Professional Documents
Culture Documents
Sties Dal 2018
Sties Dal 2018
DOI: 10.1103/PhysRevLett.121.103601
Engineering effective interactions between individual been extensively studied for a variety of systems [23],
optical photons is essential for applications in classical and the observation of antibunching is accepted as a
and quantum computation, communication, and metrology characteristic fingerprint of a single-photon source [24],
[1–3]. The most established approach to achieve two- while bound states of two photons have been observed as
photon interaction is strong coupling of light to individual strong bunching feature in Rydberg polariton systems [25].
quantum emitters, either in resonators [4–7] and wave- To study the interaction between three photons, it is natural to
guides [8–13] or through tight focusing in free space turn to third-order correlations gð3Þ [26–28]. While any two-
[14–16]. A complementary approach combines electro- body correlation will also induce a signal in the three-body
magnetically induced transparency and strong interaction correlation function, a natural approach is to subtract these
between atoms in Rydberg states to convert photons into trivial contributions via the cumulant expansion to identify
interacting Rydberg polaritons in an extended medium the pure three-body correlations. This approach leads to the
[17–19]. Here we exploit a combination of the two concepts connected part of the three-body correlation function:
by using the Rydberg blockade mechanism [20] to convert X
ð3Þ
an atomic ensemble containing N individual atoms into a gc ðs1 ; s2 ; s3 Þ ¼ 2 þ gð3Þ ðs1 ; s2 ; s3 Þ − gð2Þ ðsi ; sj Þ: ð2Þ
single effective two-level quantum system, a Rydberg i<j
superatom, with strongly enhanced coupling to a single ð3Þ
photonic mode [21,22]. Note that gc vanishes if one photon is separated from the
A central concept for quantifying the influence of an other two. Furthermore, for any classical Gaussian state of
effective photon-photon interaction in any of these systems photons, the connected part of the three-body correlation
is to study the intensity correlations imprinted by the function is zero.
interaction onto initially uncorrelated photons by the Here, we report the first experimental observation of
determination of n-body correlation functions: three-photon correlations by clear signatures in the con-
nected part of the three-body correlation function of
hE† ðs1 Þ…E† ðsn ÞEðsn Þ…Eðs1 Þi initially uncorrelated photons interacting with a single
gðnÞ ðs1 ; …; sn Þ ¼ Qn †
: ð1Þ Rydberg superatom. Our setup is based on a cold atomic
i¼1 hE ðsi ÞEðsi Þi cloud interacting with a focused photonic mode coupling
to a highly excited Rydberg state via a far detuned
Here, the operators E† ðsÞ and EðsÞ describe the creation intermediate state. The transversal size of the photonic
and annihilation, respectively, of photons at time s. The mode as well as the longitudinal extent of the atomic cloud
outgoing photon rate is related to these operators via is smaller than the Rydberg blockade volume, resulting in a
IðsÞ ¼ hE† ðsÞEðsÞi. Two-photon correlations gð2Þ have single effective emitter with collectively enhanced coupling
103601-2
PHYSICAL REVIEW LETTERS 121, 103601 (2018)
i
∂ t ρðtÞ ¼ − ½H0 ðtÞ; ρðtÞ þ ðκ þ ΓÞD½σ GW ρðtÞ
ℏ
þ γ d D½σ DW ρðtÞ þ ΓD½σ GD ρðtÞ ð4Þ FIG. 2. (a) Cut through the experimental third-order correlation
gð3Þ ðs1 ; s2 ; s3 Þ for input photon rate Rin ¼ 6.7 μs−1 along the
with the Lindblad dissipator D½σ¼σρσ †p−ðσ † † η,ffiffiffiζ averaged over the center-of-mass
ffiffiffi σρþρσ σÞ=2
relative Jacobi coordinates p
and the driving Hamiltonian H0 ðtÞ ¼ ℏ κα ðtÞσ GW þH:c: coordinate range Rrange ¼ 3 × ð2.5…3.5Þ μs. (b) Correspond-
Here, the coherent field amplitude αðtÞ is related to the ing theoretical calculation based on the single-emitter model
time-dependent mean photon rate by jαðtÞj2 ¼ Rin ðtÞ. In using the parameters extracted from the fitted curves in Fig. 1(d).
