Gross–Pitaevskii Regime

• Gross–Pitaevskii regime is defined by dilute, short-range interactions leading to an effective nonlinear theory where scattering lengths dictate the macroscopic behavior.
• In three dimensions, the scaling involves potentials like N^2V(Nx) resulting in a cubic GP functional, while in two dimensions it requires an exponentially small scattering length and logarithmic corrections.
• The regime extends to fermionic systems on the deep BEC side, where tightly bound pairs behave as bosons, and rigorous analyses detail condensation, excitation controls, and Bogoliubov corrections.

The Gross–Pitaevskii regime is a dilute, short-range scaling regime in which the microscopic many-body problem reduces, at leading order, to a nonlinear effective theory of Gross–Pitaevskii type. In the standard three-dimensional bosonic setting, this regime is realized by interactions whose range is of order $N^{-1}$ and whose scattering length is of order $N^{-1}$, so that the effective coupling in the macroscopic theory is fixed by the scattering length rather than by $\int V$ (Boccato et al., 2017). In two dimensions, the analogous regime requires an exponentially small scattering length and correspondingly stronger scaling of the potential (Caraci et al., 2020). In trapped interacting fermions on the deep BEC side of the BCS–BEC crossover, an analogous Gross–Pitaevskii regime appears when tightly bound pairs emerge from the microscopic BCS functional and behave effectively as bosons (Calignano et al., 2022). Taken together, these works use the term for a class of scale-separation limits in which strong microscopic correlations survive only through a small number of effective parameters in the macroscopic theory.

1. Microscopic definitions and canonical scalings

The defining microscopic feature of the Gross–Pitaevskii regime is the simultaneous presence of short interaction range, strong local correlations, and a macroscopic limit in which the leading energy remains of order $N$ or of the relevant macroscopic scale. In three-dimensional bosonic systems on a unit torus, the Hamiltonian is

$$H_N = \sum_{j=1}^N -\Delta_{x_j} + \kappa \sum_{1\le i<j\le N} N^2 V\big(N(x_i-x_j)\big),$$

with $V\in L^3(\mathbb R^3)$ non-negative, spherically symmetric, and compactly supported; the rescaled interaction has scattering length $a_N=a_0/N$ (Boccato et al., 2017). In trapped three-dimensional systems, the same scaling appears with an external potential,

$$H_N=\sum_{j=1}^N\Big(-\Delta_{x_j}+V_{\mathrm{ext}}(x_j)\Big)+\sum_{1\le j<k\le N}N^2V\big(N(x_j-x_k)\big),$$

again with scattering length $a/N$ for the rescaled potential (Nam et al., 2020).

In two dimensions, the Gross–Pitaevskii regime is qualitatively different because of logarithmic scattering. On the two-dimensional unit torus, the relevant Hamiltonian is

$$H_N=\sum_{j=1}^N-\Delta_{x_j}+\sum_{1\le i<j\le N} e^{2N}V\big(e^N(x_i-x_j)\big),$$

or equivalently $N^{-1}$0. In this scaling, the interaction range is of order $N^{-1}$1, and the scattering length of the rescaled potential is exponentially small in $N^{-1}$2 (Caraci et al., 2022, Caraci et al., 2020).

In the fermionic setting considered in the trapped BCS functional, the regime is parametrized not by $N^{-1}$3 but by a micro-to-macro scale ratio $N^{-1}$4. The particles interact through a short-range attractive potential with a two-body bound state, the trap scale is fixed at $N^{-1}$5, the interaction acts on scale $N^{-1}$6, and the chemical potential is tuned as

$N^{-1}$7

The resulting pairs have size $N^{-1}$8, and the leading term of the BCS ground-state energy is described by a Gross–Pitaevskii functional for a bosonic order parameter (Calignano et al., 2022).

Setting	Microscopic regime	Effective object
3D bosons, box or trap	$N^{-1}$9, scattering length $\int V$0	cubic GP functional with coupling $\int V$1 (Boccato et al., 2017, Nam et al., 2020)
2D bosons, unit torus	$\int V$2, exponentially small scattering length	constant condensate, leading energy $\int V$3, Bogoliubov spectrum (Caraci et al., 2020, Caraci et al., 2022)
Trapped fermions	$\int V$4, bound-state pairing, $\int V$5	GP functional for the pair field $\int V$6 (Calignano et al., 2022)

2. Effective Gross–Pitaevskii theories

In the three-dimensional bosonic regime, the effective macroscopic theory is the Gross–Pitaevskii functional

$\int V$7

with a unique non-negative minimizer $\int V$8 solving

$\int V$9

for some chemical potential $N$0 (Nam et al., 2020). In the homogeneous unit torus, the GP functional reduces to

$N$1

and the unique minimizer is the constant orbital $N$2 (Boccato et al., 2017).

