Defect-induced supersolidity with soft-core bosons
F. Cinti,1, 2, ∗ T. Macr`ı,1 W. Lechner,3 G. Pupillo,4 and T. Pohl1 1Max Planck Institute for the Physics of Complex Systems, N¨othnitzer Str. 38, 01187 Dresden, Germany 2National Institute for Theoretical Physics (NITheP), Stellenbosch 7600, South Africa 3IQOQI and Institute for Theoretical Physics, University of Innsbruck, Austria 4IPCMS (UMR 7504) and ISIS (UMR 7006), Universit´ede Strasbourg and CNRS, Strasbourg, France More than 40 years ago, Andreev, Lifshitz, and Chester suggested the possible existence of a peculiar solid phase of matter, the microscopic constituents of which can flow superfluidly without resistance due to the formation of zero-point defects in the ground state of self-assembled crystals. Yet, a physical system where this mechanism is unambiguously established remains to be found, both experimentally and theoretically. Here we investigate the zero-temperature phase diagram of two- dimensional bosons with finite-range soft-core interactions. For low particle densities, the system is show to feature a solid phase in which zero-point vacancies emerge spontaneously and give rice to superfluid flow of particles through the crystal. This provides the first example of defects-induced, continuous-space supersolidity consistent with the Andreev-Lifshitz-Chester scenario.
Spontaneous symmetry breaking is a focal principle of condensed matter physics – yet, simultaneous break- ing of fundamentally different symmetries represents a rare phenomenon. A prime example is the so-called su- persolid phase [1], which displays both crystalline and superfluid properties, that is, the simultaneous breaking of continuous translational and global gauge symmetry. The first mentioning of such a state goes back to Gross [2], who predicted the possibility of a density-modulated superfluid phase of weakly interacting bosons described by a classical field. Later, Andreev, Lifshitz [3], and Chester (ALC) [4] conjectured a microscopic mechanism for strongly interacting systems based on two key as- sumptions: first, that the ground state of a bosonic crys- FIG. 1: Crystallization due to singular and soft-core tal contains defects such as vacancies and interstitials, interactions. (a) For singular potentials with pure power- and, second, that these defects can delocalize, thereby, law repulsion, indicated by the blue areas, particles assemble giving rise to superfluidity. However, the physical real- into a commensurate and insulating solid. (b) Upon remov- izability of this scenario has since remained under active ing the singularity, particles can cluster and, under proper debate [5]. conditions, form an incommensurate crystal with more parti- cles than lattice sites. Defect delocalization due to inter-site In 2004, torsional oscillator experiments [6, 7] provided tunneling can then promote a finite superfluid response of the 4 first suggestive evidence for superfluidity in solid He self-assembled crystalline ground state. through a rapid drop of the resonant oscillation period below a critical temperature, viewed indicative for su- perfluid decoupling of a fraction of the He crystal. This finding has sparked a host of new experimental activity stiffening of bulk solid He, and later found no signature [8–18], that, however, challenged the original interpre- of superfluidity on avoiding this effect [25]. As a result, tation and pointed out several artefacts causing a non- there now seems to be consistent experimental and the- supersolid origin of the observations. Theoretical work oretical evidence for the absence of the long-sought su- has established that crystal incommensurability is a nec- persolid phase in He. The mere existence of continuous- essary condition for superfluidity [19] and that zero-point
arXiv:1302.4576v2 [cond-mat.quant-gas] 6 Feb 2014 space supersolidity induced by zero-point defects thus re- defects in ground state solid He are prevented by a large mains an open question. activation energy [20, 21]. In addition, Boninsegni et al. In this Article, we show that bosonic particles in- [21] Rota and Boronat [22], Ma et al. [23], and Lech- teracting via soft-core potentials (see Fig.1) provide a ner and Dellago [24] have shown that point-like defects prototype system for addressing this question. Using experience an effective attraction that results in defect- exact numerical techniques, we determine the underlying clustering and phase separation, ruling out the possibility zero-temperature phase diagram, which reveals the emer- of defect-induced supersolidity [21] as in the ALC sce- gence of defect-induced supersolidity in the vicinity of nario. Several experiments [11, 14–16] have shown that commensurate solid phases, as conjectured by ALC [3, 4]. the original observations were caused by shear modulus 2
