Abstract
We consider systems of N bosons trapped on the two-dimensional unit torus, in the Gross-Pitaevskii regime, where the scattering length of the repulsive interaction is exponentially small in the number of particles. We show that low-energy states exhibit complete Bose–Einstein condensation, with almost optimal bounds on the number of orthogonal excitations.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
We consider \(N \in {\mathbb {N}}\) bosons trapped in the two-dimensional box \(\varLambda = [-1/2;1/2]^2\) with periodic boundary conditions. In the Gross-Pitaevskii regime, particles interact through a repulsive pair potential, with a scattering length exponentially small in N. The Hamilton operator is given by
and acts on a dense subspace of \(L^2_s (\varLambda ^N)\), the Hilbert space consisting of functions in \(L^2 (\varLambda ^N)\) that are invariant with respect to permutations of the N particles. We assume here \(V \in L^3 ({\mathbb {R}}^2)\) to be compactly supported and pointwise non-negative (i.e. \(V(x) \ge 0\) for almost all \(x \in {\mathbb {R}}^2\)).
We denote by \({{\mathfrak {a}}}\) the scattering length of the unscaled potential V. We recall that in two dimensions and for a potential V with finite range \(R_0\), the scattering length is defined by
where \(R > R_0\), \(B_R\) is the disk of radius R centered at the origin and the infimum is taken over functions \(\phi \in H^1(B_R)\) with \(\phi (x)=1\) for all x with \(|x|=R\). The unique minimizer of the variational problem on the r.h.s. of (2) is non-negative, radially symmetric and satisfies the scattering equation
in the sense of distributions. For \(R_0 < |x| \le R\), we have
By scaling, \(\phi _N (x) := \phi ^{(e^N R)} (e^N x)\) is such that
We have
for all \(x \in {\mathbb {R}}^2\) with \(e^{-N} R_0 < |x| \le R\). Here \({{\mathfrak {a}}}_N= e^{-N} {{\mathfrak {a}}}\).
The spectral properties of trapped two dimensional bosons in the Gross-Pitaevskii regime (in the more general case where the bosons are confined by external trapping potentials) have been first studied in [13, 14, 16]. These results can be translated to the Hamilton operator (1), defined on the torus, with no external potential. They imply that the ground state energy \(E_N\) of (1) is such that
Moreover, they imply Bose–Einstein condensation in the zero-momentum mode \(\varphi _0 (x) = 1\) for all \(x \in \varLambda \), for any approximate ground state of (1). More precisely, it follows from [13] that, for any sequence \(\psi _N \in L^2_s (\varLambda ^N)\) with \(\Vert \psi _N \Vert = 1\) and
the one-particle reduced density matrix \(\gamma _N = \mathrm{tr}_{2, \ldots , N} |\psi _N \rangle \langle \psi _N |\) is such that
for a sufficiently small \({\bar{\delta }}>0\). The estimate (5) states that, in many-body states satisfying (4) (approximate ground states), almost all particles are described by the one-particle orbital \(\varphi _0\), with at most \(N^{1-\delta } \ll N\) orthogonal excitations.
Similar results have been obtained starting from a three dimensional Bose gas, trapped by a potential which is strongly confining in one direction, so that the system becomes effectively two-dimensional [22]. Finally, let us also mention [5, 10], where rigorous results on the time-evolution in the two-dimensional Gross-Pitaevskii regime have been established (in [5], the focus is on the dynamics of a three-dimensional gas, with strong confinement in one direction).
For \(V \in L^3({\mathbb {R}}^2)\), our main theorem improves (3) and (5) by providing more precise bounds on the ground state energy and on the number of excitations.
Theorem 1
Let \(V \in L^3 ({\mathbb {R}}^2)\) have compact support, be spherically symmetric and pointwise non-negative. Then there exists a constant \(C > 0\) such that the ground state energy \(E_N\) of (1) satisfies
Furthermore, consider a sequence \(\psi _N \in L^2_s (\varLambda ^N)\) with \(\Vert \psi _N \Vert = 1\) and such that
for a \(K > 0\). Then the reduced density matrix \(\gamma _N = \mathrm{tr}_{2,\ldots , N} | \psi _N \rangle \langle \psi _N |\) associated with \(\psi _N\) is such that
for all \(N \in {\mathbb {N}}\) large enough.
Remark
We expect that the bounds of Theorem 1 can be extended to two-dimensional systems of bosons trapped by an external potential (in three dimensions, similar estimates have been recently established in [7, 19]). In this case, the system exhibits condensation in the minimizer of the Gross-Pitaevskii energy functional, as shown in [13, 14, 16].
It is interesting to compare the Gross-Pitaevskii regime with the thermodynamic limit, where a Bose gas of N particles interacting through a fixed potential with scattering length \({{\mathfrak {a}}}\) is confined in a box with area \(L^2\), so that \(N, L \rightarrow \infty \) with the density \(\rho =N/L^2\) kept fixed. Let \(b=|\log (\rho {{\mathfrak {a}}}^2)|^{-1}\). Then, in the dilute limit \(\rho {{\mathfrak {a}}}^2 \ll 1\), the ground state energy per particle in the thermodynamic limit is expected to satisfy
with \(\gamma \) the Euler’s constant. The leading order term on the r.h.s. of (9) has been first derived in [21] and then rigorously established in [15], with an error rate \(b^{-1/5}\). The corrections up to order b have been predicted in [1, 18, 20]. To date, there is no rigorous proof of (9). Some partial result, based on the restriction to quasi-free states, has been recently obtained in [9, Theorem 1].
Extrapolating from (9), in the Gross-Pitaevskii regime we expect \(|E_N - 2\pi N| \le C\). While our estimate (6) captures the correct lower bound, the upper bound is off by a logarithmic correction. Eq. (8), on the other hand, is expected to be optimal (but of course, by (6), we need to choose \(K = C \log N\) to be sure that (7) can be satisfied). This bound can be used as starting point to investigate the validity of Bogoliubov theory for two dimensional bosons in the Gross-Pitaevskii regime, following the strategy developed in [3] for the three dimensional case; we plan to proceed in this direction in a separate paper.
The proof of Theorem 1 follows the strategy that has been recently introduced in [4] to prove condensation for three-dimensional bosons in the Gross-Pitaevskii limit. There are, however, additional obstacles in the two-dimensional case, requiring new ideas. To appreciate the difference between the Gross-Pitaevskii regime in two- and three-dimensions, we can compute the energy of the trivial wave function \(\psi _N \equiv 1\). The expectation of (1) in this state is of order \(N^2\). It is only through correlations that the energy can approach (6). Also in three dimensions, uncorrelated many-body wave functions have large energy, but in that case the difference with respect to the ground state energy is only of order N (\(N \widehat{V} (0)/2\) rather than \(4\pi \mathfrak {a} N\)). This observation is a sign that correlations in two-dimensions are stronger and play a more important role than in three dimensions (this creates problems in handling error terms that, in the three dimensional setting, were simply estimated in terms of the integral of the potential).
The paper is organized as follows. In Sect. 2 we introduce our setting, based on a description of orthogonal excitations of the condensate on a truncated Fock space. Factoring out the condensate, we introduce an excitation Hamiltonian \(\mathcal{L}_N\), unitarily equivalent to \(H_N\). In Sects. 3 and 4 we define two additional unitary maps, modelling the correlation structure characterising low-energy states. The first map is a generalized Bogoliubov transformation, given by the exponential of an anti-symmetric operator B, quadratic in creation and annihilation operators, see Eq. (33). Its action on \(\mathcal{L}_N\) leads to a second excitation Hamiltonian \(\mathcal{G}_{N,\alpha }\), whose vacuum expectation matches (6), at leading order. Unfortunately, \(\mathcal{G}_{N,\alpha }\) is not coercive enough to directly show Bose–Einstein condensation. To overcome this difficulty, we conjugate the main part of \(\mathcal{G}_{N,\alpha }\) (later denoted by \(\mathcal{G}_{N,\alpha }^{\text {eff}}\)) with a second unitary map, given by the exponential of an operator A, cubic in creation and annihilation operators, see Eq. (44). This defines a renormalized excitation Hamiltonian \(\mathcal{R}_{N,\alpha }\), where the singular interaction is regularized. In Sect. 5 we combine the bounds on \(\mathcal{G}_{N,\alpha }\) and \(\mathcal{R}_{N,\alpha }\) with a localization argument proposed in [11] for the number of excitations to conclude the proof of Theorem 1. Section 6 and App. 1 are devoted to the proof of the bounds on \(\mathcal{G}_{N,\alpha }\) and on \(\mathcal{R}_{N,\alpha }\) stated in Sects. 3 and 4, respectively. Finally, in App. 1, we establish some properties of the solution of the Neumann problem associated with the two-body potential V.
2 The Excitation Hamiltonian
Low-energy states of (1) exhibit condensation in the zero-momentum mode \(\varphi _0\) defined by \(\varphi _0 (x) = 1\) for all \(x \in \varLambda = [-1/2;1/2]^2\). Similarly as in [2, 4, 11], we are going to describe excitations of the condensate on the truncated bosonic Fock space
constructed on the orthogonal complement \(L^2_\perp (\varLambda )\) of \(\varphi _0\) in \(L^2 (\varLambda )\). To reach this goal, we define a unitary map \(U_N : L^2_s (\varLambda ^N) \rightarrow \mathcal{F}_+^{\le N}\) by requiring that \(U_N \psi _N = \{ \alpha _0, \alpha _1, \ldots , \alpha _N \}\), with \(\alpha _j \in L^2_\perp (\varLambda )^{\otimes _s j}\), if
With the usual creation and annihilation operators, we can write
for all \(\psi _N \in L^2_s (\varLambda ^N)\). It is then easy to check that \(U_N^* : \mathcal{F}_{+}^{\le N} \rightarrow L^2_s (\varLambda ^N)\) is given by
and that \(U_N^* U_N = 1\), i.e. \(U_N\) is unitary.
With \(U_N\), we can define the excitation Hamiltonian \(\mathcal{L}_N := U_N H_N U_N^*\), acting on a dense subspace of \(\mathcal{F}_+^{\le N}\). To compute the operator \(\mathcal{L}_N\), we first write the Hamiltonian (1) in momentum space, in terms of creation and annihilation operators \(a_p^*, a_p\), for momenta \(p \in \varLambda ^* = 2\pi {\mathbb {Z}}^2\). We find
where
is the Fourier transform of V, defined for all \(k \in {\mathbb {R}}^2\) (in fact, (1) is the restriction of (10) to the N-particle sector of the Fock space). We can now determine \(\mathcal{L}_N\) using the following rules, describing the action of the unitary operator \(U_N\) on products of a creation and an annihilation operator (products of the form \(a_p^* a_q\) can be thought of as operators mapping \(L^2_s (\varLambda ^N)\) to itself). For any \(p,q \in \varLambda ^*_+ = 2\pi {\mathbb {Z}}^2 \backslash \{ 0 \}\), we find (see [11]):
where \(\mathcal{N}_+ = \sum _{p\in \varLambda ^*_+} a_p^* a_p\) is the number of particles operator on \(\mathcal{F}_+^{\le N}\). We conclude that
with
where we introduced generalized creation and annihilation operators
for all \(p \in \varLambda ^*_+\).
