linking efimov physics, few-fermion universality, and the 3n...

44
Linking Efimov physics, few-fermion universality, and the 3n and 4n systems Chris Greene, Purdue University with Michael Higgins, Alejandro Kievsky and Michele Viviani (on the 3n and 4n systems) Thanks to the NSF for support! and on the 3 unitary fermions, with Yu-Hsin Chen

Upload: others

Post on 20-Feb-2021

0 views

Category:

Documents


0 download

TRANSCRIPT

  • Linking Efimov physics, few-fermion

    universality, and the 3n and 4n systems

    Chris Greene, Purdue University

    with Michael Higgins, Alejandro Kievsky and Michele Viviani

    (on the 3n and 4n systems)

    Thanks to the

    NSF for support!

    and on the 3 unitary fermions,

    with Yu-Hsin Chen

  • Today’s Talk Outline

    1.Theoretical search for a low energy bound or

    resonant trineutron or tetraneutron, and

    connections with Efimov physics and

    universality (see arXiv: 2005.04714, now in

    press at PRL)

    2. Investigation of alternative examples of three-

    fermion systems with short range forces at

    unitarity, and manifestations of Efimov physics

    (preliminary, unpublished)

  • The premise which got me interested in studying the 4n and 3n

    systems, after I learned about the fascinating 4n problem from Emiko

    Hiyama at a 2016 KITP Program we ran on “Universality in Few-Body

    Physics”, was the following thought, based on extensive experience

    with atomic & molecular systems:

    WE KNOW RESONANCES

    and

    WE KNOW UNITARITY of 4 FERMIONS

    And

    We have a hyperspherical coordinate toolkit that makes

    existence (or nonexistence) of resonances visually clear

  • Next…

    1. Review some earlier successful predictions of shape

    resonances in few-body systems using the adiabatic

    hyperspherical coordinate framework

    2. Discuss the collision problem with N=4 or N=3 particles in the

    open continuum above the total breakup threshold

    3. Present quantitative calculations of hyperspherical potentials

    for the 4-neutron and 3-neutron systems, based on the several

    state of the art inter-nucleon potentials (plus three-body

    terms), to explore the possible existence of resonances or

    bound states in those systems.

    MAIN CONCLUSIONS

    a. There are no low energy 3n or 4n bound states or

    resonances that are consistent with known n-n force fields

    b. There is a strong connection with Efimov

    physics/universality which guarantees a strong enhancement of

    the 3n and 4n density of states at E→0

  • Fossez, Rotureau, Michel, Ploszajczak Phys. Rev. Lett. 119, 032501 (2017)

    conclusion: while the energy (4n) …may be compatible with expt… its width must

    be larger than the reported upper limit ➔ (probably) …reaction process too short to

    form a nucleus

    Gandolfi, Hammer, Klos, Lynn, Schwenk Phys. Rev. Lett. 118, 232501 2017

    conclusion: a three-neutron resonance exists below a four-neutron

    resonance in nature and is potentially measurable

    First: a brief review of the recent 3n, 4n literature

    Expt: a 4n candidate published in PRL 116, 052501 (2016), Kisamori et al.

    conclusion: energy is

    And an upper limit on its width is quoted to be

    And a Nature News & Views by Bertulani & Zelevinsky, > 2000 page views

    Theory: Hiyama, Lazauskas, Carbonell, Kamimura 2016 Phys. Rev. C.

    conclusion: “…a remarkably attractive 3N force would be required…”

    Shirokov,Papadimitriou,Mazur,Mazur, Roth,Vary Phys. Rev. Lett. 117, 182502

    (2016) conclusion: 4n resonance, E=0.8 MeV,

    NO!

    YES!

    YES!

    YES!

  • Other recent (2018) work on the 3n and 4n systems from

    a collision physics point of view:

    Tetraneutron: Rigorous continuum calculation

    A.Deltuva → PhysicsLetters B 782(2018) 238–241

    “... This indicates the absence of an observable 4n

    resonance...”

