Abstract
Entangled quantum states in highdimensional space show many advantages compared with entangled states in twodimensional space. The former enable quantum communication with higher channel capacity, enable more efficient quantuminformation processing and are more feasible for closing the detection loophole in Bell test experiments. Establishing highdimensional entangled memories is essential for longdistance communication, but its experimental realization is lacking. We experimentally established highdimensional entanglement in orbital angular momentum space between two atomic ensembles separated by 1 m. We reconstructed the density matrix for a threedimensional entanglement and obtained an entanglement fidelity of (83.9±2.9)%. More importantly, we confirmed the successful preparation of a state entangled in more than threedimensional space (up to sevendimensional) using entanglement witnesses. Achieving highdimensional entanglement represents a significant step toward a highcapacity quantum network.
Similar content being viewed by others
Introduction
Quantum entanglement distributed in different nodes is essential for realizing longdistance quantum communications^{1}. By introducing a quantum repeater protocol, the problem of the exponential scaling of the error rates with channel length in quantum communication can be overcome by using entanglement storage and swapping operations^{2}. Usually, the stored photons are encoded in a twodimensional space, such as polarization, whichpath and timebin, which results in information being carried by a photon as a qubit. Recently, another photonic degree of freedom, orbital angular momentum (OAM)^{3, 4, 5}, has attracted much interest because the OAM states of a photon could belong to a highdimensional space, which would enable encoding with inherent infinite degrees of freedom, thereby enhancing the channel capacity and significantly improving the efficiency of a network^{6, 7}.
Constructing quantum networks based on OAM involves the preparation of highdimensional entangled photons and the realization of highdimensional entangled memories. The repeaters for a highdimensional quantum communication network are based on many highdimensional entangled memory modes, and longdistance communication can be realized via a swapping operation between adjacent entangled memories. Thus, establishing highdimensional entangled memories is a critical step. Progress has been made in preparing highdimensional OAM entangled states, including the confirmation of a threedimensional entanglement between a delayed atomic spin wave and a photon^{8} prepared in spontaneous Raman scattering (SRS), an 11dimensional entanglement^{9}, a highdimensional image entanglement^{10} and a 100 × 100 OAM entanglement^{11} between two photons generated by spontaneous parametric downconversion. In recent years, many groups have explored the realization of highdimensional entangled memories with some success^{12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23}, but all work to date has only displayed approaches to this goal. The experimental works on establishing entangled memories using different physical systems have focused on the entanglement in a twodimensional space^{24, 25, 26, 27, 28, 29}. Until now, the storage of a highdimensional OAM state ondemand in any physical system, which is essential for the realization of quantum repeaters for longdistance quantum communications, and the establishment of a highdimensional OAM entanglement between different quantum memories have not been reported. Trying to solve this problem is the main motivation for this work. A previous work^{24} reported the creation of a twodimensional OAM entanglement between two atomic ensembles, but the realization of a highdimensional entanglement between different quantum memories is nontrivial and is not a straightforward extension of establishing a twodimensional entanglement between two atomic ensembles. There are many challenges, such as in proving the highdimensional entanglement and determining the dimensionality of entanglement. Usually, we can characterize storage based on entanglement fidelity by comparing the density matrixes before and after storage, which are reconstructed via quantum tomography. For a highdimensional entangled state, the number of measurements required to reconstruct a density matrix scales quadratically with dimensionality. Therefore, reconstructing the density matrices for a highdimensional (>3) entangled state is impractical because the amount of data that must be measured increases significantly. Moreover, balancing the distinct efficiencies when creating and storing different OAM modes in a high dimension is a challenge, which becomes more serious with increasing dimensions.
In this study, we report the experimental establishment of highdimensional OAM entanglement between two quantum memories, making a primary step toward the building of a highdimensional quantum network. In our experiment, we entangled the highdimensional OAM states of a photon and a collective spinexcited state in a cold atomic ensemble via SRS^{8} and then sent the photon to be stored in another cold atomic ensemble using the Raman protocol^{30}. In this way, we established highdimensional OAM entanglement between two atomic ensembles. We confirmed the entanglement by mapping the spinexcited states in the two ensembles to two photons and checked their entanglement. We reconstructed the density matrices of the threedimensional OAM entangled photons. The entanglement fidelity was calculated to be (83.9±2.9)%. We then used an entanglement witness to characterize the nature of the higher dimensional entanglement and concluded that there is at least a fourdimensional entanglement within the two memories. Using the dimensionality witness, we confirmed a sevendimensional entanglement between these memories.