RESULTS
Supersolidity with Soft-core bosons. We consider a two-dimensional ensemble of N bosons with density ρ, interacting via a pair potential of the type
V0 V = γ γ . (1) r + Rc
γ This interaction approaches a constant value V0/Rc as the inter-particle distance, r, decreases below the soft- core distance Rc, and drops to zero for r > Rc. The limiting case γ → ∞ yields the soft-disc model [26], while γ = 3 and γ = 6 correspond to soft-core dipole-dipole [27] and van der Waals [28] interactions that can be realized with ultracold atoms [28, 29] or polar molecules [30, 31]. Here, we focus on the latter case (γ = 6), for which the Hamiltonian reads FIG. 2: Zero-temperature phase diagram of two- N N X ∇2 X U dimensional soft-core bosons. The phase diagram dis- Hˆ = − i + , (2) 2 1 + r6 plays the emergence of superfluid (SF) and different solid (NS) i=1 i
FIG. 3: Characterization of the different phases across the phase diagram. Panels (a)-(f) show the radial pair 2 correlation function g2(r) and the particle density distribution n(r) for ρR0U = 32 and different values of Aρ, indicated in (g). Panel (g) displays the superfluid fraction fs as a function of Aρ, covering all phases indicated in Fig.2. Global superfluidity is not well defined in the phase-separation region (green area) and therefore omitted. The horizontal dotted line shows the ρ → ∞ limit of fs for a density-wave supersolid, obtained from a mean-field evaluation following the approach of Leggett [58].
phase with exactly two particles per site (upper part of the figure), and a superfluid phase (lower part). Intermediate densities. A distinctive consequence We have carefully checked that the occurrence of this of the soft-core interaction is that the energy cost for phase-separated state is not an artifact of the simula- forming close particle pairs is bound by V0. This poten- tions, by performing accurate annealing and by choosing tially enables the formation of crystalline phases with different initial conditions, such as random and different N > Ns above a critical density where doubly occupied crystalline configurations. lattice sites become energetically favorable on increasing Above incommensurate lattice occupations the lattice constant. As expected for a triangular crys- √ N/Ns & 1.5, the direct liquid-solid quantum phase tal, the lattice constant decreases as a = ( 3ρ/2)−1/2 transition is replaced by a first order transition from a at small densities. However, around a ≈ 1.4 Rc it superfluid to a supersolid phase. The supersolid phase increases again and settles to a density-independent is approximately found to occur between the two hyper- value of a ' 1.6 R upon further increase of ρ. The 2 0 c bola defined by Rc ρU = α, with α ≈ 28 and α ≈ 38, p 2 corresponding volume of the unit cell A = 3/4a0 respectively (see dotted lines in Fig.2). These two lines provides a measure of the lattice occupancy N/Ns = Aρ, are derived from the weak interaction limit (U → 0 which is also shown in Fig.2. The transition to cluster and ρ → ∞ with α = const.), where mean-field theory crystals occurs at Aρ ≈ 1.5. This indeed coincides with predicts a transition to a density-wave supersolid that is the critical density for crystallization of the reentrant determined only by the value of α (Ref. [28]). While this superfluid phase. mean-field prediction becomes exact in the high-density Around this density, a thin region of phase separation limit (see Fig.2), the situation is dramatically different is found to lay in between the cluster crystal and the at moderate densities where the discrete nature of the superfluid phase. Fig.3(b) shows a typical example for particles plays a significant role. This gives rise to the the particle density distribution in this region. Two emergence of supersolid regions with a lobe structure, distinct coexisting phases can be recognized: a crystal 4
fraction of the crystal, which can assume sizable values of fs = 0.3 for N/Ns ≈ Aρ = 2.5.
