Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

# Sorting Fermionization from Crystallization in Many-Boson Wavefunctions

## Abstract

Fermionization is what happens to the state of strongly interacting repulsive bosons interacting with contact interactions in one spatial dimension. Crystallization is what happens for sufficiently strongly interacting repulsive bosons with dipolar interactions in one spatial dimension. Crystallization and fermionization resemble each other: in both cases – due to their repulsion – the bosons try to minimize their spatial overlap. We trace these two hallmark phases of strongly correlated one-dimensional bosonic systems by exploring their ground state properties using the one- and two-body density matrix. We solve the N-body Schrödinger equation accurately and from first principles using the multiconfigurational time-dependent Hartree for bosons (MCTDHB) and for fermions (MCTDHF) methods. Using the one- and two-body density, fermionization can be distinguished from crystallization in position space. For N interacting bosons, a splitting into an N-fold pattern in the one-body and two-body density is a unique feature of both, fermionization and crystallization. We demonstrate that this splitting is incomplete for fermionized bosons and restricted by the confinement potential. This incomplete splitting is a consequence of the convergence of the energy in the limit of infinite repulsion and is in agreement with complementary results that we obtain for fermions using MCTDHF. For crystalline bosons, in contrast, the splitting is complete: the interaction energy is capable of overcoming the confinement potential. Our results suggest that the spreading of the density as a function of the dipolar interaction strength diverges as a power law. We describe how to distinguish fermionization from crystallization experimentally from measurements of the one- and two-body density.

## Introduction

The physics of the ultracold Bose gas in one spatial dimension is strongly different from that of its three-dimensional counterpart1,2. In one spatial dimension, quantum effects are much more prominent. When the interactions are strong, quantum fluctuations are enhanced. Experimentally, in cold atom systems, the dimensionality can be manipulated using a tight transversal confinement that essentially freezes the radial degrees of freedom3,4. Such quasi-one dimensional systems display intriguing physics that cannot be realized for three-dimensional systems: Fermionization, occurs for strongly interacting bosons with contact interactions5,6,7,8,9,10,11,12 and crystallization emerges for sufficiently strongly interacting bosons with dipole-dipole interactions13,14,15,16,17,18,19. For bosons with contact interactions, fermionization leads to the formation of the Tonks-Girardeau (TG) gas when the interaction strength tends to infinity. This is a consequence of the Bose-Fermi mapping7,20,21,22 which implies that strongly interacting bosons and non-interacting spinless fermions have the same (one-body) density in position space. With increasing interaction strength, not only the density, but also the energy of fermionized bosons saturates to the energy of non-interacting fermions.

In the case of dipolar interactions, the remarkable phenomenon of crystallization occurs when the interaction strength is sufficiently large. Bosons interacting via a dipole-dipole interaction potential have become the primary cold atom system to investigate the many-body physics triggered as consequence of long-range interactions23,24,25,26,27,28. The long-ranged and anisotropic nature of the dipolar interaction potential results in a variety of interesting effects and phenomena29, like crystallization in one- and two-dimensional systems13,14,15,16,17,18,30,31,32, that are completely different from the emergent phenomena in the case of strong contact interactions. Crystallization is a consequence of the repulsive and long-ranged tail of the dipolar interactions dominating the physics33,34: the bosons maximally separate and minimize their spatial overlap. Unlike in the fermionization of bosons with strong contact interactions, the energy of crystallized bosons does not saturate. We note here that it is formally possible to define and measure an order parameter which is a function of the eigenvalues of the reduced one-body density matrix that allows to unequivocally identify the crystal phase of one-dimensional dipolar bosons34. Furthermore, we note that the formation of a crystal state is a generic feature of many-body systems of particles with long-ranged interactions. Fermions with long-ranged interactions, for instance, form a so-called Wigner crystal19,35.

In this work, we analyze the differences between fermionized and crystallized bosons’ wavefunctions using the energy as well as the one-body and two-body reduced density matrix. We demonstrate how fermionization can be distinguished from crystallization by quantifying the (experimentally accessible) spread of the one-body and two-body densities. The different spreading characteristics of the one-body and two-body densities for fermionized as compared to crystallized bosons are a direct consequence of the different behavior of the energy as a function of interaction strengths. For dipolar interactions the energy as a function of interaction strength is unbounded; this is in stark contrast to the bounded energy as a function of interactions for contact interactions. Current experimental setups, for instance, for Erbium36, Erbium-Erbium molecules37, or Sodium-Potassium molecules38, enable the experimental exploration of systems with dominant dipole-dipole interactions needed for probing the physics of crystallization.

The (momentum) densities of bosons with dipolar interactions have been compared to those of fermions with dipolar interactions in Ref. 18. Going beyond Ref. 18, we compare and quantify the spreading of the full density matrices of bosons with dipolar interactions to the spreading of the density matrices of bosons with contact interactions. We note that Ref. 39 discusses and compares the physics of spin-1/2 fermions with contact and with long-ranged interactions in lattices using a Hubbard description. Our work complements the findings in Ref. 39 by providing a comparison of single-component (“spin-0”) bosons with contact and long-ranged dipolar interactions in continuous space without a lattice and without resorting to a Hubbard-tight-binding-description.

Fermionization and crystallization entail the breakdown of mean-field approaches like the time-dependent Gross-Pitaevskii (GP) equation40,41,42. To go beyond the GP approximation, multiconfigurational methods are employed43,44,45,46,47,48,49. Variational calculations using parametrized Gaussian functions as single-particle states have been successfully applied to investigate the crystallization of few particles in two dimensions30,50,51. Here, we use the multiconfigurational time-dependent Hartree for bosons (MCTDHB)43,44 and fermions (MCTDHF)46 methods implemented in the MCTDH-X software package46,52,53,54 to compute the ground state of the few-particle Schrödinger equation, see Ref. 49 for a Review. While the MCTDHB method aims at solving the time-dependent Schrödinger equation for a many-body system, using imaginary time-propagation provides the ground-state of the system variationally, equivalent to the work of Ref. 55.