    Tri-neutron: Three-neutron resonance study using

    transition operators

    A. Deltuva, PRC 97, 034001 (2018)

    “...There are no physically observable three-neutron

    resonant states consistent with presently accepted

    interaction models”

    NO!

    NO!

  • Our main theoretical tool: formulate

    the problem in hyperspherical

    coordinates, treating the hyperradius

    R adiabatically

    The hyperradius R (squared) is a

    coordinate proportional to total moment

    of inertia of any N-particle system, i.e.:

    Here ri is the distance of the i-th particle

    from the center-of-mass. All other

    coordinates of the system are 3N-4

    hyperangles.

    And then the rest of the

    problem comes down to

    calculating energy levels as a

    function of R, which we call

    “hyperspherical potential

    curves”, and their mutual

    couplings, which can then be

    used to compute bound state

    and resonance properties,

    scattering and

    photoabsorption behavior,

    nonperturbatively

    This follows the formulation of the N-body problem in

    the adiabatic hyperspherical representation, as

    pioneered by Macek, Fano, Lin, Klar, and others

    also r2

  • Strategy of the adiabatic hyperspherical representation: FOR ANY NUMBER OF

    PARTICLES, convert the partial differential Schroedinger equation into an

    infinite set of coupled ordinary differential equations:

    To solve:

    First solve the fixed-R

    Schroedinger equation, for

    eigenvalues Un(R):

    Next expand the desired solution

    into the complete set of

    eigenfunctions with unknowns F(R)

    And the original T.I.S.Eqn. is transformed into the following

    set which can be truncated on physical grounds, with the

    eigenvalues interpretable as adiabatic potential curves, in

    the Born-Oppenheimer sense.

  • • Various methods can be used to compute the needed

    potential curves and couplings: diagonalization in a

    basis set of hyperspherical harmonics, or correlated

    Gaussians, or Monte Carlo techniques, etc.

    • A theorem exists about the truncation to a single

    potential curve, i.e. the adiabatic approximation, namely:

    Notes

  • Next: a few previous results about

    shape resonances and

    hyperspherical potential curves

  • PRL 35, 1150 (1975)

    Hyperspherical potential

    curves for H- 1PoScattering phaseshift versus E

    While we normally think of a strong resonance as having its phaseshift

    increase by p over a narrow energy range, here is an example, where the

    rise versus E is only about 2 radians, which is not unusual.

  • Botero&CHG, 1986 PRL

    C D Lin, 1975 PRL

    2016 Nature

    Commun. by

    Michisio et al.

    An expt on Ps-

    photodetachment

    H-

    H-

    Ps-

    Bryant et al, LAMPF

    experiment, 1977

    PRL H-

    photodetachment

    compared with Broad

    & Reinhardt’s

    calculation

  • Note: d= dimension of

    the relative Jacobi

    coordinates of the

    system, i.e. d=6 for 3

    particles, d=9 for 4

    particles, etc. d=3N-3

    For the 4n problem, d=9, , so for this

    symmetry, the centrifugal barrier at large R in the

    lowest channel is

    An important aside: The d-dimensional Laplacian operator is:

    This Laplacian operator acts on

    the full wavefunction, so like

    one normally does in d=3, we

    can rescale the radial

    wavefunction, i.e. set

    Nonadiabatic

    coupling terms

  • Some results about N-body elastic scattering,

    from Mehta et al., PRL 103, 153201 (2009)

    and this simplifies, if a single scattering

    matrix eigenchannel dominates, to:

    This quantity is in principle an observable, which would

    be relevant for the thermalization of a gas of neutrons that

    are out of thermal equilibrium, even if it is not a typical

    observable that could be readily seen in any standard

    nuclear physics experiment today....

