letters to nature

7. Wootters, W. K. & Zurek, W. H. A single quantum cannot be cloned. Nature 299, 802–803 (1982). surprising as precise specification of a general qubit state requires an
8. Nielsen, M. A. & Chuang, I. J. Quantum Computation and Quantum Information (Cambridge Univ.
Press, Cambridge, 2000).
infinite amount of classical information. Aside from the obvious
9. Gottesman, D. & Chuang, I. L. Demonstrating the viability of universal quantum computation using differences in the various experimental demonstrations, the basic
teleportation and single-qubit operations. Nature 402, 390–393 (1999). teleportation protocol is the same1. Alice is in possession of a qubit
10. Furusawa, A. et al. Unconditional quantum teleportation. Science 282, 706–709 (1998). (here labelled 2) that is in an unknown state jwl2 ; aj " l2 þ bj # l2 ;
11. Nielsen, M. A., Knill, E. & Laflamme, R. Complete quantum teleportation using nuclear magnetic
resonance. Nature 396, 52–55 (1998).
where j # l and j " l denote eigenstates of the qubit in the measure-
12. Barrett, M. D. et al. Quantum teleportation with atomic qubits. Nature (this issue). ment basis. In addition, Alice and Bob each possess one qubit of a
13. Schmidt-Kaler, F. et al. How to realize a universal quantum gate with trapped ions. Appl. Phys. B 77, two-qubit entangled pair that we take to be a singlet jSl1;3 ; j " l1 j # l3
789–796 (2003).
2j # l1 j " l3 (where, for simplicity, we omit normalization factors).
14. Roos, C. F. et al. Bell states of atoms with ultra long lifetimes and their tomographic state analysis.
Phys. Rev. Lett. (in the press) Preprint at khttp://arXiv.org/abs/ physics/0307210l (2003). Therefore, Alice possesses qubits 1 and 2, while Bob holds qubit 3.
15. Massar, S. & Popescu, S. Optimal extraction of information from finite quantum ensembles. Phys. Rev. Alice wishes to transmit the state of qubit 2 to Bob’s qubit using only
Lett. 74, 1259–1263 (1995). classical communication. The initial joint state of all three qubits is
16. Gisin, N. Nonlocality criteria for quantum teleportation. Phys. Lett. A 210, 157–159 (1996).
17. Kielpinski, D., Monroe, C. & Wineland, D. J. Architecture for a large-scale ion-trap quantum
computer. Nature 417, 709–711 (2002). jFl ¼ jSl1;3 ^jwl2 : ð1Þ
18. Gulde, S. et al. Implementation of the Deutsch-Jozsa algorithm on an ion-trap quantum computer.
Nature 412, 48–50 (2003).
19. Hahn, E. L. Spin Echoes. Phys. Rev. 80, 580–594 (1950). This state can be rewritten using an orthonormal basis of Bell states14
jWk l1;2 ðk ¼ 1–4Þ for the first two qubits and unitary transformations
Acknowledgements We thank H. Briegel and P. Zoller for a critical reading of the manuscript. We
U k acting on jwl3 ð¼ aj " l3 þ bj # l3 Þ so that jFl ¼
gratefully acknowledge support by the Austrian Science Fund (FWF), by the European
Commission (QUEST, QUBITS and QGATES networks), by the Institut für S4k¼1 jWk l1;2 ðU k jwl3 Þ: A measurement in the Bell-state basis {jWk l}
Quanteninformation, and by the Los Alamos LDRD Program. This material is based upon work by Alice then leaves Bob with one of the four possibilities U k jwl3 :
supported in part by the US Army Research Office. H.H is funded by the Marie-Curie program of Once Bob learns of Alice’s measurement outcome (through classical
the European Union.
communication), he can recover the original unknown state by
applying the appropriate unitary operator, U 21 k ; to his state
Competing interests statement The authors declare that they have no competing financial
interests.
U k jwl3 : We note that Alice’s Bell-state measurement can be accom-
plished by transforming from the basis {jWk l1;2 } into the measure-
Correspondence and requests for materials should be addressed to R.B. ([email protected]) ment basis {j "" l1;2; j "# l1;2 ; j #" l1;2 ; j ## l1;2 } before the measurement.