We illustrate our findings with computations for N = 4 bosons in a parabolic trapping potential and trace the complete range of dipolar and contact interaction strengths by obtaining highly accurate results with MCTDHB.

This paper is structured as follows: in Sec. 2, we discuss the Hamiltonian and quantities of interest, in Sec. 3, we introduce the numerical method, MCTDHB, that we use for obtaining solutions of the few-body Schrödinger equation, in Sec. 4 we analyze fermionized and crystallized bosonic few-body states and discuss how they can be sorted from each other and we conclude our paper in Sec. 5. Results for other observables and an assessment of the accuracy of our computations with the exact diagonalization and MCTDHF approaches are collected in the Appendices 6 and 7.

## Hamiltonian, One- and Two-Body Density

In order to discuss the stationary properties of the ground state (GS) of crystalline and fermionized bosons, we consider the time-independent many-body Schrödinger equation,

$$\hat{H}|\Psi \rangle =E|\Psi \rangle .$$
(1)

Here, |Ψ〉 is the many-body ground state, E its energy, and $$\hat{H}$$ the N-particle Hamiltonian in dimensionless units56,

$$H=\mathop{\sum }\limits_{i=1}^{N}\,(-\frac{1}{2}\frac{{\partial }^{2}}{\partial {x}_{i}^{2}}+V({x}_{i}))+\sum _{i < j}\,W({x}_{i}-{x}_{j}),$$
(2)

where we set $$V({x}_{i})=\frac{1}{2}{x}_{i}^{2}$$ to be the external harmonic trap. The term W(xi − xj) is the interaction potential. All quantities are dimensionless and expressed in harmonic oscillator units. To ensure that the system is in the quasi-1D regime, we assume strong confinement in the transversal direction, providing a cigar-shaped atomic density. The contact interactions read,

$$W({x}_{i}-{x}_{j})=\lambda \delta ({x}_{i}-{x}_{j}),$$
(3)

where λ is the interaction strength determined by the scattering length as and the transverse confinement frequencies57. For long-ranged dipolar interactions, we have

$$W({x}_{i}-{x}_{j})=\frac{{g}_{d}}{|{x}_{i}-{x}_{j}{|}^{3}+\alpha },$$
(4)

where gd is the dipolar interaction strength and α is a short-range cut-off to avoid the divergence at xi = xj. Repulsive interactions could be obtained in a quasi-1D BEC by imposing an external magnetic field to align all the dipole moments of the atoms18. This simple approximation to the one-dimensional dipole-dipole interaction potential in Eq. (4) is justified for the moderate to large interaction strengths and large inter-particle distance with respect to the harmonic-length of the transversal confinement14,15,18,19,58, that we focus on in the present work. For such interaction strengths the dipole-dipole interaction potential is well-approximated by the |xi − xj|−3 tail in Eq. (4), see59. Moreover, we have verified the consistency of the approximation in Eq. (4) for the same choice of cutoff parameter, α = 0.05, by a direct comparison to a dipole-dipole interaction augmented with an additional contact interaction potential, see Ref. 34. A rigorous discussion of the dipole-dipole interaction potential in one and two spatial dimensions can be found in Ref. 59. Here, for the sake of simplicity, we will focus on quasi-one-dimensional systems, with N = 4 interacting bosons for all our calculations and consider repulsive interactions, i.e., λ > 0 and gd > 0, exclusively.

In the following we discuss the reduced one-body density matrix, defined as

$${\rho }^{(1)}(x,x^{\prime} )=\langle \Psi |{\hat{\psi }}^{\dagger }(x^{\prime} )\hat{\psi }(x)|\Psi \rangle .$$
(5)

Its diagonal,

$$\rho (x)\equiv {\rho }^{(1)}(x,x^{\prime} =x)$$
(6)

is simply the one-body density. As a precursor of correlation effects that may be present in the state |Ψ〉 of the system, we use the eigenvalues ρi(NO) of the reduced one-body density matrix ρ(1) in Eq. (5). For this purpose, we write ρ(1) in its eigenbasis:

$${\rho }^{\mathrm{(1)}}(x,x^{\prime} )=\sum _{i}\,{\rho }_{i}^{(NO)}{\Phi }_{i}^{\ast }(x^{\prime} ){\Phi }_{i}(x\mathrm{).}$$
(7)

The eigenvalues ρi(NO) and eigenfunctions Φi(x) are referred to as natural occupations and natural orbitals, respectively. If only a single eigenvalue ρi(NO) is macroscopic, then the state |Ψ〉 describes a Bose-Einstein condensate60. The case when multiple eigenvalues ρi(NO) are comparable to the number of particles N is referred to as fragmentation43,44,53,54,61,62,63.

In the following, we will also use the two-body density ρ(2) to characterize crystallization and fermionization. It is defined as

$${\rho }^{(2)}({x}_{1},{x}_{2})=\langle \Psi |{\hat{\psi }}^{\dagger }({x}_{1}){\hat{\psi }}^{\dagger }({x}_{2})\hat{\psi }({x}_{1})\hat{\psi }({x}_{2})|\Psi \rangle .$$
(8)

The two-body density quantifies the probability to detect two particles at positions x1 and x2.