Materials and methods
The experimental media were optically thick atomic ensembles of Rubidium 85 (^{85}Rb) that were trapped in two twodimensional magnetooptical traps (MOT)^{30} and were separated by 1 m but worked independently. A schematic of the energy levels involved and the experimental setup are shown in Figure 1a and 1b. By inputting Gaussian pulse modes with a pulse width of 30 ns, we generated in MOT 1 via SRS an OAM entanglement between the signal1 photon and the collective spinexcited state of the atomic ensemble. In this process, the laser pulse of pump 1 was bluedetuned by 70 MHz to the atomic transition and signal1 was bluedetuned by 70 MHz to the atomic transition . Because this is a SRS process that conserves momentum, the initial state of the system has zero linear and angular momentum; thus, the resulting joint state of the signal1 photon and the atomic spinexcited state has zero total angular momentum, which induces OAM correlations between them. The established entanglement is written as , where m〉 denotes the OAM state of the m quantum eigenmode; the subscripts s1 and a1 refer to the signal1 photon and the atomic ensemble in MOT 1, respectively; and c_{m}^{2} is the excitation probability for different OAM modes. Next, the signal1 photons were sent to be stored as a collective spinexcited state of the atomic ensemble trapped in MOT 2 using the Raman scheme^{31}. Hence, the two atomic ensembles in different MOTs are in a highdimensional entanglement of , where o_{m}^{2} is the amplitude probability for the different modes m and subscript a2 refers to the atomic collective spinexcited states in MOT 2. The power of the coupling laser is 40 mW, with a beam waist of 2 mm, corresponding to a Rabi frequency of 10.6 Γ (Γ is the decay rate of level 5P_{1/2}(F′=3)). The storage time T_{2} in MOT 2 should not be longer than the storage time of the spin wave T_{1} in MOT 1 to demonstrate the storage of the entanglement. In our experiment, the condition T_{2}=T_{1}=100 ns was taken, that is, the coupling laser and the pump2 laser were opened at the same time. Furthermore, we applied the method detailed in ref. 31 to match the bandwidth between the signal1 photons and the memory for high storage efficiency, thus enabling a storage efficiency of 26.8% to be achieved experimentally. The memory efficiency was calculated as the ratio of the coincidence counts before and after storage. After a delayed time of 100 ns, we used the pump2 laser with a square pulse width of 250 ns, resonant with the atomic transition 1>(5S_{1/2}(F=2))→4>(5P_{3/2}(F′=3)), to read the spin wave to generate the signal2 photon, which corresponds to the transition 4>(5P_{3/2}(F′=3))→3>(5S_{1/2}(F=3)). Simultaneously, we switched the coupling laser on to read the signal1 photon emitted out of the atomic ensemble in MOT 1. The delay time was measured by comparing the peaks of the twophoton coincidence with and without delayed pumps 1 and 2. Pumps 1 and 2 had a power of 0.5 and 4 mW, respectively. A check of the entanglement between the signal1 and signal2 photons verified the highdimensional entanglement between the atomic ensembles.
Results and discussion
Before verifying the highdimensional entanglement, the correlations in the OAM space between the two ensembles before and after storage were first measured. The results are shown in Figure 2, with panels a and b showing the OAM correlations between photons of signal 1 and signal 2 without/with storing signal 1. This also characterizes the correlations between the signal1 photon and the atomic collective spinexcited state in MOT 1. The differences in OAM correlations shown in Figure 2a arise from different OAM distributions in the nonlinear SRS process (Figure 2c). Because the efficiencies were different when storing the various OAM modes m>, the resulting correlation matrix after storage (Figure 2b) was slightly different from that before storage (Figure 2a); this is similar to the result that was obtained when using a weak coherent light^{6}. The different storage efficiencies measured for different OAM modes m are shown in Figure 2d.