Our implementation uses atomic qubits (9Beþ ions) that are
confined in a linear radiofrequency Paul trap similar to that used in
ref. 15. The control electrodes are segmented into eight sections as
shown schematically in Fig. 1, providing a total of six trapping zones
.............................................................. (centred on electrode segments 2 to 7). Potentials applied to these
electrodes can be varied in time to separate ions and move them to
Deterministic quantum teleportation different locations. The qubits are composed of the ground-state
hyperfine levels j " l ; jF ¼ 1; m ¼ 21l and j # l ; jF ¼ 2; m ¼ 22l;
of atomic qubits which are separated by q0 ø 2p £ 1:25 GHz: These states are
coupled through stimulated Raman transitions16–18 from two laser
M. D. Barrett1*, J. Chiaverini1, T. Schaetz1, J. Britton1, W. M. Itano1,
J. D. Jost1, E. Knill2, C. Langer1, D. Leibfried1, R. Ozeri1 & D. J. Wineland1
1
Time and Frequency Division, NIST, Boulder, Colorado 80305, USA
2
Mathematical and Computational Sciences Division, NIST, Boulder, Colorado
80305, USA
* Present address: Department of Physics, University of Otago, PO Box 56, Dunedin, New Zealand
.............................................................................................................................................................................
Quantum teleportation1 provides a means to transport quantum
information efficiently from one location to another, without the
physical transfer of the associated quantum-information carrier.
This is achieved by using the non-local correlations of previously
distributed, entangled quantum bits (qubits). Teleportation is
expected to play an integral role in quantum communication2
and quantum computation3. Previous experimental demon-
strations have been implemented with optical systems that
used both discrete and continuous variables4–9, and with liquid-
state nuclear magnetic resonance10. Here we report uncondi-
tional teleportation5 of massive particle qubits using atomic Figure 1 Schematic representation of the teleportation protocol. The ions are numbered
(9Be1) ions confined in a segmented ion trap, which aids left to right, as indicated at the top, and retain their order throughout. Positions, relative
individual qubit addressing. We achieve an average fidelity of to the electrodes, are shown at each step in the protocol. The widths of the electrodes
78 per cent, which exceeds the fidelity of any protocol that vary, with the width of the separation electrode (6) being the smallest at 100 mm. The
does not use entanglement11. This demonstration is also import- spacing between ions in the same trap is about 3 mm, and laser-beam spot sizes (in
ant because it incorporates most of the techniques necessary traps 5 and 6) at the position of the ions are approximately 30 mm. In step 1 we prepare
for scalable quantum information processing in an ion-trap the outer ions in an entangled (singlet) state and the middle ion in an arbitrary state
system12,13. (equation (1)). Steps 2–4 constitute a measurement in a Bell-basis for ions 1 and 2 (Alice’s
Quantum teleportation1 provides a means for transporting a qubits), teleporting the state of ion 2 onto ion 3 (Bob’s qubit), up to unitary operations
quantum state between two separated parties, Alice and Bob, that depend on the measurement outcomes. In step 5 we invoke these conditional
through the transmission of a relatively small amount of classical operations, recovering the initial state. Interspersed are spin-echo pulses applied in trap 6
information. For the case of a two-state quantum system or ‘qubit’, that protect the state from de-phasing due to fluctuating magnetic fields but do not affect
only two bits of classical information are needed, which seems the teleportation protocol.
NATURE | VOL 429 | 17 JUNE 2004 | www.nature.com/nature
©2004 Nature Publishing Group 737
letters to nature
beams, which are used to implement the single-qubit rotations state of ion 1 to j " l1 ; a subsequent simultaneous detection of ions 1
Rðv; fÞ ¼ cosðv=2ÞI þ i sinðv=2ÞcosðfÞjx þ i sinðv=2ÞsinðfÞjy ; ð2Þ and 2 effectively measures the state of ion 2 with error less than 1%
due to the presence of ion 1.
