Abstract
Research on entangled multipartite systems with controllable wave functions has attracted significant interest in the field of quantum optics. For quantum communications and quantum information processing, linear and nonlinear optical susceptibilities govern high-order correlations and entangled multiple-photon resources. In single- and double-dressing quadphoton correlations, we have observed the evolution of linear and nonlinear optical responses in the group delay and Rabi oscillation regimes. In the group delay regime, when linear susceptibility is evident, the quadphoton coincidence counting rate exhibits a rectangular profile. In the Rabi oscillation regime, the enhanced nonlinear susceptibility induced by strong laser dressing effects control quadphoton wave packets based on damped Rabi oscillation. Additionally, at different delay times, some photons exist in the group delay regimes, while others exist in the Rabi oscillation regimes, suggesting a coexistence mechanism. Additionally, there is a transition regime in which a portion of the photons are in both the group delay and Rabi oscillation regimes. By varying the power of the dressing field and optical depth, we realized the evolution between these two regimes for entangled quadphotons. Additionally, we demonstrate the shortening of coherence times under double-dressing conditions compared to single-dressing conditions. These results can help improving the length of coherence time and information capacity, which have great significance for the future development of long-distance and long coherent time quantum communication and quantum storage.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Over the past 50 years, non-classical light has become a powerful tool for exploring fundamental quantum optics theory. Quantum correlations and entanglements among multiparticles have great significance in terms of efficient quantum information transmission. There has been significant research on correlated and entangled multiple-photon resources [1–6]. Furthermore, strong dressing effect may enhance the efficiency of six-wave mixing (SWM) and eight-wave mixing (EWM) [7,8] and also the generation efficiency of tirphotons and quadphotons [9,10]. Additionally, for quantum transmission research, controllable wave function systems are of great significance. For single photons, quantum waveforms can be manipulated using external modulations [11,12]. Subsequently, Du et al. demonstrated a technique for shaping the temporal quantum waveforms of narrowband biphotons both theoretically and experimentally, and generated a cold atomic ensemble via four-wave mixing by periodically modulating two input classical lasers [13,14]. They also proposed the use of spatial light modulation to shape biphoton waveforms [15]. Recently, dressing laser beams have been used to alternate both the linear and nonlinear susceptibilities of the hot rubidium vapor, resulting in the modification of biphoton temporal correlation functions [16].
In this study, we observed the competition and coexistence of linear and nonlinear optical responses through an EWM process in 85Rb atomic vapor. Nonlinear susceptibility plays a major role in determining the wave packets of quadphotons when the effective coupling Rabi frequencies and linewidths are significantly smaller than the phase-matching and electromagnetically induced transparency (EIT) bandwidths. In the presence of multimodal quadphoton beating or destructive interference, the wave function of the quadphoton exhibits periodic Rabi oscillation. When the effective coupling Rabi frequency and linewidths are significantly larger than both the phase-matching and EIT bandwidths, linear susceptibility governs the shape of quadphoton wave packets in the group delay regime. When the phase-matching bandwidth is smaller than the EIT bandwidth, the wave function exhibits a rectangular profile. Otherwise, it exhibits a rectangular attenuation profile. These results have many potential applications in long-distance and long coherent time quantum communication, quantum information processing, and quantum storage.
2. Experimental setup and basic theory
A simplified experimental setup and atomic energy-level diagrams are shown in Figs. 1(a) and 1(b), respectively. Here, a 85Rb atomic vapor medium is used to generate a quadphoton. A thin and long cylindrical volume with a length of L confines the rubidium vapor. Initially, the four-level atoms are prepared at ground level |1 > . All four beams are then coupled and focused onto the center of the Rb atomic medium by optical lenses. Two beams with wavelengths of 795 nm act as weak pump beams E1 (frequency ω1, wave vector k1, detuning Δ1, and Rabi frequency G1) and E3 (ω3, k3, Δ3 and G3). The detuning, which is expressed as Δi=Ωi−ωi, is defined as the difference between the resonant transition frequency Ωi and the laser frequency ωi(Ei). The pump beams E1 and E3 have the same frequency ω and detuning parameter Δ. However, there is a small angle between the wave vectors k1 and k3. The atomic transitions |1>→|3 > and |2>→|3 > are pumped by E1 and E3, respectively. Additionally, there are two coupling beams E2 (ω2, k2, Δ2, G2, 780 nm) and E4 (ω4, k4, Δ4, G4, 780 nm). The strong coupling beam E2 counter-propagates with E1 near the resonant frequency of the atomic transition |2>→|4 > . Additionally, the coupling beam E4 propagates along E1 and is applied to the atomic transition |2>→|4 > with a detuning parameter Δ4. The EWM process occurs spontaneously and satisfies the phase-matching condition k1 + k2 + k3 + k4 = kS1 + kS2 + kS3 + kS4. Therefore, correlated quadphotons (ES1, ES2, ES3, and ES4) represented by the four small blue circles in Fig. 1(c) can be generated and detected by four single-photon counting modules (SPCMs). Let ωSi(i = 1,2,3,4) and δi(i = 1,2,3,4) represent the frequency and detuning of the generated photons, respectively. The entangled correlated quadphotons satisfy the conservation of total energy (i.e., ω1 + ω2 + ω3 + ω4 = ωS1 + ωS2 + ωS3 + ωS4, and δ1 + δ2 + δ3 + δ4 = 0). According to the energy conservation condition, coherent channels represented by the large orange area in Fig. 1(c) are interrelated and interfere with each other. Here, polarization effects, Doppler broadening, and quantum Langevin noise are ignored.