    And an important tool for resonance analysis on the real

    energy axis is the collisional time delay:

  • Next, consider the

    4-fermion problem

  • Some previous work from our group on 4-body hyperspherical

    studies of 4 equal mass fermions or bosons (reasonable agreement

    with 2004 Petrov, Salomon, Shlyapnikov results)

    The system of two spin up, two spin down fermions at large 2-body

    scattering lengths a is important for the theory of the BCS-BEC

    crossover

    Hyperspherical potential curves

    Dimer-dimer scattering length

    (Re and Im parts)

    D’Incao et al, PRA 79, 030501 (2009)

    a>0

  • For the 4-neutron system, since there are no bound

    subsystems, this is simplest to treat in the H-type Jacobi tree:

    2 spin up neutrons, p-wave Y1m(13)

    2 spin down neutrons, p-wave Y1m’ (24)

    s-wave Y00 in the motion

    of the two pairs about each

    other

    So we consider the L=0, S=0, even parity symmetry, which corresponds to

    K=2, 4, 6, … and the lowest channel asymptotically should have a zeroth

    order potential curve

  • Rakshit and Blume, Phys. Rev. A 86, 062513 (2012) found

    that as a-> - infinity, the hyperspherical potentials are

    entirely repulsive, at |a|>>r :

    First Conclusion: The true potential for 4n in this symmetry is

    expected to be less attractive than the lower of these two

    potential curves, making the possibility of a bound state for this

    symmetry unlikely.

    Whereas in the noninteracting limit the asymptotic

    potential for two spin-up and two spin-down identical

    fermions is known to be:

    Considerations about the UNITARY limit

    Aside: This reduction of the coefficient of the asymptotic 1/r2

    potential in the unitary limit is analogous to the Efimov effect...

    a(nn)=-18.98 fm for the Argonne

    AV8’ potential, similar for AV18.

  • Jp =0+ Hyperspherical potentials for 4 fermions:

    both noninteracting and unitary limit a→-infinity

    Non-interacting

    unitary limit

    a(nn)=-18.98 fm for the Argonne

    AV8’ potential, similar for AV18.

  • First attempt to compute the 4n 0+ hyperspherical potential curves

    using a hyperspherical harmonic basis set, for several different

    K(max) up to 140 (from Kievsky and Viviani HH codes in Pisa)

    Potential curves are still not

    converged, but they keep getting

    more attractive at r>20 fm as

    K(max) is increased in the basis

    set

  • After failing to get satisfactory

    convergence at very long range, we

    decided to implement our stochastic

    correlated Gaussian hyperspherical

    basis set method, using the AV8’

    potential, which had been fitted to

    Gaussians by Emiko Hiyama and

    provided to us.

    Side note: in order to be thorough, we also treated other nuclear force models,

    i.e. AV8’, AV18, Minnesota potential, a chiral EFT potential (NV2-Ia) , and with

    or without the 3-body n-n-n term from the Urbana and Illinois (e.g. IL)

    interaction. Minimal differences are observed with these different interactions

  • Improved convergence is apparent in this black curve, obtained for the

    AV8’ interaction using the correlated Gaussian hyperspherical basis

    set method

    Lowest 0+ adiabatic hyperspherical potential for 4 neutrons

  • 0+

    The most attractive hyperspherical potential curves for the 4n and 3n

    systems, obtained using the AV8’ n-n interaction potentials (magenta)

    The converged potentials are clearly totally repulsive, with no sign

    of a local maximum that can trap probability in a resonance.

    HH expansion, unconverged at

    large r

  • Next consider the scattering phaseshift

    in the lowest 4n potential

    Note the very sharp rise of the

    phaseshift close to zero energy, which

    might lead some to interpret this as a

    resonance! However, I now show that

    this is a consequence of the long-

    range potential curve near unitarity.