where I is the identity operator, j x, jy and j z denote the Pauli spin To complete the teleportation, we apply unitary operations to ion
matrices in the {j " l; j # l} basis (k " j ; ð1; 0Þ; k # j ; ð0; 1Þ), v is 3 that depend on the measurement outcomes for ions 1 and 2. We
proportional to the duration of the Raman pulse, and f is the first move ions 1 and 2 to trap 2 and ion 3 to trap 5, where unitary
relative phase between the Raman beams at the position of the ion. operations consisting of a p/2-pulse, R(p/2,p/2), followed by
The Raman beams are also used to generate entanglement between the operators j x, j y, I, j z for the measurement outcomes j "" l1;2 ;
two qubits by implementing the phase gate18 j "# l1;2 ; j #" l1;2 ; j ## l1;2 respectively, are applied. As noted above, the
aj "" l þ bj "# l þ cj #" l þ dj ## l ! aj "" l inclusion of the spin-echo pulses does not fundamentally change
ð3Þ the teleportation protocol; however for fSE ¼ p=2; we must reorder
2 ibj "# l 2 icj #" l þ dj ## l: the operations following the p/2-pulse, R(p/2,p/2), to I, j z, j x, j y
respectively. A complete experiment is about 4 ms in duration,
Our teleportation scheme is shown schematically in Fig. 1. We predominantly due to three elements: the cooling of all three axial
highlight the key elements of the protocol in bold lettering and also modes to the ground state (1 ms), implementing the ion separations
include the auxiliary ‘spin-echo’ pulses15,18 (R(p,f SE)) applied to and movements (2 ms), and the three detection durations (0.6 ms).
ions in trap 6. These pulses are required in the experiment to prevent In the future, use of smaller trap electrodes to speed up ion-
dephasing caused by variations in the ambient magnetic field on a separation and gate operations, coupled with better detection,
timescale longer than the duration between the spin-echo pulses should considerably increase the speed of the teleportation process.
and, with an appropriate choice of f SE, can compensate phase To demonstrate the full protocol we first teleport the basis states
accumulation due to the presence of a static magnetic-field gradi- j " l2 and j # l2 ; and achieve a fidelity of about 80% (78 ^ 3% for
ent. As they do not fundamentally affect the teleportation, we omit j " l and 84 ^ 2% for j # l for the data taken in the same run as that
their effects in the following discussion. shown in Fig. 2). We also perform Ramsey experiments where the
We first prepare the state jSl1;3 ^j # l2 in two steps: starting first p/2 pulse (having a variable phase f) is applied to ion 2
from the state j ### l1;2;3 we combine the gate in equation (3) (starting in the j # l2 state) and the second pulse (with a fixed phase)
applied to ions 1 and 3 with rotations to generate18 the state is applied to ion 3 after the teleportation is implemented. That is,
ðj ## l1;3 2 ij "" l1;3 Þ^j # l2 ; followed by implementing individual R(p/2,f) is applied to ion 2 and R(p/2,f fixed) is applied to ion 3
ion rotations as discussed in ref. 19 to produce jSl1;3 from j ## l1;3 after teleportation is completed. Ramsey fringes obtained in this
2ij "" l1;3 (see methods section). For state jSl1;3 ^j # l2 ; ions 1 and 3 way are shown in Fig. 2. We perform these experiments for ffixed ¼ 0
are in the singlet, which is invariant under a global rotation. and p/2. From this data, we can extract the teleportation fidelities of
Therefore, a global rotation R(v,f)1,2,3 to all three ions rotates the the states j ^ Xl; j ^ Yl; which are eigenstates of the operators j x
middle ion without affecting the singlet state of ions 1 and 3, and and j y respectively. From these fidelities and those for the states j " l
allows us to produce the state of equation (1) for any a and b with and j # l; we determine an average fidelity kFl ¼ 78 ^ 2% for the
appropriate choices of v and f. teleportation process. Furthermore, if we perform the teleportation
To teleport the state of ion 2 to ion 3, we start by implementing a without the conditional operations, the Ramsey fringes disappear
Bell-state measurement on Alice’s qubits, ions 1 and 2. All three ions and the teleportation fidelity drops to 1/2, equivalent to Bob making
are transferred to trap 6 and then separated, with ions 1 and 2 going a random guess for the teleported state. There are three dominant
to trap 5 and ion 3 to trap 7. A phase gate (equation (3)) followed by mechanisms limiting the final fidelity; imperfect preparation of the
a p/2-pulse, R(p/2,0), is then applied to ions 1 and 2 in trap 5. initial state jSl1;3 ^j # l2 ; imperfections in the second phase gate due
Our previous experiments15 showed a significant amount of to heating accrued in the separation process, and dephasing of the