According to perturbation theory, the Hamiltonian interaction describes the EWM process and provides a clear picture of the quadphoton generation mechanism [9]. Here, the quadphoton state is expressed as
According to the internal dressing effect of the field E2, the seventh-order nonlinear susceptibility can be expressed as follows [9,17–20]:
The linear susceptibilities of the generated quadphotons (ES1, ES2, ES3, and ES4) with a single-dressing field are defined as follows:
Similarly, the linear susceptibilities of the generated quadphoton with double-dressing (E2 and E4) are defined as follows:
The bandwidth and time determined by the group delay velocity (vsi) are referred to as the phase-matching bandwidth (Δωgi) and group delay time (τgi), respectively. The phase-matching bandwidth is one of the key factors controlling the profile of quadphoton coincidence counting. The expressions for Δωgi and τgi are defined as follows:
Additionally, Δωtr is referred to as the EIT bandwidth. The specific expressions for Δωtr2 and Δωtr4 are defined as follows:According to Eq. (2), the properties of the quadphoton amplitude are dependent on both the nonlinear parametric coupling coefficient $\kappa$ and longitudinal detuning function Φ. Although the effective coupling Rabi frequency Ωe and linewidth Γe are far smaller than the phase-matching bandwidth Δωg and EIT bandwidth Δωtr, the phase mismatch Δk is approximately equal to zero. In this case, the nonlinear susceptibility $\chi _{S2}^{(7)}$ plays a significant role in determining the spectral width of the quadphoton. Additionally, the effective coupling Rabi frequency Ωe can result in multimodal EWM processes, thereby generating multimodal quadphotons. Based on the beating and destructive interference generated by multimodal quadphotons, the corresponding wave function exhibits the form of damped Rabi oscillation. Conversely, when Δωg and Δωtr are far smaller than Γe and Ωe, the nonlinear parametric coupling coefficient $\kappa$ is constant and the linear susceptibility ${\chi _{Si}}$(i = 1, 2, 3, 4) determines the spectral width of the quadphotons. In this case, when the phase-matching bandwidth is smaller than the EIT bandwidth (Δωg < Δωtr), the wave function of the quadphoton exhibits a rectangular profile and rectangular attenuation profile.
3. Counting measurement
In this section, we investigate the competing and coexisting of single- and double-dressing linear and nonlinear optical responses in quadphoton correlation. Three different cases are discussed below.
3.1 Group delay regime
As suggested by Balić et al. [3] and Kolchin [22], and demonstrated by Du et al. [23], when the linewidth Γe is greater than the phase-matching bandwidth Δωg, the quadphoton coincidence counting rate lies within the group delay regime. In this case, the quadphoton bandwidth depends on the linear susceptibility. The EIT profile controls the wave function through a transparency window and slow light.
