SciPost Submission Page
Quantum BoseFermi droplets
by Debraj Rakshit, Tomasz Karpiuk, Mirosław Brewczyk, Mariusz Gajda
This is not the latest submitted version.
This Submission thread is now published as
Submission summary
Authors (as registered SciPost users):  Miroslaw Brewczyk · Mariusz Gajda · Tomasz Karpiuk · Debraj Rakshit 
Submission information  

Preprint Link:  https://arxiv.org/abs/1801.00346v4 (pdf) 
Date submitted:  20190220 01:00 
Submitted by:  Brewczyk, Miroslaw 
Submitted to:  SciPost Physics 
Ontological classification  

Academic field:  Physics 
Specialties: 

Approach:  Theoretical 
Abstract
We study stability of a zero temperature mixture of attractively interacting degenerate bosons and spinpolarized fermions in the absence of confinement. We demonstrate that higher order corrections to the standard mean field energy of the system can lead to a formation of BoseFermi liquid droplets  selfbound incompressible systems in a threedimensional space. The stability analysis of the homogeneous system is supported by numerical simulations of finite systems by explicit inclusion of surface effects. Our results indicate that BoseFermi droplets can be realized experimentally.
Author comments upon resubmission
We thank both referees for their comments on the manuscript. Both referees agree that the topic we consider is timely and interesting. Unfortunately they express worries that our approach is not valid because of applying a mean fieldapproach and pseudowavefunction formalism which, as they suspect, cannot account for a correct treatment of fermionic systems in a droplet.
The Referee 1 points two major defects of the manuscript:
1The methods used are not valid to describe fermion dynamics. 2The methods used are not valid to describe fermions in selfbound droplets.
Referee’s 2 criticism is based upon the same arguments. The referee says:
 The use of a single complex function to represent all of the fermions in this system is not physically reasonable. There are significant issues both for the ground state and for dynamics.
The arguments on the necessity of accounting for a discrete spectrum of bound fermions are given in the two reports. In addition the referees point out to problems of the unphysical phase of the pseudowavefunction for fermions.
The Referee 2 requests the following changes:
 Use a methodology that allows for the small number of discrete modes available to the fermions.
 With that methodology show the bounds on the number of bosons and fermions for a droplet to exist and how that varies with interaction strength.
 For reasonable interaction strengths, then give corresponding peak densities in SI units to allow comment on threebody loss.
In the response below we show that the above mentioned worries are not justified and we substantiate our results by additional arguments and results based on numerical calculations. Because both reports address similar issues we will write a single answer where we carefully address all points raised by the two referees.
Resubmitted version of our manuscript is substantially modified. In particular:

We added a new section no. 4, entitled: Finite system analysis – atomicorbital approach, where we show results based on the HartreeFock formalism for fermions which, contrary to the mean field description, accounts for a discreet character of the fermionic spectrum a. In the first part we present dynamical calculations for the ground state with a discreet singleparticle basis for fermionic atoms (the size of the basis equals to the number of fermions). The base functions are modified adiabatically, starting from noninteracting trap system evolving towards interacting one without any trapping potential eventually. We consider both bound and unbound states. New figure 5 is added. b. In the second part we present results of extensive numerical diagonalization of the system Hamiltonian in basis of 2600 oscillatory functions. We investigate two cases: unbound and selfbound systems. In both cases we show singleparticle spectrum of fermionic subsystem. Figures 6 and 7 are added.

We added discussion of losses and new Fig.4 illustrating how number of bosons in droplets decreases due to the threebody collisions. We give physical values of parameters for the CesiumLithium mixture as used in the Chin Cheng’s group experiment [1].
We did not included in the resubmitted manuscript the request #2 of the second Referee: `With that methodology show the bounds on the number of bosons and fermions for a droplet to exist and how that varies with interaction strength.’ We have to admit that this is indeed a very interesting problem. We have shown that droplets exist even for as little as 35 fermions and 350 bosons. However systematic studies using the HartreeFock method require extensive, time consuming numerical work. In our opinion comprehensive answer to the referee’s question deserves a separate publication.
Below we present more elaborate discussion in reply to the referees’ concerns.
 Comment on the referees’ criticism of using the pseudowavefunction formalism and on the meanfield description of fermionic systems.