Note the different scales on the colorbars, the contrast of the
addition to the intrinsic collectively enhanced decay κ into experimentally observed correlations is reduced due to finite
the photonic mode, we include phenomenologically the photon detection efficiency (see Supplementary Material [38]).
spontaneous decay of the Rydberg level with rate Γ as well
as the dephasing of the superatom state jWi into the emitted photons never interact with the system again [40].
manifold of dark states with rate γ d . The outgoing pffiffiffi electric For other input photon states, a more sophisticated quantum
field operator is determined by EðtÞ ¼ αðtÞ − i κσ GW ðtÞ. field formalism can be employed. The theoretical third-
Fitting these parameters by comparing the theoretical order correlations based on the parameters extracted from
predictions with the experimental outgoing photon time the time trace fits and averaged over the time range Rrange
traces [orange lines in Fig. 1(d)] yields a single set of are shown in Fig. 2(b), reproducing to very good agreement
parameters: κ ¼0.55μs−1 , Γ¼0.14μs−1 , and γ d ¼1.49 μs−1 . the bunching and antibunching features observed in the
Next, we calculate the third-order correlations experimental data.
gð3Þ ðs1 ; s2 ; s3 Þ [Eq. (1)] from the outgoing photon traces. To quantify pure three-body correlations in the outgoing
A natural choice for the visualization of these correlations pffiffiffi photon stream, we now extract the connected third-order
ð3Þ
is to transform to Jacobi coordinates
pffiffiffiffiffiffiffiffi R ¼ ðs 1 þs 2 þs 3 Þ= 3, correlation function gc by subtracting the two-body
pffiffiffi
η ¼ ðs1 − s2 Þ= 2, and ζ ¼ 2=3½ðs1 þ s2 Þ=2 − s3 . Since correlations for all pairwise combinations of time coor-
we are investigating the response of the superatom to a dinates s1 , s2 , and s3 from gð3Þ , as defined in Eq. (2).
ð3Þ
pulsed probe, the correlation function is not stationary and Figures 3(a)–3(c) show cuts gc ðη; ζÞ through the measured
thus depends on the center-of-mass value R as well as on connected three-body correlation function for all three
the relative coordinates η and ζ. Nevertheless, this choice of investigated photon input rates Rin ¼ 3.4 μs−1 , Rin ¼
coordinates lets us average the gð3Þ function over a limited 6.7 μs−1 , and Rin ¼ 15.2 μs−1 averaged over Rrange, deter-
range of R, corresponding to times ðs1 ; s2 ; s3 Þ within the mined from the measured second- and third-order corre-
flattop part of the Tukey pulse. In Fig. 2(a), we show the lation functions. Even for low photon numbers, we find a
third-order correlation function gð3Þ ðη; ζÞ for the input clear signal of three-photon correlations in the connected
−1
photon rate pffiffiffi Rin ¼ 6.7 μs averaged over the time range part of gð3Þ with a three-photon bunching at short distances,
Rrange ¼ 3 × ð2.5…3.5Þ μs [with respect to the time axis accompanied by an antibunching at intermediate separa-
shown in Fig. 1(d)]. While the averaging over R certainly tions, followed by another ring of bunching. This sequence
reduces the visibility of the three-body correlations, it is of bunching and antibunching features increases with
essential to extract a significant signal from the few-photon increasing photon numbers. Note that gð3Þ for a transla-
data for a realistic number of repetitions of the experiment. tionally invariant system exhibits a sixfold symmetry in
Within the theoretical model described above any multi- Jacobi coordinates. The reduction to a threefold symmetry,
time correlation functions for a coherent input pulse are visible in particular in Fig. 3(c), is a consequence of the
conveniently calculated using the quantum regression finite length of our probe pulse. In Figs. 3(d) and 3(e), we
theorem; alternative approaches using Keldysh formalisms show the corresponding theoretical predictions from our
have also been developed [39]. Note that, for a single quantitative model, which well reproduce the observed
superatom, the quantum regression theorem is exact, as the structure in the experimental data.
103601-3
PHYSICAL REVIEW LETTERS 121, 103601 (2018)
(a) (d)
(a) (b)
(b) (e)
Q
interested in an incoming state ψ in ¼ i ψðsi Þ for n
FIG. 3. Connected part of the three-photon correlation function photons in the single photon mode ψ with width τ.