## Numerical Method

The computation of the exact many-body wave function is a difficult problem. To attack the many-body Schrödinger equation, Eq. (1), we use the time-dependent Schrödinger equation,

$$i{\partial }_{t}|\Psi \rangle =\hat{H}|\Psi \rangle ,$$
(9)

with a Wick’s rotation t → −, i.e., a propagation with imaginary time. We expand the many-body wavefunction |Ψ〉 of N interacting bosons in a complete set of time-dependent permanents $$|\overrightarrow{n};t\rangle =|{n}_{1},\,\mathrm{...,}{n}_{M};t\rangle$$ with at most M single-particle states or orbitals. The MCTDHB ansatz for the many-body wave function is thus

$$|\Psi (t)\rangle =\sum _{\overrightarrow{n}}\,{C}_{\overrightarrow{n}}(t)|\overrightarrow{n};t\rangle .$$
(10)

Here, the permanents $$|\overrightarrow{n};t\rangle$$ are symmetrized bosonic many-body states that are also referred to as “configurations”. The sum in Eq. (10) runs on all configurations $$\overrightarrow{n}$$ of N particles in M orbitals. The number of permanents and coefficients $${C}_{\overrightarrow{n}}(t)$$ is Nconf = $$(\begin{array}{c}N+M-1\\ N\end{array})$$. In second quantized representation the permanents are given as

$$|\bar{n};t\rangle =|{n}_{1},\,...\,{n}_{M};t\rangle =(\mathop{\prod }\limits_{i=1}^{M}\,\frac{{({\hat{b}}_{i}^{\dagger }(t))}^{{n}_{i}}}{\sqrt{{n}_{i}!}})|vac\rangle .$$
(11)

Here $${\hat{b}}_{k}^{\dagger }(t)$$ is the bosonic creation operator which creates a boson in the time-dependent single particle state $${\varphi }_{k}(x,t)$$. Equation (10) spans the full N-body Hilbert space in the limit of M → ∞. For practical computations, we restrict the number of orbitals and require the convergence of our observables, like the one- and two-body density matrix, with respect to the number of single-particle states M.

A set of coupled equations of motion for, both, the time-dependent expansion coefficients $${C}_{\overrightarrow{n}}(t)$$ and the time-dependent orbitals $${\varphi }_{k}(x,t)$$ are obtained by requiring the stationarity of the action of the time-dependent Schrödinger equation43,44 under variations of $${C}_{\overrightarrow{n}}(t)$$ and $${\varphi }_{k}(x,t)$$. Using MCTDHB, both, the coefficients and orbitals are variationally optimized64. MCTDHB is thus fundamentally different from exact diagonalization, i.e., an ansatz built with time-independent orbitals. It can be demonstrated that MCTDHB delivers solutions of the Schrödinger equation at a significantly increased accuracy in comparison to exact diagonalization approaches when the same number of single-particle basis states is employed, see Refs. 53,65. for a demonstration with the harmonic interaction model and Appendix A for a demonstration with dipole-dipole interactions, i.e., the Hamiltonian in Eqs. (2) and (4). Despite the accuracy of MCTDHB for weakly interacting particles, for strong interactions a large number of orbitals is required to describe the system accurately. In the case λ → ∞, the Bose-Fermi mapping provides analytical solutions that can be compared to the numerical results. These difficulties to converge MCTDHB results for strong contact interactions may be an instance of the discussion provided in Ref. 66.

We solve the set of coupled MCTDHB equations using the MCTDH-X software46,52,53,54. When imaginary time is used, the propagation of an initial guess function converges to the ground state of the system, and the stationary properties of the system can be investigated.

## Fermionization vs Crystallization

We now discuss our findings on the fermionized and the crystalline state of parabolically trapped one-dimensional ultracold bosons. We first independently characterize fermionization and crystallization from a “many-body point of view”, see Sec. 4.1 and Sec. 4.2, respectively. Thereafter, we investigate how to sort the one, fermionization, from the other, crystallization, in Sec. 4.3. Here and in the following, we used the term “many-body point of view” to highlight that our considerations go beyond an effective single-particle or mean-field description of the state.

### Fermionization

Bosons fermionize when they feel an infinitely repulsive contact interaction in one spatial dimension. For fermionized bosons, the total energy E and the density [Eq. (6)] of the system become exactly equal to the energy and the density of non-interacting spinless fermions, respectively. For our showcase of few-bosons systems (N = 2 to N = 5) in a harmonic trap with frequency one, $$V(x)=\frac{1}{2}{x}^{2}$$, the limiting value is thus $${E}_{\lambda \to \infty }^{N}=\frac{{N}^{2}}{2}$$.

We start our investigation with the one-body density as a function of the interaction strength λ [Fig. 1(a–b)]. For comparatively weak repulsion, the density is clustered at the center of the trap, but becomes flatter and broader when λ increases. For stronger repulsion, the density gradually acquires modulations and the number of humps finally saturates to the number of bosons in the system; four humps for N = 4 bosons are clearly visible when the interaction strength goes above λ ~ 10. The emergence of N maxima in the density indicates that the TG regime is approached. The density modulations/humps are more pronounced in the center of the trap, where the potential is close to zero. For larger distances from the origin, the humps in the density are less pronounced due to the non-zero value of the confinement potential. Importantly, the outermost density modulation also becomes less pronounced if the number of particles N is increased. See also Appendix B for a direct comparison of the relative height of innermost and outermost peaks for different particle numbers N.

We note that the density’s maxima in the Tonks-Girardeau regime are distinct but not isolated. We also observe that, once the TG regime is reached, the density does not broaden further with increasing values of λ for all particle numbers. We also provide a direct comparison with the ground state properties of non-interacting fermions computed with the multiconfigurational time-dependent Hartree method for fermions (MCTDHF), see Fig. 2. We note that the results for non-interacting fermions can be obtained analytically, i.e., here, we use the heavy MCTDHF method only for the sake of computational convenience.

