function q(x)(0) de- malization Group analysis of certain quantum lattice fined{ on the− real} interval ∋ [0, 1]. This is achieved by intro- Hamiltonian models, specifically, the ITF model (Ising ducing the variable x [0, 1] defined as x := i∆x with in a Transverse Field) [24],[25]. The basic idea of a QRG ∆x := 1 . Thus, in the∈ limit D (∆x 0), we get method is: i/ elimination of high energy states plus, ii/ D−1 → ∞ → iterative process. This is precisely what happens in a a probability density as q(0)(i∆x) q(0)(x)dx. It is also i → quantum distillation process which we have seen in the normalized as preceeding sections, achieving a purification of a mixed state by means of discarding states and a recursive pro- 1 cedure. This relationship can be made even closer if (0) q (x)dx =1. (48) we briefly recall what a quantum renormalization group Z0 (QRG) method is. The subject of the distillation is a Likewise, we can take the continuum limit of the general mixed state operator ρ, while that of the renormaliza- recursion equation (46) in order to obtain the probability tion is a quantum Hamiltonian operator H. A summary 9

Quantum Distillation Quantum RG Mixed State ρ Quantum Hamiltonian H Computational Basis Local Site Basis Bell Basis Energy Basis Alice & Bob Tensor Product Blocking Method L.O.C.C. Truncation Operator Maximum Fidelity Minimum Energy RG-Flow Diagram Distillation-Flow Diagram

TABLE II: Summary of the comparative analysis between the quantum distillation process and a quantum renormalization group method for lattice Hamiltonians.

Initial chain with 8 sites

FIG. 9: Block decomposition of the Heisenberg chain in 3-site blocks.

of these relations is presented in Table II that will be QRG deduced along the way. The easiest way to present PSfrag replacements F0 F0 F0 F0 F0 F0 F0 F0

F1 F1 F1 F1 FIG. 11: Example of renormalization QRG process for an initial chain with 8 sites in 3 steps (making blocks of 2 sites

F2 F2 each).

PSfrag replacements F3 of the distillation process, as shown in Fig. 10. This is to be compared with the similar iterative process in the QRG method in Fig. 11. The block Hamiltonian is then

FIG. 10: Example of distillation process starting with 8 pairs HB = J(S S + S S ) 1 · 2 2 · 3 of mixed states for Alice (and the same amount for Bob). J (51) After 3 steps the original ﬁdelity F is improved up to a ﬁnal = (S + S + S )2 S2 (S + S )2 0 2 1 2 3 − 2 − 1 3 value of F3 (we assume full success for simplicity). The label B here stands for Block and not for Bob. The diagonalization of HB is straightforward using the the QRG method is with an example of quantum lattice Clebsch-Gordan decomposition of the tensor product of 1 Hamiltonian like the isotropic Heisenberg model on a 1D 3 irreducible representations of spin S = 2 , chain: 1 1 1 = 1 1 3 . (52) N 2 ⊗ 2 ⊗ 2 2 ⊕ 2 ⊕ 2 H = J Si Si , (50) · +1 In particular, the ground state (GS) is given by i=0 X 1 S 1 i GS = [2 ] , (53) with i spin- 2 operators at site of the chain. The local |⇑i √6 | ↑↓↑i−| ↓↑↑i−| ↑↑↓i site basis , corresponds to the computational basis 0 , {|1 ↓i. Much| ↑i} like this latter basis is not enough which is a spin doublet (with a similar expresion for the {| i | i} for doing the distillation, the local site basis needs to be other state GS, with the spins reversed). This fact is complemented with another type of basis. To see this, peculiar of the| ⇓i 3-site block and it is the main underlying let us start the RG process with the block decomposition reason for using a block of that size in the QRG (this fact of the chain in blocks of nB = 3 sites as shown in Fig. 9. is model dependent: for the ITF model, the blocking is This blocking method in QRG corresponds to the tensor with nB = 2 sites [24],[25], Fig. 11). In the energy basis, product of Alice and Bob’s shared states at the begining the block Hamiltonian is diagonal and this corresponds 10

1/ Block Decomposition: H = HB + HBB . 2/ Diagonalization of HB. 3/ Truncation within each Block: O. ′ † ′ † 4/ Renormalization: HB = OHBO , HBB = OHBB O ′ ′ ′ 5/ Iteration: Go to 1/ with H = HB + HBB. FIG. 12: QRG renormalization of HB.