Next, we verified the entanglement with d=3. We projected the signal1 and signal2 photons onto SLM 1 and SLM 2, respectively, with nine different phase states ψ_{1−9}> corresponding to states L>, G>, R>, (G>+L>)/2^{1/2}, (G>+R>)/2^{1/2}, (G>+iL>)/2^{1/2}, (G>iR>)/2^{1/2}, (L>+R>)/2^{1/2} and (L>+iR>)/2^{1/2} (refs 6, 32), where L>, G> and R> are states corresponding to a welldefined OAM of −ћ, 0 and +ћ, respectively. With this projection, the mirrors and lens should be considered due to the photon’s transformation during the imaging process. We reconstructed the density matrix before storage (Figure 3a and 3b) by converting the spinexcited state in MOT 1 into a signal2 photon. The signal1 photon was then stored for a while in MOT 2 following a Raman protocol. This established the entanglement between the two atomic ensembles, for which the reconstructed density matrix is given in Figure 3c and 3d. The difference between the states before and after storage was due to the different storage efficiencies for the R> (L>) mode and G> mode (Figure 2d). To check the entangled state before and after storage, we used pump 2 and the coupling lasers to read the atom–atom entangled state into the retrieved signal 1–signal 2 entangled state and then checked the entanglement between these photons. The reconstructed density matrices are shown in Figure 3a–3d. Using the formula , where x represents the input and output and ρ_{ideal} is the density matrix of the ideal threedimensional OAM entangled state of Ψ_{ideal}>=(R>_{a1}R>_{a2}+G>_{a1}G>_{a2}+L>_{a1}L>_{a2})/3^{1/2}, we calculated the fidelity of the reconstructed density matrix before and after storage, which was (76.7±2.8)% and (71.7±2.8)%, respectively. All error bars in this experiment were estimated using Poisson statistics and performing Monte Carlo simulations using Mathematica software. Both exceeded the threshold of 2/3 (refs 8, 33) for a maximally entangled state of Schmidt rank 3, which confirms that the density matrix cannot be decomposed into an ensemble of pure states of Schmidt rank 1 or 2, that is, the Schmidt number of the density matrix must be equal to or >3 both before and after storage. We calculated the fidelity of entanglement with , which yielded (83.9±2.9)%.
Finally, we focused on the main part of this study, that is, the establishment of a highdimensional entanglement between two memories. In principle, the density matrices of the higherdimensional entanglement can be reconstructed using the above method, but in practice, there are some experimental challenges in its realization. For example, for a ddimensional entangled state, the amount of data needed is of the order d^{4}, which makes the reconstruction of the density matrix impractical. Basically, there are three methods for checking whether a system is in highdimensional entanglement: (1) using unbiased basis states that span the whole subspace^{34, 35}; (2) checking inequalities in higher dimensions directly^{36, 37}; or (3) finding a violation that is stronger than allowed within a twodimensional state space, thereby hinting at entanglement in (untested) higher dimensions. Here we used method 3 to characterize the entanglement. We used the entanglement witness^{38, 39} to prove whether there was a highdimensional entanglement and the dimensionality witness^{40, 41, 42} to characterize the dimensionality of the entanglement. These witnesses determine the level of entanglement using a minimum number of measurements. To calculate the witnesses, we only need to measure all of the states entangled in a twodimensional OAM subspace, that is, the correlations in three mutually unbiased bases, including diagonal/antidiagonal, left/right and horizontal/vertical bases, need to be measured, with the number of data points needed being reduced to 3d(d−1) (ref. 43). For example, to reconstruct the density matrix, the amount of data that must be measured for 3D is 81, and for 4D, it is 256. If we use the witness, then the amount of data needed is reduced significantly to 18 and 36, respectively. This greatly shortens the experimental time. The entanglement and dimensionality witnesses can be calculated from the sum of the visibilities M=V_{x}+V_{y} and N=V_{x}+V_{y}+V_{z}, respectively, in each 2 × 2 subspace, where the visibilities are defined as , i=x,y,z. Here σ_{x}, σ_{y}, σ_{z} represent the measurements in the diagonal/antidiagonal, left/right and horizontal/vertical bases, respectively. The superposition is calculated by adding equal amounts of the two modes, and the phase is calculated just from the argument of the resultant complex^{44}. Figure 4a shows an example of mutually unbiased bases formed from the OAM modes m=5 and m=−1. For a separable state within a ddimensional subspace, a product state of a (d−1)dimensional maximum entangled state and a single state maximizes the sum of the visibilities. Because the allowed maximum visibility of entanglement in a twodimensional subspace is 2 (M=V_{x}+V_{y}=2,V_{x}=V_{y}=1), the allowed maximum visibilities can be calculated as (d−1)(d−2) for a (d−1)dimensional entanglement. The maximum visibilities for the remaining separable state are (d−1) (ref. 39). Hence, the maximum bound for highdimensional entanglement is given as
If there is a ddimensional entanglement, the maximum bound of M_{d} should be violated. For a state comprising m=2, 1, 0, −1, the maximum bound is M_{4}=9. The measured M′ is 9.30±0.06 and 9.19±0.06 before and after storage, respectively. These values clearly suggest that there is at least a fourdimensional entanglement between these distant memories.