ð3Þ
gc in the Jacobi coordinates η, ζ. (a)–(c) show experimental results Then, the outgoing n-photon wave function after interact-
for Rin ¼ 3.4 μs−1, Rin ¼ 6.7 μs−1 , and Rin ¼ 15.2μs−1 , respec- ing with a single superatom at times s1 ≥ ≥ sn reduces
tively. (d)–(f) show the corresponding theoretical predictions. to (see Supplemental Material [38])
P
ðnÞ
ψ out ðs1 ;…;sn Þ ¼ ∂ α1 …∂ αn eðκ=2Þ i ðsi −2αi Þ ϕðs1 − α1 Þ
While this agreement between the theory and experiment
suggests that our simple single-emitter model captures the Y n
physics of the superatom-light interaction and the effective × ½ϕðsi −αi Þ −ϕðsi−1 þαi−1 Þ ð5Þ
i¼2 αi ¼0
photon-photon interaction mediated through the superatom
very well, it is not straightforward to understand the R
with ϕðsÞ ¼ s∞ dtψðtÞe−κt=2 ; the wave function for all
microscopic origin for the appearance of three-photon
values of fsi g is obtained by requiring the bosonic
correlations from this model. In order to provide a micro-
symmetry of the wave function. For three photons in a
scopic and qualitative understanding of these correlations,
wide incoming mode τκ ≫ 1, the wave function in the
we turn to the theoretical study of an idealized setup, which
center of the pulse reduces to (s1 ≥ s2 ≥ s3 )
allows for a fully analytical solution. For this purpose, we
point out that the Hamiltonian Eq. (3) is exactly solvable ð3Þ
ψ out ¼ 1 þ 12e−κ½ðs1 −s3 Þ=2 − 4ðe−κ½ðs1 −s2 Þ=2 þ e−κ½ðs2 −s3 Þ=2 Þ;
via the Bethe ansatz [41]. While such an approach ignores
the additional dephasing and spontaneous emission of the ð6Þ
excited state, it allows us to gain a microscopic under-
standing of the possible correlations induced by a single from which we can analytically derive the three-body
superatom. A first important aspect is that the exact correlation functions shown in Fig. 4. We find a very
eigenstates for three photons can be characterized as a strong three-photon bunching around η ¼ ζ ¼ 0. The
three-photon bound state, a combination of a two-body contribution of the three-body bound state to this signal
bound state with an additional scattering photon, and can be observed by the decomposition of the wave function
ð3Þ
finally pure scattering states [41]. Especially, the three- ψ out ¼ 4e−κðs1 −s3 Þ þ ψ sc . The first term describes the three-
photon bound state naturally provides a nontrivial contri- body bound state, while ψ sc accounts for the remaining
bution to the connected part of the correlation function. A contributions of scattering states and two-photon bound
second important aspect is the relation between the out- states with ψ sc ¼ 1 at s1 ¼ s2 ¼ s3 . Therefore, the three-
going wave function and the incoming photon wave body bound state provides the dominant contribution to
function, which for the considered problem can be derived the three-photon bunching signal, but contributions of
in a closed form [42]. For the present setup, we are the remaining states are still significant. Especially, the
103601-4
PHYSICAL REVIEW LETTERS 121, 103601 (2018)
exponential decrease of the bunching signal at s2 ¼ s3 [4] J. M. Raimond, M. Brune, and S. Haroche, Rev. Mod. Phys.
exhibits a decay κ instead of the faster decay 2κ expected 73, 565 (2001).
from the three-body bound state wave function. [5] I. Shomroni, S. Rosenblum, Y. Lovsky, O. Bechler, G.
Comparing this result with the experimentally observed Guendelman, and B. Dayan, Science 345, 903 (2014).
correlation functions, we find that the idealized setup exhibits [6] T. G. Tiecke, J. D. Thompson, N. P. de Leon, L. R. Liu, V.