We now move to discuss the two-body densities ρ(2) of bosons with contact interactions [see Fig. 3(a)]. For weak interaction strength, λ = 0.1, the bosons are clustered near the center, i.e, at x1 = x2 = 0 [Fig. 3(a)]. As the interaction strength increases, ρ(2) spreads out to the off-diagonal (x1 ≠ x2) while the diagonal ($${x}_{1} \sim {x}_{2}$$) is depleted [see Fig. 3(a) for λ = 1].

For stronger repulsion a so-called “correlation hole” in the two-body density forms on the diagonal, ρ(2)(x, x) → 0 [see Fig. 3(a) for λ = 10 and λ = 30]. The probability of finding two bosons at the same position tends towards zero. In the limit of infinite repulsion the correlation hole persists in ρ(2). In analogy, however, to the boundedness of the energy as a function of the interaction strength, the width of two-body density on its anti-diagonal [ρ(2)(x, −x)] is also bounded, i.e., the spread of ρ(2) converges in the fermionization limit when λ → ∞.

Similar to the one-body density, the maxima which are formed in the off-diagonal of the two-body density are distinct but not isolated [see Fig. 3(a) for λ = 10 and λ = 30].

We infer that the correlation hole along the diagonal and the confined spread are the unique signatures of the two-body density of a fermionized state.

### Crystallization

For bosons with dipole-dipole interactions, crystallization occurs when the long-range tail of the interaction [see Eq. (4)] becomes dominant34: the bosons form a lattice structure which allows them to minimize their mutual overlap. To characterize crystallization we analyze the one-body and two-body density for bosons with dipolar interaction of strength gd. We choose the cut-off parameter α = 0.05 in Eq. (4) such that the effective interaction features the same physical beahavior as the “real” dipolar interaction that additionally contains a contact-interaction term (see Ref. 34 for a direct comparison).

We plot the one-body density of N = 4 bosons as a function of gd in Fig. 1(c,d). The system is condensed at the center of the trap for small gd. As gd increases, the density starts to exhibit a four-hump structure (see Fig. 1(c,d) for $${g}_{d}\in [\sim \,1,\sim \,\mathrm{5]}$$) similar to the density observed for the fermionization of bosons with contact interactions [Fig. 1(a)].

This attempted fermionization results from a dominant contribution of the short-range part of the dipolar interaction potential, see also Ref. 18. However, this fermionization-like behavior is only a precursor to the crystal transition that takes place when the long-range nature of the interaction starts to dominate the physics of the system for larger interaction strengths [Fig. 1(c,d) for $${g}_{d}\gtrsim 10$$]. For crystallized dipolar bosons at sufficiently large gd, the value of the density at its minima between the humps tends to zero while the spreading of the density profile diverges as gd increases, see Fig. 4. At gd = 30.0, we observe four well-isolated peaks heralding the crystallization of the N = 4 bosons. We collect results for other numbers of bosons (N = 2, 3, 5, 6) with dipole-dipole interactions – including the relative height of the peaks in the density that shows that the peaks are well-isolated in comparison to particles with contact interactions – in Appendix B. A comparison of MCTDHB results with exact diagonalization is shown in Appendix A.

We now analyze the two-body density for dipole-dipole interactions [Fig. 3(b)]. For small interaction strength, gd = 0.1, the atoms are clustered together at the center of the trap. As gd increases, a correlation hole develops: ρ(2)(x, x) tends to zero [Fig. 3(b) for gd ≥ 1]. Thus, due to the long-range interaction, the probability of finding two bosons in the same place is strongly reduced. In the crystalline phase [Fig. 3(b) for gd ≥ 10]: the bosons escape their spatial overlap entirely and even the off-diagonal peaks of ρ(2) become isolated. We term this behavior the formation of an off-diagonal correlation hole. For crystallized bosons, the spread of the anti-diagonal of the two-body density, ρ(2)(x, −x), is diverging as gd is increasing [compare Fig. 3(b) for gd = 10 to Fig. 3(b) for gd = 30].

We assert that the correlation hole along the diagonal and the off-diagonal and the unbounded spreading are the unique signatures of the two-body density of a crystalline state of dipolar bosons.

### Sorting crystallization from fermionization

We now discuss how to distinguish fermionized from crystallized many-body states. One clear distinction is given by the spread of the one- and two-body densities: for bosons with contact interactions it is bounded, while for bosons with dipole-dipole interactions it diverges as a function of the interaction strength. We assert that, (1) the bounded spreading of the density for contact interactions is a consequence of the bounded energy as the interaction strength tends to infinity. Similarly, we assert, (2) that the unbounded spreading of the density for dipole-dipole interactions is a consequence of the unbounded energy as the interaction strength gd tends to infinity. To validate the assertions (1) & (2), we quantify the spreading of the density as a function of interactions and plot the position of its outermost peak as a function of the interaction strength in Fig. 5(a) and compare it to the energy in Fig. 5(b), for N = 4.

From fitting the energy in Fig. 5(b) we can infer that the energy as a function of contact interaction strength approaches the fermionization limit exponentially, a power law does not fit as accurately the data. For very large interactions, in the limit of λ−1 → 0, our results are in agreement with the analysis in Ref. 67, see Appendix B. For dipolar interactions, the growth of the energy as a function of the interaction strength is fitting well to a power law.

Indeed, the comparison of Fig. 5(a,b) corroborates our assertions (1) & (2), and holds for different number of particles.

We thus conclude that the crystalline phase can be distinguished from the TG regime gas by virtue of the behavior of its density profile as a function of the strength of the interparticle interactions: The width of the density distribution converges for an increasing strength of contact interactions, but it continuously spreads for an increasing strength of long-range interactions [compare Fig. 1(b) with Fig. 1(d) as well as Fig. 5(a) with Fig. 5(b)]. In Appendix B, we demonstrate that the exponent of the power law of the spreading of the density as a function of the strength of the interaction is independent of the particle number N.