High densities. For higher densities Aρ > 3, the scenario described above changes considerably. As shown in Fig.3(g) the superfluid fraction approaches a constant, density-independent value fs ≈ 0.24 with increasing ρ. In particular, for an average commensurate filling N/Ns = 4 there is no direct phase transition between a superfluid and a solid insulating phase, and instead the supersolid phase persists with no significant difference to the case of incommensurate lattice occupancies N/Ns 6= 4. This behavior signals a crossover to the regime where the supersolid phase can be understood in terms of a density modulated superfluid [2, 26–28, 39], where the discrete nature of the particles becomes irrelevant. In this limit, the FIG. 4: Defect-induced superfluidity. Panel (a) shows superfluid-supersolid quantum phase transition is well the superfluid fraction in a doubly occupied solid (N/Ns = 2) captured by a mean-field description [26, 28, 40]. It as a function of the fraction of defects in the form of singly predicts a transition point at α = R2ρU = 28.2 as well occupied sites for fixed U = 31 and varying particle number c between N = 139 and N = 144. For N = 144 the sys- as a superfluid fraction that is solely determined by the tem forms an insulating commensurate solid with Ns = 72 value α, and yields fs = 0.23 for α = 32. As shown doubly occupied sites. Decreasing the particle number does in Figs.2 and 3(g) both predictions are well confirmed not affect Ns but leads to the formation of a small fraction by our Monte Carlo results for Aρ & 3.5, suggesting fdef = (2Ns − N)/Ns of singly occupied lattice sites. Panel that the transition to density-wave supersolidity takes (b) shows the particle density (colour code) for fdef = 0.03, place at a surprisingly small number of only N/Ns ≈ 3.5 obtained from a Monte Carlo configuration along with the particles per lattice site. particle positions (spheres) for a single imaginary-time slice. The two initially adjacent vacancies (red and green sphere) delocalize, as indicated by the corresponding imaginary-time Defect delocalization. The most interesting be- trajectories. Accordingly, the vacancy-vacancy pair correla- havior takes place around the superfluid-solid quantum tion function (red), shown in (c) for fdef = 0.07, resembles the phase transition at N/Ns = 2. Figure 4 provides a more g2(r) of the underlying crystal (blue), as expected for a delo- detailed look at the transition between the insulating calized, very dilute gas of repulsive bosons. The error bars in crystal and the supersolid phase, that is, for U = 31 and panel a represent statistical uncertainties of the Monte Carlo sampling. N/Ns ≈ 2. Starting from the insulating solid with dou- bly occupied lattice sites, we successively remove a small number of particles from randomly chosen sites and mon- itor the superfluid fraction of the resulting new ground which vanish at commensurate lattice occupations state obtained from our simulations. Removing a small N/Ns = 2 and N/Ns = 3. There, we find a direct tran- number of particles does not cause structural changes sition between a superfluid and an insulating solid phase. of the ground state but rather creates a small fraction fdef = (2Ns − N)/Ns of zero-point crystal defects in the This behavior is illustrated in Fig.3, where the form of singly occupied sites. An analysis of the Monte superfluid fraction, fs, is shown as a function of Aρ Carlo configurations shows that defects do not cluster for α = 32, that is, in between the two dotted lines in and instead delocalize, as illustrated in Fig.4b. This is Fig.2. As shown in Fig.3(c) and (e), at Aρ = 2 and also confirmed by the vacancy-vacancy pair correlation Aρ = 3 one finds a commensurate crystal with exactly function, shown in Fig.4c. For r & a0 it closely resem- N/Ns = 2 [Fig.3(c)] and N/Ns = 3 [Fig.3(e)] particles bles the g2(r) of the underlying solid, as expected for a per lattice site, respectively, and vanishing superflu- very dilute gas of repulsive bosons. Indeed, we find a idity [Fig.3(g)]. Importantly, the crystal structure in finite superfluid fraction even for small defect concentra- between these two densities is practically unchanged, tions, which increases linearly with fdef . We have verified as seen by comparing the particle density distributions that this finding is pertinent to the ground state and not n(r) and density-density correlation functions g2(r) to a metastable configuration by performing simulations in Figs.3(c)-(e). However, the incommensurate lattice with different initial conditions, including clustered de- filling N/Ns and the resulting fluctuations of individual fects. The observed behavior is, thus, consistent with site occupations enables particles to tunnel between defect-induced supersolidity according to the ALC sce- the sites. This gives rise to a non-vanishing superfluid nario, and constitutes the central result of this work. 5
DISCUSSION molecules by external fields. Moreover, far off-resonant excitation of high lying Rydberg states [28, 29, 50] in degenerate atomic gases [48, 49, 51] was shown to real- Supersolidity in this system is the consequence of two ize interactions of the type of equation (1). Following unique features of soft-core bosons. First, the energy cost Maucher et. al [29], such a Rydberg dressing of 87Rb for forming close particle pairs is bound by V0, which condensates to Rb(35p3/2) states [52] with a laser detun- facilitates the formation of cluster crystals that natu- ing of ∼ 500MHz and an intensity of 100kW/cm2 would rally entail zero-point defects. Second, the dynamics and produce a sizeable interaction strength of U ∼ 35. While interaction of these defects differs fundamentally from we have focussed here on zero-temperatures, we have those of conventional solids. In the latter case, vacan- also performed finite-temperature simulations, showing cies and interstitials induce displacement fields that lead that these parameters will permit the experimental ob- to purely attractive defect interactions [21, 23, 24, 44] servation of defect-induced supersolid phases for temper- 8 −2 and, therefore, prevent a delocalization of defects [21]. In atures T . 10nK, around typical densities ∼ 10 cm cluster solids, on the other hand, defect interactions are and with a condensate lifetime of ∼ 30ms, limited by ra- purely repulsive, since they interact via the same under- diative decay of the weakly admixed Rydberg state. Re- lying particle interaction V (r) of eq. (1). In the present cent experimental breakthroughs reporting the first ob- case, the transition between these two regimes is con- servation of Rydberg interaction effects in a laser-driven trolled by the particle density. For Aρ & 1.5 delocalized Bose-Einstein condensate [53] hold high promise for the zero-point defects allow for the formation of supersolid near-future realization of the setting described in this phases. Below this density particles do not explore the work. soft core part of the interaction potential, such that de- fects are attractive and supersolidity is absent consistent with the results of Boninsegni et. al. [21]. Around the transition region Aρ ≈ 1.5 neither picture applies, and METHODS one observes separation between a superfluid and a dou- bly occupied, insulating cluster solid. Preliminary cal- culations based on path integral Langevin dynamics [45] Numerical Details. Our numerical results were obtained suggest that in this region structural and dynamical het- from path-integral Monte Carlo simulations [54] based on the erogeneity can give rise to a quantum glass phase at finite continuous-space worm algorithm [55] to determine the equi- temperature. librium properties of