Vuletić, and M. D. Lukin, Nature (London) 508, 241 (2014).
the characteristic features of three-photon bunching at
[7] B. Hacker, S. Welte, G. Rempe, and S. Ritter, Nature
short distances and antibunching at intermediate distances (London) 536, 193 (2016).
observed in the experiment for low photon numbers. We [8] A. F. van Loo, A. Fedorov, K. Lalumière, B. C. Sanders, A.
expect, therefore, that the microscopic origin of the three- Blais, and A. Wallraff, Science 342, 1494 (2013).
body correlations in the experiment are well captured by an [9] J. Petersen, J. Volz, and A. Rauschenbeutel, Science 346, 67
understanding of the idealized model as the combination of (2014).
a three-photon bound state, scattering states, and two-photon [10] S. Faez, P. Türschmann, H. R. Haakh, S. Götzinger, and V.
bound states. Note that the width of these signals is increased Sandoghdar, Phys. Rev. Lett. 113, 213601 (2014).
due to the reduced losses in the idealized setup. Furthermore, [11] P. Lodahl, S. Mahmoodian, and S. Stobbe, Rev. Mod. Phys.
the appearance of oscillations for higher number of photons 87, 347 (2015).
can be understood as the single emitter undergoing Rabi [12] A. Goban, C.-L. Hung, J. D. Hood, S.-P. Yu, J. A. Muniz, O.
oscillations, which gives rise to a characteristic beating for Painter, and H. J. Kimble, Phys. Rev. Lett. 115, 063601
(2015).
an increasing photon number.
[13] A. Sipahigil, R. E. Evans, D. D. Sukachev, M. J. Burek, J.
In conclusion, the experimental observation of three-
Borregaard, M. K. Bhaskar, C. T. Nguyen, J. L. Pacheco,
photon correlations imprinted by a single Rydberg H. A. Atikian, C. Meuwly, R. M. Camacho, F. Jelezko, E.
superatom on an initially uncorrelated photonic state is Bielejec, H. Park, M. Lončar, and M. D. Lukin, Science 354,
well accounted for by a single-emitter model, where the 847 (2016).
dephasing of the collectively excited state as well as the [14] A. N. Vamivakas, M. Atatüre, J. Dreiser, S. T. Yilmaz, A.
spontaneous decay are included. On a microscopic level, Badolato, A. K. Swan, B. B. Goldberg, A. Imamoglu, and
the appearance of the three-photon correlations is well M. S. Ünlü, Nano Lett. 7, 2892 (2007).
understood in an idealized setup, which suggests that the [15] M. K. Tey, Z. Chen, S. A. Aljunid, B. Chng, F. Huber, G.
observed three-photon bunching signal here cannot be Maslennikov, and C. Kurtsiefer, Nat. Phys. 4, 924 (2008).
purely attributed to a three-photon bound state in contrast [16] A. Maser, B. Gmeiner, T. Utikal, S. Götzinger, and V.
to recent observations on Rydberg polaritons [28]. Sandoghdar, Nat. Photonics 10, 450 (2016).
Observing such three-photon correlations imprinted by [17] J. D. Pritchard, D. Maxwell, A. Gauguet, K. J. Weatherill,
M. P. A. Jones, and C. S. Adams, Phys. Rev. Lett. 105,
a single two-level systems sheds light on the fundamental
193603 (2010).
processes of absorption and emission at the quantum level
[18] T. Peyronel, O. Firstenberg, Q. Liang, S. Hofferberth, A.
and highlights the potential for experimentally realizing Gorshkov, T. Pohl, M. Lukin, and V. Vuletic, Nature
photonic strongly correlated many-body systems in quan- (London) 488, 57 (2012).
tum nonlinear optical systems. Besides further improving [19] O. Firstenberg, C. S. Adams, and S. Hofferberth, J. Phys. B
the emitter-light coupling and investigating intrinsic 49, 152003 (2016).
superatom dephasing mechanisms, we envision the scal- [20] M. D. Lukin, M. Fleischhauer, R. Cote, L. M. Duan, D.
ing of our system to more complex arrangements of Jaksch, J. I. Cirac, and P. Zoller, Phys. Rev. Lett. 87, 037901
multiple superatoms for implementing quantum optical (2001).
networks [1,3]. [21] Y. O. Dudin and A. Kuzmich, Science 336, 887 (2012).
[22] A. Paris-Mandoki, C. Braun, J. Kumlin, C. Tresp, I.