In the case of long-ranged interactions, the unbounded spreading of densities as a function of increasing interaction strength and the formation of well-isolated peaks are in sharp contrast to the bounded spreading of densities and the non-isolated peaks in the case of contact interactions in the TG regime [cf. Figs 1 and 5(a–b)].

We now turn to analyze the eigenvalues of the reduced one-body density matrix, the so-called natural occupations60, as a function of the interaction strength between the particles [Fig. 5(c,d)]. As expected8,33,34, when the value of the interaction strength increases, the occupation of the first natural orbital decreases while the other orbitals start to be occupied. For contact interactions, mostly one natural occupation, n1, dominates, while the other occupations nk, k > 1 remain comparatively small even for large values of λ: depletion emerges as the fermionized state is reached [Fig. 5(c)], see also Ref. 8. For long-range interactions, however, all occupations ρk(NO) for k ≤ N contribute on an equal footing for large values of gd. This full-blown N-fold fragmentation emerges as the crystal state is reached [Fig. 5(d)], see also Ref. 34. In the crystal state the bosons behave similar to distinguishable particles68, and the particle statistics does not influence the physical observables considered in Fig. 5(b,d): the energy and the natural occupations for bosons and fermions converge to the same values. Thus, the finding of Ref. 68 for two particles may be extended to larger number of particles.

The emergence of complete fragmentation is a consequence of long-ranged interactions and in sharp contrast to the emergent depletion in the case of contact interactions.

## Conclusions

In this paper we highlight the key characteristics of the many-body wavefunction that reveal the difference between the fermionized bosons with contact interactions and crystallized bosons with dipolar interactions.

In the case of fermionization, the one-(two-)body density shows a modulation with a number of maxima corresponding to the number of particles. The maxima are confined but not completely separated. The incomplete separation is a consequence of the representability of momentum distribution of fermionized bosons using a basis set: infinitely many basis states are necessary to accurately resolve the cusp – a fact that is reflected by the depletion of the state which we quantified by the eigenvalues of the reduced one-body density matrix. We found that the peaks in the density as well as the energy as a function of the interaction strength approach the fermionization limit exponentially.

In the case of crystallization, the one-(two-)body density shows well-separated peaks whose distances diverge as a function of the interaction strength as a power law. This completed separation is the consequence of the formation of a Mott-insulator-alike many-body state where the “lattice potential” is replaced by the long-ranged interparticle interactions and the “lattice constant” is dictated by the strength of the interparticle interactions.

We close by stating that all the signatures that distinguish crystalline bosons from fermionized bosons can be measured experimentally using single-shot absorption imaging69,70,71,72,73. From experimental absorption images, the one-body and two-body density are available as averages of many single-shot images. Thus, a direct verification of our results for the spread of the one-body and two-body density can be performed. Furthermore, Refs. 34,63 suggest that the natural occupations can be inferred from the integrated variance of single-shot images, at least at zero temperature. It is, of course, an open question how thermal fluctuations affect the variance in absorption images and up to which temperature it is still possible to determine the fragmentation of the system.

## Appendix A: Comparison of MCTDHB and exact diagonalization

In this Appendix, we demonstrate that MCTDHB yields solutions to the Schrödinger equation at a larger accuracy as compared to the exact diagonalization approach (ED). As is conventional, we use the eigenfunctions of the non-interacting system as the single-particle basis states for the ED. We solve the same system as shown in Fig. 1(c),(d) for an interaction strength of gd = 30 and compare the energies obtained with MCTDHB and ED, see Fig. 6. Due to the variationally optimized single-particle basis in MCTDHB computations it features a much smaller error than the ED computations with an unoptimized single-particle basis for the same number of orbitals. This observation is in agreement with other works that benchmark the MCTDHB and the MCTDHF approaches against ED, see Ref. 65 and Ref. 46, respectively.

## Appendix B: Different Particle Numbers

In this Appendix, we corroborate our results in the main text by studying different particle numbers.

### Contact interactions

The results of the manuscript have been obtained with MCTDHB with M = 12, 14, 20, 22 orbitals for N = 2, 3, 4, 5 bosons, respectively, with a contact interaction strength up to λ = 1000. In Fig. 2 of the main text, our results are consistent with fits of an exponential function AN[exp(−λ/BN) − 1], see caption of Fig. 2 for the fitting parameters AN and BN. Furthermore, we assess the convergence of the spread of the density as a function of the interaction strength to the spread of the density of the non-interacting fermionic system, see arrows labeled “2F”,“3F”,“4F”, and “5F” in Fig. 2.

To compare our results for the energy in the fermionization limit to analytical predictions for very large contact interaction strengths in Ref. 67, we plot the energies as a function of −λ−1 in Fig. 7. We find that our results are consistent with the linear limit for the energy as a function of −λ−1 of Ref. 67.

We now turn to the relative height of the innermost and outermost peak(s),

$$\Delta \rho (x)=\frac{{\rho }_{\max }-{\rho }_{\min }}{{\rho }_{\max }+{\rho }_{\min }}$$
(B1)

in the density. Here, ρmax refers to the value of the density ρ(x) at the peak position and ρmin refers to the value of the density ρ(x) at the position of the minimum to the left to the considered peak. See Fig. 8 for a plot of Δρ(x) for N = 2, 3, 4, 5 bosons. It is clearly seen that, for fixed N, the outermost peaks’ relative height is much smaller than the relative height of the innermost peaks.

### Dipolar interactions

Here, we assess the validity of the power-law-like unbounded spreading of the density as a function of the strength of dipole-dipole interactions, that we have shown in Fig. 4 of the main text for N = 4 particles. In Fig. 4 We plot the spread of the density for N = 2, 3, 4, 5, 6 dipolar bosons obtained with MCTDHB with M = 16, 16, 22, 28, 26 orbitals, respectively, and fit it with a power law $${C}_{N}{g}_{x}^{{D}_{N}}$$. We find that the exponent in the power law is almost identical for all particle numbers studied here, i.e., DN ≈ 0.16 for N = 2, 3, 4, 5, 6.