When the phase-matching bandwidth Δωg is smaller than the EIT bandwidth Δωtr, quadphoton wave packets appear as rectangular functions. In this case, we only consider the real part of Δk. By calculating Eq. (2), the amplitude functions of generated photons with single- and double-dressing fields can be expressed uniformly as follows:
For a single-dressing field E2, the numerical simulation described by Eq. (12) is presented in Fig. 2(a). The axes τ2, τ3, and τ4 represent the coherence times of coincidence counting between photons ES2, ES3, ES4, and ES1. Here, the value of OD is 300, P2 represents the power of dressing field G2 and it is equal to 10 mW. The linewidths Γei and EIT bandwidths Δωtri are greater than the phase-matching bandwidths Δωgi in all directions. Therefore, the quadphoton wave packets exhibit rectangular profiles in all three directions. From Fig. 2(a), we can obtain sectional views of four-dimensional photon counting measurements, as shown in Figs. 2(b) to 2(d). In Fig. 2(b), the sectional view represents the τ2 and τ3 axes. The coherence time in the direction of τ2 is approximately 150 ns. In the direction of τ3, it is approximately 108 ns. In Fig. 2(c), the sectional view represents the τ2 and τ4 axes. The coherence time in the direction of τ4 is approximately 150 ns. In Fig. 2(d), the sectional view represents the τ3 and τ4 axes.
Similar to Fig. 2, we present a simulated diagram of the quadphoton coincidence counting rate with double-dressing effect in Fig. 3. During our simulations, we ensured the consistency of parameters between Figs. 2 and 3. According to Fig. 3(a), in the directions of τ2, τ3, and τ4, the coherence times are 75 ns, 54 ns, and 75 ns, respectively. In our previous description, the double-dressing field is actually cascade double-dressing. In fact, we can also implement another kind of double-dressing case as shown in Fig. 3(b), namely nested double-dressing. According to Fig. 3(b), in the directions of τ2, τ3, and τ4, the coherence times are 195 ns, 140 ns, and 195 ns, respectively. It is obvious that the coherence time of each direction in Fig. 3(b) is much longer than that in Fig. 3(a). In view of the fact that these two cases are similar except for the length of the coherence time, in the following discussion, we will choose cascaded double-dressing as an example to compare with the single-dressing case. We refer to cascade double-dressing as double -dressing by default.
Furthermore, when the EIT bandwidth Δωtr is smaller than the phase-matching bandwidth Δωg, quadphoton wave packets exhibit rectangular attenuation profiles. In this case, by extending Eq. (2), the amplitude functions for generated quadphotons with single- and double-dressing fields can be expressed uniformly as follows:
3.2 Coexistence mechanism for different delay times
With different delay times, the optical properties of quadphoton amplitude [Eq. (2)] are determined by both nonlinear and linear optical responses. In such cases, the quadphoton coincidence counting rate is subject to a coexistence mechanism between the Rabi oscillation and group delay regime. In other words, a portion of the photons are in the Rabi oscillation regime (Γe << Δωg), while others lie in the group delay regime (Γe >> Δωg).
To avoid repetition, we will only consider the condition with a single-dressing field as a representative example. In this case, a photon S2 is generated in the Rabi oscillation regime and other photons are generated in the group delay regime. The coincidence counts of the quadphotons can be rewritten as
A double-dressing field with the same parameters discussed above is presented in Fig. 4(b). The coherence times in the directions of τ3 and τ4 are 80 ns and 115 ns, respectively. Based on Eqs. (7) to (10), we can obtain the following expression via comparisons:
Similar to Fig. 4, although Δωtr is smaller than Δωg, the corresponding direction exhibits a rectangular attenuation function. Simulation results for single- and double-dressing fields are presented in Figs. 5(a) and 5(b), respectively. In the directions of τ3 and τ4, the linear optical response still dominates. Because Δωtr is smaller than Δωg, the wave packets of τ3 and τ4 exhibit rectangular attenuation functions. In the direction of τ2, Γe is still greater than Δωg and the nonlinear optical response plays a dominant role. Similar to Fig. 4(c), two-dimensional cross-sectional comparison results in the direction of τ3 are presented in Fig. 5(c). The coherence times of τ3 are 80 ns and 160 ns for curves A and B, respectively. However, the width of τgs3 is still twice that of τgd3.
Similar to Fig. 4(a), but one direction still lies in the group delay regime and the others lie in the Rabi oscillation regime. When the directions of τ3 and τ4 lie in the Rabi oscillation regime, the coincidence counting rate of the quadphotons can be rewritten as
Figure 6(b) is similar to Fig. 6(a), except that it represents a double-dressing field. In the directions of τ3 and τ4, the photon coincidence counts exhibit periodic Rabi oscillations. The coincidence count in the direction of τ2 exhibits a rectangular profile. To avoid repetition, we will not describe these factors in detail. Under these conditions, comparison results for the τ2 axis are presented in Fig. 6(c). Curves A and B correspond to the two-dimensional cross sections in the cases of double- and single-dressing, respectively. The coherence times τ3 are 90 and 230 ns for curves A and B, respectively. Unlike the parameter values in Fig. 4, because the values of P2 and P4 in Fig. 6 are no longer equal, τgs2 is no longer twice as large as τgd2, as suggested by Eq. (15). However, τgs2 is still larger than τgd2.