motional-mode heating during the separation process, and the teleported state due to fluctuating magnetic fields. We have inves-
separation was achieved with only a 95% success rate. Aided by a tigated these issues in independent experiments and find that each
smaller separation electrode in the current trap, we can separate the results in a loss of 8 ^ 2% in the fidelity of the final state, consistent
ions in the desired manner with no detectable failure rate. More
importantly, the heating has been significantly reduced and we find
that after the separation, the stretch mode of the two ions in trap 5 is in
the ground state and the centre-of-mass mode has a mean quantum
number of about 1. This enables us to implement the phase gate
(equation (3)) between ions 1 and 2 with fidelity greater than 90% and
without the need for sympathetic recooling12,13,17. Ideally (and in the
absence of the spin-echo pulses) this leaves the ions in the state
j "" l1;2 ^Rðp=2; 2p=2Þjx jwl3 þ j "# l1;2 ^Rðp=2; 2p=2Þjy jwl3
þij #" l1;2 ^Rðp=2; 2p=2ÞIjwl3 2 j ## l1;2 ^Rðp=2; 2p=2Þjz jwl3 ;

ð4Þ
where jwl3 ¼ aj " l3 þ bj # l3 :
To complete the Bell-state measurement, we need to detect the
states of ions 1 and 2 individually. We recombine all three ions in
trap 6 and separate them, with ion 1 being transferred to trap 5 and
ions 2 and 3 transferred to trap 7. Detection on ion 1 is then Figure 2 Ramsey fringes demonstrating the teleportation protocol. The two curves
achieved through state-dependent resonance fluorescence measure- correspond to the second Ramsey pulse having ffixed ¼ 0 (circles) and ffixed ¼ p=2
ments19 (j # l strongly fluoresces whereas j " l does not), after which (triangles) as discussed in the text. We plot the probability P #,3 of observing ion 3 in the
we optically pump the ion back to the state j # l1 : All three ions are j # l3 state versus the phase of the first Ramsey pulse. Solid curves are best-fit sinusoidal
then recombined in trap 6 and separated again. For this separation, functions to the data. The oscillations of the Ramsey fringes have an amplitude jr#" j
ions 1 and 2 are transferred to trap 5 and ion 3 is returned to trap 7. where r#" ¼ ðr"# Þ* is the off-diagonal element of the density matrix of the teleported
As the most recent spin-echo pulse applied in trap 6 transferred the state. The fidelity of the teleported state is then given by F ¼ 1=2 þ jr#" j:
738 ©2004 Nature Publishing Group NATURE | VOL 429 | 17 JUNE 2004 | www.nature.com/nature
 letters to nature
with the quoted result for teleportation. In the methods section, we described in the text. Furthermore, imperfect state preparation of ion 2 contributes to an
overall loss in fidelity. In a similar way, by simulation, we have also examined the effects of
discuss the effects of imperfections in the state preparation and likely imperfections in the teleporting process and conclude that these imperfections only
teleportation process. reduce the measured value of kFl: Therefore, the experimentally measured value of kFl
The average fidelity of 78 ^ 2% achieved by our implemented indicates that entanglement between bits 1 and 3 was required.
quantum teleportation using atomic qubits, exceeds the value 2/3 Received 17 March; accepted 4 May 2004; doi:10.1038/nature02608.
necessary to establish the presence of entanglement11, and is accom-
1. Bennett, C. H. et al. Teleporting an unknown quantum state via dual classical and Einstein-Podolsky-
plished on demand and without post-selection of the data. Although Rosen channels. Phys. Rev. Lett. 70, 1895–1899 (1993).
teleportation has been demonstrated in other systems, our demon- 2. Briegel, H.-J., Dür, W., Cirac, J. I. & Zoller, P. Quantum repeaters: The role of imperfect local
stration incorporates the protocol into a simple experiment in such a operations in quantum communication. Phys. Rev. Lett. 81, 5932–5935 (1998).
3. Gottesman, D. & Chuang, I. L. Demonstrating the viability of universal quantum computation using
way that it can be viewed as a subroutine of a quantum algorithm, teleportation and single-qubit operations. Nature 402, 390–393 (1999).
here, a Ramsey experiment using two separated qubits. Furthermore, 4. Bouwmeester, D. et al. Experimental quantum teleportation. Nature 390, 575–579 (1997).
our demonstration incorporates most of the important features 5. Furusawa, A. et al. Unconditional quantum teleportation. Science 282, 706–709 (1998).
required for large-scale quantum information processing using 6. Boschi, D., Branca, S., De Martini, F., Hardy, L. & Popescu, S. Experimental realization of teleporting
an unknown pure quantum state via dual classical and Einstein-Podolsky-Rosen channels. Phys. Rev.