We want to stress that the paper can be divided into twoparts: a semianalytic (to some extend) approach where we specify conditions for equilibrium of a selfbound system with a free surface. This method, valid in the limit of infinite system, is general, and contrary to the approximate approach of D. Petrov does not base upon diagonalization of a quadratic form. It looks as if the referees are ignoring this part. In our opinion this part is very important. Results of this part were supported by numerical calculations based on the hydrodynamic approach.
In a stationary situation, both infinite system and finite system case, our approach is simply the standard ThomasFermi method with the Weizsacker correction included.
Evidently, the ThomasFermi model, introduced to describe electronic cloud and binding energies of multielectron atoms is far from the accuracy of sophisticated quantumchemist approaches to multielectron systems. We do not claim that we are such accurate. But the TF approach in not totally wrong. It gives quite reasonable estimation (with 10% accuracy) of the binding energies of atoms, quite small systems though. Our predictions prove that the quantum fluctuations are able to stabilize the BoseFermi mixtures and can lead to formation of liquid droplets. We are convinced that in a case of large systems, the approach used by us is quite accurate, and we are not aware of any better than meanfield approach for systems having about 1000, 10000, or 100000 fermions or more, as we show in Fig. 2. No discreet treatment is possible for the system this big!
The static approach is generalized then to a dynamical situation by introducing a velocity field. The corresponding hydrodynamic equations are brought to the form of the Schoedingerlike equation for the pseudowavefunction which results from a kind of “complexification” of the density and the velocity “potential”. The transformation assumes that the velocity field is irrotational, or more precisely that it is defined on a simply connected support, and a velocity potential – a phase, can be used instead. We explain the issue of complex pseudowavefunction for fermions in details in the comments to the reports and we cannot add anything substantial to this discussion without repeating the same arguments. So we believe that our approach is correct as long as ThomasFermi model is justified.
 Elaborated discussion of the arguments against our approach bringing the controversy on continuous versus discreet approach for many fermion system.
The referees say, that selfbound BoseBose droplets have only few, or none, bound excited states. It’s true, all bosons can occupy the same state so one bound singeparticle state can support BoseBosedroplets, this is not enough in the case of fermionic component because of the Pauli principle. For fermions number of bound singe particle states gives an estimation of number of fermions which can be “trapped” in the droplet.
The referees, based on observation of [2] claim that no more than 10 oneparticle states can be trapped in bosonic cloud, thus discreetness of the fermionic spectrum is crucial for bound systems. The referee 2 gives also example of BoseBose droplets, which have only few excited states.
The referee’s observation is correct. Number of bound states in effective potential formed by bosonic atoms must not be smaller than the number of fermions. We want to stress that we deal here with 3D situation, contrary to the effective 1D case of [2]. Note that in 3D the energy states are highly degenerate, every angular momentum state L, is (2L+1)fold degenerate. Our approach, leading to the meanfield energy of fermionic component is based on estimation of the number of bound states in a uniform potential. And this estimation is correct up to the leading order in the Fermi energy. For a spherically symmetric harmonic oscillator, the number of bound states of energies not larger them $m \hbar \omega$ grows as $m^3$. The TF model recovers the same scaling.
Because this issue seems to be controversial, what is expressed also in the comment of Mathew Davis, in the corrected version of the manuscript we included the subsection entitled “Finite system analysis – atomic orbital approach” devoted exclusively to justification of the mean field method. We support the hydrodynamic results by results obtained in the HartreeFock method accounting for discreetness of fermionic orbitals.
In the first part of this section we defined the HartreeFock formalism equivalent to the energy functional used by us. Then we find densities of droplet by adiabatic following of the ground state of 35 fermions and 350 bosons starting from decoupled system and gradually increasing the mutual coupling. We used the basis of 35 fermionic states, which were dynamically modified according to the HartreeFock equations coupled to the extended GrossPitaevskii equation. Resulting densities of droplets agree very well with those obtained by the pseudowavefunction formalism. In addition we showed that for too weak BoseFermi attraction the system is not bound and its radius spreads in time after releasing from the trap.
In the second part we used a huge basis of oscillatory wavefunctions to find fermionic singleparticle states in the effective bosonic potential for a small system of 35 fermions and 350 bosons. We want to stress that these are severalmonthlasting calculations. Within this approach we show that number of states bound by bosonic cloud is exactly equal to the number of fermions.