In Fig. 7, similar to Fig. 6, Δωtr2 is smaller than Δωg2, resulting in a rectangular attenuation profile for the photon coincidence counts in the direction of τ2. The wave packets exhibit a clear Rabi oscillation waveform in the directions of both τ3 and τ4. The parameters used here are the same as those used in Fig. 6. Figures 7(a) and 7(b) present the quadphoton coincidence counts under the coexistence mechanism with single- and double-dressing fields, respectively. Here, we compare the coherence times based on the example of the τ2 axis. Comparison results are presented in Fig. 7(c). The comparison results are similar to those in Fig. 6(c) and will not be described in detail here.
3.3 Transition regime between linear and nonlinear optical responses
When the phase-matching bandwidth Δωg and Rabi frequency Ωe are approximately equal, a transition regime between linear and nonlinear optical responses is suggested by the quadphoton wave packets. The quadphoton coincidence counting rate with a single-dressing field can be expressed as follows:
A simulated diagram of the quadphoton coincidence counting rate with a single-dressing effect is presented in Fig. 8(a). Here, the value of OD is 200, and P2 and P4 are both 10 mW. Under these conditions, Δωg2 is much smaller than Γe2, and the direction of τ2 lies in the group delay regime. Additionally, Δωg2 is smaller than Δωtr2, so wave packets exhibit a rectangular profile. The coherence time is 40 ns. Δωg3 and Δωg4 are approximately equal to Ωe3 and Ωe4, respectively. Therefore, these directions both lie in the transition regime. In other words, in these two directions, the wave packets exhibit a mixture of a rectangular profile and Rabi oscillation. In the direction of τ3, there are two periods of 2π / Δ1 and 2π / Δ4. In the direction of τ4, the periods of the photon coincidence counts are 2π / Δ1, 2π / Δ4, and 2π / (Δ4– Δ1).
To analyze the transition regime in greater detail, we consider the direction of τ3 as an example. We can control the proportion of linear and nonlinear optical responses by adjusting the value of the dressing field and OD. In the following analysis, we adjust these two parameters separately to observe the resulting change trends in the wave packets.
According to our previous analysis, dressing fields affect both linear and nonlinear optical effects. By changing the power of the dressing field, we can induce the evolution process of the transition regime, as shown in Fig. 9. Here, the value of OD is 200. Figure 9(ai) represents a single-dressing field. Figures 9(a1) to 9(a6) correspond to powers of P2 = 1, 5, 10, 15, 20, and 30 mW. As shown in Fig. 9(a1), the coincidence counts of the two-dimensional cross sections lie in the transition regime and the linear optical response plays a dominant role. Under these conditions, the total coherence time τc is the sum of the linear coherence time τg and nonlinear coherence time τe (τc = τg + τe). One can observe a very clear rectangular profile with a width of approximately 100 ns and a τc value greater than 200 ns. According to Eq. (10), τg∝OD/G22 and τe∝G2. Therefore, we obtain the following expression: τg/τe∝OD/G23. Combined with the results of Fig. 9(ai), as the power of G2 increases gradually, the effect of the linear optical response is diminished and τg decreases. Additionally, the effect of the nonlinear optical response is enhanced and τe increases. Until Fig. 9(a6), the wave packets exhibit rough Rabi oscillation and τc is approximately 120 ns. As τg decreases with an increasing value of τe, the total coherence time τc decreases. It can be predicted that the effect of the nonlinear optical response will be far stronger than that of the linear optical response when G2 increases significantly. Under these conditions, the waveform will take on the form of pure Rabi oscillation.
Similarly, Fig. 9(bi) represents a double-dressing field. The changing trend of P2 is the same as that above, but P4 = 10 mW. Under these conditions, the trend of the wave packets is similar to that with a single-dressin The changing trend of P2 is the same as that above, but P4 = 10 mW.g field. The main difference compared to the previous cause lies in the coherence time. We obtained the following expression: τg /τe∝OD/ (G43+G4G22). As the power of the dressing field increases gradually, the role of the nonlinear optical response becomes more significant. Based on the double-dressing effect, the total coherence time τc is shorter than that in the case with single-dressing.