equation (2) in the canonical ensemble, Having identified a physical system that facilitates that is, at a fixed temperature temperature T and a fixed defect-induced supersolidity, we hope that this work will particle number, chosen between N = 100 and N = 400. provide useful guidance for future experiments and initi- From these simulations we obtain, for example, density pro- P ate further theoretical explorations. An important ques- files, n(r) = h i δ(r − ri(t))it, and pair correlation functions, −1 P P tion concerns the general features of the interaction po- g2(r) = [2πn(N − 1)r] h i j6=i δ(r − rij (t))it as well as tential that are required to maintain the type of super- the superfluid fraction fs, computed from the area estima- solid states described in this work. While the emer- tor, as described in Sindzingre et al. [56] and Pollock et gence of density-wave supersolids is largely insensitive to al. [57]. Here, rij = |rj − ri|, ri are the positions of the the detailed shape of the soft-core interaction, the low- i = 1, ..., N particles, and h..it denotes an average of the cor- density physics described in this work may be strongly responding imaginary time trajectories ri(t). The properties affected. In fact, it seems reasonable to expect an inter- of the system ground state were obtained by extrapolating to esting competition between intra- and inter-site interac- the limit of zero temperature, that is, by lowering the tem- tions within the self-assembled crystal that will strongly perature until observables, such as the total energy, superfluid depend on the long-range tail of the particle interactions. fraction and pair-correlations did not change upon further de- Moreover, the role of the dimensionality and confined ge- crease of T . Moreover, the size of our two-dimensional sim- ometries of finite systems represent another outstanding ulation box with periodic boundary conditions was varied to issue, and in particular their role for frustration effects assure insensitivity to system size. The calculated observables with regards to defect delocalization. were used to construct the phase diagram. The first order While the considered interactions do not straightfor- superfluid-normal solid transition is detected by an abrupt wardly occur in natural crystals, they can be designed in increase of the maximum value (Smax) of the static structure R ikr ultracold atom experiments. Recent experiments with factor S(k) = 1 + ρ dre (g2(r) − 1) and a simultaneous Bose-Einstein condensates in optical cavities have al- vanishing of the superfluid fraction. The superfluid-supersolid ready demonstrated a density-wave supersolid due to the transition is characterized by a jump of Smax and an abrupt breaking of a discrete translational symmetry [46, 47] decrease of the superfluid fraction from fs ≈ 1 to a finite value and theoretical work has devised several schemes [30, 31] fs > 0, while the supersolid-normal solid transition is signaled for manipulation of long-range interactions between polar by the vanishing of fs. 6
[24] Lechner, W. & Dellago, C., Defect interactions in two- dimensional colloidal crystals: vacancy and interstitial strings, Soft Matter 5, 2752-2758 (2009). ∗ Electronic address: [email protected] [25] Kim, D. Y. & Chan, M. H. W., Absence of Supersolidity [1] Boninsegni, M. & Prokof’ev, N. V., Colloquium: Super- in Solid Helium in Porous Vycor Glass, Phys. Rev. Lett. solids: What and where are they?, Rev. Mod. Phys. 84, 109, 155301 (2012). 759-776 (2012). [26] Pomeau, Y. & Rica, S., Dynamics of a model of super- [2] Gross, E. P., Unified Theory of Interacting Bosons, Phys. solid, Phys. Rev. Lett. 72, 2426-2430 (1994). Rev. 106, 161 (1957). [27] Cinti, F. et al., Supersolid Droplet Crystal in a Dipole- [3] Andreev, A. F. & Lifshitz, I. M., Quantum theory of Blockaded Gas, Phys. Rev. Lett. 105, 135301 (2010). defects in crystals, JETP 29, 1107-1113 (1969). [28] Henkel, N., Nath, R. & Pohl, T., Three-Dimensional Ro- [4] Chester, G. V., Speculations on Bose-Einstein Conden- ton Excitations and Supersolid Formation in Rydberg- sation and Quantum Crystals, Phys. Rev. A 2, 256-258 Excited Bose-Einstein Condensates, Phys. Rev. Lett. (1970). 104, 195302 (2010). [5] Balibar, S., The enigma of supersolidity, Nature 464, [29] Maucher, F. et al., Rydberg-Induced Solitons: Three- 176-182 (2010). Dimensional Self-Trapping of Matter Waves, Phys. Rev. [6] Kim, E. & Chan, M. H. W., Probable observation of a Lett. 106, 170401 (2011). supersolid helium phase, Nature 427, 225-227 (2004). [30] B¨uchler, H. P. et al., Strongly Correlated 2D Quantum [7] Kim, E. & Chan, M. H. W., Observation of superflow in Phases with Cold Polar Molecules: Controlling the Shape solid helium, Science 305, 1941-1944 (2004). of the Interaction Potential, Phys. Rev. Lett. 98, 060404 [8] Day, J. & Beamish, J., Pressure-Driven Flow of Solid (2007). Helium, Phys. Rev. Lett. 96, 105304 (2006). [31] Micheli, A., Pupillo, G., B¨uchler, H. P. & Zoller, P., Cold [9] Rittner, A. S. & Reppy, J., Observation of classical rota- polar molecules in two-dimensional traps: Tailoring in- tional inertia and nonclassical supersolid signals in solid teractions with external fields for novel quantum phases, He-4 below 250 mK, Phys. Rev. Lett. 97, 165301 (2006). Phys. Rev. A 76, 043604 (2007). [10] Aoki, Y., Graves, J. C. & Kojima, H., Oscillation Fre- [32] Likos, C. N., Lang, A., Watzlawek, M. & L¨owen, H., Cri- quency Dependence of Nonclassical Rotation Inertia of terion for determining clustering versus reentrant melting 4 Solid He, Phys. Rev. Lett. 99, 015301 (2007). behavior for bounded interaction potentials, Phys. Rev. [11] Day, J. & Beamish, J., Low-temperature shear modu- E 63, 031206 (2001). 4 lus changes in solid He and connection to supersolidity, [33] Mladek, B. M. et al., Formation of Polymorphic Cluster Nature 450, 853-856 (2007). Phases for a Class of Models of Purely Repulsive Soft [12] Hunt, B., et al., Evidence for a Superglass State in Solid Spheres, Phys. Rev. Lett. 96, 045701 (2006). 4 He, Science 324, 632-636 (2009). [34] Coslovich, D., Strauss, L. & Kahl, G., Hopping and mi- [13] West, J. T. , Lin, X., Cheng, Z. G. & Chan, M. H. W., croscopic dynamics of ultrasoft particles in cluster crys- Supersolid Behavior in Confined Geometry, Phys. Rev. tals, Soft Matter 7, 2127-2137 (2011). Lett. 102, 185302 (2009). [35] Sepulveda, N., Josserand, C. & Rica, S., Superfluid den- [14] Day, J., Syshchenko,O. & Beamish J., Intrinsic and sity in a two-dimensional model of supersolid, Eur. Phys. dislocation-induced elastic behavior of solid helium, J. B 78, 439-447 (2010). Phys. Rev. B 79, 214524 (2009). [36] Saccani, S., Moroni, S. & Boninsegni, M., Excitation [15] Day, J., Syshchenko, O., & Beamish, J., Nonlinear Spectrum of a Supersolid, Phys. Rev. B 83, 092506 Elastic Response in Solid Helium: Critical Velocity or (2011). Strain?, Phys. Rev. Lett. 104, 075302 (2010). [37] Kunimi, M. & Kato, Y., Mean-field and stability analyses [16] Reppy, J. D., Nonsuperfluid Origin of the Nonclassical of two-dimensional flowing soft-core bosons modeling a 4 Rotational Inertia in a Bulk Sample of Solid He, Phys. supersolid, Phys. Rev. B 86, 060510 (2012). Rev. Lett. 104, 255301 (2010). [38] Mason, P., Josserand, C. & Rica, S., Activated Nucle- [17] Choi, H., Takahashi, D., Choi, W., Kono, K. & Kim, ation of Vortices in a Dipole-Blockaded Supersolid Con- E., Staircaselike Suppression of Supersolidity under Ro- densate, Phys. Rev. Lett. 109, 045301 (2012). tation, Phys. Rev. Lett. 108, 105302 (2012). [39] Henkel, N., Cinti, F., Jain, P., Pupillo, G. & Pohl, T., [18] Mi, X. & Reppy, J. D., Anomalous Behavior of Solid Supersolid Vortex Crystals in Rydberg-Dressed Bose- 4 He