TABLE III: Steps of the quantum renormalization group method (QRG) for lattice Hamiltonians. to the Bell basis for the mixed state ρ in the distillation process. n n + 1 Now, the truncation of states amounts to retaining the r l state of lowest energy (doublet) and discarding the re- Sr Sl maining 2 excited states. This reduction scheme is ofPSfrag replacements the form 23 = 8 2. This truncation corresponds to discarding unwanted−→ states of non-coincidences in the distillation process. The new effective site is again a spin- 1 2 site as shown in Fig. 12. The RG-truncation is implemented by means of a truncation operator O constructed from the lowest energy eigenvalues of HB retained during n n + 1 the renormalization process. In this example, O is constructed from the lowest energy doublet in the Clebsch- FIG. 13: QRG renormalization of the interblock Hamiltonian Gordan decomposition (52), namely (53). Similarly, in HBB . the distillation process we have that the tensor product of Alice and Bob’s states can be decomposed into states with coincident qubits in the target, denoted by ρa (4), C n n and states with non-coincident qubits in the target, de- where we denote by and + 1 two successive blocks noted by ρa (4), i.e., in the original lattice (see Fig. 13) that become two suc- NC cessive sites (see Fig. 12) in the new lattice after the renormalization. We can collect all these steps in Ta- ρ ρ = ( ρa ) ( ρa ), (54) AB AB C NC ble III [26], [24]. This table should be contrasted with ⊗ a ⊕ a X X the similar table for the distillation process that can be where the sum in a runs over a certain number of mixed formed with the steps explained in Sect. II. states of 4 parties. Notice that this stage is similar to the The outcome of the RG-method is that we obtain the RG-stage represented by equation (52). Next, an elimi- correct RG-flow for the coupling constant J 0, sig- −→ nation process similar to the RG-truncation is performed nalling a gapless system plus an approximate estima- by means of LOCC operations (measurements and classi- tion for the ground state energy, which by means of the cal communication) that retains only the bipartite states variational principle, it is an upper bound for the ex- a embedded in the ρC states. act energy. Therefore, the QRG is an energy minimiza- Then, the renormalization of the block Hamiltonian is tion procedure. Likewise, the purification process pro- simply duces a protocol for fidelity maximization along with a distillation-flow diagram. This completes the relation- † HB′ = OHBO = E0 = J. (55) ship established in Table II between quantum distillation − and quantum renormalization. Similarly, we could have expressed the state-elimination of the distillation in previous sections in terms of an truncation operator, say OD, such that the new mixed state VII. CONCLUSIONS ′ ρAB is obtained as

′ † The ﬁeld of quantum distillation protocols has become ρ = OD(ρAB ρAB)O . (56) AB ⊗ D very active in the theory of quantum information due to the central role played by entanglement in the quantum In the case of the quantum Hamiltonian, we still need communication procedures and its tendency to degrada- extra work since there are interaction links between tion. blocks (see Fig. 9). These are absent in the distillation In this work we have been interested in several exten- protocol. However, the renormalization of the interblock sions of the puriﬁcation protocols when dealing with mul- Hamiltonian H follows also the same prescription as BB tilevel systems (qudits) instead of the more usual qubit in (55) and we arrive at protocols. We have seen the various advantages of having 2 distillation methods for qudits systems as compared with JSn Sn+1 RG J( )2S′ S′ , (57) r · l −−→ 3 n · n+1 the simple case of qubits. We have also obtained the gen- 11