Similar to the effect of varying the dressing field, the transition regime between linear and nonlinear optical responses can also be controlled by varying OD. However, OD only affects the linear optical response. In other words, the coherence time τe of the nonlinear optical response is constant. Figure 10 illustrates how the wave packets of two-dimensional cross sections change with increases in OD. Figure 10(a) represents a single-dressing field. Here, the power of P2 is 10 mW. Figures 10(a1) to 10(a6) correspond to the values of OD = 10, 100, 200, 300, 400, and 500, respectively. In Fig. 10(a1), the wave packets exhibit rough Rabi oscillation. As OD increases, τg also increases. Therefore, the proportion of τg in the total coherence time τc increases. Under these conditions, the linear effect becomes stronger, but there is no variation in the nonlinear effect. Figure 10(a6) presents a clear rectangular function with little oscillation. The experimental parameters are the same as those in Fig. 4(a), except that OD = 800. In Fig. 4(a), the linear optical response is dominant in the direction of τ3. This can be regarded as an extreme case in which OD increases significantly, causing the linear optical responses to grow much larger than the nonlinear optical responses. Under these conditions, the wave packets exhibit a pure rectangular profile.
Figure 10(bi) is similar to Fig. 10(ai), except that the dressing fields contain G2 and G4, which both have a power of 20 mW. Under these conditions, the trend of the two-dimensional cross sections’ wave packets changes in the same manner as that above, but the coherence time is shorter than that in Fig. 10(a). Figure 4(b) can be considered as an extreme case in which OD increases significantly. Under these conditions, the wave packets exhibit a pure rectangular function. By combining Figs. 9 and 10, we can reach the following conclusions. Based on the double-dressing effect, the total coherence time τc is significantly smaller than that in the case with single-dressing. Here, the cascaded scheme is similar to the parallel-cascade scheme [17]. This enhances the overall effect of the dressing field, as well as the group velocity, and the total coherence time is reduced. However, if the cascaded scheme takes on the form of a nested cascade, the situation is reversed. The cascading scheme should be selected according to specific application requirements.
4. Conclusion
In this study, we observed the competition and coexistence of linear and nonlinear optical responses in single- and double-dressing quadphoton correlations. The quadphoton coincidence counting rate exhibits a rectangular function when the linear optical response plays a dominant role. The quadphoton coincidence counting rate exhibits Rabi oscillation when strong dressing effects enhance seventh-order nonlinear susceptibility. By increasing or decreasing the temperature, it is possible to increase or decrease OD, thereby enhancing or diminishing the linearity of quadphotons. Additionally, the power of the dressing field is closely related to the nonlinearity of quadphotons. By varying the power of the dressing field and OD, we simulated the evolution process from a dominant linear optical response to a dominant nonlinear response. We also proved that the double-dressing effect reduces coherence time compared to the single-dressing effect. This theoretical research further provides physical insights into the experimental research of the linear and nonlinear optical response competition. These results also have great significance for the widespread development of long distance quantum communications and quantum storage.
Funding
National Key Research and Development Program of China (2017YFA0303700, 2018YFA0307500); National Natural Science Foundation of China (11604256, 11804267, 11904279, 61605154, 61975159).
Disclosures
The authors declare no conflicts of interest.
References
1. S. E. Harris, M. K. Oshman, and R. L. Byer, “Observation of Tunable Optical Parametric Fluorescence,” Phys. Rev. Lett. 18(18), 732–734 (1967). [CrossRef]
2. H. Yan, S. Zhang, J. F. Chen, M. M. T. Loy, G. K. L. Wong, and S. Du, “Generation of Narrow-Band Hyperentangled Nondegenerate Paired Photonsm,” Phys. Rev. Lett. 106(3), 033601 (2011). [CrossRef]
3. V. Balic, D. A. Braje, P. Kolchin, G. Y. Yin, and S. E. Harris, “Generation of Paired Photons with Controllable Waveforms,” Phys. Rev. Lett. 94(18), 183601 (2005). [CrossRef]
4. H. Hübel, D. R. Hamel, A. Fedrizzi, S. Ramelow, K. J. Resch, and T. Jennewein, “Direct generation of photon triplets using cascaded photon-pair sources,” Nature 466(7306), 601–603 (2010). [CrossRef]
5. D. R. Hamel, L. K. Shalm, H. Hübel, A. J. Miller, F. Marsili, V. B. Verma, R. P. Mirin, S. W. Nam, K. J. Resch, and T. Jenneweinm, “Direct generation of three-photon polarization entanglement,” Nat. Photonics 8(10), 801–807 (2014). [CrossRef]
6. C. Shu, P. Chen, T. K. A. Chow, L. Zhu, Y. Xiao, M. M. T. Loy, and S. Du, “Subnatural-linewidth biphotons from a Doppler-broadened hot atomic vapour cell,” Nat. Commun. 7(1), 12783 (2016). [CrossRef]
7. Y. Zhang, B. Anderson, and M. Xiao, “Coexistence of four-wave, six-wave and eight-wave mixing processes in multi-dressed atomic systems,” J. Phys. B: At., Mol. Opt. Phys. 41(4), 045502 (2008). [CrossRef]
8. Z. Wu, C. Yuan, Z. Zhang, H. Zheng, S. Huo, R. Zhang, R. Wang, and Y. Zhang, “Observation of eight-wave mixing via electromagnetically induced transparency,” Europhys. Lett. 94(6), 64005 (2011). [CrossRef]
9. Y. Feng, K. Li, I. Ahmed, Y. Li, W. Li, G. Lan, and Y. Zhang, “Generation of a Quadphoton via Seventh-Order Nonlinear Susceptibility,” Ann. Phys. 531(10), 1900072 (2019). [CrossRef]
10. Y. Li, Y. Feng, W. Li, K. Li, Y. Liu, Y. Lan, and Y. Zhang, “Double-dressing tri and quad-photon correlations in spontaneous six and eight-wave mixing,” Phys. Scr. 94(10), 105802 (2019). [CrossRef]
11. M. Keller, B. Lange, K. Hayasaka, W. Lange, and H. Walther, “Continuous generation of single photons with controlled waveform in an ion-trap cavity system,” Nature 431(7012), 1075–1078 (2004). [CrossRef]
12. P. Kolchin, C. Belthangady, S. Du, G. Y. Yin, and S. E. Harris, “Electro-Optic Modulation of Single Photons,” Phys. Rev. Lett. 101(10), 103601 (2008). [CrossRef]
13. S. Du, J. Wen, and C. Belthangady, “Temporally shaping biphoton wave packets with periodically modulated driving fields,” Phys. Rev. A 79(4), 043811 (2009). [CrossRef]
14. J. F. Chen, S. Zhang, H. Yan, M. M. T. Loy, G. K. L. Wong, and S. Du, “Shaping Biphoton Temporal Waveforms with Modulated Classical Fields,” Phys. Rev. Lett. 104(18), 183604 (2010). [CrossRef]
15. L. Zhao, X. Guo, Y. Sun, Y. Su, M. M. T. Loy, and S. Du, “Shaping the Biphoton Temporal Waveform with Spatial Light Modulation,” Phys. Rev. Lett. 115(19), 193601 (2015). [CrossRef]
16. X. Li, D. Zhang, D. Zhang, L. Hao, H. Chen, Z. Wang, and Y. Zhang, “Dressing control of biphoton waveform transitions,” Phys. Rev. A 97(5), 053830 (2018). [CrossRef]
17. Y. Zhang, A. W. Brown, and M. Xiao, “Opening Four-Wave Mixing and Six-Wave Mixing Channels via Dual Electromagnetically Induced Transparency Windows,” Phys. Rev. Lett. 99(12), 123603 (2007). [CrossRef]
18. Y. Zhang, U. Khadka, B. Anderson, and M. Xiao, “Temporal and Spatial Interference between Four-Wave Mixing and Six-Wave Mixing Channels,” Phys. Rev. Lett. 102(1), 013601 (2009). [CrossRef]
19. Y. Zhang, Z. G. Wang, Z. Q. Nie, C. B. Li, H. X. Chen, K. Q. Lu, and M. Xiao, “Four-Wave Mixing Dipole Soliton in Laser-Induced Atomic Gratings,” Phys. Rev. Lett. 106(9), 093904 (2011). [CrossRef]
20. Z. Nie, H. Zheng, P. Li, Y. Yang, Y. Zhang, and M. Xiao, “Interacting multiwave mixing in a five-level atomic system,” Phys. Rev. A 77(6), 063829 (2008). [CrossRef]
21. S. E. Harris, “Nonlinear optics at low light levels,” Phys. Rev. Lett. 82(23), 4611–4614 (1999). [CrossRef]
22. P. Kolchin, “Electromagnetically-induced-transparency-based paired photon generation,” Phys. Rev. A 75(3), 033814 (2007). [CrossRef]
23. S. Du, P. Kolchin, C. Belthangady, G. Y. Yin, and S. E. Harris, “Generation of Paired Photons with Controllable Waveforms,” Phys. Rev. Lett. 100(18), 183603 (2008). [CrossRef]