By assuming the correlations , as in ref. 11, we also sum the visibilities N for each of the bases to calculate a witness value W to determine the dimensionality of highdimensional entanglement. All experimentally measured visibilities N are shown in Figure 4b and 4c, corresponding to quantities before and after storage, respectively. The dimensional witness value^{11} is given by
where D is the number of OAM modes in the measurement. If W>W_{d} holds, the memories are entangled in at least d+1 dimensions. In our experiment, the measured number of modes was 11 (m=−5→5); the obtained W of 123.9±0.8 for the input state and 112.8±0.8 for the output state violated the bound of 110 for an input of d=6 and 99 for an output of d=5, both by 17 s.d.’s, implying that there strongly exists a sixdimensional entanglement between the memories. The obtained W also violated the bound of 121 for an input of d=7 and 110 for an output of d=6, both by 3 s.d’s, demonstrating that there exists a sevendimensional entanglement and indeed a highdimensional entanglement in our memories.
In demonstrating threedimensional entanglement, the fidelity was affected by the distinct storing efficiencies (Figure 2d) for different OAM modes^{6}, narrowing the spiral bandwidth of the OAM modes. We can improve the entanglement fidelity by purifying the entangled state^{10, 45}. However, achieving a balance in storing different OAM modes is a big challenge^{46}. In all experiments, entanglement was verified by checking the entangled photon readout from the atomic ensembles, which were found to be aposteriori entangled. Both signal photons were completely covered by the pump and coupling laser beams in our experiment; hence, it is reasonable to assume that the readout efficiencies for different OAM modes are the same. Hence, the photonic entangled state can be regarded as a postselected entangled state of the atomic ensembles.
The main reasons affecting the entanglement dimension are as follows: (1) the distinct efficiencies in storing the different OAM modes; (2) the noise associated with storage, which is mainly from the scattering from the coupling laser, resulting in a low signaltonoise ratio (SNR); (3) the stability of the whole system over long experimental periods (the total measured time for the storage and retrieval processes was ~100 h). To increase the OAM entanglement dimension, we believe four problems need to be solved: (a) generating a maximal highdimensional entanglement between the signal1 photon and the spin wave using purification, as done in ref. 9; (b) balancing the storage efficiency for different OAM modes by, for example, smoothing the transverse distribution of the coupling laser beam in the cold atomic ensemble and the atomic density; (c) reducing the background noise by using more strict filtering to achieve a better SNR; (d) improving the working state of the whole system over long experimental periods.
There are many limiting factors for atomic storage time, including the residual magnetic field and atomic motion. In general, memory time can be improved by compensating for the magnetic field or by using magnetic fieldinsensitive states. By reducing atomic motion with an optical lattice, a millisecond, even a hundred millisecond storage time can be achieved. Moreover, the dynamic decoupling method can also be used to improve the storage time. In the present experiment, the storage time was also limited by the experimental time sequence, which was performed within hundreds of nanoseconds. The storage time can be improved further by optimizing the time sequence.