We believe that this extensive and timeconsuming calculations are convincing for the two referees, moreover they justify usage of the mean field approach as well as the pseudowavefunction formalism.
Finally, as the referee 2 requested, we assumed realistic values of densities of CesiumLithium systems [1], for which we calculated the peak atomic densities, estimated the loss rate and showed the results of numerical simulations of droplet dynamics with losses included. We showed that the lifetime of droplet is sufficiently long there. To illustrate this analysis we included new figure.
We think that we gave answered to all the concerns of the both referees. We modified the manuscript to account for all their criticisms and we hope that our manuscript, in the present form, will be accepted for publication.
References [1] B.J. DeSalvo et al. Phys. Rev. Lett. 119, 233401 (2017) [2] M. Wentzel et al. Physica Scripta 93, 104004 (2018)
List of changes
1. We added a new section no. 4, entitled: Finite system analysis – atomicorbital approach, where we show results based on the HartreeFock formalism for fermions which, contrary to the mean field description, accounts for a discreet character of the fermionic spectrum
a. In the first part we present dynamical calculations for the ground state with a discreet singleparticle basis for fermionic atoms (the size of the basis equals to the number of fermions). The base functions are modified adiabatically, starting from noninteracting trap system evolving towards interacting one without any trapping potential eventually. We consider both bound and unbound states. New figure 5 is added.
b. In the second part we present results of extensive numerical diagonalization of the system Hamiltonian in basis of 2600 oscillatory functions. We investigate two cases: unbound and selfbound systems. In both cases we show singleparticle spectrum of fermionic subsystem. Figures 6 and 7 are added.
2. We added discussion of losses and new Fig.4 illustrating how number of bosons in droplets decreases due to the threebody collisions. We give physical values of parameters for the CesiumLithium mixture as used in the Chin Cheng’s group experiment.
Current status:
Reports on this Submission
Anonymous Report 3 on 2019329 (Invited Report)
 Cite as: Anonymous, Report on arXiv:1801.00346v4, delivered 20190329, doi: 10.21468/SciPost.Report.894
Strengths
1. The work is very timely. Currently, selfbound quantum droplets represent
one the most exciting topics in the field of ultracold atoms.
2. The results are interesting and they could inspire new experiments, and also further theoretical investigation.
Weaknesses
Some points need to be clarified (see the requested changes).
Report
I would be pleased to recommend the publication of this manuscript after the authors have addressed the points mentioned below.
Requested changes
1) The author state in the abstract that "BoseFermi liquid droplets – selfbound incompressible systems" are formed,
but then the (in)compressibility of the system is never discussed. The authors should clarify this point, are these systems
really incompressible?
1) The choice of the time unit  ħ/(mBaB2)  is very uncommon. Why not using milliseconds or something more
immediate to read? I would recommend to change it or explain clearly in the text the reason of this choice.
2) The content of Fig. 4 does not justify the need of a figure. Its meaning could be easily explained by adding a text line.
3) Which is the physical origin of the oscillations in Fig. 4right?
Have the authors checked that this is not just a numerical effect?
Please explain in the text.
4) The English may need some revision. Please check carefully the (missing) articles.
Anonymous Report 2 on 2019325 (Invited Report)
 Cite as: Anonymous, Report on arXiv:1801.00346v4, delivered 20190324, doi: 10.21468/SciPost.Report.891
Strengths
1. The work is still very topical.
2. The calculations are generally well justified, with two methods being compared.
Weaknesses
1. Comparison of expansion dynamics using the two computational methods is not possible due to different assumptions.
2. Information of three body loss is overly precise given uncertainties and not clearly laid out.
Report
I thank the authors for their resubmission. The work in the new Section 4 using fermionic orbitals provides very helpful justification of the earlier hydrodynamic calculations.
I recommend the manuscript is published, after quickly addressing two issues.
Requested changes
1. Dynamics with one Fermi wavefunction
I am still concerned that the solid curve in Fig. 3(left) may not be approximately correct, as it uses a single wavefunction with a single phase for an expanding gas. The behaviour of the solid line Fig. 3(left) cannot be easily compared to the black curve in Fig. 5(right) as the quench, timescales and number of atoms are quite different. In Fig. 5(right), please include dynamical results from the hydrodynamical approach with the same parameters as for the atomicorbital approach. This will allow easy comparison, as we have for equilibrium results in Fig. 5(left).