in Porous Vycor Glass, Phys. Rev. Lett. 108, 225305 Einstein Condensates, Phys. Rev. Lett. 108, 265301 (2012). (2012). [19] Prokof’ev, N. & Svistunov, B., Supersolid State of Mat- [40] Macr`ı,T., Maucher, F., Cinti, F. & Pohl, T., Elemen- ter, Phys. Rev. Lett. 94, 155302 (2005). tary excitations of ultracold soft-core bosons across the [20] Ceperley, D. M. & Bernu, B., Ring Exchanges and the superfluid-supersolid phase transition, Phys. Rev. A 87, 4 Supersolid Phase of He, Phys. Rev. Lett. 93, 155303 061602 (2013). (2004). [41] Astrakharchik, G. E., Boronat, J., Kurbakov, I. L. & [21] Boninsegni, M. et al., Fate of Vacancy-Induced Superso- Lozovik, Y. E., Quantum Phase Transition in a Two- 4 lidity in He, Phys. Rev. Lett. 97, 080401 (2006). Dimensional System of Dipoles, Phys. Rev. Lett. 98, [22] Rota, R. & Boronat, J., Onset Temperature of Bose- 060405 (2007). 4 Einstein Condensation in Incommensurate Solid He, [42] Mora, C., Parcollet, O. & Waintal, X., Quantum melting Phys. Rev. Lett. 108, 045308 (2012). of a crystal of dipolar bosons, Phys. Rev. B 76, 064511 [23] Ma, P.N., Pollet, L., Troyer, M. & Zhang, F.-C., A clas- (2007). sical picture of the role of vacancies and interstitials in [43] Osychenko, O. N., Astrakharchik, G. E., Lutsyshyn, Y., Helium-4, J. of Low Temp. Phys., 152, 156-163 (2008). Lozovik, Yu. E. & Boronat, J., Phase diagram of Rydberg 7
atoms with repulsive van der Waals interaction, Phys. [53] Balewski J. B. et al., Coupling a single electron to a Rev. A 84, 063621 (2011). Bose-Einstein condensate, Nature, 502 664-667 (2013). [44] Lechner, W. & Dellago, C., Point defects in two- [54] Ceperley, D. M., Path integrals in the theory of con- dimensional colloidal crystals: simulation vs. elasticity densed helium, Rev. Mod. Phys. 67, 279-355 (1995). theory, Soft Matter 5, 646-659 (2009). [55] Boninsegni, M., Prokof’ev, N. & Svistunov, B., Worm Al- [45] Markland, T. E., et al., Theory and simulations of quan- gorithm for Continuous-Space Path Integral Monte Carlo tum glass forming liquids, J. Chem. Phys. 136, 074511 Simulations, Phys. Rev. Lett. 96, 070601 (2006). (2012). [56] Sindzingre, P., Klein, M. L. & Ceperley, D. M., Path- [46] Baumann, K., Guerlin, C., Brennecke, F. & Esslinger, T., integral Monte Carlo study of low-temperature 4He clus- Dicke quantum phase transition with a superfluid gas in ters, Phys. Rev. Lett. 63, 1601-1605 (1989). an optical cavity, Nature 464, 1301-1306 (2010). [57] Pollock, E. L. & Ceperley, D. M., Path-Integral Compu- [47] Mottl, R., et al., Roton-Type Mode Softening in a Quan- tation of Superfluid densities, Phys. Rev. B 36, 8343-8352 tum Gas with Cavity-Mediated Long-Range Interactions, (1987). Science 336, 1570-1573 (2012). [58] Leggett, A. J., Can a Solid Be ”Superfluid”?, Phys. Rev. [48] Heidemann, R., et al., Rydberg Excitation of Bose- Lett. 25, 1543-1547 (1970). Einstein Condensates, Phys. Rev. Lett. 100, 033601 (2008). [49] Viteau, M., et al., Rydberg Excitations in Bose-Einstein Condensates in Quasi-One-Dimensional Potentials and ACKNOWLEDGEMENT Optical Lattices, Phys. Rev. Lett. 107, 060402 (2011). [50] Pupillo, G., Micheli, A., Boninsegni, M., Lesanovsky, I. & Zoller, P., Strongly Correlated Gases of Rydberg-Dressed We thank M. Boninsegni, S. Pilati, N. V. Prokof’ev, Atoms: Quantum and Classical Dynamics, Phys. Rev. and S. G. S¨oylerfor valuable discussions. This work was Lett. 104, 223002 (2010). supported by the EU through the ITN COHERENCE. [51] Schauß P. et al., Observation of spatially ordered struc- tures in a two-dimensional Rydberg gas, Nature 491, 87- G. P. is supported by the ERC-St Grant “ColDSIM” 91 (2012). (grant agreement 307688) and EOARD. W.L. acknowl- [52] Tong, D., et al., Local Blockade of Rydberg Excitation edges support by the Austrian Science Fund through P in an Ultracold Gas, Phys. Rev. Lett. 93, 063001 (2004). 25454-N27.