We now discuss the relative peak height Δρ(x), see Eq. (B1), of the outermost peak as a function of the dipolar interaction strength, see Fig. 9 for a plot for N = 2, 3, 4, 5, 6. As hinted by Fig. 1(c) in the main text, the relative peak height for the case of the dipole-dipole interactions converges towards unity as the strength of interactions gd increases, because the values of the minimum, ρmin in Eq. (B1), tends to zero: the peaks in the crystal state are well-isolated in comparison to the peaks in the fermionization limit for bosons with contact interactions (compare magnitude of relative peak heights in Figs. 8 and 9).

## References

1. 1.

Petrov, D. S., Shlyapnikov, G. V. & Walraven, J. T. M. Regimes of Quantum Degeneracy in Trapped 1D Gases. Phys. Rev. Lett. 85, 3745 (2000).

2. 2.

Dunjko, V., Lorent, V. & Olshanii, M. Bosons in Cigar-Shaped Traps: Thomas-Fermi Regime, Tonks-Girardeau Regime, and In Between. Phys. Rev. Lett. 86, 5413 (2001).

3. 3.

Görlitz, A. et al. Realization of Bose-Einstein Condensates in Lower Dimensions. Phys. Rev. Lett. 87, 130402 (2001).

4. 4.

Greiner, M. et al. Exploring Phase Coherence in a 2D Lattice of Bose-Einstein Condensates. Phys. Rev. Lett. 87, 160405 (2001).

5. 5.

Zürn, G. et al. Fermionization of Two Distinguishable Fermions. Phys. Rev. Lett. 108, 075303 (2012).

6. 6.

Jacqmin, T., Armijo, J., Berrada, T., Kheruntsyan, K. V. & Bouchoule, I. Sub-Poissonian Fluctuations in a 1D Bose Gas: From the Quantum Quasicondensate to the Strongly Interacting Regime. Phys. Rev. Lett. 106, 230405 (2011).

7. 7.

Girardeau, M. Relationship between Systems of Impenetrable Bosons and Fermions in One Dimension. J. Math Phys 1, 516 (1960).

8. 8.

Zöllner, S., Meyer, H.-D. & Schmelcher, P. Correlations in ultracold trapped few-boson systems: Transition from condensation to fermionization. Phys. Rev. A 74, 063611 (2006).

9. 9.

Alon, O. E. & Cederbaum, L. S. Pathway from Condensation via Fragmentation to Fermionization of Cold Bosonic Systems. Phys. Rev. Lett. 95, 140402 (2005).

10. 10.

Zöllner, S., Meyer, H.-D. & Schmelcher, P. Composite fermionization of one-dimensional Bose-Bose mixtures. Phys. Rev. A 78, 013629 (2008).

11. 11.

Roy, R. et al. Phases, many-body entropy measures, and coherence of interacting bosons in optical lattices. Phys. Rev. A 97, 043625 (2018).

12. 12.

Kościk, P. Fermionized Dipolar Bosons Trapped in a Harmonic Trap. Few-Body Syst 58, 59 (2017).

13. 13.

Arkhipov, A. S., Astrakharchik, G. E., Belikov, A. V. & Lozovik, Y. E. Ground-state properties of a one-dimensional system of dipoles. JETP Lett 82, 39 (2005).

14. 14.

Zöllner, S., Bruun, G. M., Pethick, C. J. & Reimann, S. M. Bosonic and Fermionic Dipoles on a Ring. Phys. Rev. Lett. 107, 035301 (2011).

15. 15.

Zöllner, S. Bosonic and Fermionic Dipoles on a Ring. Ground states of dipolar gases in quasi-one-dimensional ring traps. Phys. Rev. A 84, 063619 (2011).

16. 16.

Astrakharchik, G. E., Morigi, G. E., De Chiara, G. & Boronat, J. Ground state of low-dimensional dipolar gases: Linear and zigzag chains. Phys. Rev. A 78, 063622 (2008).

17. 17.

Astrakharchik, G. E. & Lozovik, Y. E. Super-Tonks-Girardeau regime in trapped one-dimensional dipolar gases. Phys. Rev. A 77, 013404 (2008).

18. 18.

Deuretzbacher, F., Cremon, J. C. & Reimann, S. M. Ground-state properties of few dipolar bosons in a quasi-one-dimensional harmonic trap. Phys. Rev. A 81, 063616 (2010).

19. 19.

Sowiński, T. & García-March, M. Á. One-dimensional mixtures of several ultracold atoms: a review. Rep. Prog. Phys. 82, 104401 (2019).

20. 20.

Paredes, B. et al. Tonks–Girardeau gas of ultracold atoms in an optical lattice. Nature 429, 277 (2004).

21. 21.

Deuretzbacher, F. et al. Evolution from a Bose-Einstein condensate to a Tonks-Girardeau gas: An exact diagonalization study. Phys. Rev. A 75, 013614 (2007).

22. 22.

Kinoshita, T., Wenger, T. & Weiss, D. S. Observation of a One-Dimensional Tonks-Girardeau Gas. Science 305, 1125 (2004).

23. 23.

Santos, L., Shlyapnikov, G. V. & Lewenstein, M. Roton-Maxon Spectrum and Stability of Trapped Dipolar Bose-Einstein Condensates. Phys. Rev. Lett. 90, 250403 (2003).

24. 24.

Andre, A. et al. A coherent all-electrical interface between polar molecules and mesoscopic superconducting resonators. Nature Phys 2, 636 (2006).