(n) eral form of the solution to the distillation recursion rela- Thus, gkj are unnormalized probability weights. The re- tions and several particular solutions have been studied cursion relations they satisfy can be read as follows (A1): explicitly. We have developed the relationship between (n) for a fixed second index j, the unnormalized weights gkj quantum distillation protocols and quantum renormal- at the step n of the distillation process are obtained as the ization group methods, something which is interesting in convolution over the first indices k of the unnormalized itself and could serve as a guide for possible extension of (n−1) weights gkj in an earlier step. This fact calls for the purification methods. introduction of the Fourier transform in order to analyze We would like to mention that the possibility of work- the relations (A1). Let us introduce the new variables ing with qudits systems has become quite realistic in the R(n) recent years. For instance, it is possible to realize mul- j defined as tilevel systems in terms of the orbital angular momen- D−1 ˆ tum of photons, instead of the more standard polariza- (n) 2πikk (n) R := e D g . (A3) tion (qubit) degree of freedom [27], [28], [29]. Yet another kjˆ kj k possibility is to use the so called multiport beam splitters X=0 [30], [31], [32], [33]. There are several ways in which this work can be ex- Now, using the properties of the convolution and the tended. One is the consideration of noise as as source of Fourier transform it is inmediate to arrive at a simpler errors during the distillation protocol itself. Another one recursion relation is to allow the possibility of having these distillation pro- 2 tocols for qudits be embedded into a quantum repeater R(n) = R(n−1) , (A4) kjˆ kjˆ protocol [15], [16]. h i Acknowledgments. This work is partially supported by which can be iterated all the way down to the initial step the DGES under contract BFM2000-1320-C02-01.

(2n) n APPENDIX A: GENERAL SOLUTION OF THE R( ) = R(0) . (A5) kjˆ kjˆ DISTILLATION RECURSION RELATIONS h i Fourier transforming back to the unnormalized variables, In this appendix, we look for more general solutions we get to the general distillation recursion relations (23) than n those studied in section III. To this end, it is convenient D−1 D−1 (2 ) ˆ ˆ ′ (n) (n) 1 2πikk 2πikk (0) to introduce auxiliary variables g deﬁned by − D D kj gkj = e e qk′j , (A6) D ′ kˆ=0 "k =0 # D−1 X X (n) (n−1) (n−1) (0) (0) g g ′ g ′ , g q , kj = k⊖k j k j kj = kj (A1) from which we also obtain the normalized probability ′ k =0 (n) X weights qkj upon normalization (A2). In particular, for (n) D so that the real weights q are related to these auxiliary the case of qubits treated in Sect. II, = 2 and if we also kj q F, q variables as restrict ourselves to weights of the form 00 = 01 = 1 F, q = q = 0, we again obtain from the general − 10 11 g(n) solution (A6) the simple recursion relation in equation q(n) := kj . (A2) (7). kj D−1 (n) l,i=0 gli P

[1] M. Lewenstein, D. Bruss, J.I. Cirac, B. Kraus, M. Kus, J. [5] N. Gisin, “Hidden quantum nonlocality revealed by local Samsonowicz, A. Sanpera, R. Tarrach, “Separability and filters”, Phys. Lett. A210, 151, (1996). distillability in composite quantum systems -a primer-”, [6] P.G. Kwiat, S. Barraza-Lopez, A. Stefanov, N. Gisin, J. Mod. Opt. 47, 2841 (2000). “Experimental entanglement distillation and ’hidden’ [2] D. Bruss, J.I. Cirac, P. Horodecki, F. Hulpke, B. Kraus, non-locality,” Nature 409, 1014 (2001). M. Lewenstein, A. Sanpera, “Reflections upon separabil- [7] P.W. Shor, “Scheme for reducing decoherence in quan- ity and distillability”, J. Mod. Opt. 49, 1399-1418 (2002). tum computer memory”, Phys. Rev. A52, 2493-2496, [3] C.H. Bennett, G. Brassard, S. Popescu, B. Schumacher, (1995). J.A. Smolin, W.K. Wootters, “Purification of Noisy En- [8] A.M. Steane, “Error correcting codes in quantum the- tanglement and Faithful Teleportation via Noisy Chan- ory”, Phys. Rev. Lett. 77, 793, (1996). nels”, Phys. Rev. Lett. 76, 722-725 (1996). [9] D. Gottesman, “A theory of fault-tolerant quantum com- [4] J.W. Pan, C. Simon, C. Brukner, A. Zeilinger, “Feasi- putation,” Phys. Rev. A 57, 127-137 (1998). ble Entanglement Purification for Quantum Communi- [10] J. Preskill, “Fault-tolerant quantum computation.” cation”, Nature 410, 1067 (2001). quant-ph/9712048. 12