We thank Florian Christaller and Simon W. Ball for Mirgorodskiy, F. Christaller, H. P. Büchler, and S.
contributing to the construction of the experimental setup. Hofferberth, Phys. Rev. X 7, 041010 (2017).
This work is supported by the European Union under [23] M. O. Scully and S. M. Zubairy, Quantum Optics
the ERC consolidator grants SIRPOL (Grant No. 681208) (Cambridge University Press, Cambridge, England, 1997).
and RYD-QNLO (Grant No. 771417) and the Deutsche [24] M. D. Eisaman, J. Fan, A. Migdall, and S. V. Polyakov, Rev.
Forschungsgemeinschaft (DFG) under SPP 1929 GiRyd Sci. Instrum. 82, 071101 (2011).
Project No. HO 4787/3-1. A. P-M. acknowledges support [25] O. Firstenberg, T. Peyronel, Q. Liang, A. V. Gorshkov, M. D.
from UNAM-PAPIIT IA101718 RA101718. Lukin, and V. Vuletic, Nature (London) 502, 71 (2013).
[26] M. Aßmann, F. Veit, M. Bayer, M. van der Poel, and J. M.
Hvam, Science 325, 297 (2009).
[27] M. Koch, C. Sames, M. Balbach, H. Chibani, A. Kubanek,
[1] H. J. Kimble, Nature (London) 453, 1023 (2008). K. Murr, T. Wilk, and G. Rempe, Phys. Rev. Lett. 107,
[2] D. E. Chang, V. Vuletić, and M. D. Lukin, Nat. Photonics 8, 023601 (2011).
685 (2014). [28] Q.-Y. Liang, A. V. Venkatramani, S. H. Cantu, T. L.
[3] A. Reiserer and G. Rempe, Rev. Mod. Phys. 87, 1379 Nicholson, M. J. Gullans, A. V. Gorshkov, J. D. Thompson,
(2015). C. Chin, M. D. Lukin, and V. Vuletić, Science 359, 783 (2018).
103601-5
PHYSICAL REVIEW LETTERS 121, 103601 (2018)
[29] S. Weber, C. Tresp, H. Menke, A. Urvoy, O. Firstenberg, [36] A. Gaëtan, Y. Miroshnychenko, T. Wilk, A. Chotia, M.
H. P. Büchler, and S. Hofferberth, J. Phys. B 50, 133001 Viteau, D. Comparat, P. Pillet, A. Browaeys, and P.
(2017). Grangier, Nat. Phys. 5, 115 (2009).
[30] R. H. Lehmberg, Phys. Rev. A 2, 883 (1970). [37] Y. O. Dudin, L. Li, F. Bariani, and A. Kuzmich, Nat. Phys.
[31] A. A. Svidzinsky, J.-T. Chang, and M. O. Scully, Phys. Rev. 8, 790 (2012).
Lett. 100, 160504 (2008). [38] See Supplemental Material at http://link.aps.org/
[32] D. Porras and J. I. Cirac, Phys. Rev. A 78, 053816 supplemental/10.1103/PhysRevLett.121.103601 for addi-
(2008). tional details about the theoretical results presented in the
[33] Y. Miroshnychenko, U. V. Poulsen, and K. Mølmer, Phys. main paper and the experimental sequence.
Rev. A 87, 023821 (2013). [39] A. Grankin, E. Brion, R. Boddeda, S. Ćuk, I. Usmani, A.
Ourjoumtsev, and P. Grangier, Phys. Rev. Lett. 117, 253602
[34] J. Kumlin, S. Hofferberth, and H. P. Büchler, Phys. Rev.
(2016).
Lett. 121, 013601 (2018).
[40] T. Shi, D. E. Chang, and J. I. Cirac, Phys. Rev. A 92, 053834
[35] E. Urban, T. A. Johnson, T. Henage, L. Isenhower, D. D.
(2015).
Yavuz, T. G. Walker, and M. Saffman, Nat. Phys. 5, 110
[41] V. I. Yudson, Sov. Phys. JETP 61, 1043 (1985).
(2009).
[42] V. I. Yudson and P. Reineker, Phys. Rev. A 78, 052713 (2008).
103601-6