  • A key point: The lowest energy behavior

    is controlled by the longest range

    portions of the potential curveRecall what we learned from Efimov physics,

    treated in the adiabatic hyperspherical framework

    (e.g. Zhen & Macek, 1988 Phys. Rev. A):

    For three identical bosons, with finite particle-

    particle scattering length a, the asymptotic

    hyperradial potential was shown to behave for

    large |a| as

    Noninteracting term,

    For 3 identical bosons

    But in the UNITARITY LIMIT,

    a→infinity, as Efimov taught

    us, the potential changes to

    the following form:

  • Here are the values of these parameters for our 4n and 3n systems:

    Our

    numerical

    result

    Yin and

    Blume,

    2015 PRA

  • Here is a test of our asymptotic hyperspherical potentials for the 4n system

  • Implications of the attractive a/r3 potential at long range

    One can readily derive (e.g. using the Born approximation)

    that the limiting low energy phaseshift in such a potential of

    the form,

    is:

    4n 0+

    3n 3/2-

  • Next, consider further implications of the low

    energy phaseshift behavior from the perspective

    of a Wigner-Smith time delay analysis.

    →(in the single-

    channel limit)

    Moreover, Q(E) divided by is the density of states

    We can conclude that both the 3n and 4n

    systems have a divergent density of states

    proportional to 1/sqrt(E) at E→0, because

    the phaseshift is proportional to sqrt(E).

    E. P. Wigner, Lower limit for the energy derivative of the scattering phase shift, Phys. Rev. 98, 145 (1955).F. T. Smith, Lifetime matrix in collision theory, Phys.Rev. 118, 349 (1960).

  • 4n 0+

    4n 0+

    3n 3/2-

    3n 3/2-

    Rescaled time delays for the 3n and 4n

    systems, showing that they are finite at

    E→0 when multiplied by sqrt(E)

  • Main conclusions about the tetraneutron and trineutron study:

    (a) There is strong attraction in the system at each hyperradius that lowers the

    potential energy, associated with a(nn) ~ -19 fm (in the singlet channel)

    (b) The attraction diminishes with increasing hyperradius, such that the

    tetraneutron always experiences an outward force, as shown in the cartoon

    (c) The enhanced density of states associated with the long range attractive potential could

    help to explain the enhanced low energy events observed by Kisamori et al.

  • 4n →4n elastic scatteringDerivation: Mehta et al. PRL 103,

    153201 (2009), Eq.6

    Rescaled elastic

    cross section4n density of

    states in the

    lowest channel

    Recall: textbook applications of

    the “Fermi golden rule”, include a

    density of final states factor (from

    integration over the energy-

    conserving Dirac delta function).

    The transition probability per unit

    time is the familiar formula:

    Kisamori et al.

    2016 PRL

  • Evidence that these 3n and 4n systems are in

    the universal regime→ i.e. where the scattering length a is much larger than every

    other length scale in the system

    To address universality in cold atom systems, one typically

    chooses a very short-range two-body potential, like a Gaussian

    or a square well or a delta function, and tunes the potential depth

    to get any desired scattering length.

    e.g. This is the approach that was used in the important study by

    Petrov, Salomon, and Shlyapnikov (2004 PRL) who first showed

    that: Re[ a(dimer-dimer) ] ~ 0.6 a(atom-atom)

    and that Im[ a(dimer-dimer) ] ~ a-2.55 very important for

    studies of the BEC-BCS crossover studied extensively in the

    ultracold atom experiments and theory

  • To see how well the 3n and 4n systems are described at low energy by

    universality, we have adopted a single Gaussian n-n interaction potential

    of short range, which looks VERY DIFFERENT from the true n-n

    interaction potentials, but which gives the same singlet scattering length.

    Next, we recalculated everything, i.e. hyperspherical adiabatic potential

    curves, scattering phaseshifts, time delays, using the SG(single Gaussian)

    potential, and compared to the full AV8’ results

    Comparison between the central AV8’

    potentials and the single Gaussian potentials

  • 0+

    The most attractive hyperspherical potential curves for the 4n and 3n

    systems, obtained using the AV8’ n-n interaction potentials (magenta), and

    the open circles are the potential curves for the SINGLE GAUSSIAN model

    This agreement between calculated results with the AV8’ potential and the SINGLE

    GAUSSIAN potential shows that the 3n and 4n systems are clearly in the universal

    regime, insensitive to details of the short range forces

    HH expansion, unconverged at

    large r

  • 4n 0+

    4n 0+

    3n 3/2-

    3n 3/2-

    This graph (shown earlier) also confirms that phaseshifts and

    time delays in the 3n and 4n systems are controlled almost

    exclusively by the singlet n-n scattering length, and are not

    dependent on the short range details of the n-n forces

    Magenta = AV8’