On states exhibiting complete Bose–Einstein condensation in the zero-momentum mode \(\varphi _0\), we have \(a_0 , a_0^* \simeq \sqrt{N}\) and we can therefore expect that \(b_p^* \simeq a_p^*\) and that \(b_p \simeq a_p\). From the canonical commutation relations for the standard creation and annihilation operators \(a_p, a_p^*\), we find
Furthermore,
for all \(p,q,r \in \varLambda _+^*\); this implies in particular that \([b_p , \mathcal{N}_+] = b_p\), \([b_p^*, \mathcal{N}_+] = - b_p^*\). It is also useful to notice that the operators \(b^*_p, b_p\), like the standard creation and annihilation operators \(a_p^*, a_p\), can be bounded by the square root of the number of particles operators; we find
for all \(\xi \in \mathcal{F}^{\le N}_+\). Since \(\mathcal{N}_+ \le N\) on \(\mathcal{F}_+^{\le N}\), the operators \(b_p^* , b_p\) are bounded, with \(\Vert b_p \Vert , \Vert b^*_p \Vert \le (N+1)^{1/2}\).
3 Quadratic Renormalization
From (13) we see that conjugation with \(U_N\) extracts, from the original quartic interaction in (10), some large constant and quadratic contributions, collected in \(\mathcal{L}^{(0)}_N\) and \(\mathcal{L}^{(2)}_N\) respectively. In particular, the expectation of \(\mathcal{L}_N\) on the vacuum state \(\varOmega \) is of order \(N^2\), this being an indication of the fact that there are still large contributions to the energy hidden among cubic and quartic terms in \(\mathcal{L}^{(3)}_N\) and \(\mathcal{L}^{(4)}_N\). Since \(U_N\) only removes products of the zero-energy mode \(\varphi _0\), correlations among particles remain in the excitation vector \(U_N \psi _N\). Indeed, correlations play a crucial role in the two dimensional Gross-Pitaevskii regime and carry an energy of order \(N^2\).
To take into account the short scale correlation structure on top of the condensate, we consider the solution \(f_{\ell }\) of the equation
associated with the smallest possible eigenvalue \(\lambda _\ell \), on the ball \(|x| \le e^N \ell \), with Neumann boundary conditions and normalized so that \(f_{\ell }(x) = 1\) for \(|x|= e^N\ell \). Here and in the following we omit the N-dependence in the notation for \(f_\ell \) and for \(\lambda _\ell \). By scaling, we observe that \(f_{\ell }(e^N\cdot )\) satisfies
on the ball \(|x| \le \ell \). We choose \(\ell < 1/2\), so that the ball of radius \(\ell \) is contained in the box \(\varLambda = [-1/2 ; 1/2]^2\). We extend then \(f_\ell (e^N.)\) to \(\varLambda \), by setting \(f_{N,\ell } (x) = f_\ell (e^Nx)\), if \(|x| \le \ell \) and \(f_{N,\ell } (x) = 1\) for \(x \in \varLambda \), with \(|x| > \ell \). Then, assuming also that \(R_0 e^{-N} < \ell \) (later we will choose \(\ell = N^{-\alpha }\), so this condition is satisfied, for all N large enough),
where \(\chi _\ell \) is the characteristic function of the ball of radius \(\ell \). The Fourier coefficients of the function \(f_{N,\ell }\) are given by
for all \(p \in \varLambda ^*\). We introduce also the function \(w_\ell (x)=1-f_\ell (x)\) for \(|x| \le e^N\ell \) and extend it by setting \(w_\ell (x)=0\) for \(|x| >e^N\ell \). Its re-scaled version is defined by \(w_{N,\ell }: \varLambda \rightarrow {\mathbb {R}}\) \(w_{N,\ell }(x)=w_\ell (e^Nx)\) if \(|x| \le \ell \) and \(w_{N,\ell }=0\) if \(x \in \varLambda \) with \(|x| > \ell \).
The Fourier coefficients of the re-scaled function \(w_{N,\ell }\) are given by
We find \(\widehat{f}_{N,\ell }(p) = \delta _{p,0}- e^{-2N}{{\widehat{w}}}_{\ell }(e^{-N}p)\). From the Neumann problem (16) we obtain
where we used the notation \({\widehat{\chi }}_\ell \) for the Fourier coefficients of the characteristic function on the ball of radius \(\ell \). Note that \({\widehat{\chi }}_\ell (p)= \ell ^2\, \widehat{\chi }(\ell p)\) with \({{\widehat{\chi }}}(p)\) the Fourier coefficients of the characteristic function on the ball of radius one.
In the next lemma, we collect some important properties of the solution of (15).
Lemma 1
Let \(V\in L^3({\mathbb {R}}^2)\) be non-negative, compactly supported (with range \(R_0\)) and spherically symmetric, and denote its scattering length by \({{\mathfrak {a}}}\). Fix \(0<\ell <1/2\), N sufficiently large and let \(f_{\ell }\) denote the solution of (16). Then
-
(i)
$$\begin{aligned} 0 \le f_{\ell }(x) \le 1 \qquad \forall \, |x| \le e^N \ell \,. \end{aligned}$$
-
(ii)
We have
$$\begin{aligned} \left| \lambda _{\ell } - \frac{2}{(e^N\ell )^2 \log (e^N\ell /{{\mathfrak {a}}})} \right| \le \frac{C}{(e^N\ell )^2 \log ^2(e^N\ell /{{\mathfrak {a}}})} \end{aligned}$$(19) -
(iii)
There exists a constant \(C>0\) such that
$$\begin{aligned} \left| \int \text {d}x\, V(x) f_{\ell }(x) - \frac{4\pi }{\log (e^N\ell /{{\mathfrak {a}}})} \right| \le \frac{C}{\log ^2(e^N\ell /{{\mathfrak {a}}})} \end{aligned}$$(20) -
(iv)
There exists a constant \(C>0\) such that
$$\begin{aligned} \begin{aligned} |w_{\ell }(x)|&\le \left\{ \begin{array}{ll} C \qquad &{}\text {if } |x| \le R_0 \\ C \, \frac{\log (e^N\ell /|x|) }{\log (e^N\ell /{{\mathfrak {a}}})} \quad &{} \text {if } R_0 \le |x|\le e^N \ell \end{array} \right. \\ |\nabla w_{\ell }(x)|&\le \frac{C}{\log (e^N\ell /{{\mathfrak {a}}})} \frac{1}{|x| + 1} \qquad \text {for all } \, |x| \le e^N \ell \end{aligned} \end{aligned}$$(21) -
(v)
Let \(w_{N,\ell }= 1- f_{N,\ell }\) with \(f_{\ell , N}=f_\ell (e^N x)\). Then the Fourier coefficients of the function \(w_{N,\ell }\) defined in (17) are such that
$$\begin{aligned} |{{\widehat{w}}}_{N,\ell }(p)| \le \frac{C}{p^2 \log (e^N\ell /{{\mathfrak {a}}})}. \end{aligned}$$(22)
Proof
The proof of points (i)–(iv) is deferred in Appendix B. To prove point v) we use the scattering equation (18):
Using the fact that \( e^{2N}\lambda _\ell \le C \ell ^{-2} |\ln (e^N\ell /{{\mathfrak {a}}})|^{-1}\) and that \(0 \le f_\ell \le 1\), we end up with
\(\square \)
We now define \({\check{\eta }} : \varLambda \rightarrow {\mathbb {R}}\) through
With (21) we find
and in particular, recalling that \( e^{-N}R_0 < \ell \le 1/2\),
for all \(x \in \varLambda \). Using (24) we find
In the following we choose \(\ell =N^{-\alpha }\), for some \(\alpha >0\) to be fixed later, so that
This choice of \(\ell \) will be crucial for our analysis, as commented below. Notice, on the other hand, that the \(H^1\)-norms of \(\eta \) diverge, as \(N \rightarrow \infty \). From (23) and Lemma 1, part iv) we find
for \(N \in {\mathbb {N}}\) large enough. We denote with \(\eta : \varLambda ^* \rightarrow {\mathbb {R}}\) the Fourier transform of \({\check{\eta }}\), or equivalently
With (22) we can bound (since \(\ell = N^{-\alpha }\))
for all \(p \in \varLambda _+^*=2\pi {\mathbb {Z}}^2 \backslash \{0\}\), and for some constant \(C>0\) independent of N, if N is large enough. From (26) we also have
Moreover, (18) implies the relation
or equivalently, expressing also the other terms through the coefficients \(\eta _p\),
We will mostly use the coefficients \(\eta _p\) with \(p\ne 0\). Sometimes, however, it will be useful to have an estimate on \(\eta _0\) (because Eq. (31) involves \(\eta _0\)). From (27) and Lemma 1, part iv) we find
With the coefficients (27) we define the antisymmetric operator
and we consider the unitary operator
We refer to operators of the form (34) as generalized Bogoliubov transformations. In contrast with the standard Bogoliubov transformations
defined in terms of the standard creation and annihilation operators, operators of the form (34) leave the truncated Fock space \(\mathcal{F}_+^{\le N}\) invariant. On the other hand, while the action of standard Bogoliubov transformation on creation and annihilation operators is explicitly given by
there is no such formula describing the action of generalized Bogoliubov transformations.
Conjugation with (34) leaves the number of particles essentially invariant, as confirmed by the following lemma.
Lemma 2
Assume B is defined as in (33), with \(\eta \in \ell ^2 (\varLambda ^*)\) and \(\eta _p = \eta _{-p}\) for all \(p \in \varLambda ^*_+\). Then, for every \(n \in {\mathbb {N}}\) there exists a constant \(C > 0\) such that, on \(\mathcal{F}_+^{\le N}\),
as an operator inequality on \(\mathcal{F}^{\le N}_+\).
The proof of (36) can be found in [6, Lemma 3.1] (a similar result has been previously established in [23]).
With the generalized Bogoliubov transformation \(e^{B} : \mathcal{F}_+^{\le N} \rightarrow \mathcal{F}^{\le N}_+\), we define a new, renormalized, excitation Hamiltonian \(\mathcal{G}_{N,\alpha } : \mathcal{F}^{\le N}_+ \rightarrow \mathcal{F}^{\le N}_+\) by setting
In the next proposition, we collect important properties \(\mathcal{G}_{N,\alpha }\). We will use the notation
for the kinetic and potential energy operators, restricted on \(\mathcal{F}_+^{\le N}\), and \(\mathcal{H}_N = \mathcal{K}+ \mathcal{V}_N\). We also introduce a renormalized interaction potential \(\omega _N\in L^\infty (\varLambda )\), which is defined as the function with Fourier coefficients \({\widehat{\omega }}_N \)
for any \(p \in \varLambda ^*_+\), and
with \({{\widehat{\chi }}}(p)\) the Fourier coefficients of the characteristic function of the ball of radius one. From (19) and \(\ell =N^{-\alpha }\) one has \(|g_N|\le C\). Note in particular that the potential \({{\widehat{\omega }}}_N(p)\) decays on momenta of order \(N^\alpha \), which are much smaller than \(e^N\). From Lemma 1 parts (i) and (iii) we find
Proposition 1
Let \(V\in L^3({\mathbb {R}}^2)\) be compactly supported, pointwise non-negative and spherically symmetric. Let \(\mathcal{G}_{N,\alpha }\) be defined as in (37) and define
Then there exists a constant \(C > 0\) such that \(\mathcal{E}_{\mathcal{G}}= \mathcal{G}_{N,\alpha } - \mathcal{G}^\text {eff}_{N,\alpha }\) is bounded by
for all \(\alpha >1\), \(\xi \in \mathcal{F}_+^{\le N}\) and \(N\in {\mathbb {N}}\) large enough.