25. 25.

Büchler, H. P. et al. Strongly Correlated 2D Quantum Phases with Cold Polar Molecules: Controlling the Shape of the Interaction Potential. Phys. Rev. Lett. 98, 060404 (2007).

26. 26.

Baranov, M. A. Theoretical progress in many-body physics with ultracold dipolar gases. Phys. Rep 464, 71 (2008).

27. 27.

Griesmaier, A., Werner, J., Hensler, S., Stuhler, J. & Pfau, T. Bose-Einstein Condensation of Chromium. Phys. Rev. Lett. 94, 160401 (2005).

28. 28.

Beaufils, Q. et al. All-optical production of chromium Bose-Einstein condensates. Phys. Rev. A 77, 061601 (2008).

29. 29.

Lahaye, T., Menotti, C., Santos, L., Lewenstein, M. & Pfau, T. The physics of dipolar bosonic quantum gases. Rep. Prog. Phys. 72, 126401 (2009).

30. 30.

Romanovsky, I., Yannouleas, C. & Landman, U. Crystalline Boson Phases in Harmonic Traps: Beyond the Gross-Pitaevskii Mean Field. Phys. Rev. Lett. 93, 230405 (2004).

31. 31.

Kościk, P. Quantum correlations in one-dimensional Wigner molecules. Eur. Phys J. D 71, 286 (2017).

32. 32.

Kościk, P. The von Neumann entanglement entropy for Wigner-crystal states in one dimensional N-particle systems. Phys. Lett. A 379, 293 (2015).

33. 33.

Chatterjee, B., Brouzos, I., Cao, L. & Schmelcher, P. Ultracold dipolar few-boson ensembles in a triple-well trap. J. Phys. B: At. Mol. Opt. Phys 46, 085304 (2013).

34. 34.

Chatterjee, B. & Lode, A. U. Order parameter and detection for a finite ensemble of crystallized one-dimensional dipolar bosons in optical lattices. J. Phys. Rev. A 98, 053624 (2018).

35. 35.

Wigner, E. On the Interaction of Electrons in Metals. Phys. Rev 46, 1002 (1934).

36. 36.

Aikawa, K. et al. Bose-Einstein Condensation of Erbium. Phys. Rev. Lett. 108, 210401 (2012).

37. 37.

Frisch, A. et al. Ultracold Dipolar Molecules Composed of Strongly Magnetic Atoms. Phys. Rev. Lett. 115, 203201 (2015).

38. 38.

Park, J. W., Will, S. A. & Zwierlein, M. W. Ultracold Dipolar Gas of Fermionic 23Na40K Molecules in Their Absolute Ground State. Phys. Rev. Lett. 114, 205302 (2015).

39. 39.

Xu, Z., Li, L., Xianlong, G. & Chen, S. Wigner crystal versus fermionization for one-dimensional Hubbard models with and without long-range interactions. J. Phys.: Condens. Matter 25, 055601 (2013).

40. 40.

Pethick, C. J. & Smith, H. Bose-Einstein Condensation in Dilute Gases (Cambridge University Press, Cambridge UK, 2002).

41. 41.

Bogoliubov, N. N., Selected Works II: Quantum and Statistical Mechanics (Gordon and Breach, New York, 1991).

42. 42.

Pitaevskii, L. P. & Stringari, S., Bose-Einstein Condensation (Clarendon Press, Oxford, 2003).

43. 43.

Alon, O. E., Streltsov, A. I. & Cederbaum, L. S. Multiconfigurational time-dependent Hartree method for bosons: Many-body dynamics of bosonic systems. Phys. Rev. A 77, 033613 (2008).

44. 44.

Streltsov, A. I., Alon, O. E. & Cederbaum, L. S. Role of Excited States in the Splitting of a Trapped Interacting Bose-Einstein Condensate by a Time-Dependent Barrier. Phys. Rev. Lett. 99, 030402 (2007).

45. 45.

Cao, L., Krönke, S., Vendrell, O. & Schmelcher, P. The multi-layer multi-configuration time-dependent Hartree method for bosons: Theory, implementation, and applications. J. Chem. Phys. 139, 134103 (2013).

46. 46.

Fasshauer, E. & Lode, A. U. Multiconfigurational time-dependent Hartree method for fermions: Implementation, exactness, and few-fermion tunneling to open space. J. Phys. Rev. A 93, 033635 (2016).

47. 47.

Léveque, C. & Madsen, L. B. Time-dependent restricted-active-space self-consistent-field theory for bosonic many-body systems. New Journal of Physics 19, 043007 (2017).

48. 48.

Léveque, C. & Madsen, L. B. Multispecies time-dependent restricted-active-space self-consistent-field theory for ultracold atomic and molecular gases. Journal of Physics B 51, 155302 (2018).

49. 49.

Lode, A. U. J., Léveque, C., Madsen, L. B., Streltsov, A. I. & Alon, O. E. Multiconfigurational time-dependent Hartree approaches for indistinguishable particles, arXiv:1908.03578.

50. 50.

Mese, A. I., Capuzzi, P., Aktas, S., Akdeniz, Z. & Okan, S. E. Condensation of two-dimensional harmonically confined bosons with Bessel-type interactions. Phys. Rev. A 84, 043604 (2011).

51. 51.

Yannouleas, C. & Landman, U. Symmetry breaking and quantum correlations infinite systems: studies of quantum dots andultracold Bose gases and related nuclear andchemical methods. Rep. Prog. Phys. 70, 2067 (2007).

52. 52.

Lode, A. U. J. et al. MCTDH-X: The time-dependent multiconfigurational Hartree for indistinguishable particles software, http://ultracold.org (2019).

53. 53.

Lode, A. U. J. Tunneling Dynamics in Open Ultracold Bosonic Systems, Springer Theses, (Springer, Heidelberg, 2014).