The effective coupling is not the bare integral of the interaction but the scattering length. In three dimensions this is the origin of the coefficient $N$3 in the quartic term (Nam et al., 2020). In two dimensions, the same role is played by the logarithmic dependence on the exponentially small scattering length. On the unit torus, the leading energy is $N$4, and the constant mode $N$5 plays the role of the condensate orbital (Caraci et al., 2020). A plausible implication is that the two-dimensional regime is still Gross–Pitaevskii in the sense of being scattering-controlled, but with logarithmic renormalization replacing the three-dimensional $N$6 law.

For interacting fermions in a trap, the effective macroscopic theory is

$N$7

with minimization over

$N$8

and coupling

$N$9

Here $$H_N = \sum_{j=1}^N -\Delta_{x_j} + \kappa \sum_{1\le i<j\le N} N^2 V\big(N(x_i-x_j)\big),$$0 is the bound-state wave function of $$H_N = \sum_{j=1}^N -\Delta_{x_j} + \kappa \sum_{1\le i<j\le N} N^2 V\big(N(x_i-x_j)\big),$$1, $$H_N = \sum_{j=1}^N -\Delta_{x_j} + \kappa \sum_{1\le i<j\le N} N^2 V\big(N(x_i-x_j)\big),$$2 is the trap, and the chemical-potential shift $$H_N = \sum_{j=1}^N -\Delta_{x_j} + \kappa \sum_{1\le i<j\le N} N^2 V\big(N(x_i-x_j)\big),$$3 enters as the effective bosonic chemical potential (Calignano et al., 2022).

3. Condensation and excitation control

One of the central structural results in the Gross–Pitaevskii regime is that low-energy states exhibit complete Bose–Einstein condensation with quantitative control on depletion. In the three-dimensional translation-invariant box, if a normalized state $$H_N = \sum_{j=1}^N -\Delta_{x_j} + \kappa \sum_{1\le i<j\le N} N^2 V\big(N(x_i-x_j)\big),$$4 satisfies

$$H_N = \sum_{j=1}^N -\Delta_{x_j} + \kappa \sum_{1\le i<j\le N} N^2 V\big(N(x_i-x_j)\big),$$5

then

$$H_N = \sum_{j=1}^N -\Delta_{x_j} + \kappa \sum_{1\le i<j\le N} N^2 V\big(N(x_i-x_j)\big),$$6

so the number of particles outside the condensate is uniformly bounded in $$H_N = \sum_{j=1}^N -\Delta_{x_j} + \kappa \sum_{1\le i<j\le N} N^2 V\big(N(x_i-x_j)\big),$$7 (Boccato et al., 2017). The same paper proves an operator inequality equivalent to

$$H_N = \sum_{j=1}^N -\Delta_{x_j} + \kappa \sum_{1\le i<j\le N} N^2 V\big(N(x_i-x_j)\big),$$8

with $$H_N = \sum_{j=1}^N -\Delta_{x_j} + \kappa \sum_{1\le i<j\le N} N^2 V\big(N(x_i-x_j)\big),$$9 the excitation number operator (Boccato et al., 2017).

In trapped three-dimensional systems, the analogous quantitative statement is

$V\in L^3(\mathbb R^3)$0

and for any $V\in L^3(\mathbb R^3)$1-body state $V\in L^3(\mathbb R^3)$2,

$V\in L^3(\mathbb R^3)$3

For ground states, the number of particles outside the GP mode is therefore bounded by a constant independent of $V\in L^3(\mathbb R^3)$4, and the depletion fraction is $V\in L^3(\mathbb R^3)$5 (Nam et al., 2020).

In two dimensions on the unit torus, low-energy states obey the same qualitative pattern. If

$V\in L^3(\mathbb R^3)$6

then

$V\in L^3(\mathbb R^3)$7

This yields an $V\in L^3(\mathbb R^3)$8 bound on the number of orthogonal excitations and establishes complete Bose–Einstein condensation with almost optimal bounds in the two-dimensional GP regime (Caraci et al., 2020).

A further strengthening is available for trapped three-dimensional gases. For every low-energy state $V\in L^3(\mathbb R^3)$9, there exists $a_N=a_0/N$0 sufficiently small such that

$a_N=a_0/N$1

which implies the large-deviation bound

$a_N=a_0/N$2

This is exponential control of excitations, rather than merely polynomial control of moments (Behrmann et al., 27 Jan 2025).

4. Bogoliubov theory, excitation spectra, and energy corrections

The Gross–Pitaevskii regime is not exhausted by leading-order condensation. A large body of work identifies the low-energy excitation Hamiltonian and the subleading energy corrections by rigorous Bogoliubov methods. In the three-dimensional homogeneous box, the low-energy spectrum of $a_N=a_0/N$3 is asymptotically generated by quasiparticles with dispersion

$a_N=a_0/N$4

and the low-energy eigenvalues are finite sums

$a_N=a_0/N$5

up to errors vanishing with $a_N=a_0/N$6 (Cenatiempo, 2019). In trapped systems, the corresponding Bogoliubov operator is

$a_N=a_0/N$7

and the many-body low-energy spectrum converges to sums of its eigenvalues (Brennecke, 2022).

At the level of the ground-state energy, the three-dimensional GP regime admits a progressively sharper asymptotic expansion. For integrable repulsive potentials in the unit torus,

$a_N=a_0/N$8

which identifies the order-$a_N=a_0/N$9 correction beyond the GP term and recovers the Lee–Huang–Yang contribution in the thermodynamic interpretation (Basti et al., 2022). For hard spheres of radius $$H_N=\sum_{j=1}^N\Big(-\Delta_{x_j}+V_{\mathrm{ext}}(x_j)\Big)+\sum_{1\le j<k\le N}N^2V\big(N(x_j-x_k)\big),$$0, a second-order upper bound with the same structure is proved up to an error $$H_N=\sum_{j=1}^N\Big(-\Delta_{x_j}+V_{\mathrm{ext}}(x_j)\Big)+\sum_{1\le j<k\le N}N^2V\big(N(x_j-x_k)\big),$$1, using a trial state that is the product of a Jastrow factor and a Bogoliubov-type wave function (Basti et al., 2022).

The third-order correction has also been resolved for the translation-invariant three-dimensional GP regime. The ground-state energy satisfies

$$H_N=\sum_{j=1}^N\Big(-\Delta_{x_j}+V_{\mathrm{ext}}(x_j)\Big)+\sum_{1\le j<k\le N}N^2V\big(N(x_j-x_k)\big),$$2

so the next term in the GP expansion is of order $$H_N=\sum_{j=1}^N\Big(-\Delta_{x_j}+V_{\mathrm{ext}}(x_j)\Big)+\sum_{1\le j<k\le N}N^2V\big(N(x_j-x_k)\big),$$3 and depends on the potential only through its scattering length $$H_N=\sum_{j=1}^N\Big(-\Delta_{x_j}+V_{\mathrm{ext}}(x_j)\Big)+\sum_{1\le j<k\le N}N^2V\big(N(x_j-x_k)\big),$$4 (Caraci et al., 2023).

The two-dimensional GP regime has its own Bogoliubov structure. On the unit torus with $$H_N=\sum_{j=1}^N\Big(-\Delta_{x_j}+V_{\mathrm{ext}}(x_j)\Big)+\sum_{1\le j<k\le N}N^2V\big(N(x_j-x_k)\big),$$5, the low-energy excitation spectrum converges to a free quasiparticle Fock spectrum with dispersion

$$H_N=\sum_{j=1}^N\Big(-\Delta_{x_j}+V_{\mathrm{ext}}(x_j)\Big)+\sum_{1\le j<k\le N}N^2V\big(N(x_j-x_k)\big),$$6

and the derivation requires, in addition to quadratic and cubic conjugations, a quartic conjugation specific to the two-dimensional regime (Caraci et al., 2022).

5. Time-dependent formulations and fluctuation theory

The dynamical Gross–Pitaevskii regime concerns the persistence of condensation under many-body Schrödinger evolution and the emergence of effective nonlinear dynamics for the condensate wave function. For three-dimensional bosons with two-body interactions $$H_N=\sum_{j=1}^N\Big(-\Delta_{x_j}+V_{\mathrm{ext}}(x_j)\Big)+\sum_{1\le j<k\le N}N^2V\big(N(x_j-x_k)\big),$$7, the macroscopic dynamics is governed by the GP equation

$$H_N=\sum_{j=1}^N\Big(-\Delta_{x_j}+V_{\mathrm{ext}}(x_j)\Big)+\sum_{1\le j<k\le N}N^2V\big(N(x_j-x_k)\big),$$8

while the orthogonal excitations are described, after factoring out microscopic correlations, by a time-dependent Bogoliubov dynamics with quadratic generator (Caraci et al., 2023).

A strong form of this statement is an $$H_N=\sum_{j=1}^N\Big(-\Delta_{x_j}+V_{\mathrm{ext}}(x_j)\Big)+\sum_{1\le j<k\le N}N^2V\big(N(x_j-x_k)\big),$$9-norm approximation of the full many-body evolution. For initial data of the form

$a/N$0

with suitable bounds on $a/N$1, one has

$a/N$2

for $a/N$3 large. The limiting fluctuation dynamics is quasi-free, and a central limit theorem holds for fluctuations of bounded one-body observables around the GP evolution (Caraci et al., 2023).

The dynamical GP program has also been extended to genuine three-body interactions. For a Hamiltonian

$a/N$4

which is the three-body Gross–Pitaevskii scaling with $a/N$5, condensation is preserved under the many-body evolution and the condensate wave function solves the quintic GP equation

$a/N$6

The effective coupling is universal and depends only on a three-body scattering hypervolume $a/N$7 (Triay et al., 6 Feb 2026). This extends the GP paradigm from cubic to quintic nonlinearities while retaining the same scattering-based notion of universality.

6. Fermionic realizations and the scope of the regime

A distinctive feature of the modern literature is that the Gross–Pitaevskii regime is not confined to elementary bosons. In trapped interacting fermions at $a/N$8, with an attractive two-body potential supporting a bound state and chemical potential

$a/N$9

the BCS ground-state energy satisfies

$$H_N=\sum_{j=1}^N-\Delta_{x_j}+\sum_{1\le i<j\le N} e^{2N}V\big(e^N(x_i-x_j)\big),$$0

and approximate BCS minimizers have pairing kernel

$$H_N=\sum_{j=1}^N-\Delta_{x_j}+\sum_{1\le i<j\le N} e^{2N}V\big(e^N(x_i-x_j)\big),$$1

with $$H_N=\sum_{j=1}^N-\Delta_{x_j}+\sum_{1\le i<j\le N} e^{2N}V\big(e^N(x_i-x_j)\big),$$2 an approximate GP minimizer and $$H_N=\sum_{j=1}^N-\Delta_{x_j}+\sum_{1\le i<j\le N} e^{2N}V\big(e^N(x_i-x_j)\big),$$3 the two-body bound-state wave function (Calignano et al., 2022). This is the rigorous realization of tightly bound fermion pairs behaving as bosons in a trap.

The same deep-BEC picture appears in effective-field-theory form for cold Fermi gases with tunable attraction. In the deep BEC regime, at the level of the Bogoliubov approximation, the dimer condensate is identical to an elementary-boson GP condensate with mass $$H_N=\sum_{j=1}^N-\Delta_{x_j}+\sum_{1\le i<j\le N} e^{2N}V\big(e^N(x_i-x_j)\big),$$4, density $$H_N=\sum_{j=1}^N-\Delta_{x_j}+\sum_{1\le i<j\le N} e^{2N}V\big(e^N(x_i-x_j)\big),$$5, and Bogoliubov dispersion

$$H_N=\sum_{j=1}^N-\Delta_{x_j}+\sum_{1\le i<j\le N} e^{2N}V\big(e^N(x_i-x_j)\big),$$6

This equivalence fails rapidly as one moves toward the BCS regime, and even in the deep BEC regime there remains an intrinsic difference beyond Bogoliubov: the effective action becomes a coupled two-component GP theory rather than a single elementary GP field (Rivers et al., 2016).

These fermionic results clarify the conceptual scope of the Gross–Pitaevskii regime. It is not a statement about bosonic statistics at the microscopic level; it is a statement about a macroscopic limit in which the relevant low-energy degrees of freedom are condensed bosonic modes, whether elementary or composite. In that sense, the regime unifies dilute Bose gases, molecular condensates on the BEC side of the BCS–BEC crossover, and even three-body condensate dynamics under a single scattering-controlled effective description (Calignano et al., 2022, Triay et al., 6 Feb 2026).