trapped ions12,13: We (a) reliably select qubits from a group and Lett. 80, 1121–1125 (1998).
move them to separate trap zones while maintaining their entangle- 7. Kim, Y.-H., Kulik, S. P. & Shih, Y. Quantum teleportation of a polarization state with a complete Bell
ment, (b) manipulate and detect qubits without the need for strongly state measurement. Phys. Rev. Lett. 86, 1370–1373 (2001).
focused laser beams, and (c) perform quantum logic operations 8. Bowen, W. P. et al. Experimental investigation of continuous-variable quantum teleportation. Phys.
Rev. A. 67, 032302 (2003).
conditioned on ancilla measurement outcomes. Finally, we note 9. Zhang, T. C., Goh, K. W., Chou, C. W., Lodahl, P. & Kimble, H. J. Quantum teleportation of light
that the University of Innsbruck has also implemented teleportation beams. Phys. Rev. A 67, 033802 (2003).
with the use of three Caþ ions in a linear Paul trap20. A 10. Nielsen, M. A., Knill, E. & Laflamme, R. Complete quantum teleportation using nuclear magnetic
resonance. Nature 396, 52–55 (1998).
11. Massar, M. & Popescu, S. Optimal extraction of information from finite quantum ensembles. Phys.
Methods Rev. Lett. 74, 1259–1263 (1995).
State preparation 12. Wineland, D. J. et al. Experimental issues in coherent quantum-state manipulation of trapped atomic
ions. J. Res. Natl Inst. Stand. Technol. 103, 259–358 (2003).
Following the protocol outlined in ref. 18, three ions held in trap 5 are first laser-cooled,
13. Kielpinski, D., Monroe, C. & Wineland, D. J. Architecture for a large-scale ion-trap quantum
leaving all three axial modes in the ground state with 99% efficiency21. The internal states
computer. Nature 417, 709–711 (2002).
are then initialized to j ### l by optical pumping. A motional phase-space ‘displacement’
14. Nielsen, M. A. & Chuang, I. L Quantum Computation and Quantum Information (Cambridge Univ.
pulse, inserted in a spin-echo sequence, is applied to the stretch mode of ions 1 and 3 by
Press, Cambridge, 2000).
Raman beams that have a relative detuning of Dq ¼ qs þ d; where q s is the frequency of
15. Rowe, M. A. et al. Transport of quantum states and separation of ions in a dual rf ion trap. Quant. Inf.
the stretch mode and d is a small detuning as described in ref. 18. The displacement pulse Comp. 2, 257–271 (2002).
implements the transformation in equation (3) to ions 1 and 3. The spin-echo sequence 16. Wineland, D. J. et al. Quantum information processing with trapped ions. Phil. Trans. R. Soc. Lond. A
acts on all three ions: a p/2-pulse, a delay T s in which the displacement pulse acts, a 361, 1349–1361 (2003).
p-pulse, an equal time delay Ts, and a final p/2-pulse. As the amplitude of motion for the 17. Barrett, M. D. et al. Sympathetic cooling of 9Beþ and 24Mgþ for quantum logic. Phys. Rev. A 68,
middle ion is zero for the stretch mode, the entangling displacement pulse as described has 042302 (2003).
no effect on this ion. Thus the spin-echo pulse sequence leaves the middle ion in the state 18. Leibfried, D. et al. Experimental demonstration of a robust, high-fidelity geometric two ion-qubit
j # l; while the outer ions are affected as described in ref. 18. For an appropriate choice of phase gate. Nature 422, 412–415 (2003).
detuning, d, and pulse duration, T ¼ 2p=d; we produce the state ðj ## l1;3 2 ij "" l1;3 Þ^j # l2 19. Rowe, M. A. et al. Experimental violation of a Bell’s inequality with efficient detection. Nature 409,
(in the experiment, T ¼ 9:6 ms). 791–794 (2001).