54. 54.

Lode, A. U. J. Multiconfigurational time-dependent Hartree method for bosons with internal degrees of freedom: Theory and composite fragmentation of multicomponent Bose-Einstein condensates. Phys. Rev. A 93, 063601 (2016).

55. 55.

Heimsoth, M. & Bonitz, M. Interacting bosons beyond the Gross–Pitaevskii mean field. Physica E 42, 420 (2010).

56. 56.

We divide the dimensional Hamiltonian by ħ²/(mL²) where m is the mass of the considered bosons and L a conveniently chosen length scale.

57. 57.

Olshanii, M. Atomic Scattering in the Presence of an External Confinement and a Gas of Impenetrable Bosons. Phys. Rev. Lett. 81, 938 (1998).

58. 58.

Sinha, S. & Santos, L. Cold Dipolar Gases in Quasi-One-Dimensional Geometries. Phys. Rev. Lett. 99, 140406 (2007).

59. 59.

Cai, Y., Rosenkranz, M., Lei, Z. & Bao, W. Mean-field regime of trapped dipolar Bose-Einstein condensates in one and two dimensions. Phys. Rev. A 82, 043623 (2010).

60. 60.

Penrose, O. & Onsager, L. Bose-Einstein Condensation and Liquid Helium. Phys. Rev 104, 576 (1956).

61. 61.

Nozieres, P. & James, D. S. Particle vs. pair condensation in attractive Bose liquids. J. Phys. (France) 43, 1133 (1982).

62. 62.

Spekkens, R. W. & Sipe, J. E. Spatial fragmentation of a Bose-Einstein condensate in a double-well potential. Phys. Rev. A 59, 3868 (1999).

63. 63.

Lode, A. U. J. & Bruder, C. Fragmented Superradiance of a Bose-Einstein Condensate in an Optical Cavity. Phys. Rev. Lett. 118, 013603 (2017).

64. 64.

Kramer, P. & Saracen, M. Geometry of the time-dependent variational principle, (Springer, Berlin, 1981).

65. 65.

Lode, A. U. J., Sakmann, K., Alon, O. E., Cederbaum, L. S. & Streltsov, A. I. Numerically exact quantum dynamics of bosons with time-dependent interactions of harmonic type. Phys. Rev. A 86, 063606 (2012).

66. 66.

Gwak, Y., Marchukov, O. V., Fischer, U. R., Benchmarking multiconfigurational Hartree by the exact wavefunction of two harmonically trapped bosons with contact interaction. arXiv:1811.04705 [cond-mat.quant-gas].

67. 67.

Zinner, N. T., Volosniev, A. G., Fedorov, D. V., Jensen, A. S. & Valiente, M. Fractional energy states of strongly interacting bosons in one dimension. Eur. Phys. Lett. 107, 60003 (2014).

68. 68.

Kościk, P. & Sowiński, T. Exactly solvable model of two trapped quantum particles interacting via finite-range soft-core interactions. Scientific Reports 8, 48 (2018).

69. 69.

Sakmann, K. & Kasevich, M. Single shot simulations of dynamic quantum many-body systems. Nature Phys 12, 451 (2016).

70. 70.

Javanainen, J. & Yoo, S. M. Quantum Phase of a Bose-Einstein Condensate with an Arbitrary Number of Atoms. Phys. Rev. Lett. 76, 161 (1996).

71. 71.

Castin, Y. & Dalibard, J. Relative phase of two Bose-Einstein condensates. Phys. Rev. A 55, 4330 (1997).

72. 72.

Dziarmaga, J., Karkuszewski, Z. P. & Sacha, K. Images of the dark soliton in a depleted condensate. J. Phys. B 36, 1217 (2003).

73. 73.

Dagnino, D., Barberán, N. & Lewenstein, M. Vortex nucleation in a mesoscopic Bose superfluid and breaking of the parity symmetry. Phys. Rev. A 80, 053611 (2009).

Download references

## Acknowledgements

S. Bera acknowledges DST (Govt. of India) for the financial support through INSPIRE fellowship [2015/IF150245] and Rhombik Roy for some useful discussions. B. Chakrabarti acknowledges FAPESP (grant No. 2016/19622-0). AG and MCT acknowledge FAPESP and AG thanks CNPq for financial support. B. Chatterjee acknowledges financial support from the Department of Science and Technology, Government of India under the DST Inspire Faculty fellowship. AUJL and CL acknowledge financial support by the Austrian Science Foundation (FWF) under grant Nos. P 32033 and M 2653, respectively, and the Wiener Wissenschafts- und TechnologieFonds (WWTF) project No. MA16-066.

## Author information

Authors

### Contributions

S.B. conducted the numerical and theoretical investigations, S.B., C.L. and A.U.J.L. wrote the manuscript, A.U.J.L. and C.L. performed complementary calculations, S.B., B. Chak., A.G., M.C.T., M.L.L., B. Chat., C.L., and A.U.J.L. conceived the idea for the project and supported the writeup.

### Corresponding author

Correspondence to A. U. J. Lode.

## Ethics declarations

### Competing interests

The authors declare no competing interests.

## Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

## About this article

### Cite this article

Bera, S., Chakrabarti, B., Gammal, A. et al. Sorting Fermionization from Crystallization in Many-Boson Wavefunctions. Sci Rep 9, 17873 (2019). https://doi.org/10.1038/s41598-019-53179-1

Download citation

• Received:

• Accepted:

• Published:

## Further reading

• ### Impact of the transverse direction on the many-body tunneling dynamics in a two-dimensional bosonic Josephson junction

• Anal Bhowmik
• , Sudip Kumar Haldar
•  & Ofir E. Alon

Scientific Reports (2020)

## Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

## Search

### Quick links

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing