How Three Dimensions Learn to Behave Like One

There is a familiar story about dimensional reduction. Make a trap very tight in two directions, leave one direction weakly confined, and the particles become one-dimensional. The statement sounds geometric: squeeze the cloud until it resembles a line.

That picture is useful, but it misses the interesting physics. A particle may have no real transverse excitations and still visit them virtually during a collision. Those virtual excursions change the scattering amplitude, generate confinement-induced resonances, and sometimes leave behind effective-range terms that matter to few-body binding. The frozen directions are absent from the final wavefunction, but not from the low-energy theory.

There is a second route that is stranger. Do not build a narrow external waveguide at all. Instead, engineer a strongly anisotropic interaction. If the interaction confines the relative orientation of two particles around an attractive axis, angular zero-point motion can create an effective repulsive core. The pair then moves as if it were in a one-dimensional channel written by the interaction itself. In a sufficiently impenetrable channel, bosons and fermions can even share nearly the same bound-state spectrum.

This article develops these two mechanisms side by side. The first follows the analysis of heteronuclear atoms in species-dependent quasi-one- and quasi-two-dimensional traps by Shi and Cui. The second follows the microwave-shielded polar-molecule clusters of Shi, Wang, and Cui. Both papers ask what remains after high-energy motion is removed. Their answers look different, but the method is the same:

1. identify the modes that appear frozen;
2. integrate out their virtual response rather than deleting them;
3. express the result through a small set of low-energy scattering data;
4. ask which few-body states are fixed by those data.

The reward is not merely a simpler Hamiltonian. It is a sharper meaning of universality.

1. A waveguide changes the collision

Begin with two particles interacting through a short-range potential in three dimensions. Far below the inverse range of the potential, an $s$-wave collision is summarized by the effective-range expansion

Here $a_s$ is the three-dimensional scattering length and $r_e$ is the effective range. A narrow Feshbach resonance is often written in the equivalent energy-dependent form

where $\mu$ is the reduced mass and $R^{\ast }$ measures the resonance width. We use $\hslash =1$ temporarily in this section. Positive and large $R^{\ast }$ means that energy dependence cannot be ignored over the experimentally relevant window.

Now place the particles in a harmonic waveguide. The collision no longer has one channel. It has an open channel, in which both particles occupy the transverse ground state, and infinitely many closed channels containing excited transverse oscillators. At energies below the first transverse threshold, the closed channels cannot appear as outgoing states. They can nevertheless appear between two interaction events:

This is why the naive instruction “project onto the transverse ground state” fails. Projection before solving the collision removes precisely the virtual processes that renormalize the interaction.

The formal statement is a Feshbach projection. Let $P$ project onto the open transverse channel and $Q=1-P$. For the full Hamiltonian $H$, the energy-dependent operator acting in the open sector is

The second term carries the entire memory of closed transverse motion. Its poles and rapid energy dependence produce a confinement-induced resonance. In the simplest equal-frequency waveguide this is the mechanism found by Olshanii; the interpretation of transverse modes as closed Feshbach channels was made especially explicit by Bergeman, Moore, and Olshanii.

This already gives the first lesson:

Low dimension is an infrared description, not permission to erase the ultraviolet route by which the particles collide.

2. Two species refuse to share a centre-of-mass coordinate

The standard separation into centre-of-mass and relative motion hides an assumption. It works cleanly when both species experience the same trap frequency. The heteronuclear problem of Shi and Cui is more general.

Let a heavy atom $h$ and a light atom $l$ have masses $m_h$ and $m_l$. In one of the tightly confined directions, take

Define

and the usual coordinates

Their inverse is

Substitution gives

with

The mixed term is the important one. Centre-of-mass and relative motion separate only when

Equal oscillator lengths are not enough. If

then $m_h{\omega }_h=m_l{\omega }_l$, so

For unequal masses the frequencies are unequal and the $Zz$ coupling survives. This point matters experimentally because species-selective optical potentials naturally produce unequal frequencies.

The consequence is more than an inconvenient basis. A contact collision changes the relative coordinate at $z=0$, while the confinement mixes that collision with molecular centre-of-mass excitations. Several closed channels can approach threshold and generate several resonant features. A single Olshanii-style number cannot describe the entire problem.

3. The correct low-energy objects

The full confined scattering calculation is complicated, but its low-energy output is compact. In quasi-two dimensions, with in-plane relative momentum $q$, Shi and Cui write the on-shell amplitude as

The logarithm is characteristic of two-dimensional scattering. The parameter $a_{2D}$ is a length, while $R_{2D}$ has dimensions of length squared.

In quasi-one dimension the corresponding expansion is

Now $a_{1D}$ is again a length, but $R_{1D}$ has dimensions of length cubed. Calling both $R_{2D}$ and $R_{1D}$ simply “the effective range” can obscure this dimensional difference. They are better regarded as effective-range coefficients in the appropriate low-dimensional expansion.

These formulae do several jobs at once. They specify the scattering phase, locate shallow dimers, and provide the two-body input for the few-body integral equations. They also state the domain of the low-energy theory. Terms omitted from the denominators must be small at the momenta carried by the bound state.

That last condition is easily forgotten. Matching $a_d$ and $R_d$ does not make the effective theory exact at arbitrary binding energy. If a cluster probes momenta comparable to the transverse scale or the microscopic interaction range, higher shape parameters and explicit closed channels re-enter.

3.1 Poles become dimers

A bound state has energy

and is found by continuing $q$ to $i\kappa$. In quasi-two dimensions the pole condition becomes

In quasi-one dimension it is

The dimensions provide a quick check. Each term in the first equation is dimensionless. Each term in the second has dimensions of length.

If the effective-range term is negligible, the quasi-one-dimensional equation gives $\kappa =1/a_{1D}$ for $a_{1D}>0$. Once $R_{1D}{\kappa }^2$ becomes comparable to $a_{1D}$, the dimer is no longer controlled by the scattering length alone. This is exactly where a narrow Feshbach resonance or strong centre-of-mass coupling can matter.

3.2 Why a molecular centre-of-mass basis helps

At contact, ${\mathbf{r}}_h={\mathbf{r}}_l$. It is therefore natural to expand the short-range pair amplitude in oscillator states of an artificial molecule with mass $M$. In the confined direction one introduces

where

This does not restore an exact separation of centre-of-mass and relative motion. It provides an efficient basis for the pair amplitude at coincidence. The matrix-valued closed-channel propagator can then be evaluated and expanded near the open-channel threshold. Diagonal structures that would be spread over many product oscillator states become much easier to resolve.

The calculation returns $a_{1D}$ or $a_{2D}$ together with the appropriate $R_d$. It also reveals multiple confinement-induced resonances. Those resonances are not mysterious extra interactions. They are different centre-of-mass-dressed molecular channels crossing the scattering threshold.

4. One light atom can bind several heavy fermions

We now use the low-dimensional amplitude rather than the full transverse Hamiltonian. Consider $N$ identical heavy fermions and one distinguishable light atom. Only heavy-light pairs interact in the $s$-wave channel. Heavy-heavy $s$-wave scattering is forbidden by antisymmetry.

The simplest cluster beyond a dimer is a trimer containing two heavy atoms and one light atom. A useful momentum-space representation introduces a spectator function $F(\mathbf{p})$. Up to normalization conventions, its zero-range equation has the structure

Here $d=1$ or $2$, ${\mathcal{T}}_d$ is the effective two-body amplitude, and

is the reduced mass of the spectator heavy atom relative to the heavy-light pair. The minus sign comes from exchanging the identical heavy fermions. Detailed conventions move constants between $F$, ${\mathcal{T}}_d$, and the kernel, but the physical content is stable: two-body propagation dresses the left-hand side, while exchange of the light atom binds the heavy particles on the right.

For a tetramer, the spectator amplitude depends on two heavy momenta and must be antisymmetric under their exchange. The equation is larger, not conceptually different. In both cases the mass ratio matters. A light atom can rearrange rapidly between slowly moving heavy atoms, producing an induced attraction. Fermionic antisymmetry opposes binding by requiring nodes or nontrivial exchange parity. Increasing $m_h/m_l$ strengthens the first effect relative to the second.

The stability measure used by Shi and Cui is

For the trimer, $E_N$ is the dimer energy; for the tetramer, it is the trimer energy. A negative ${\mathrm{\Delta }}_{1+N,N}$ means that the larger cluster lies below the threshold for removing one heavy fermion.

4.1 What “universal” means here

The word universal needs a complete sentence after it. In this problem it means:

Within the low-energy zero-range description and its regime of validity, the cluster energy is fixed by the mass ratio and the effective low-dimensional two-body parameters, without an additional short-range three- or four-body input.

This is not the statement that every microscopic confinement gives the same spectrum. Different traps give different $a_d$ and $R_d$. It is not the statement that arbitrarily deep clusters are universal. Deep states probe momenta at which the two-term expansion of $t_q$ fails. Nor is it automatically an Efimov statement. The low-dimensional clusters considered here do not require the discrete scale invariance of a three-dimensional resonant $1/R^2$ channel.

Universality is a controlled loss of microscopic memory. One must say which memory has been lost and which low-energy parameters still carry it.

5. What the Li-Cr and Li-K calculations predict

Shi and Cui apply the construction to two experimentally relevant mixtures. For ${}^6\mathrm{Li}$-${}^{53}\mathrm{Cr}$ the mass ratio is about $8.8$, and the resonance-width scale used in the calculation is $R^{\ast }\simeq 6000a_0$. For ${}^6\mathrm{Li}$-${}^{40}\mathrm{K}$ the corresponding values are about $6.7$ and $R^{\ast }\simeq 2400a_0$. The calculations use an equal confinement length

which, as we saw, still leaves centre-of-mass and relative motion coupled.

For ${\ell }_{ho}=700a_0$, the paper identifies magnetic-field intervals in which the trimer and tetramer are both bound by experimentally relevant energies. Representative ranges are:

geometry and mixture	interaction window $R^{\ast }/a_s$	trimer separation ${\mathrm{\Delta }}_{3,2}/h$	tetramer separation ${\mathrm{\Delta }}_{4,3}/h$
quasi-2D Li-Cr	$-69.9$ to $-61.8$	$-0.28$ to $-0.92$ kHz	$-0.18$ to $-0.50$ kHz
quasi-2D Li-K	$-15.3$ to $-12.4$	$-0.17$ to $-0.80$ kHz	$-0.10$ to $-0.42$ kHz
quasi-1D Li-Cr	$-194.0$ to $-127.1$	$-1.79$ to $-7.85$ kHz	$-0.35$ to $-1.13$ kHz
quasi-1D Li-K	$-53.3$ to $-25.7$	$-1.41$ to $-8.81$ kHz	$-0.21$ to $-1.23$ kHz

These numbers should not be read as four immutable predictions. Each row refers to the paper’s specified confinement and chosen useful interval. The robust comparison is that quasi-one-dimensional confinement gives substantially larger absolute trimer binding in these examples. The deepest trimer separation reaches several kilohertz rather than remaining below one kilohertz.

There is a subtlety in the tetramer comparison. Its absolute separation also grows in quasi-one dimension, but its size relative to the dimer or trimer binding need not grow. A spectrum can become deeper while two neighboring thresholds become less widely separated in fractional terms. Experimental resolvability depends on absolute frequency, lifetime, temperature, and threshold crowding, not on a single dimensionless ratio.

The exact quasi-two-dimensional dimer calculation agrees with the pole obtained from

over the broad negative-$R^{\ast }/a_s$ region used for the cluster study. That comparison is an important internal check: the few-body calculation is not being fed an effective two-body model outside the range in which it reproduces the confined dimer.

5.1 What would falsify the effective description?

Suppose an observed trimer energy disagrees with the low-dimensional integral equation. Several explanations are logically distinct:

• the extracted $a_d$ or $R_d$ is inaccurate because too few confinement channels were retained;
• the cluster momentum is high enough that the next shape parameter matters;
• finite-range or inelastic physics introduces a new few-body scale;
• the experimental trap is not harmonic or not separable in the unconfined directions;
• the observed feature is a different centre-of-mass-excited cluster.

A discrepancy would not by itself prove “nonuniversality.” It would identify missing information in the proposed low-energy data set.

6. A second route: let the interaction build the waveguide

The first mechanism uses an external trap to freeze coordinates. Microwave-shielded polar molecules offer a more unusual possibility. Their relative interaction can itself confine the angular motion.

Ultracold polar molecules have a practical problem. The dipole-dipole interaction is long-ranged and useful, but attractive collision paths can carry the molecules to short range, where chemical reaction or complex formation causes loss. Microwave shielding dresses rotational states so that approaching molecules encounter a repulsive adiabatic barrier. The shielding idea and its coupled-channel basis were developed, among other works, by Karman and Hutson. Experiments have since observed long-lived field-linked molecular pairs, including the tetratomic complexes reported in Nature 626, 283 (2024).

The word tetratomic is literal here. Each polar molecule is diatomic, so a bound state of two molecules contains four atoms. A three-molecule cluster is hexatomic.

Shi, Wang, and Cui study a microwave-dressed interaction at ellipticity $\xi =\pi /4$. With the microwave propagating along $z$ and its linear axis along $y$, their effective potential is

The coefficients are

and

where $d$ is the molecular dipole moment, $\mathrm{\Omega }$ is the microwave Rabi frequency, and ${\delta }_r=|\delta |/\mathrm{\Omega }$ is the detuning-to-Rabi ratio. The calculations in the paper take ${\delta }_r=0.2$.

Along the $y$ axis, $\theta =\pi /2$ and $\varphi =\pi /2$. The potential is then

The explicit $C_6/r^6$ term vanishes on the axis. If we simply set the angular coordinates to their minimum, the potential falls without a repulsive core. This seems to predict collapse, not shielding.

That conclusion repeats the earlier mistake. It freezes a coordinate before accounting for its zero-point motion.

7. Zero-point motion creates the wall

Write small angular deviations from the attractive $y$ direction as

To quadratic order,

At fixed $r$, the two angular deviations are harmonic modes. The angular part of the relative kinetic energy near the axis is

where $m$ is the mass of one molecule and the pair reduced mass is $m/2$. Comparing each angular Hamiltonian with a harmonic oscillator, the sum of the two zero-point energies is

The effective radial potential along the attractive axis is therefore

This formula contains the central mechanism of the molecular paper. At short distance the $C_6$ term dominates inside the square root, giving

The repulsive $r^{-4}$ wall is not present in the bare potential on the $y$ axis. It is generated by the increasing angular zero-point cost of keeping the molecules inside an ever narrower attractive cone. The interaction first attracts the particles toward an axis and then charges them quantum kinetic energy for remaining there.

This is interaction-induced dimensional reduction. The radial coordinate is slow; the two angular coordinates are fast and localized. Integrating out the fast modes produces a one-dimensional potential with a fluctuation-induced core.

7.1 The size of the bound pair

Define the microwave length

In the regime where the short-distance $r^{-4}$ wall balances the $r^{-3}$ attraction, minimizing the asymptotic form gives

up to the weak factor $(1+{\delta }_r^2{)}^{1/4}$ that is close to one at ${\delta }_r=0.2$. Thus the Rabi frequency controls the molecular separation:

This scale is not a van der Waals core inserted by hand. It follows from the competition between long-range microwave-dressed attraction and angular zero-point motion.

The other useful scale is the dipole length

The paper chooses the unit ${\ell }_u={\ell }_d/20$ and $E_u={\hslash }^2/(m{\ell }_u^2)$. For ${}^{23}{\mathrm{Na}}^{40}\mathrm{K}$, ${\ell }_u$ is about $550a_0$. A Rabi frequency $\mathrm{\Omega }=2\pi \times 10\mathrm{MHz}$ corresponds to $\hslash \mathrm{\Omega }/E_u\simeq 52$ in the parameters quoted by the authors.

8. Why bosons and fermions can share a spectrum

The short-distance $r^{-4}$ core suppresses particle crossing. This divides one-dimensional configuration space into disconnected ordering sectors. For $N$ particles, one such sector is

If the wavefunction vanishes whenever $y_i=y_j$, a bosonic eigenfunction can be converted into a fermionic one by multiplying by

Thus

Away from coincidence, $\mathcal{A}$ is locally constant, so the kinetic and potential energies act on ${\mathrm{\Psi }}_F$ exactly as they act on ${\mathrm{\Psi }}_B$. At coincidence both states obey the same hard-core boundary condition. The energies and position-space probabilities are therefore identical:

This is the logic behind the one-dimensional Bose-Fermi mapping introduced by Girardeau. It does not mean that bosons become fermions. The sign structure changes. Momentum distributions, off-diagonal coherence, and operators sensitive to exchange phases need not agree.

In the microwave-shielded molecule problem the mapping is approximate for two reasons. First, the repulsive core is large but finite, so the wavefunction has a small amplitude in crossing regions. Second, the physical motion remains three-dimensional, with a narrow but nonzero angular width. Exact three-dimensional calculations show that the reduced one-dimensional wavefunction has an overlap above $97\mathrm{\%}$ in the strong-shielding regime $\hslash \mathrm{\Omega }/E_u>10$. The bosonic and fermionic bound-state energies differ by less than about $5\mathrm{\%}$ already for $\hslash \mathrm{\Omega }/E_u>2$ in the reported calculations.

The approximation fails first for shallow states near threshold. Bosons approach threshold through an $s$-wave channel, while identical fermions require odd exchange symmetry and are naturally associated with a $p$-wave channel in three dimensions. A large, deeply localized pair can forget much of this distinction after angular freezing. A diffuse threshold state cannot.

The claim should therefore be precise:

Strong microwave shielding produces an approximate Bose-Fermi spectral duality for the localized few-molecule bound states, not an exact equality of the complete three-dimensional scattering theories.

9. From a tetratomic pair to a hexatomic cluster

For three identical molecules it is convenient to use Jacobi coordinates

and

The remaining pair separations can be written as

The attractive part of the effective one-dimensional three-body potential is the sum over pairs,

The angular deviations of the three particles are coupled. Diagonalizing their quadratic fluctuation Hamiltonian produces a collective zero-point term $u^{(3)}(y_r,y_{\rho })$. The resulting surface is

This is not merely the sum of three isolated two-body effective potentials. The fast angular modes are shared by the whole configuration, so integrating them out can generate nonadditive terms.

The three-molecule ground state is nevertheless well approximated by a product of neighboring pair states in an ordering sector:

This has a transparent geometric meaning. The hexatomic cluster resembles a short chain with adjacent molecules separated by $r_m$. Its pair-correlation function has a nearest-neighbor feature near $r_m$ and a next-nearest-neighbor feature near $2r_m$.

The hexatomic binding is more negative than twice the tetratomic binding. The excess does not require a mysterious irreducible three-body force. The two end molecules attract through the long-range $-1/r^3$ tail, and the collective angular zero-point energy is not exactly pairwise additive. The paper identifies a negative correlation contribution from these effects.

9.1 The large-$N$ extrapolation

If a long chain maintains spacing $r_m$, the attraction between non-neighboring molecules contributes approximately

For large $N$,

The correlation energy is extensive because the $1/s^3$ sum converges. This supports the possibility of an elongated self-bound molecular droplet or quantum crystal.

It remains an extrapolation. The two- and three-molecule calculations establish the mechanism and the first correlations; they do not by themselves solve the thermodynamic many-body problem. At larger $N$, collective phonons, finite temperature, loss, residual transverse motion, and competing planar structures can alter the phase diagram.

10. Ask the Born-Oppenheimer question carefully

Three-body binding often invites the word Efimov. The safe question is not “are there three particles?” It is “does the slow hyperradial motion see an attractive scale-invariant potential?”

For two heavy particles separated by $R$, let the light or fast degrees of freedom adjust adiabatically. Their ground energy defines a Born-Oppenheimer potential $V_{BO}(R)$. An Efimov channel requires, over a substantial interval,

The $R^{-2}$ form is special because kinetic and potential energies scale identically. It permits discrete scale invariance after a short-distance boundary condition is supplied.

In the deep, interaction-induced-one-dimensional regime analyzed in the 2026 Bose-Fermi-duality paper, the computed three-molecule Born-Oppenheimer surface does not display this $-R^{-2}$ interval. The cluster size is controlled by $r_m$, and the state resembles a molecular chain built from the shielded pair minimum. Calling this state Efimovian would erase the mechanism that actually binds it.

There is, however, no general prohibition against Efimov physics in microwave-shielded polar molecules. A later 2026 study, Deng et al., “Efimov Effect in Ultracold Microwave-Shielded Polar Molecules”, finds an Efimov window near a tunable scattering resonance and predicts a universal three-body parameter in dipolar units. The two results concern different regimes:

• the chain-like clusters live near the interaction-induced minimum at a scale of order $r_m$;
• the Efimov states emerge near resonance and explore a much larger scale-invariant interval.

Deep cluster and shallow Efimov state are not rival names for the same wavefunction. They are different solutions supported by different parts of the potential landscape. As the microwave parameters are tuned, an experiment may move between regimes or encounter avoided crossings, but the distinction should be made before interpreting a spectrum.

11. Two mechanisms, two meanings of universality

The common structure is easiest to see in a comparison.

question	species-dependent atomic waveguide	microwave-shielded molecules
What is frozen?	transverse oscillator excitations	angular deviations from an attractive axis
What freezes it?	an external harmonic trap	the anisotropic interaction itself
What survives virtually?	closed transverse and molecular centre-of-mass channels	angular zero-point fluctuations
What does elimination generate?	$a_d$, effective-range coefficients, multiple confinement resonances	an effective $-r^{-3}+r^{-4}$ potential and a noncrossing core
What binds the cluster?	exchange of a light atom between heavy fermions	pair attraction plus collective angular confinement
What is universal?	few-body energies fixed by low-dimensional scattering data and mass ratio	localized cluster structure set mainly by ${\ell }_d$ and ${\ell }_{\mathrm{\Omega }}$
Principal failure mode	cluster momentum reaches transverse or microscopic scales	angular localization weakens or a shallow threshold channel dominates

The two cases also clarify a useful distinction.

Parameter universality means that many microscopic potentials reduce to the same few low-energy numbers, such as $a_{2D}$ and $R_{2D}$. Once those numbers are matched, the low-energy observables agree.

Mechanism universality means that a robust balance of long-range terms fixes a shape or scale. In the microwave problem the attractive dipolar tail and fluctuation-induced core determine a pair minimum in terms of ${\ell }_d$ and ${\ell }_{\mathrm{\Omega }}$. Detailed short-range chemistry is screened from the state.

Neither form of universality means independence from all parameters. A universal prediction is often highly tunable. It is universal because its dependence is restricted and calculable.

12. The adiabatic approximation has a price

Both reductions use a separation between slow and fast motion. It is worth writing the neglected term.

Suppose the fast Hamiltonian $H_f(x)$ depends on a slow coordinate $x$, with eigenstates

Expand the full wavefunction as

Acting with the slow kinetic energy differentiates both ${\psi }_n$ and the basis state. The channel equations contain derivative couplings

and scalar Born-Huang terms built from $⟨{\partial }_{x}m|{\partial }_{x}n⟩$. Keeping only the fast ground state therefore gives more than ${\epsilon }_0(x)$. It also gives a geometric correction

The reduction is reliable when the slow kinetic scale and derivative couplings are small compared with the fast excitation gap. In the molecular problem, stronger microwave shielding narrows the angular state and generally improves the effective one-dimensional overlap, but an abrupt variation of the angular eigenstate can still amplify nonadiabatic corrections. A follow-up analysis, “Interaction-induced Dimension Reduction for Ultracold Polar Molecules”, maps this validity region and includes higher-order angular fluctuations.

The same logic appears in the atomic waveguide in a less geometric form. The closed-channel denominator $E-QHQ$ must remain well separated from unretained thresholds except where their effect has been encoded in the fitted scattering parameters. Near a new transverse threshold, a single-channel low-energy expansion loses uniform accuracy.

13. What an experiment would have to resolve

The theory suggests several measurements, but they test different layers of the argument.

For the heteronuclear atomic clusters, radio-frequency or magnetic-field association spectroscopy can locate the dimer, trimer, and tetramer thresholds. A convincing test should vary the confinement strength. The resonance positions and effective ranges must move with ${\ell }_{ho}$ in the manner predicted by the two-body confined calculation, while the cluster energies should collapse onto the low-dimensional few-body theory when expressed through the extracted $a_d$ and $R_d$.

Species-dependent centre-of-mass coupling can be tested by changing the ratio ${\omega }_h/{\omega }_l$ at fixed low-energy scattering length. If the spectrum depended only on a single geometric mean confinement length, such a change would do little. The matrix theory predicts otherwise because the mixed $Zz$ term and closed-channel composition change.

For microwave-shielded molecules, spectroscopy as a function of $\mathrm{\Omega }$ can test

Real-space pair correlations would probe the peaks near $r_m$ and $2r_m$ for the three-molecule cluster. Comparing bosonic and fermionic isotopologues would be especially sharp. Spectral lines and diagonal density correlations should approach each other in the strong-shielding regime, while momentum distributions should retain their statistical distinction.

The crossover to an Efimov regime requires a different scan. One must approach a scattering resonance and look for a geometric family of shallow three-body features, rather than infer Efimov physics from the existence of a single deep hexatomic state. Size, scaling with detuning, and the relation between successive states carry more information than the particle count.

Loss remains part of the physics. A state can be universal and still be difficult to observe if its preparation path crosses an unshielded region or if nonadiabatic transitions expose short-range chemistry. Binding energy and lifetime must be reported together.

14. A worked chain of deductions

This section condenses the argument into five deductions. Each one is short enough to verify, but together they connect the two papers.

Deduction A: equal trap lengths need not decouple motion

Given ${\ell }_h={\ell }_l$, we have $m_h{\omega }_h=m_l{\omega }_l$. The centre-of-mass coupling is

For $m_h\ne m_l$, equal lengths imply ${\omega }_h\ne {\omega }_l$, hence $V_{Zz}\ne 0$. Equal lengths make the spatial widths similar; they do not make the oscillation dynamics identical.

Deduction B: the effective range can control a shallow state without being microscopic

In quasi-one dimension the pole equation is

Compare the last two terms. Effective-range physics matters when

This condition can be met while $\kappa$ is still below the inverse microscopic range, because $R_{1D}$ may be enhanced by a narrow resonance or confinement-channel mixing. “Shallow” does not always mean “scattering-length only.”

Deduction C: the repulsive molecular wall is quantum mechanical

On the attractive axis the bare potential has no $C_6/r^6$ contribution. Yet angular fluctuations have stiffness

The angular kinetic coefficient scales as $1/r^2$, so the oscillator frequency scales as

At short range $K\sim r^{-6}$, and therefore

The wall is a zero-point energy, not a classical repulsive path along the axis.

Deduction D: spectral duality does not imply identical correlations

The mapping ${\mathrm{\Psi }}_F=\mathcal{A}{\mathrm{\Psi }}_B$ leaves $|\mathrm{\Psi }{|}^2$ invariant. Any diagonal position observable

therefore has the same expectation value in the exactly hard-core mapped states. The one-body density matrix is off-diagonal:

The sign factors at $y$ and $y^{\prime }$ do not generally cancel. Momentum distributions, which are Fourier transforms of ${\rho }_1$, can differ even when the energies coincide.

Deduction E: a deep chain state is not diagnosed by Efimov counting

A chain state has a preferred spacing $r_m$ and a wavefunction concentrated near $y_{i+1}-y_i\simeq r_m$. Scaling all separations changes the balance between $r^{-3}$ attraction and $r^{-4}$ repulsion. The Hamiltonian is not scale invariant there.

An Efimov state lives in an interval where the effective potential is approximately $-1/R^2$. It has no preferred classical minimum inside that interval; its scale is fixed by a boundary condition and discrete scaling. The two mechanisms can coexist in a multiscale potential, but their wavefunctions and parameter dependence are distinguishable.

15. Problems

Problem 1: derive the coupled oscillator exactly

Starting from

derive the expression in $Z$ and $z$. Then answer:

1. Which condition eliminates the mixed term?
2. Can a linear canonical transformation diagonalize the quadratic Hamiltonian even when ${\omega }_h\ne {\omega }_l$?
3. If it can, why does that not solve the contact-scattering problem once and for all?

Hint

The quadratic noninteracting Hamiltonian can always be written in normal modes. The interaction is local in the physical relative coordinate $z_h-z_l$, not necessarily in either normal-mode coordinate.

Solution

Substitute $z_h=Z+(m_l/M)z$ and $z_l=Z-(m_h/M)z$ and collect terms. The cross coefficient is $\mu ({\omega }_h^2-{\omega }_l^2)$. Thus equal frequencies eliminate it.

A symplectic or mass-weighted rotation can diagonalize the quadratic part for arbitrary positive frequencies. In fact, the original coordinates $z_h$ and $z_l$ already diagonalize the noninteracting trap. The difficulty is that the interaction imposes a boundary condition at $z_h=z_l$, a line oblique to generic normal-mode axes. A basis that diagonalizes propagation makes contact complicated; a basis that makes contact simple leaves propagation coupled. The molecular centre-of-mass basis is designed around the latter requirement.

Problem 2: count the information in a pole

For the quasi-one-dimensional pole equation

introduce $x=\kappa a_{1D}$ and $\rho =R_{1D}/a_{1D}^3$. Show that

Find the first correction to $x=1$ for $|\rho |\ll 1$. Discuss why more than one positive root can appear and which root should be trusted by a low-energy expansion.

Solution

Let $x=1+\delta$. To first order,

$$1-(1-\delta )+\rho (1+2\delta )=0,$$Copy LaTeX

so $\delta \simeq -\rho$ and

$$\kappa \simeq \frac{1}{a_{1D}}(1-\frac{R_{1D}}{a_{1D}^3}).$$Copy LaTeX

The cubic equation can have multiple positive roots. The deepest root may satisfy $\kappa {\ell }_{\perp }\gtrsim 1$ or probe the microscopic range, where terms beyond $R_{1D}{q}^2$ are no longer small. Root counting inside a truncated amplitude is not state counting in the full Hamiltonian. The trustworthy root is the one that lies inside the momentum window used to obtain the expansion and that remains stable when the next-order term is included.

Problem 3: reconstruct the molecular minimum

Use

to find $r_m$. Substitute the definitions of $C_3$ and $C_6$ and show that the result is proportional to ${\ell }_{\mathrm{\Omega }}$. What assumption is hidden in using only the short-distance form of the square root?

Solution

Setting $U^{\prime }(r_m)=0$ gives

$$12C_3{r}_m^{-4}-16\hslash \sqrt{\frac{C_6}{m}}{r}_m^{-5}=0,$$Copy LaTeX

hence

$$r_m=\frac{4\hslash }{3C_3}\sqrt{\frac{C_6}{m}}=4\sqrt{2}(1+{\delta }_r^2{)}^{1/4}{\ell }_{\mathrm{\Omega }}.$$Copy LaTeX

For ${\delta }_r=0.2$ the final factor is close to one. The approximation assumes that the $C_6/r^8$ term dominates over the $C_3/r^5$ term inside the zero-point square root near the minimum. If the minimum moves outward, the $C_3$ contribution and higher angular corrections change its position.

Problem 4: identify what the Bose-Fermi map preserves

For two hard-core particles in one dimension, write

Show that the pair distribution is identical. Then use the definition of the one-body density matrix to explain why its off-diagonal elements need not be identical. Which of the following are necessarily shared: energy, density profile, pair correlation, momentum distribution, contact value at coincidence?

Solution

The mapping changes only a sign, so every diagonal probability in position space is unchanged. For an exact hard-core mapping, energy, density profile, and pair correlation agree. The contact probability at coincidence is zero for both. The off-diagonal density matrix contains wavefunctions evaluated at two different coordinates; the sign products are then different functions of the integrated spectator coordinates. The momentum distribution need not agree.

Problem 5: design a regime discriminator

Imagine that spectroscopy finds one three-molecule bound state below the dimer-plus-molecule threshold. Propose a parameter scan that distinguishes an interaction-induced chain state from an Efimov state. Your answer should use at least two of the following: size, successive energy ratios, dependence on $\mathrm{\Omega }$, dependence on scattering length, and pair correlations.

Discussion

A chain state should retain a pronounced pair-correlation peak near $r_m\propto {\mathrm{\Omega }}^{-1/2}$, and its energy should track the balance between the $r^{-3}$ attraction and $r^{-4}$ core. Near an Efimov resonance, a shallow state expands with the scattering length, successive states approach a geometric energy ratio within the scale-invariant window, and no fixed pair spacing tied to the minimum is expected. A single energy line is insufficient. Measuring its size or pair correlation while scanning both $\mathrm{\Omega }$ and the resonant detuning would separate the two mechanisms much more cleanly.

16. Further reading

The following sequence is chosen to follow the logic of the article rather than chronology.

1. M. Olshanii, “Atomic Scattering in the Presence of an External Confinement and a Gas of Impenetrable Bosons”. The canonical derivation of confinement-induced one-dimensional scattering.
2. T. Bergeman, M. G. Moore, and M. Olshanii, “Atom-Atom Scattering under Cylindrical Harmonic Confinement”. The closed transverse modes are displayed as the Feshbach structure behind the resonance.
3. T. Shi and X. Cui, “Effective scatterings and universal clusters of heteronuclear ultracold mixtures in quasi-low dimensions”. The primary source for the species-dependent confinement, effective scattering coefficients, and Li-Cr/Li-K cluster predictions discussed here. As of this article’s date, this is a preprint.
4. M. Girardeau, “Relationship between Systems of Impenetrable Bosons and Fermions in One Dimension”. The original hard-core Bose-Fermi mapping.
5. T. Karman and J. M. Hutson, “Microwave Shielding of Ultracold Polar Molecules”. A foundation for the dressed interaction used to suppress short-range loss.
6. X.-Y. Chen et al., “Field-linked resonances of polar molecules”. Experimental observation of long-lived field-linked tetratomic states.
7. T. Shi, H. Wang, and X. Cui, “Universal Bound States with Bose-Fermi Duality in Microwave-Shielded Ultracold Molecules”. The primary source for the fluctuation-induced $r^{-4}$ wall, tetratomic and hexatomic clusters, and approximate spectral duality.
8. H. Wang, T. Shi, and X. Cui, “Interaction-induced Dimension Reduction for Ultracold Polar Molecules”. A broader analysis of the effective one-dimensional regime and higher-order angular fluctuations.
9. F. Deng et al., “Efimov Effect in Ultracold Microwave-Shielded Polar Molecules”. The resonant, large-scale three-body regime that must be distinguished from the deep chain-like clusters.

17. What to carry forward

The frozen coordinate is often the coordinate doing the most conceptual work. In an atomic waveguide, virtual transverse excitations turn a three-dimensional resonance into low-dimensional scattering data and can couple relative motion to molecular centre-of-mass channels. In a microwave-shielded molecule, angular zero-point motion creates the repulsive wall that makes the effective one-dimensional problem possible.

The few-body state is universal only after the surviving data have been named. For the heteronuclear clusters those data are the mass ratio and the low-dimensional amplitude, including its effective-range coefficient when necessary. For the molecular clusters they are long-range dressed scales such as ${\ell }_d$ and ${\ell }_{\mathrm{\Omega }}$, together with the condition that angular localization remains adiabatic.

Finally, statistics is not erased by dimensional reduction. It can be hidden from the spectrum when configuration sectors cease to communicate, while remaining visible in phase-sensitive and momentum-space observables. That is a more useful statement than saying that bosons and fermions become the same. The geometry has made one class of measurements forget their difference and left another class able to remember it.