2. Three body loss
The calculation for Fig. 4 is too approximate to be useful. The quench of scattering lengths and trap in an experiment will lead to oscillations which will affect loss, there being a tradeoff between a fast quench (desirable due to short lifetimes) but stronger oscillations. The experiment will also have noise which has not been added in the calculations. I suggest removing Fig. 4 and coming up with a broad estimate based on the rate coefficient.
Also the paragraph on three body loss is confusing. Please state clearly what you get from where. You are considering a case of $a_B=250a_0$ and $a_{BF}=3.6a_B=900a_0$? The bosonic density of $n_B = 0.0009/a_B^3 = 3 \times 10^{14}\mathrm{cm}^{3}$ is taken from your calculations (the manuscript says 'From the rate equation')? I see you have used $K_3=\Gamma/n_B^2$, but where did your $\Gamma=10/\mathrm{s}$ come from? Please do not state $K_3$ to three significant figures. Please state clearly how you get $\Gamma=50/\mathrm{s}$ from Fig. 4(b) of [29] including how you allow for your values of $a_B$ and $a_{BF}$ and for your increased density.
Please also discuss how 'The loss rate exceeds the thermalization rate at $a_{BF}=520a_0$, above which the system no longer reaches thermal equilibrium' [29] relates to your system.
Anonymous Report 1 on 2019321 (Invited Report)
 Cite as: Anonymous, Report on arXiv:1801.00346v4, delivered 20190321, doi: 10.21468/SciPost.Report.883
Strengths
This a timely and relevant study on BoseFermi selfbound droplets.
This manuscript now considers the discrete and finite number of bound fermion excited states.
Weaknesses
For part of their work, the authors employ a single pseudowavefunction to describe all fermions of the system. The ‘phase’ of this fermion pseudowavefunction is then used to describe [via Eqs. (10)] the nonadiabatic dynamics of an expanding droplet in Fig. 3. The authors do not adequately justify the use of this apparently unphysical term in (10). However, I do recognize that in this figure the most important result is probably the question of whether the droplet is stable or unstable [which might be better described by Eqs. (10)], rather than unstable dynamics itself.
Report
There is still one issue that remains. In Eqs. (10) the authors use a single pseudowavefunction to describe all fermions of a system, which they use for nonadiabatic dynamic simulations in Fig. 3. In this equation there is a term which involves the ‘phase’ of this fermion pseudowavefunction. The authors later claim
“… the hydrodynamic equations in Sec. 3 can be safely used as long as the hydrodynamic velocity field is irrotational.”. However, this claim seems overly strong, is unsubstantiated, and I can see no situation where such a pseudowavefunction ‘phase’ should describe many fermions in an actual physical dynamic process. Even after I mentioned this last time, the authors have not been able to explain this. However, I do recognize that the dynamics of the fermion pseudowavefunction is not a central part of this manuscript. Even for the problematic figure 3, which presents results for nonadiabatically expanding droplet, the most interesting part (in my opinion) is the instability itself, not the postinstability dynamics. For these reasons I do not request any further changes to the calculations or simulations themselves, as I do not wish to delay publication, but the authors should better address these concerns in their text. The authors should either admit that the term in Eqs. (10) which involves the ‘phase’ of the fermion pseudowavefunction is unphysical when used to describe many fermions, or they should explain what physical process it actually represents, and give evidence for this.
Requested changes
Text change only: The authors should either admit that the term in Eqs. (10) which involves the ‘phase’ of the single fermion pseudowavefunction (which they use to describe the nonadiabatic dynamics of many fermions) is unphysical, or they should explain what physical process it actually represents, and give evidence for this.
Matthew Davis on 20190401 [id 478]
This opening paragraph was left out of report 1 in error by the author, and communicated to the editor separately:
"The manuscript now addresses the crucial fact that there are only a finite number of bound excitations within selfbound droplets by employing a kind of HartreeFock theory. Clearly this is an important effect since only one fermion can occupy each bound state. While many unanswered questions remain, I believe that these can be left for future work and I am now willing to recommend this manuscript for publication as long as the authors address my remaining concern below, which only regards a text change."