    Open Circles=SG

  • Another conclusion from this universality study:

    Even if the n-n force is made significantly more

    attractive, such that the singlet scattering length

    diverges to infinity, there would still be no bound or

    resonant trimer, and no bound or resonant tetramer

    Other tests we have made:

    1. Inclusion of 3-body force terms, often highly important in few-nucleon

    systems: for the 3n and 4n systems, they are weak and repulsive,

    and have only a small effect on our results, so most of our

    calculations have not included them.

    2. Inclusion of multichannel coupling of more than one adiabatic

    hyperspherical potential curve: These also make only a small change

    in the computed eigenphaseshifts

  • Lowest 5 adiabatic hyperspherical potential curves for the

    4n 0+ system (excluding parity-unfavored channels)

    Note that all potential curves are fully repulsive

  • Channel coupling test for the trineutron system

  • Channel coupling test for the 4n system

    Comparison of the rescaled 4n 0+

    lowest potential with and without the

    IL7 3-body force term included

  • Gandolfi, et al. PRL 118, 232501 (2017), extrapolated resonance positions

    4n, 2.2 MeV

    3n, 1 MeV

    Our conclusion: Extrapolations of bound state calculations to predict resonance

    existence/position is dangerous if you don’t have a full theory of the analytic

    continuation

  • PRL 2016Our results

    using the AV8’

    potential

    Interestingly, the Shirokov et al. phaseshift resembles our calculated

    4n-4n phaseshift, but they reached a different conclusion. We

    conclude from this that a rapid rise of the phaseshift is not enough to

    establish existence of a resonance, because such a rise can also be

    contributed non-resonantly by a long range potential in the system

  • In summary, we have a new treatment, quite complementary to

    other theoretical approaches, to address the question of whether

    the tetraneutron has a low lying bound state or resonance state.

    Here is the tally so far:

    NO - Hiyama et al. 2016 PRC:

    NO - Deltuva et al. 2018 PRC, PL

    NO - Our work (Higgins et al. 2020 PRL in press)

    YES – Gandolfi et al. 2017 PRL

    YES -- Fossey et al. 2017 PRL

    YES -- Shirokov et al. 2016 PRL

    While the results in this table alone might not seem to strongly

    support our conclusion that there is no tetraneutron low-lying

    resonance or bound state, we believe that our visual evidence of

    total repulsion settles the issue once and for all.

    It remains desirable to understand the very low energy 4n events

    observed by Kisamori et al., and here we believe that the

    divergent zero energy density of states enhancement could

    provide a helpful clue.

  • Other recent work, not discussed today:

    Fractional Quantum Hall Studies with the few-body hyperspherical toolkit:

    K. Daily, R. Wooten, CHG Phys. Rev. B 92, 125427 (2015)

    R. Wooten, Bin Yan, CHG Phys. Rev. B 95, 035150 (2017)

    Bin Yan, R. Biswas, CHG, Phys. Rev. B 99, 035153 (2019)

    Few-body physics with spinor systems

    Three-Body Physics in Strongly Correlated Spinor Condensates

    (with V. Colussi and J. P. D’Incao) PRL 113, 045302 (2014);

    & J. Phys. B 2016

    Spin current generation and relaxation in a

    quenched spin-orbit-coupled Bose-Einstein

    condensate:

    Nature Commun. 10, 375 (2019); Chuan-Hsun Li, Chunlei Qu,

    Robert J. Niffenegger, Su-Ju Wang, Mingyuan He,

    David B. Blasing, Abraham J. Olson, CHG, Yuli Lyanda-Geller,

    Qi Zhou, Chuanwei Zhang & Yong P. Chen