We emphasize that achieving highdimensional entanglement between different quantum memories is nontrivial and is not a straightforward extension of establishing a twodimensional entanglement between two atomic ensembles. There are many challenges both in its creation and verification. For distant memories entangled in a twodimensional space^{24}, the entanglement can be well characterized by reconstructing the density matrix and checking the Belltype inequality. In contrast, characterizing a highdimensional entanglement is more complex and difficult. As noted earlier, reconstructing the density matrices using the method for twodimensional entanglement is impractical for a highdimensional entanglement (>3) because the amount of data that must be measured significantly increases. Therefore, we sought a different way to characterize it, using a witness instead. Moreover, it is not easy to balance the distinct efficiencies when storing different OAM modes. This is not a problem in the twodimensional case because any two OAM modes with the same value but opposite sign serve as a twodimensional space. Another significant difference is that a series of 4F imaging systems have been designed subtly and constructed meticulously to detect the highdimensional entanglement, instead of the direct projection onto SLM for a low azimuthal index of ±1 used in ref. 24; otherwise, we could not obtain the correct results for reconstructing the density matrix and calculating the witness. We believe these are the main reasons why there have been no reports of experimental progress until now. Indeed, in one important aspect, that is, the storage of a highdimensional OAM state on demand in any physical system and the establishment of a highdimensional OAM entanglement between different quantum memories, our work represents primary progress and a significant step forward in this field.
Lightcarried OAM cannot be transmitted in a commercial optical fiber; therefore, OAMbased quantum networks may be more suitable for work in a free space system. Recently, Zeilinger’s group realized the distribution of OAM entanglement between two sites separated by 3 km in Vienna in 2015 (ref. 47). We also note that light with OAM can transmit along some special fibers over kilometers^{48}. Currently, many groups and people are working in this field; therefore, a quantum network based on OAM may be realized in the future.
Moreover, we note that ref. 49 reported a quantum storage of a threedimensional entanglement in solid crystal. Compared with that work, in addition to the media for memory being different, our work reports a higherdimensional entanglement storage. Most importantly, the memory we achieved can work ondemand; this is the key point for realizing a longdistance quantum communication based on a quantum repeater. The technique of a twolevel atomic frequency comb used in that work cannot work ondemand; the storage time is predetermined. There is another difference between these two works: we achieved highdimensional entanglement between two physical systems, not the OAM entanglement between a delayed photon and the atomic excitation achieved in solid storage.
Conclusions
In summary, quantum memories entangled in highdimensional space between two 1mdistant atomic ensembles were experimentally established for the first time. The density matrices for the threedimensional entanglement were reconstructed, giving (83.9±2.9)% entanglement fidelity. For a higherdimensional case, we proved that at least a fourdimensional entanglement existed between two memories using an entanglement witness. After verifying the dimension of the entangled memories, the experimental data showed that there was a sevendimensional entanglement within the two atomicensemble memories. The experiment to establish highdimensional entangled memories is an important step toward highdimensional quantum communications.
References
Kimble HJ . The quantum internet. Nature 2008; 453: 1023–1030.
Briegel HJ, Dür W, Cirac JI, Zoller P . Quantum repeaters: the role of imperfect local operations in quantum communication. Phys Rev Lett 1998; 81: 5932.
Mair A, Vaziri A, Weihs G, Zeilinger A . Entanglement of the orbital angular momentum states of photons. Nature 2001; 412: 313–316.
FrankeArnold S, Allen L, Padgett M . Advances in optical angular momentum. Laser Photon Rev 2008; 2: 299–313.
Yao AM, Padgett MJ . Orbital angular momentum: origins, behavior and applications. Adv Opt Photon 2011; 3: 161–204.
Ding DS, Zhang W, Zhou ZY, Pan JS, Xiang GY et al. Toward highdimensionalstate quantum memory in a cold atomic ensemble. Phys Rev A 2014; 90: 042301.
Wang J, Yang JY, Fazal IM, Ahmed N, Yan Y et al. Terabit freespace data transmission employing orbital angular momentum multiplexing. Nat Photon 2012; 6: 488–496.
Inoue R, Yonehara T, Miyamoto Y, Koashi M, Kozuma M . Measuring qutritqutrit entanglement of orbital angular momentum states of an atomic ensemble and a photon. Phys Rev Lett 2009; 103: 110503.
Dada AC, Leach J, Buller GS, Padgett MJ, Andersson E . Experimental highdimensional twophoton entanglement and violations of generalized Bell inequalities. Nat Phys 2011; 7: 677–680.
Edgar MP, Tasca DS, Izdebski F, Warburton RE, Leach J et al. Imaging highdimensional spatial entanglement with a camera. Nat Commun 2012; 3: 984.