To create the state jSl1;3 ^j # l2 ; a second spin-echo sequence is applied with the phase 20. Riebe, M. et al. Deterministic quantum teleportation with atoms. Nature (this issue).
of the pulses shifted by p/4 with respect to the previous sequence. In addition, for the 21. King, B. E. et al. Cooling the collective motion of trapped ions to initialize a quantum register. Phys.
duration of the middle p-pulse, the axial confinement is changed. The resulting change in Rev. Lett. 81, 1525–1528 (1998).
each ion’s position gives rise to a relative phase of 2p/4, 0, and þp/4 for ions 1, 2 and 3,
respectively19. For ion 2, the phase of the p-pulse is the same as the first and last p/2-pulses Acknowledgements This work was supported by ARDA/NSA and NIST. We thank J. Bollinger
and the sequence leaves the ion in the state j # l2 : The outer two ions are initially in the state and J. Martinis for helpful comments on the manuscript. T.S. acknowledges a Deutsche
j ## l1;3 2 ij "" l1;3 and the first p/2-pulse yields the state j #" l1;3 þ j "# l1;3 : The p-pulse, with Forschungsgemeinschaft research grant. This paper is a contribution of the National Institute of
the p/2 phase difference on ions 1 and 3, results in the singlet state j "# l1;3 2 j #" l1;3 : This Standards and Technology and is not subject to US copyright.
state, being invariant under a global rotation, remains unchanged by the final p/2-pulse.
The three-ion state is then the desired state jSl1;3 ^j # l2 : Auxiliary experiments establish a
singlet fidelity F S ¼ Tr2 ð1;3 kSjrexp jSl1;3 Þ < 0:92ð1Þ and a fidelity for the initial state of ion 2 Competing interests statement The authors declare that they have no competing financial
of Tr1;3 ð2 k # jrexpj # l2 Þ < 0:95ð1Þ: interests.

Correspondence and requests for materials should be addressed to D.J.W. ([email protected])

Imperfect state preparation and teleportation operations
As the operations in the experiment are imperfect, we must examine these effects to show
that the observed teleportation fidelity that exceeds 2/3 could not be caused by these
imperfections. In particular, as the initial state of ions 1 and 3 is not a perfect singlet, when
we rotate all three qubits to prepare the state jwl2 ¼ aj " l2 þ bj # l2 ; we could potentially
encode this information onto bit 3 even before the teleportation process is started. We
address this issue as follows: We can verify that the gate in the state preparation does not
..............................................................
entangle bit 2 with bits 1 and 3. Therefore the most general input state is given by
jWlinitial ¼ ða0 j ## l1;3 þ a1 j #" l1;3 þ a2 j "# l1;3 þ a3 j "" l1;3 Þ^ðb0 j # l2 þ b1 j " l2 Þ; ð5Þ
In situ observation of colloidal
where, ideally, a1 ¼ 2a2 ¼ 221=2 and b0 ¼ 1: Here we assume that the subsequent rotation
to prepare jwl2 and the teleportation operations are perfect. However, in the experiment,
monolayer nucleation driven
after preparation of jWlinitial ; we switch Raman beams from a geometry where the beams
propagate at 908 to each other to a geometry where the beams co-propagate. Owing to by an alternating electric field
fluctuating differences in the optical path lengths between these two sets of beams, there is
a random phase difference from experiment to experiment between the corresponding Ke-Qin Zhang & Xiang Y. Liu
Raman pulses. Over many experiments, some of the coherence terms between all three
qubits average out due to this phase randomization, leading to a simplification of the Department of Physics, National University of Singapore, 2 Science Drive 3,
calculated average fidelity with respect to the imperfections in state preparation.
The average fidelity kFl is evaluated from the expression
Singapore 117542
.............................................................................................................................................................................
kFl ¼ ðFðj " lÞ þ Fðj # lÞ þ Fðj þ XlÞ þ Fðj 2 XlÞ þ Fðj þ YlÞ þ Fðj 2 YlÞÞ=6; ð6Þ
The nucleation of crystalline materials is a hotly debated subject
where the arguments correspond to the initial state, jwl2 . We find
in the physical sciences1. Despite the emergence of several
2 2
kFl ¼ ð2 2 jb0 j Þ=3 þ 2ð2jb0 j 2 1ÞF s =3: ð7Þ theories in recent decades2–7, much confusion still surrounds
When F s # 1=2; kFl # 2=3; and we conclude that the teleportation fidelity cannot the dynamic processes of nucleation5–7. This has been due in part
exceed 2/3 unless the teleportation has occurred through the entangled singlet state as to the limitations of existing experimental evidence. Charged
NATURE | VOL 429 | 17 JUNE 2004 | www.nature.com/nature
©2004 Nature Publishing Group 739