[11] C.H. Bennett, D.P. DiVincenzo, J. Smolin, W.K. Woot- [23] C.H. Bennett, H.J. Berstein, S. Popescu, B. Schumacher, ers, “Mixed-state entanglement and quantum error cor- “Concentrating partial entanglement by local opera- rection”, Phys. Rev. A54, 3824 (1996). tions”, Phys. Rev. A53, 2046 (1996). [12] D. Deutsch, A. Ekert, R. Jozsa, C. Macchiavello, S. [24] J. González, M.A. Mart´ın-Delgado, G. Sierra, A.H. Voz- Popescu, A. Sanpera, “Quantum Privacy Amplification mediano, Quantum Electron Liquids and High-Tc Super- and the Security of Quantum Cryptography over Noisy conductivity, Lecture Notes in Physics, Monographs vol. Channels”, Phys. Rev. Lett. 77, 2818-2821 (1996). 38. Springer-Verlag 1995. [13] C. Macchiavello, “On the analytical convergence of the [25] M.A. Mart´ın-Delgado, G. Sierra; “Analytic Formulations QPA procedure”, Phys. Lett. A246, 385 (1998). of the Density Matrix Renormalization Group”; Int. J. [14] C. Macchiavello, “Quantum Privacy Amplification”, in Mod. Phys. A11, 3147-3174 (1996). D. Bouwmeester, A. Ekert, A. Zeilinger (eds.), The [26] S.R. White, “Density-matrix algorithms for quantum physics of quantum information, Springer-Verlag 2000. renormalization groups” Phys. Rev. B48, 10345 (1993). [15] W. Dür, H.-J. Briegel, J.I. Cirac, P. Zoller, “Quantum re- [27] A. Mair, A. Vaziri, G. Weihs, A. Zeilinger, “Entangle- peaters based on entanglement purification”, Phys. Rev. ment of Orbital Angular Momentum States of Photons”, A. 59, 169 (1999). Nature 412, 3123-316 (2001). [16] W. Dür, H.-J. Briegel, J.I. Cirac, P. Zoller, “Quantum [28] H.H. Arnaut, G.A. Barbosa, “Orbital and Intrinsic An- repeaters for communication”, quant-ph/9803056. gular Momentum of Single Photons and Entangled Pairs [17] M. Horodecki, P. Horodecki, “Reduction criterion of sep- of Photons Generated by Parametric Down-Conversion” arability and limits for a class of distillation protocols”, Phys. Rev. Lett. 85, 286 (2000). Phys. Rev. A59, 4206-4216, (1999). [29] S. Franke-Arnold, S.M. Barnett, “Two-photon entangle- [18] G. Alber, A. Delgado, N. Gisin, I. Jex, “Generalized ment of orbital angular momentum states”, Phys. Rev. quantum XOR-gate for quantum teleportation and state A65, 033823 (2002). purificationin arbitrary dimensional Hilbert spaces”, [30] M. Reck, A. Zeilinger, H.J. Bernstein, P. Bertani, “Ex- quant-ph/000802. perimental realization of any discrete unitary operator”, [19] G. Alber, A. Delgado, N. Gisin, I. Jex, “Efficient bipar- Phys. Rev. Lett. 73, 58-61 (1994). tite quantum state purification in arbitrary dimensional [31] M. Zukowski, A. Zeilinger, M.A. Horne, “Realizable Hilbert spaces”, quant-ph/0102035. higher-dimensional two-particle entanglements via multi- [20] H. Bechmann-Pasquinucci, A. Peres, “Quantum Cryp- port beam splitters”, Phys. Rev. A 55, 2564-2579 (1997). tography with 3-State Systems”, Phys. Rev. Lett. 85, [32] G. Molina-Terriza, J.P. Torres, L. Torner, “Management 3313-3316, (2000). of the Angular Momentum of Light: Preparation of Pho- [21] H. Bechmann-Pasquinucci, W. Tittel, “Quantum cryp- tons in Multidimensional Vector States of Angular Mo- tography using larger alphabets”, Phys. Rev. A 61, mentum”, Phys. Rev. Lett. 88, 013601 (2002). 062308 (2000). [33] J. Leach, M.J. Padgett, S.M. Barnett, S. Franke-Arnold, [22] M. Bourennane, A. Karlsson, G. Björk, “Quantum key J. Courtial, “Measuring the Orbital Angular Momentum distribution using multilevel encoding”, Phys. Rev. A 64, of a Single Photon”, Phys. Rev. Lett. 88, 257901 (2002). 012306 (2001).