Krenn M, Huber M, Fickler R, Lapkiewicz R, Ramelow S et al. Generation and confirmation of a (100 × 100)dimensional entangled quantum system. Proc Natl Acad Sci USA 2014; 111: 6243–6247.
Moretti D, Felinto D, Tabosa JWR . Collapses and revivals of stored orbital angular momentum of light in a coldatom ensemble. Phys Rev A 2009; 79: 023825.
Vudyasetu PK, Camacho RM, Howell JC . Storage and retrieval of multimode transverse images in hot atomic rubidium vapor. Phys Rev Lett 2008; 100: 123903.
Shuker M, Firstenberg O, Pugatch R, Ron A, Davidson N . Storing images in warm atomic vapor. Phys Rev Lett 2008; 100: 223601.
Wu JH, Liu Y, Ding DS, Zhou ZY, Shi BS et al. Light storage based on fourwave mixing and electromagnetically induced transparency in cold atoms. Phys Rev A 2013; 87: 013845.
Ding DS, Wu JH, Zhou ZY, Shi BS, Zou XB et al. Multiple image storage and frequency conversion in a cold atomic ensemble. Phys Rev A 2013; 87: 053830.
Heinze G, Rudolf A, Beil F, Halfmann T . Storage of images in atomic coherences in a rareearthiondoped solid. Phys Rev A 2010; 81: 011401(R).
Ding DS, Wu JH, Zhou ZY, Liu Y, Shi BS et al. Multimode image memory based on a cold atomic ensemble. Phys Rev A 2013; 87: 013835.
Higginbottom DB, Sparkes BM, Rancic M, Pinel O, Hosseini M et al. Spatialmode storage in a gradientecho memory. Phys Rev A 2012; 86: 023801.
Glorieux Q, Clark JB, Marino AM, Zhou ZF, Lett PD . Temporally multiplexed storage of images in a gradient echo memory. Opt Express 2012; 20: 12350–12358.
Veissier L, Nicolas A, Giner L, Maxein D, Sheremet AS et al. Reversible optical memory for twisted photons. Opt Lett 2013; 38: 712–714.
Ding DS, Zhou ZY, Shi BS, Guo GC . Singlephotonlevel quantum image memory based on cold atomic ensembles. Nat Commun 2013; 4: 2527.
Nicolas A, Veissier L, Giner L, Giacobino E, Maxein D et al. A quantum memory for orbital angular momentum photonic qubits. Nat Photon 2014; 8: 234–238.
Ding DS, Zhang W, Zhou ZY, Shi S, Xiang GY et al. Quantum storage of orbital angular momentum entanglement in an atomic ensemble. Phys Rev Lett 2015; 114: 050502.
Saglamyurek E, Sinclair N, Jin J, Slater JA, Oblak D et al. Broadband waveguide quantum memory for entangled photons. Nature 2011; 469: 512–515.
Choi KS, Deng H, Laurat J, Kimble HJ . Mapping photonic entanglement into and out of a quantum memory. Nature 2008; 452: 67–71.
Zhang H, Jin XM, Yang J, Dai HN, Yang SJ et al. Preparation and storage of frequencyuncorrelated entangled photons from cavityenhanced spontaneous parametric downconversion. Nat Photon 2011; 5: 628–632.
Clausen C, Usmani I, Bussières F, Sangouard N, Afzelius M et al. Quantum storage of photonic entanglement in a crystal. Nature 2011; 469: 508–511.
Dai HN, Zhang H, Yang SJ, Zhao TM, Rui J et al. Holographic storage of biphoton entanglement. Phys Rev Lett 2012; 108: 210501.
Liu Y, Wu JH, Shi BS, Guo GC . Realization of a twodimensional magnetooptical trap with a high optical depth. Chin Phys Lett 2012; 29: 024205.
Ding DS, Zhang W, Zhou ZY, Shi S, Shi BS et al. Raman quantum memory of photonic polarized entanglement. Nat Photon 2015; 9: 332–338.
Thew RT, Nemoto K, White AG, Munro WJ . Qudit quantumstate tomography. Phys Rev A 2002; 66: 012303.
Sanpera A, Bruß D, Lewenstein M . Schmidtnumber witnesses and bound entanglement. Phys Rev A 2001; 63: 050301(R).
Spengler C, Huber M, Brierley S, Adaktylos T, Hiesmayr BC . Entanglement detection via mutually unbiased bases. Phys Rev A 2012; 86: 022311.
Hiesmayr BC, Löffler W . Complementarity reveals bound entanglement of two twisted photons. New J Phys 2013; 15: 083036.
Collins D, Gisin N, Linden N, Massar S, Popescu S . Bell inequalities for arbitrarily highdimensional systems. Phys Rev Lett 2002; 88: 040404.
Romero J, Leach J, Jack B, Barnett SM, Padgett MJ et al. Violation of Leggett inequalities in orbital angular momentum subspaces. New J Phys 2010; 12: 123007.
Gühne O, Tóth G . Entanglement detection. Phys Rep 2009; 474: 1–75.
Fickler R, Krenn M, Lapkiewicz R, Ramelow S, Zeilinger A . Realtime imaging of quantum entanglement. Sci Rep 2013; 3: 1914.
Bruß D . Characterizing entanglement. J Math Phys 2002; 43: 4237–4251.
Gühne O, Hyllus P, Bruß D, Ekert A, Lewenstein M et al. Detection of entanglement with few local measurements. Phys Rev A 2002; 66: 062305.
Agnew M, Salvail JZ, Leach J, Boyd RW . Generation of orbital angular momentum bell states and their verification via accessible nonlinear witnesses. Phys Rev Lett 2013; 111: 030402.
Agnew M, Leach J, Boyd RW . Observation of entanglement witnesses for orbital angular momentum states. Eur Phys J D 2012; 66: 156.
FrankeArnold S, Leach J, Padgett MJ, Lembessis VE, Ellinas D et al. Optical ferris wheel for ultracold atoms. Opt Express 2007; 15: 8619–8625.
Bennett CH, Bernstein HJ, Popescu S, Schumacher B . Concentrating partial entanglement by local operations. Phys Rev A 1996; 53: 2046–2052.
GrodeckaGrad A, Zeuthen E, Sørensen AS . Highcapacity spatial multimode quantum memories based on atomic ensembles. Phys Rev Lett 2012; 109: 133601.
Krenn M, Handsteiner J, Fink M, Fickler R, Zeilinger A . Twisted photon entanglement through turbulent air across Vienna. Proc Natl Acad Sci USA 2015; 112: 14197–14201.
Bozinovic N, Yue Y, Ren YX, Tur M, Kristensen P et al. Terabitscale orbital angular momentum mode division multiplexing in fibers. Science 2013; 340: 1545–1548.
Zhou ZQ, Hua YL, Liu X, Chen G, Xu JS et al. Quantum Storage of ThreeDimensional orbitalangularmomentum entanglement in a crystal. Phys Rev Lett 2015; 115: 070502.
Acknowledgements
We thank Miles J. Padgett and Alison M. Yao for kindly helping us solve the different OAM phase superposition states. We also thank YongJian Han for helpful discussions and GuoYong Xiang for loaning two SLMs. This work was supported by the National Fundamental Research Program of China (Grant No. 2011CBA00200) and the National Natural Science Foundation of China (Grant Nos. 11174271, 61275115, 61435011 and 61525504).
Author information
Authors and Affiliations
Corresponding authors
Ethics declarations
Competing interests
The authors declare no conflict of interest.
Rights and permissions
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
About this article
Cite this article
Ding, DS., Zhang, W., Shi, S. et al. Highdimensional entanglement between distant atomicensemble memories. Light Sci Appl 5, e16157 (2016). https://doi.org/10.1038/lsa.2016.157
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/lsa.2016.157
Keywords
This article is cited by

648 Hilbertspace dimensionality in a biphoton frequency comb: entanglement of formation and Schmidt mode decomposition
npj Quantum Information (2021)

Advances in highdimensional quantum entanglement
Nature Reviews Physics (2020)

Broad spiral bandwidth of orbital angular momentum interface between photon and memory
Communications Physics (2019)

Twisted photons: new quantum perspectives in high dimensions
Light: Science & Applications (2017)

Deterministic secure quantum communication using a single dlevel system
Scientific Reports (2017)