The proof of Proposition 1 is very similar to the proof of [3, Prop. 4.2]. For completeness, we discuss the changes in Appendix A.
4 Cubic Renormalization
Conjugation with the generalized Bogoliubov transformation (35) renormalizes constant and off-diagonal quadratic terms on the r.h.s. of (42). In order to estimate the number of excitations \(\mathcal{N}_+\) through the energy and show Bose–Einstein condensation, we still need to renormalize the diagonal quadratic term (the part proportional to \(N \widehat{V} (0) \mathcal{N}_+\), on the first line of (42)) and the cubic term on the second line of (42). To this end, we conjugate \(\mathcal{G}_{N,\alpha }^{\text {eff}}\) with an additional unitary operator, given by the exponential of the anti-symmetric operator
with \(\eta _p\) defined in (27).
An important observation is that while conjugation with \(e^A\) allows to renormalize the large terms in \(\mathcal{G}_{N,\alpha }\), it does not substantially change the number of excitations. The following proposition can be proved similarly to [4, Proposition 5.1].
Proposition 2
Suppose that A is defined as in (44). Then, for any \(k\in {\mathbb {N}}\) there exists a constant \(C >0\) such that the operator inequality
holds true on \(\mathcal{F}_+^{\le N}\), for any \(\alpha > 0\) (recall the choice \(\ell = N^{-\alpha }\) in the definition (27) of the coefficients \(\eta _r\)), and N large enough.
We will also need to control the growth of the expectation of the energy \(\mathcal{H}_N\) with respect to the cubic conjugation. This is the content of the following proposition, which is proved in Sect. 6.1.
Proposition 3
Let A be defined as in (44). Then there exists a constant \(C > 0\) such that
for all \(\alpha \ge 1\), \(s \in [0;1]\) and \(N \in {\mathbb {N}}\) large enough.
We use now the cubic phase \(e^{A}\) to introduce a new excitation Hamiltonian, obtained by conjugating the main part \(\mathcal{G}_{N, \alpha }^{\text {eff}}\) of \(\mathcal{G}_{N,\alpha }\). We define
on a dense subset of \(\mathcal{F}_+^{\le N}\). Conjugation with \(e^{A}\) renormalizes both the contribution proportional to \(\mathcal{N}_+\) (in the first line on the r.h.s. of (42)) and the cubic term on the r.h.s. of (42), effectively replacing the singular potential \(\widehat{V} (p/e^N)\) by the renormalized potential \({{\widehat{\omega }}}_N(p)\) defined in (39). This follows from the following proposition.
Proposition 4
Let \(V\in L^3({\mathbb {R}}^2)\) be compactly supported, pointwise non-negative and spherically symmetric. Let \(\mathcal{R}_{N,\alpha }\) be defined in (46) and define
Then for \(\ell =N^{-\alpha }\) and \(\alpha >2\) there exists a constant \(C>0\) such that \(\mathcal{E}_\mathcal{R}= \mathcal{R}_{N,\alpha }- \mathcal{R}_{N,\alpha }^{\text {eff}}\) is bounded by
for \(N \in {\mathbb {N}}\) sufficiently large.
The proof of Proposition 4 will be given in Sect. 6. We will also need more detailed information on \(\mathcal{R}_{N,\alpha }^{\text {eff}}\), as contained in the following proposition.
Proposition 5
Let \(\mathcal{R}_{N,\alpha }^{\text {eff}}\) be defined in (47). Then, for every \(c > 0\) there is a constant \(C > 0\) (large enough) such that
for all \(\alpha >2\) and \(N \in {\mathbb {N}}\) large enough.
Moreover, let \(f,g : {\mathbb {R}}\rightarrow [0;1]\) be smooth, with \(f^2 (x) + g^2 (x) =1\) for all \(x \in {\mathbb {R}}\). For \(M \in {\mathbb {N}}\), let \(f_M := f(\mathcal{N}_+/M)\) and \(g_M:= g(\mathcal{N}_+/M)\). Then there exists \(C > 0\) such that
with
for all \(\alpha > 2\), \(M \in {\mathbb {N}}\) and \(N \in {\mathbb {N}}\) large enough.
Proof
From (47), using that \(| {{\widehat{\omega }}}_N(0) |\le C\) we have
For the cubic term on the r.h.s. of (51), with
we can bound
As for the off-diagonal quadratic term on the r.h.s of (51), we combine it with part of the kinetic energy to estimate. For any \(0< \mu < 1\), we have
since \(a_p^* a_p - b_p^* b_p = a_p^* (\mathcal{N}_+ / N) a_p\). With (14), we conclude that
With the choice \(\mu = C / \log N\) and with (52), we obtain
To bound the first terms on the r.h.s. of the last equation, we use the term \({\widehat{\omega }}_N (0) \mathcal{N}_+\), in (51). To this end, we observe that, with (41),
for every \(p \in \varLambda ^*_+\) (notice that \(|p| \ge 2\pi \), for every \(p \in \varLambda ^*_+\)) and for N large enough (recall the choice \(\mu = C / \log N\)). Inserting (53) and (55) in (51) and using the kinetic energy \(\mu \mathcal{K}= C (\log N)^{-1} \mathcal{K}\) (remaining after subtracting the term \((1-\mu ) \mathcal{K}\) needed on the l.h.s. of (55)) to bound the r.h.s. of (53), we find
Let us now consider the second term on the r.h.s more carefully. Using that, from (39), \({\widehat{\omega }}_N (p) = g_N {\widehat{\chi }} (p/N^\alpha )\), we can bound, for any fixed \(K > 0\),
With \(|{\widehat{\omega }}_N (p) - {\widehat{\omega }}_N (0)|\le C |p|/ N^\alpha \), we obtain
For \(q \in {\mathbb {R}}^2\), let us define \(h(q) = 1/p^2\), if q is contained in the square of side length \(2\pi \) centered at \(p \in \varLambda ^*_+\) (with an arbitrary choice on the boundary of the squares). We can then estimate, for K large enough,
For q in the square centered at \(p \in \varLambda ^*_+\), we bound
Hence
Inserting in (57), we conclude that
Combining the last bound with (41) (and noticing that the contribution proportional to \(\log N\) cancels exactly), from (56) we obtain
which proves (49).
Next we prove (50). From (47), with \(|\widehat{\omega }_N(0)|\le C\), the bound (53) and since, by (52),
it follows that
where for arbitrary \( \delta > 0\), there exists a constant \(C>0\) such that
We now note that for \(f: {\mathbb {R}}\rightarrow {\mathbb {R}}\) smooth and bounded and \(\theta _{N,\alpha }\) defined above, there exists a constant \(C>0\) such that
for all \(\alpha >2\) and \(N \in {\mathbb {N}}\) large enough. The proof of (60) follows analogously to the one for (59), since the bounds leading to (59) remain true if we replace the operators \(b_p^\#\), \(\#=\{ \cdot , *\}\), and \(a_p^* a_q\) with \([f(\mathcal{N}_+/M), [f(\mathcal{N}_+/M), b^\#_p ]]\) or \([f(\mathcal{N}_+/M), [f(\mathcal{N}_+/M), a^*_p a_q ]]\) respectively, provided we multiply the r.h.s. by an additional factor \(M^{-2}\Vert f'\Vert _\infty ^2\), since, for example
and \(\Vert f(\mathcal{N}_+/M) - f((\mathcal{N}_++1)/M) \Vert \le C M^{-1} \Vert f'\Vert _\infty \). With an explicit computation we obtain
Writing \(\mathcal{R}_{N,\alpha }^{\text {eff}}\) as in (58) and using (60) we get
\(\square \)
5 Proof of Theorem 1
The next proposition combines the results of Propositions 1, 4 and 5. Its proof makes use of localization in the number of particle and is an adaptation of the proof of [4, Proposition 6.1]. The main difference w.r.t. [4] is that here we need to localize on sectors of \(\mathcal{F}^{\le N}\) where the number of particles is o(N), in the limit \(N \rightarrow \infty \).
Proposition 6
Let \(V\in L^3({\mathbb {R}}^2)\) be compactly supported, pointwise non-negative and spherically symmetric. Let \(\mathcal{G}_{N,\alpha }\) be the renormalized excitation Hamiltonian defined as in (37). Then, for every \(\alpha \ge 5/2\), there exist constants \(C,c > 0\) such that
for all \(N \in {\mathbb {N}}\) sufficiently large.
Proof
Let \(f,g: {\mathbb {R}}\rightarrow [0;1]\) be smooth, with \(f^2 (x) + g^2 (x)= 1\) for all \(x \in {\mathbb {R}}\). Moreover, assume that \(f (x) = 0\) for \(x > 1\) and \(f (x) = 1\) for \(x < 1/2\). For a small \(\varepsilon >0\), we fix \(M = N^{1-\varepsilon }\) and we set \(f_M = f (\mathcal{N}_+ / M), g_M = g (\mathcal{N}_+ / M)\). It follows from Proposition 5 that
Let us consider the first term on the r.h.s. of (62). From Proposition 5, for all \(\alpha >2\) there exist \(c, C>0\) such that
On the other hand, with (58) and (59) we also find
for all \(\alpha > 2\) and N large enough. Moreover, due to the choice \(M=N^{1-\varepsilon }\), we have
With the last bound, Eq. (63) implies that
for N large enough.
Let us next consider the second term on the r.h.s. of (62). We claim that there exists a constant \(c > 0\) such that
for all N sufficiently large. To prove (66) we observe that, since \(g(x) = 0\) for all \(x \le 1/2\),
where \(\mathcal{F}_{\ge M/2}^{\le N} = \{ \xi \in \mathcal{F}_+^{\le N} : \xi = \chi (\mathcal{N}_+ \ge M/2) \xi \}\) is the subspace of \(\mathcal{F}_+^{\le N}\) where states with at least M/2 excitations are described (recall that \(M = N^{1-\varepsilon }\)). To prove (66) it is enough to show that there exists \(C > 0\) with
for all N large enough. On the other hand, using the definitions of \(\mathcal{G}_{N,\alpha }\) in (42), \(\mathcal{R}_{N,\alpha }\) and \(\mathcal{R}^{\text {eff}}_{N,\alpha }\) in (47), we obtain that the ground state energy \(E_N\) of the system is given by
with \(\mathcal{E}_{L} = \mathcal{E}_{\mathcal{R}} +e^{-A} \mathcal{E}_{\mathcal{G}}e^A\). The bounds (43) and (48), together with Propositions 2 and 3, imply that for any \(\alpha \ge 5/2\) there exists \(C>0\) such that
With (64) we obtain
and therefore, with \(\mathcal{N}_+\le N\)
From the result (3) of [13, 14, 16]
as \(N \rightarrow \infty \). If we assume by contradiction that (67) does not hold true, then we can find a subsequence \(N_j \rightarrow \infty \) with
as \(j \rightarrow \infty \) (here we used the notation \(M_j = N_j^{1-\varepsilon }\)). This implies that there exists a sequence \({\tilde{\xi }}_{N_j} \in \mathcal{F}^{\le N_j}_{ \ge M_j /2}\) with \(\Vert {\tilde{\xi }}_{N_j} \Vert = 1\) for all \(j \in {\mathbb {N}}\) such that
On the other hand, using the relation \(\mathcal{R}^{\text {eff}}_{N_j,\alpha } = e^{-A}\mathcal{G}_{N_j,\alpha } e^A - \mathcal{E}_{L,j}\) with \(\mathcal{E}_{L,j}\) satisfying the bound (68) (with \(\mathcal{N}_+ \le N_j\)), we obtain that there exist constants \(c_1, c_2, C>0\) such that
Hence for \(\xi _{N_j} = e^{A} {\tilde{\xi }}_{N_j}\) we have
Let now \(S:= \{N_j: j\in {\mathbb {N}}\} \subset {\mathbb {N}}\) and denote by \(\xi _N\) a normalized minimizer of \(\mathcal{G}_{N,\alpha }\) for all \(N\in {\mathbb {N}}\setminus S\). Setting \(\psi _N = U_N^* e^{B} \xi _N\), for all \(N \in {\mathbb {N}}\), we obtain that \(\Vert \psi _N \Vert = 1\) and that
Eq. (69) shows that the sequence \(\psi _N\) is an approximate ground state of \(H_N\). From (5), we conclude that \(\psi _N\) exhibits complete Bose–Einstein condensation in the zero-momentum mode \(\varphi _0\), and in particular that there exists \({{\bar{\delta }}} >0\) such that
Using Lemma 2, Proposition 2 and the rules (11), we observe that
as \(N \rightarrow \infty \).
On the other hand, for \(N \in S = \{ N_j : j \in {\mathbb {N}}\}\), we have \(\xi _N = \chi (\mathcal{N}_+ \ge M/2) \xi _N\) and therefore
Choosing \(\varepsilon < {\bar{\delta }}\) and N large enough we get a contradiction with (70). This proves (67), (66) and therefore also
Inserting (65) and (71) on the r.h.s. of (62), we obtain that
for N large enough. With (64), (72) implies
To conclude, we use the relation \(e^{-A}\mathcal{G}_{N,\alpha }e^A=\mathcal{R}^{\text {eff}}_{N,\alpha } + \mathcal{E}_L\) and the bound (68). We have that for \(\alpha \ge 5/2\) there exist \(c, C>0\) such that
where we used (72) and Proposition 2. \(\square \)
We are now ready to show our main theorem.
Proof of Theorem 1
Let \(E_N\) be the ground state energy of \(H_N\). Evaluating (42) and (43) on the vacuum \(\varOmega \in \mathcal{F}^{\le N}_+\) and using (40), we obtain the upper bound
Notice that we cannot reach the expected optimal upper bound \(E_N \le 2 \pi N + C\) because of the logarithmic correction in \({\hat{\omega }}_N (0)\) (see (40)). In the lower bound, this logarithmic factor is compensated by the contribution arising from the off-diagonal quadratic term, extracted starting from (54). To obtain the same term for the upper bound, we would have to modify our trial state (diagonalizing the quadratic terms in \(\mathcal{R}_{N,\alpha }\)); this, however, would produce even larger contributions arising from the potential energy.
With Eq. (61) we also find the lower bound \(E_N \ge 2\pi N - C \). This proves (6).
Let now \(\psi _N \in L^2_s (\varLambda ^N)\) with \(\Vert \psi _N \Vert =1\) and
We define the excitation vector \(\xi _N = e^{-B} U_N \psi _N\). Then \(\Vert \xi _N \Vert = 1\) and, recalling that \(\mathcal{G}_{N,\alpha } = e^{-B} U_N H_N U_N^* e^{B}\) we have, with (61),
From Eqs. (73) and (74) we conclude that
If \(\gamma _N\) denotes the one-particle reduced density matrix associated with \(\psi _N\), using Lemma 2 we obtain
which concludes the proof of (8). \(\square \)
6 Analysis of the Excitation Hamiltonian \(\mathcal{R}_{N} \)
In this section, we show Proposition 4, where we establish a lower bound for the operator \(\mathcal{R}_{N,\alpha } = e^{-A} \mathcal{G}_{N,\alpha }^\text {eff} e^A\), with \(\mathcal{G}^\text {eff}_{N,\alpha }\) as defined in (42) and with
We decompose
with \(\mathcal{K}\) and \(\mathcal{V}_N\) as in (38), and with
We will analyze the conjugation of all terms on the r.h.s. of (76) in Sects. 6.2–6.6. The estimates emerging from these subsections will then be combined in Sect. 6.6 to conclude the proof of Proposition 4. Throughout the section, we will need Proposition 3 to control the growth of the expectation of the energy \(\mathcal{H}_N = \mathcal{K}+ \mathcal{V}_N\) under the action of (75); the proof of Proposition 3 is contained in Sect. 6.1.
In this section, we will always assume that \(V \in L^3 ({\mathbb {R}}^2)\) is compactly supported, pointwise non-negative and spherically symmetric.
6.1 A Priori Bounds on the Energy
In this section, we show Proposition 3. To this end, we will need the following proposition.
Proposition 7
Let \(\mathcal{V}_N\) and A be defined in (38) and (44) respectively. Then, there exists a constant \(C > 0\) such that
where
for any \(\alpha > 0\), for all \(\xi \in \mathcal{F}^{\le N}_+\), and \(N \in {\mathbb {N}}\) large enough.
Proof
We proceed as in [4, Prop. 8.1], computing \([ a_{p+u}^* a_{q}^* a_{p}a_{q+u}, b^*_{r+v} a^*_{-r}a_{v}]\). We obtain
with
and with \(\sum ^*\) running over all momenta, except choices for which the argument of a creation or annihilation operator vanishes. We conclude that \(\delta _{\mathcal{V}_N} = \varTheta _1 + \varTheta _2 + \varTheta _3 + \text{ h.c. }\). Next, we show that each error term \(\varTheta _j\), with \(j=1,2,3\), satisfies (78). To bound \(\varTheta _1\) we switch to position space and apply Cauchy–Schwarz. We find
for any \(\xi \in \mathcal{F}_+^{\le N}\) The term \(\varTheta _3\) can be controlled similarly. We find
It remains to bound the term \(\varTheta _2\) on the r.h.s. of (79). Passing to position space we obtain, by Cauchy–Schwarz,
To bound the term in the square bracket, we write it in first quantized form and, for any \(2< q < \infty \), we apply Hölder inequality and the Sobolev inequality \(\Vert u \Vert _{q} \le C \sqrt{q} \, \Vert u \Vert _{H^1}\) to estimate (denoting by \(1< q' < 2\) the dual index to q),
we conclude that
for any \(2< q < \infty \), if \(1/q + 1/q' = 1\). Choosing \(q = \log N\), we obtain that
\(\square \)
Using Proposition 7, we can now show Proposition 3.
Proof of Proposition 3
The proof follows a strategy similar to [4, Lemma 8.2]. For fixed \(\xi \in \mathcal{F}_+^{\le N}\) and \(s\in [0; 1]\), we define
We compute
With Proposition 7, we have
with \(\delta _{\mathcal{V}_N}\) satisfying (78). Switching to position space and using Proposition 2 we find , using (25) to bound \(\Vert {\check{\eta }} \Vert _\infty \le C N\),
Together with (78) we conclude that for any \(\alpha > 1/2\)
if N is large enough. Next, we analyze the first term on the r.h.s. of (81). We compute
With (31), we write
The contribution of \(\text {T}_{11}\) can be estimated similarly as in (82); switching to position space and using (20), we obtain
for any \(\xi \in \mathcal{F}^{\le N}_+\). The second term in (85) can be controlled using that for any \(\xi \in \mathcal{F}^{\le N}_+\) and \(2 \le q < \infty \) we have
Hence, choosing \(q=\log N\),
With (86) and (88) we conclude that
for all \(\xi \in \mathcal{F}_+^{\le N}\). As for the second term on the r.h.s. of (84) we have
for any \(\xi \in \mathcal{F}^{\le N}_+\). With (89) and Proposition 2, we conclude that
Combining with Eq. (83) we obtain
With Proposition 2 we obtain the differential inequality
By Gronwall’s Lemma, we find (45). \(\square \)
6.2 Analysis of \(e^{-A} \mathcal{O}_N e^{A}\)
In this section we study the contribution to \(\mathcal{R}_{N,\alpha }\) arising from the operator \(\mathcal{O}_N\), defined in (77). To this end, it is convenient to use the following lemma.
Lemma 3
Let A be defined in (44). Then, there exists a constant \(C > 0\) such that
where
for all \(\alpha > 0\), \(\xi _1, \xi _2 \in \mathcal{F}_+^{\le N}\), \(F \in \ell ^{\infty } (\varLambda ^*_+)\), and \(N \in {\mathbb {N}}\) large enough.
Proof
The lemma is analogous to [4, Lemma 8.6]. We estimate
where we used Proposition 2. \(\square \)
We consider now the action of \(e^A\) on the operator \(\mathcal{O}_N\), as defined in (77).
Proposition 8
Let A be defined in (44). Then there exists a constant \(C>0\) such that
where
for all \(\alpha > 0\), and \(N \in {\mathbb {N}}\) large enough.
Proof
The proof is very similar to [4, Prop. 8.7]. First of all, with Lemma 3 we can bound
Moreover, for the contribution quadratic in \(\mathcal{N}_+\), we can decompose
with \(\xi _1 = e^{-A} \mathcal{N}_+ e^A \xi \) and \(\xi _2 = \mathcal{N}_+ \xi \), and estimate, again with Lemma 3,
With Proposition 2, we have \(\Vert (\mathcal{N}_+ + 1)^{1/2} \xi _1 \Vert \le C \Vert (\mathcal{N}_+ +1)^{3/2} \xi \Vert \). \(\square \)
6.3 Contributions from \(e^{-A} \mathcal{K}e^{A}\)
In Sect. 6.6 we will analyse the contributions to \(\mathcal{R}_{N,\alpha }\) arising from conjugation of the kinetic energy operator \(\mathcal{K}= \sum _{p \in \varLambda _+^*} p^2 a_p^* a_p\). To this aim we will exploit properties of the commutator \([\mathcal{K}, A]\), collected in the following proposition.
Proposition 9
Let A be defined as in (44) and \({\widehat{\omega }}_N(r)\) be defined in (39). Then there exists a constant \(C>0\) such that
where
for all \(\alpha >1\), \(\xi \in \mathcal{F}_+^{\le N}\), and \(N \in {\mathbb {N}}\) large enough. Moreover, the operator
satisfies
for all \(\alpha > 1\), \(\xi \in \mathcal{F}_+^{\le N}\), and \(N \in {\mathbb {N}}\) large enough.
Proof
To show (91) we recall from Eqs. (84), (85) that
with \(\text {T}_2\) satisfying (90). Using the definition \({\widehat{\omega }}_{N}(p)=2 N e^{2N} \lambda _\ell \widehat{\chi }_\ell (p)\) we write
Hence, \(\delta _K = T_2 + T_{122}\). To bound \(T_{122}\) we switch to position space:
To bound the term in the parenthesis, we proceed similarly as in (80). We find
for any \(q > 2\) and \(1< q' < 2\) with \(1/q+ 1/q' =1\). Choosing \(q = \log N\), we obtain
Let us now focus on (92). We have
With the commutators from the proof of Proposition 8.8 in [4], we arrive at
where
To conclude the proof of Proposition 9, we show that all operators in (93) satisfy (92). To study all these terms it is convenient to switch to position space. We recall that \({{\widehat{\omega }}}_N(p)= g_N {{\widehat{\chi }}}(\ell p)\) with \(|g_N|\le C\) and \(\ell = N^{-\alpha }\). Using (87) we find:
The expectation of \(\varUpsilon _2\) is bounded following the same strategy used to show (87). For any \(2\le q < \infty \) we have
where in the last line we chose \(q=\log N\). The term \(\varUpsilon _3\) is of lower order; using that \(\big | \sum _r {\widehat{\omega }}_N (r) \eta _r \big | \le \Vert {\widehat{\chi }} (./N^\alpha ) \Vert _2 \Vert \eta \Vert _2 \le C\) and Cauchy–Schwarz, we easily obtain
The term \(\varUpsilon _4\) can be estimated as \(\varUpsilon _1\) using (87):
The term \(\varUpsilon _5\) is bounded similarly to \(\varUpsilon _2\); with \(q=\log N\) we have
The terms \(\varUpsilon _6\) and \(\varUpsilon _7\) are of smaller order and can be bounded with Cauchy–Schwarz; we have
and
The terms \(\varUpsilon _8, \varUpsilon _{11}, \varUpsilon _{12}\) are again bounded, as \(\varUpsilon _1\), using (87). We find
It remains to bound \(\varUpsilon _9\) and \(\varUpsilon _{10}\). The term \(\varUpsilon _9\) is bounded analogously to \(\varUpsilon _2\):
As for \(\varUpsilon _{10}\), we find
Proceeding as in (80), we obtain
for any \(q > 2\), and \(q' < 2\) with \(1/q + 1/q' = 1\). Since, for an arbitrary \(q' < 2\), \(\Vert {\check{\eta }} \Vert _{q'} \le \Vert {\check{\eta }} \Vert _2 = \Vert \eta \Vert _2 \le N^{-\alpha }\), we obtain
We conclude that for any \(\alpha >1\)
\(\square \)
6.4 Analysis of \(e^{-A} \mathcal{Z}_N e^{A}\)
In this subsection, we consider contributions to \(\mathcal{R}_{N,\alpha }\) arising from conjugation of \(\mathcal{Z}_{N}\), as defined in (77).
Proposition 10
Let A be defined in (44). Then, there exists a constant \(C>0\) such that
where
for all \(\alpha > 0\), and \(N \in {\mathbb {N}}\) large enough.
Proof
We have
We compute
With (95) we write
with
To bound the first term, we observe, with (52),
The term \(\varPi _3\) can be bounded similarly to \(\varPi _1\), with (52). We find
With \(|{\widehat{\omega }}_N (r)| \le C\), we similarly obtain
Finally, we estimate, using again (52),
With (94), we conclude that
With Proposition 2, Lemma 3, we conclude that
\(\square \)
6.5 Contributions from \(e^{-A}\mathcal{C}_N e^{A}\)
In Sect. 6.6 we will analyse the contributions to \(\mathcal{R}_{N,\alpha }\) arising from conjugation of the cubic operator \(\mathcal{C}_N\) defined in (77). To this aim we will need some properties of the commutator \([\mathcal{C}_N, A]\), as established in the following proposition.
Proposition 11
Let A be defined in (44). Then, there exists a constant \(C>0\) such that
where
for all \(\alpha > 0\), \(\xi \in \mathcal{F}^{\le N}_+\), and \(N \in {\mathbb {N}}\) large enough.
Proof
We consider the commutator
As in the proof of Proposition 9, we use the commutators from the proof of Proposition 8.8 in [4] to conclude that
where
To prove the proposition, we have to show that all terms \(\varXi _j\), \(j=1,\ldots , 12\), satisfy the bound (96). We bound \(\varXi _1\) in position space, with Cauchy–Schwarz, by
We can proceed similarly to control \(\varXi _9\). We obtain
The expectations of the terms \(\varXi _3\) and \(\varXi _{12}\) can be bounded analogously:
As for \(\varXi _4\), we find
The terms \(\varXi _5\) and \(\varXi _6 \) can be bounded in momentum space, using (154). Hence,
Similarly we have
Next, we rewrite \( \varXi _7\), \(\varXi _8\) and \(\varXi _{11}\) as
Thus, we obtain
as well as
and
Collecting all the bounds above, we arrive at (96). \(\square \)
6.6 Proof of Proposition 4
With the results of Sects. 6.1–6.5, we can now show Proposition 4. We assume \(\alpha > 2\). From Eq. (76), Propositions 8 and 10 we obtain that
with
From Propositions 7, 9 and 11, we can write, for N large enough,
where
for all \(\xi \in \mathcal{F}^{\le N}_+\). From Proposions 2, 3 and recalling the definition (77) of the operator \(\mathcal{C}_N\), we deduce that
with
for \(N \in {\mathbb {N}}\) sufficiently large.
We now rewrite
With Lemma 1, part (iii) we get
and therefore, using Lemma 3 and (99)
On the other hand it is easy to check that \(e^{-sA} \text {Q}_2 e^{sA}\) is an error term; to this aim we notice that
Hence with Props. 2 and 3 we find
To handle the second term on the second line of (97), we apply Proposition 9 and then Propositions 2 and 3
As for the first term on the second line of (97), we use again Proposition 11. Using (98), (100) and (101) we have
with \(\pm \mathcal{E}_{\mathcal{R}}^{(4)} \le C N^{2-\alpha } (\mathcal{H}_N+1) + C N^{-1} (\mathcal{N}_++1)\).
Inserting the bounds (100), (101), (102) and (103) into (97) we arrive at
with
for \(N \in {\mathbb {N}}\) sufficiently large.
References
Andersen, J.O.: Ground state pressure and energy density of an interacting homogeneous Bose gas in two dimensions. Eur. Phys. J. B 28, 389 (2002)
Boccato, C., Brennecke, C., Cenatiempo, S., Schlein, B.: Complete Bose–Einstein condensation in the Gross–Pitaevskii regime. Commun. Math. Phys. 359(3), 975–1026 (2018)
Boccato, C., Brennecke, C., Cenatiempo, S., Schlein, B.: Bogoliubov Theory in the Gross–Pitaevskii limit. Acta Math. 222(2), 219–335 (2019)
Boccato, C., Brennecke, C., Cenatiempo, S., Schlein, B.: Optimal rate for Bose–Einstein condensation in the Gross–Pitaevskii regime. Commun. Math. Phys. 376, 1311–1395 (2020)
Boßmann, L.: Derivation of the 2d Gross–Pitaevskii equation for strongly confined 3d bosons. Arch. Ration. Mech. Anal. 238(2), 541–606 (2020)
Brennecke, C., Schlein, B.: Gross–Pitaevskii dynamics for Bose–Einstein condensates. Anal. PDE 12(6), 1513–1596 (2019)
Brennecke, C., Schlein, B., Schraven, S.: Bose–Einstein Condensation with Optimal Rate for Trapped Bosons in the Gross–Pitaevskii Regime. Preprint arXiv:2102.11052 (2021)
Erdős, L., Schlein, B., Yau, H.-T.: Derivation of the Gross–Pitaevskii hierarchy for the dynamics of Bose–Einstein condensate. Commun. Pure Appl. Math. 59(12), 1659–1741 (2006)
Fournais, S., Napiórkowski, M., Reuvers, R., Solovej, J.P.: Ground state energy of a dilute two-dimensional Bose gas from the Bogoliubov free energy functional. J. Math. Phys. 60, 071903 (2019)
Jeblick, M., Leopold, N., Pickl, P.: Derivation of the time dependent Gross–Pitaevskii equation in two dimensions. Commun. Math. Phys. 372(1), 1–69 (2019)
Lewin, M., Nam, P.T., Serfaty, S., Solovej, J.P.: Bogoliubov spectrum of interacting Bose gases. Commun. Pure Appl. Math. 68(3), 413–471 (2014)
Lieb, E.H., Loss, M.: Analysis. AMS (2001)
Lieb, E.H., Seiringer, R.: Proof of Bose–Einstein condensation for dilute trapped gases. Phys. Rev. Lett. 88, 170409 (2002)
Lieb, E.H., Seiringer, R.: Derivation of the Gross–Pitaevskii equation for rotating Bose gases. Commun. Math. Phys. 264, 505–537 (2006)
Lieb, E.H., Yngvason, J.: The ground state energy of a dilute two-dimensional Bose gas. J. Stat. Phys. 103, 509–526 (2001)
Lieb, E.H., Seiringer, R., Yngvason, J.: A rigorous derivation of the Gross–Pitaevskii energy functional for a two-dimensional bose gas. Commun. Math. Phys. 224(1), 17–31 (2001)
Lieb, E.H., Seiringer, R., Solovej, J.P., Yngvason, J.: The Mathematics of the Bose Gas and Its Condensation. Oberwolfach Seminars, Birkhäuser, Basel (2005)
Mora, C., Castin, Y.: Ground state energy of the two-dimensional weakly interacting Bose gas: first correction beyond Bogoliubov theory. Phys. Rev. Lett. 102, 180404 (2009)
Nam, P.T., Napiórkowski, M., Ricaud, J., Triay, A.: Optimal rate of condensation for trapped bosons in the Gross–Pitaevskii regime. To appear in Analysis & PDE
Pilati, S., Boronat, J., Casulleras, J., Giorgini, S.: Quantum Monte Carlo simulation of a two-dimensional Bose gas. Phys. Rev. A 71, 023605 (2005)
Schick, M.: Two-dimensional system of hard-core bosons. Phys. Rev. A 3, 1067 (1971)
Schnee, K., Yngvason, J.: Bosons in disc-shaped traps: from 3D to 2D. Commun. Math. Phys. 269(3), 659–691 (2007)
Seiringer, R.: The excitation spectrum for weakly interacting bosons. Commun. Math. Phys. 306, 565–578 (2011)
Simon, B.: Schrödinger semigroups. Bull. Am. Math. Soc. 7(3), 447–526 (1982)
Acknowledgements
We are thankful to A. Olgiati for discussions on the two dimensional scattering equation.
Funding
Open Access funding provided by University Zurich.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Alessandro Giuliani.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
C.C. and S.C. gratefully acknowledge the support from the GNFM Gruppo Nazionale per la Fisica Matematica - INDAM through the project “Derivation of effective theories for large quantum systems”. B. S. gratefully acknowledges partial support from the NCCR SwissMAP, from the Swiss National Science Foundation through the Grant “Dynamical and energetic properties of Bose–Einstein condensates” and from the European Research Council through the ERC-AdG CLaQS.
Appendices
Appendix A: Analysis of \( \mathcal{G}_{N,\alpha }\)
The aim of this section is to show Proposition 1. From (12) and (37), we can decompose
with
To analyse \(\mathcal{G}_{N,\alpha }\) we will need precise informations on the action of the generalized Bogoliubov transformation \(e^{B}\) with B the antisymmetric operator defined in (33), which are summarized in Sect. 1. Then, in the Sects. 1–1 we prove separate bounds for the operators \(\mathcal{G}_{N,\alpha }^{(j)}\), \(j=0,2,3,4\), which we combine in Sect. 1 to prove Proposition 1.
The analysis in this section follows closely that of [4, Sect. 7] with some slight modifications due to the different scaling of the interaction potential and the fact that the kernel \(\eta _p\) of \(e^{B}\) is different from zero for all \(p \in \varLambda ^*_+\) (in [4] \(\eta _p\) is different from zero only for momenta larger than a sufficiently large cutoff of order one). Moreover, while in three dimensions it was sufficient to choose the function \(\eta _p\) appearing in the generalized Bogoliubov transformation with \(\Vert \eta \Vert \) sufficiently small but of order one, we need here \(\Vert \eta \Vert \) to be of order \(N^{-\alpha }\) for some \(\alpha >0\) large enough. As discussed in the introduction this is achieved by considering the Neumann problem for the scattering equation in (16) on a ball of radius \(\ell =N^{-\alpha }\); as a consequence some terms depending on \(\ell \) will be large, compared to the analogous terms in [4].
1.1 Appendix A.1: Generalized Bogoliubov Transformations
In this subsection we collect important properties about the action of unitary operators of the form \(e^{B}\), as defined in (34). As shown in [2, Lemma 2.5 and 2.6], we have, if \(\Vert \eta \Vert \) is sufficiently small,
where the series converge absolutely. To confirm the expectation that generalized Bogoliubov transformation act similarly to standard Bogoliubov transformations, on states with few excitations, we define (for \(\Vert \eta \Vert \) small enough) the remainder operators
where \(q \in \varLambda ^*_+\), \( (\sharp _m, \alpha _m) = (\cdot , +1)\) if m is even and \((\sharp _m, \alpha _m) = (*, -1)\) if m is odd. It follows then from (104) that
where we introduced the notation \(\gamma _q = \cosh (\eta _q)\) and \(\sigma _q = \sinh (\eta _q)\). It will also be useful to introduce remainder operators in position space. For \(x \in \varLambda \), we define the operator valued distributions \({\check{d}}_x, {\check{d}}^*_x\) through
where \({\check{\gamma }}_x (y) = \sum _{q \in \varLambda ^*} \cosh (\eta _q) e^{iq \cdot (x-y)}\) and \({\check{\sigma }}_x (y) = \sum _{q \in \varLambda ^*} \sinh (\eta _q) e^{iq \cdot (x-y)}\). The next lemma is taken from [4, Lemma 3.4].
Lemma 4
Let \(\eta \in \ell ^2 (\varLambda _+^*)\), \(n \in {\mathbb {Z}}\). For \(p \in \varLambda _+^*\), let \(d_p\) be defined as in (106). If \(\Vert \eta \Vert \) is small enough, there exists \(C > 0\) such that
for all \(p \in \varLambda ^*_+, \xi \in \mathcal{F}_+^{\le N}\). In position space, with \({\check{d}}_x\) defined as in (107), we find
Furthermore, letting \(\check{\overline{d}}_x = {\check{d}}_x + (\mathcal{N}_+ / N) b^*({\check{\eta }}_x)\), we find
and, finally,
for all \(\xi \in \mathcal{F}^{\le n}_+\).
A first simple application of Lemma 4 is the following bound on the growth of the expectation of \(\mathcal{N}_+\).
Lemma 5
Assume B is defined as in (33), with \(\eta \in \ell ^2 (\varLambda ^*)\) and \(\eta _p = \eta _{-p}\) for all \(p \in \varLambda ^*_+\). Then, there exists a constant \(C > 0\) such that
for all \(\xi \in \mathcal{F}_+^{\le N}\).
Proof
With (106) we write
with \(\gamma _p^{(s)} = \cosh (s \eta _p)\), \(\sigma _p^{(s)} = \sinh (s \eta _p)\). Using \(|\gamma ^{(s)}_p| \le C\) and \(|\sigma _p^{(s)}| \le C |\eta _p|\), (108) in Lemma 4 we arrive at
\(\square \)
1.2 Appendix A.2: Analysis of \( \mathcal{G}_{N,\alpha }^{(0)}=e^{-B}\mathcal{L}^{(0)}_N e^{B}\)
We define \(\mathcal{E}_{N}^{(0)}\) so that
where we recall from (13) that
Proposition 12
Under the assumptions of Proposition 1, there exists a constant \(C > 0\) such that
for all \(\alpha > 0\) and \(N \in {\mathbb {N}}\) large enough.
Proof
The proof follows [4, Prop. 7.1].
We write
Hence,
To bound the first term we use (106), \(|\gamma _q^2 - 1| \le C \eta _q^2\), \(|\sigma _q| \le C |\eta _q|\), the first bound in (108), Cauchy–Schwarz and the estimate \(\Vert \eta \Vert \le C N^{-\alpha }\). To bound the second term, we use Lemma 5. We conclude that
\(\square \)
1.3 Appendix A.3: Analysis of \(\mathcal{G}_{N,\alpha }^{(2)}=e^{-B}\mathcal{L}^{(2)}_N e^{B}\)
We consider first conjugation of the kinetic energy operator.
Proposition 13
Under the assumptions of Proposition 1, there exists \(C > 0\) such that
where
for any \(\alpha >1\), \(\xi \in \mathcal{F}^{\le N}_+\) and \(N \in {\mathbb {N}}\) large enough.
Proof
We proceed as in the proof of [4, Prop. 7.2]. We write
with \(\gamma _p^{(s)} = \cosh (s \eta _p)\), \(\sigma ^{(s)}_p =\sinh (s \eta _p)\) and where \(d^{(s)}_p\) is defined as in (105), with \(\eta _p\) replaced by \(s \eta _p\). We find
with
Since \(|\big ((\gamma _p^{(s)})^2-1\big )|\le C \eta _p^2\), \((\sigma ^{(s)}_p)^2\le C \eta _p^2\), \(p^2 |\eta _p| \le C \), \(\Vert \eta \Vert _\infty \le N^{-\alpha }\), we can estimate
for any \(\xi \in \mathcal{F}_+^{\le N}\). To bound the term \(\text {G}_3\) in (114), we switch to position space:
With (111), we obtain
Finally, we consider \(\text {G}_2\) in (114). We split it as \(\text {G}_2 = \text {G}_{21} + \text {G}_{22} + \text {G}_{23} + \text {G}_{24}\), with
We consider \(\text {G}_{21}\) first. We write
where \(\mathcal{E}_{2}^K = \sum _{j=1}^3 \mathcal{E}_{2j}^K\), with
and where we introduced the notation \(\overline{d}^{(s)}_{-p} = d_{-p}^{(s)} + s \eta _p (\mathcal{N}_+ / N) b_p^*\). With (29), we find
Using \(|\gamma _p^{(s)} - 1| \le C \eta _p^2\) and (108), we obtain
To control the third term in (118), we use (30) and we switch to position space. We find
With (110) and \(|{{\check{\eta }}}(x-y)| \le C N\), we obtain
As for \(\mathcal{E}_{232}^K\), with (110) and Lemma 1 (recalling \(\ell = N^{-\alpha })\), we find
To bound the last term on the r.h.s. of (123) we use Hölder’s and Sobolev inequality \(\Vert u \Vert _q \le C q^{1/2} \Vert u \Vert _{H^1}\), valid for any \(2 \le q <\infty \). We find
Choosing \(q=\log N\), we get
Therefore, for any \(\xi \in \mathcal{F}^{\le N}_+\),
Combining the last bound with (119), (120) and (122), we conclude that
for any \(\alpha >1\), \(N \in {\mathbb {N}}\) large enough, \(\xi \in \mathcal{F}^{\le N}_+\).
The term \(\text {G}_{22}\) in (117) can be bounded using (108). We find
We split \(\text {G}_{23} = \mathcal{E}_{31}^K + \mathcal{E}_{32}^K + \text{ h.c. }\), with
With (108), we find
To estimate \(\mathcal{E}_{32}^K\), we use (30) and we switch to position space. Proceeding as we did in (121), (122), (123), we obtain
Combining the bounds for \(\mathcal{E}_{31}^K \) and \(\mathcal{E}_{32}^K \) , we conclude that, if \(\alpha > 1\),
To bound \(\text {G}_{24}\) in (117), we use (108), the bounds (28) and \(\Vert \eta \Vert ^2_{H_1} \le C N\), and the commutator (14):
Together with (117), (125), (126) and (127), this implies that
with
Combining (115), (116) and (128), we obtain (112) and (113). \(\square \)
In the next proposition, we consider the conjugation of the operator
Proposition 14
Under the assumptions of Proposition 1, there is a constant \(C > 0\) such that
where
for any \(\alpha > 1\), \(\xi \in \mathcal{F}^{\le N}_+\) and \(N \in {\mathbb {N}}\) large enough.
Proof
We write
With (106), we find
where \(\gamma _p = \cosh \eta _p\), \(\sigma _p = \sinh \eta _p\) and the operators \(d_p\) are defined in (105). Using \(|1-\gamma _p | \le \eta _p^2\), \(|\sigma _p| \le C |\eta _p|\) and using Lemma 4 for the terms on the second line, we find
with \(\pm \mathcal{E}_1^V \le C N^{1-\alpha } (\mathcal{N}_+ + 1)\).
Let us now consider the second contribution on the r.h.s. of (131). We find
with
With Lemma 2, we easily find \(\pm \mathcal{E}_2^V \le C N^{-\alpha } (\mathcal{N}_+ + 1)\).
Finally, we consider the last term on the r.h.s. of (131). With (106), we obtain
Using \(|1-\gamma _p| \le C \eta _p^2\), \(|\sigma _p| \le C |\eta _p|\), we obtain
with \(\pm \mathcal{E}_3^V \le C N^{1-\alpha }(\mathcal{N}_++1)\). As for \(\text {F}_{32}\) in (134), we divide it into four parts
We start with \(\text {F}_{321}\), which we write as
where \(\mathcal{E}_4^V = \mathcal{E}_{41}^V + \mathcal{E}_{42}^V + \mathcal{E}_{43}^V + \text{ h.c. }\), with
and with the notation \(\overline{d}_{-p} = d_{-p} + N^{-1} \eta _p \, \mathcal{N}_+ b_p^*\). Since \(|\gamma _p - 1| \le C \eta _p^2\), \(\Vert \eta \Vert _\infty \le C N^{-\alpha }\), we find easily with (108) that
Moreover
As for \(\mathcal{E}_{42}^V\), we switch to position space and we use (110). We obtain
We conclude that
To bound the term \(\text {F}_{322}\) in (136), we use (108) and \(|\sigma _p|\le C|\eta _p|\); we obtain
Let us now consider the term \(\text {F}_{323}\) on the r.h.s. of (136). We write \(\text {F}_{323} = \mathcal{E}_{51}^V + \mathcal{E}_{52}^V + \text{ h.c. }\), with
With \(|\gamma _p - 1| \le C \eta _p^2 \) and (108) we obtain
We find, switching to position space and using (109),
Hence,
To estimate the term \(\text {F}_{324}\) in (136) we use (108) and the bound
We find
Combining the last bounds, we arrive at
with
To control the last contribution \(\text {F}_{33}\) in (134), we switch to position space. With (111) and (25) we obtain
The last equation, combined with (134), (135) and (137), implies that
with
Together with (132) and with (133), and recalling that \(b_p^* b_p -N^{-1} a_p^* a_p = a_p^* a_p (1- \mathcal{N}_+ / N)\), we obtain (129) with (130). \(\square \)
1.4 Appendix A.4: Analysis of \( \mathcal{G}_{N,\alpha }^{(3)}=e^{-B}\mathcal{L}^{(3)}_N e^{B}\)
We consider here the conjugation of the cubic term \(\mathcal{L}_N^{(3)}\), defined in (13).
Proposition 15
Under the assumptions of Proposition 1, there exists a constant \(C > 0\) such that
where
for any \(\alpha > 1\) and \(N \in {\mathbb {N}}\) large enough.
Proof
This proof is similar to the proof of [4, Prop. 7.5]. Expanding \(e^{-B} a_{-p}^* a_q e^B\), we arrive at
where, as usual, \(\gamma _p = \cosh \eta (p)\), \(\sigma _p = \sinh \eta (p)\) and \(d_p\) is as in (105). We consider \(\mathcal{E}_1^{(3)}\). To this end, we write
Since \(|\gamma _{p+q}-1|\le |\eta _{p+q}|^2\) and \(\Vert \eta \Vert \le C N^{-\alpha }\), we find
As for \(\mathcal{E}^{(3)}_{12}\), we commute \(a^*_{-p}\) through \(b_{-p-q}\) (recall \(q \not = 0\)). With \(|\sigma _{p+q}| \le C |\eta _{p+q}|\), we obtain
We decompose now \(\mathcal{E}^{(3)}_{13} = \mathcal{E}^{(3)}_{131} + \mathcal{E}^{(3)}_{132}\), with
where we defined \(d^*_{p+q}= \overline{d}^*_{p+q} - \frac{(\mathcal{N}_++1)}{N}\, \eta _{p+q} b_{-p-q}\). The term \(\mathcal{E}^{(3)}_{132}\) is estimated similarly to \(\mathcal{E}_{12}^{(3)}\), moving \(a_{-p}^*\) to the left of \(b_{-p-q}\); we find \(\pm \mathcal{E}^{(3)}_{132} \le C N^{1-\alpha } (\mathcal{N}_+ + 1)\). We bound \(\mathcal{E}^{(3)}_{131}\) in position space. We find
With (140) and (141) we obtain
Next, we focus on \(\mathcal{E}^{(3)}_2\), defined in (139). With Eq. (106), we find
with \(\gamma ^{(s)}_p = \cosh (s \eta _p)\), \(\sigma ^{(s)}_p = \sinh (s \eta _p)\) and \(d^{(s)}_p\) defined as in (105), with \(\eta \) replaced by \(s \eta \). With Lemma 2, we get
Since \([b_{p},b^*_{-q}] = - a^*_{-q} a_{p} /N \) for \(p \ne -q\), we find
As for the third term on the r.h.s. of (143), we switch to position space. We find
Using the bounds (109), (110), (111) and Lemma 2 we arrive at
where \({\check{r}}\) indicates the function in \(L^2 (\varLambda )\) with Fourier coefficients \(r_p =1-\gamma _p\), and the fact that \( \Vert {\check{\eta }}\Vert , \Vert {\check{r}}\Vert , \Vert {\check{\sigma }}\Vert \le C N^{-\alpha }\). Combined with (144) and (145), the last bound implies that
To bound the last contribution on the r.h.s. of (139), it is convenient to bound (in absolute value) the expectation of its adjoint
Since \(q \ne 0\), \([b_{-p},b^*_{-p-q}] = - a^*_{-p-q} a_{-p} /N \). Thus, we can estimate
To bound the expectation of \(\mathcal{E}^{(3)}_{31}\), we switch to position space. We find
With Lemma 2, we conclude that
From (147) and (148) we obtain
Together with (139), (142) and (146), we arrive at (138). \(\square \)
1.5 Appendix A.5: Analysis of \( \mathcal{G}_{N,\alpha }^{(4)}=e^{-B}\mathcal{L}^{(4)}_N e^{B}\)
Finally, we consider the conjugation of the quartic term \(\mathcal{L}_N^{(4)}\). We define the error operator \(\mathcal{E}_N^{(4)}\) through
Proposition 16
Under the assumptions of Proposition 1 there exists a constant \(C > 0\) such that
for any \(\alpha > 1\), \(\xi \in \mathcal{F}^{\le N}_+\) and \(N \in {\mathbb {N}}\) large enough.
To show Proposition 16, we use the following lemma, whose proof can be obtained as in [4, Lemma 7.7].
Lemma 6
Let \(\eta \in \ell ^2 (\varLambda ^*)\) as defined in (27). Then there exists a constant \(C > 0\) such that
for all \(\xi \in \mathcal{F}_+^{\le N}\), \(n \in {\mathbb {Z}}\).
Proof of Proposition 16
We follow the proof of [4, Prop. 7.6]. We write
with
Let us first consider the term \(\text {W}_1\). With (106), we find
where
with
With
uniformly in \(N \in {\mathbb {N}}\), we can estimate the first term in (153) by
Using (154) and (108) we also find
For the third term in (153) we switch to position space and use (109):
Consider now the fourth term in (153). We write \(\mathcal{E}_{104}^{(4)} = \mathcal{E}_{1041}^{(4)} + \mathcal{E}_{1042}^{(4)}\), with
With \(|\gamma _q^{(s)} - 1| \le C |\eta _q|^2\), (154) and \(\Vert d^*_q \xi \Vert \le C \Vert \eta \Vert \Vert (\mathcal{N}_++1)^{1/2}\xi \Vert \), we find
As for \(\mathcal{E}_{1042}^{(4)}\), we switch to position space. Using (25) and (109), we obtain
Let us consider the last term in (153). Switching to position space and using (111) in Lemma 4 and again (25), we arrive at
Summarizing, we have shown that (152) can be bounded by
for any \(\alpha > 1\), \(\xi \in \mathcal{F}^{\le N}_+\). Next, we come back to the terms \(\text {W}_{11}, \text {W}_{12}, \text {W}_{13}\) introduced in (151). Using (154) and \(|\gamma _q^{(s)} -1| \le C \eta _q^2\), we can write
where \(\mathcal{E}_{11}^{(4)}\) is such that
Next, we can decompose the second term in (151) as
where \(\pm \mathcal{E}^{(4)}_{12} \le C N^{-\alpha } \mathcal{N}_+ + N^{1-3\alpha }\).
The third term on the r.h.s. of (151) can be written as
where \(\mathcal{E}^{(4)}_{13} = \mathcal{E}_{131}^{(4)} + \mathcal{E}_{132}^{(4)} + \mathcal{E}_{133}^{(4)} + \mathcal{E}_{134}^{(4)}\), with
With (154), we immediately find
With \(|\gamma _q^{(s)} -1| \le C \eta _q^2\), Lemma 4 and, again, (154), we also obtain
Let us now consider \(\mathcal{E}_{132}^{(4)}\). In position space, with \(\check{\overline{d}}^{(s)}_y = d^{(s)}_y + (\mathcal{N}_+ / N) b^* ({\check{\eta }}_{y})\) and using (110), we obtain
It follows that
With (155), (156), (157), (158), we obtain
where
Next, we control the term \(\text {W}_2\), from (150). In position space, we find
with \({\check{\eta }}_{x} (z) = {\check{\eta }} (x-z)\). By Cauchy–Schwarz, we have
With
and using Lemma 6, we obtain
Also for the term \(\text {W}_3\) in (150), we switch to position space. We find
and thus
With Lemma 2, we find
Using Lemma 6, we conclude that
The term \(\text {W}_4\) in (150) can be bounded similarly. In position space, we find
with \(\check{\eta ^2}\) the function with Fourier coefficients \(\eta ^2_q\), for \(q \in \varLambda ^*\), and where \(\check{\eta ^2_x}(y) := \check{\eta ^2}(x-y)\). Clearly \(\Vert \check{\eta ^2_{x}} \Vert \le C \Vert {\check{\eta }}\Vert ^2 \le C N^{-2\alpha }\). With Cauchy–Schwarz and Lemma 2, we obtain
Applying Lemma 6 and then Lemma 2, we obtain
From (159), (160), (161) and the last bound, we conclude that
where \(\mathcal{E}_{N,\alpha }^{(4)}\) satisfies (149). \(\square \)
1.6 Appendix A.6: Proof of Proposition 1
With the results established in Sects. 1–1, we cam now show Proposition 1. Propositions 12, 13, 14, 15, 16, imply that
where
for any \(\alpha >1\) and \(\xi \in \mathcal{F}^{\le N}_+\). With (31), we find
From Lemma 1 and estimating \(\Vert \widehat{\chi }_\ell \Vert = \Vert \chi _\ell \Vert \le C N^{-\alpha }\), \(\Vert \eta \Vert \le C N^{-\alpha }\) and \(\Vert {\widehat{\chi }}_\ell * \eta \Vert = \Vert \chi _\ell {\check{\eta }} \Vert \le \Vert {\check{\eta }} \Vert \le C N^{-\alpha }\), we have
and
Moreover, using (154) and the bound (32) we find
We obtain
with \(\pm \mathcal{E}_2 \le C \) for all \(\alpha \ge 1/2\). On the other hand, using (32) we have
with \(\pm {\tilde{\mathcal{E}}}_2 \le C N^{1-2\alpha }\). With the first bound in (41) we conclude that
where \(\pm \mathcal{E}_3 \le C\), if \(\alpha \ge 1/2\). Using (31), we can also handle the fourth line of (162); we find
The last two terms on the right hand side of (164) are error terms. With (32) and (154) we have
The second term on the right hand side of (164) can be bounded in position space:
The term in parenthesis can be bounded similarly as in (80). Namely,
for any \(q >2\) and \(1< q' < 2\) with \(1/q +1/q' =1\). Choosing \(q =\log N\), we get
and, from (164), we conclude that
with
if \(\alpha > 1\). Combining (162) with (163) and (165), and using the definition (39) we conclude that
with
for any \(\alpha > 1\). Observing that \(|\widehat{V} (p/e^N) - \widehat{V} (0)| \le C |p| e^{-N}\) in the second line on the r.h.s. of (166), we arrive at \(\mathcal{G}_{N,\alpha } = \mathcal{G}^\text {eff}_{N,\alpha } + \mathcal{E}_{\mathcal{G}}\), with \(\mathcal{G}^\text {eff}_{N,\alpha }\) defined as in (42) and with \(\mathcal{E}_{\mathcal{G}}\) that satisfies (43).
Appendix B: Properties of the Scattering Function
Let V be a potential with finite range \(R_0>0\) and scattering length \({{\mathfrak {a}}}\). For a fixed \(R>R_0\), we study properties of the ground state \(f_R\) of the Neumann problem
on the ball \(|x|\le R\), normalized so that \(f_{R}(x)=1\) for \(|x|=R\). Lemma 1, parts (i)–(iv), follows by setting \(R=e^N \ell \) in the following lemma.
Lemma 7
Let \(V\in L^3({\mathbb {R}}^2)\) be non-negative, compactly supported and spherically symmetric, and denote its scattering length by \({{\mathfrak {a}}}\). Fix \(R>0\) sufficiently large and denote by \(f_{R}\) the Neumann ground state of (167). Set \(w_{R}=1 - f_{R}\). Then we have
Moreover, for R large enough there is a constant \(C>0\) independent of R such that
and
Finally, there exists a constant \(C>0\) such that
for R large enough.
To show Lemma 7 we adapt to the two dimensional case the strategy used in [8, Lemma A.1] for the three dimensional problem. We will use some well known properties of the zero energy scattering equation in two dimensions, summarized in the following lemma.
Lemma 8
Let \(V \in L^3 ({\mathbb {R}}^2)\) non-negative, with \(\text {supp } V \subset B_{R_0} (0)\) for an \(R_0 > 0\). Let \({{\mathfrak {a}}}\le R_0\) denote the scattering length of V. For \(R > R_0\), let \(\phi _R : {\mathbb {R}}^2 \rightarrow {\mathbb {R}}\) be the radial solution of the zero energy scattering equation
normalized such that \(\phi _R (x) = 1\) for \(|x| = R\). Then
for all \(|x| > R_0\). Moreover, \(|x| \rightarrow \phi _R (x)\) is monotonically increasing and there exists a constant \(C > 0\) (depending only on V) such that
for all \(x \in {\mathbb {R}}^2\). Furthermore, there exists a constant \(C > 0\) such that
for all \(x \in {\mathbb {R}}^2\).
Proof
The existence of the solution of (171), the expression (172), the fact that \(\phi _R (x) \ge 0\) and the monotonicity are standard (see, for example, Theorem C.1 and Lemma C.2 in [17]). The bound (173) for \(\phi _R (0)\) follows from (172), comparing \(\phi _R (0)\) with \(\phi _R (x)\) at \(|x| = R_0\), with Harnack’s inequality (see [24, Theorem C.1.3]). Finally, (174) follows by rewriting (171) in integral form
For \(|x| \le R_0\), this leads (using that \(\phi _R (y) \le \log (R_0/{{\mathfrak {a}}}) / \log (R/{{\mathfrak {a}}})\) for all \(|y| \le R_0\) and the local integrability of \(|.|^{-3/2}\)) to
Combining with the bound for \(|x| > R_0\) obtained differentiating (172), we obtain the desired estimate. \(\square \)
Proof of Lemma 7
By standard arguments (see for example [17, proof of theorem C1]), \(f_R(x)\) is spherically symmetric and non-negative. We now start by proving an upper bound for \(\lambda _R\), consistent with (168). To this end, we calculate the energy of a suitable trial function. For \(k \in {\mathbb {R}}\) we define
with \(J_0\) and \(Y_0\) the zero Bessel functions of first and second type, respectively. Note that
and \(\psi _k (x)=0\) if \(|x| =a\). We define \(k= k(R)\) to be the smallest positive real number satisfying \(\partial _r \psi _R (x)=0\) for \(|x| = R\). One can check that
in the limit \(R\rightarrow \infty \). To prove (175), we observe that
and we expand for \(kR, k{{\mathfrak {a}}}\ll 1\) using (with \(\gamma \) the Euler constant)
With (177) one finds that (176)
The smallest solution of
is such that
in the limit of large R. Inserting in (178), we find that the r.h.s. changes sign around the value of k defined in (179). By the intermediate value theorem, we conclude that there is a \(k = k (R) > 0\) satisfying (175), such that \(\partial _r \psi _{k(R)} (x) = 0\) if \(|x| = R\).
Now, let \(\phi _R (x)\) be the solution of the zero energy scattering equation (171), with \(\phi _R(x)=1\) for \(|x| = R\). We set
with \(k = k(R)\) satisfying (175) and
With this choice we have \(m_R(x)=|x|\) outside the range of the potential; hence \(\varPsi _R (x)=\psi _{k} (x)\) for \(R_0 \le |x| \le R\). In particular, \(\varPsi _R\) satisfies Neumann boundary conditions at \(|x|=R\).
From (172), (173) and the monotonicity of \(\phi _R\), we get
and for a constant \(C > 1\), independent of R. From (174) we also get
With the notation \({{\mathfrak {h}}}= -\varDelta + \frac{1}{2} V \), we now evaluate \({\bigl \langle {\varPsi _R, {{\mathfrak {h}}} \varPsi _R}\bigr \rangle }\). To this end we note that
Let us consider the region \(|x| < R_0\). From (180) and (177) we find, first of all,
Next, we compute \(-\varDelta \varPsi _R (x)\). With
we obtain (here, we use the notation \(m'_R\) and \(m''_R\) for the radial derivatives of the radial function \(m_R\))
We note that, using the scattering equation (171),
Now we write
where \(g_R(x)= \sum _{i=1}^3 g_R^{(i)}(x)\) with
With (185), (177) and (181), (182), we find
Next, with \( | J_2(r) - r^2/8 | \le C r^4 \) we get
With (184), we can also bound
We conclude that \( |g_R(r)| \le C (1+V(x)) k^2 \) for all \(r \le R_0\) and R large enough. Finally, using Eq. (185), the expansion for \(Y_1(r)\) in Eq. (177), and the bound
we can rewrite the first term on the r.h.s. of (186) as
with \(|h_R(x)|\le C (1+V(x)) k^2\) for all \(r \le R_0\), R large enough. With the identities (186) and (187) we obtain
for all \(|x| \le R_0\) and for R sufficiently large. With (184), we conclude that, for \(0 \le |x| \le R_0\),
With (183), (188) and the upper bound
for all \(|x| \le R_0\) (which follows from (184) and (181)), we get
On the other hand, Eq.(184), together with \(m_R(x)=|x|\) for \(|x| \ge R_0\), implies the lower bound
Hence, with (175), we conclude that
in agreement with (168).
To prove the lower bound for \(\lambda _R\) it is convenient to show some upper and lower bounds for \(f_R\). We start by considering \(f_R\) outside the range of the potential. We denote \(\varepsilon _R= \sqrt{\lambda _R}\, R\). Keeping into account the boundary conditions at \(|x| = R\), we find, for \(R_0 \le |x| \le R\),
with
and
From (190), we have \(|\varepsilon _R|\le C\, |\log (R/{{\mathfrak {a}}})|^{-1/2}\). Thus, we can expand \(f_R\) for large R, using (177) and, for \(Y_0\), the improved bound
we find
which leads to
We can also compute the radial derivative
With the expansions (177) and (191) we conclude that for all \(R_0 \le |x| < R\) we have
The bound (193) shows that \(\partial _r f_R (x)\) is positive, for, say, \(R_0< |x| < R/2\). Since \(\partial _r f_R (x)\) must have its first zero at \(|x| = R\), we conclude that \(f_R\) is increasing in |x|, on \(R_0 \le |x| \le R\). From the normalization \(f_R (x) = 1\), for \(|x| = R\), we conclude therefore that \(f_R (x) \le 1\), for all \(R_0 \le |x| \le R\).
From (192) and (190) we obtain, on the other hand, the lower bound
for R sufficiently large. Let \(R_*=\max \{R_0, e {{\mathfrak {a}}}\}\). Then Eq. (194) implies in particular that, for R large enough,
for all \(R_* < |x| \le R\).
Finally, we show that \(f_R (x) \le 1\) also for \(|x| \le R_0\). First of all, we observe that, by elliptic regularity, as stated for example in [12, Theorem 11.7, part iv)], there exists \(0<\alpha <1\) and \(C > 0\) such that
With \(\Vert V f_R \Vert _2 \le \Vert V \Vert _3 \Vert f_R \Vert _6 \le C \Vert f_R \Vert _{H^1} \le C (1 + \lambda _R ) \Vert f_R \Vert _2\), we conclude that \(0 \le f_R (x) \le 1 + C \Vert f \Vert _2\) for all \(|x| \le R_0\) (because we know that \(f_R (x) \le 1\) for \(R_0 \le |x| \le R\)). To improve this bound, we go back to the differential equation (167), to estimate
This implies that \(f_R (x) + \lambda _R (1+C \Vert f \Vert _2) x^2 /2\) is subharmonic. Using (192), we find \(f_R (x) \le 1 - C \varepsilon _R^2\) for \(|x| = R_0\). From the maximum principle, we obtain therefore that
for all \(|x| \le R_0\). In particular, this implies that \(\Vert f_R \mathbf{1}_{|x| \le R_0} \Vert _2 \le C + C \lambda _R \Vert f_R \Vert _2\), and therefore that
With \(f_R (x) \le 1\) for \(R_0 \le |x| \le R\), we find, on the other hand, that \(\Vert f_R \mathbf{1}_{R_0 \le |x| \le R} \Vert _2 \le C R\). We conclude therefore that \(\Vert f_R \Vert _2 \le CR\) and, from (197), that \(f_R (x) \le 1 - C \varepsilon _R^2 + C/R \le 1\), for all \(|x| \le R_0\), if R is large enough.
We are now ready to prove the lower bound for \(\lambda _R\). We use now that any function \(\varPhi \) satisfying Neumann boundary conditions at \(|x|=R\) can be written as \(\varPhi (x)=q(x)\varPsi _R(x)\), with \(\varPsi _R(x)\) the trial function used for the upper bound and \(q>0\) a function that satisfies Neumann boundary condition at \(|x|=R\) as well. This is in particular true for the solution \(f_R(x)\) of (167). In the following we write
where \(q_R\) satisfies Neumann boundary conditions at \(|x|=R\). From (184), we find \(|\varPsi _R (x)| \ge C / \log (ka)\). The bound \(f_R(x)\le 1\) implies therefore that there exists \(c>0\) such that
From the identity
we have
Hence
With (198), we obtain
With (195) (recalling that \(R_*=\max \{R_0, e {{\mathfrak {a}}}\}\)), we bound
and, inserting in (199), we conclude that
where in the last inequality we used (175).
To prove (169) we use the scattering equation (167) to write
Passing to polar coordinates, and using that \(\varDelta f_R (x) = |x|^{-1} \partial _r |x| \partial _r f_R (x)\), we find that the first term vanishes. Hence
With the upper bound \(f_R(r) \le 1\) and with (168), we find
To obtain a lower bound for the same integral we use that \( f_{R}(r) \ge 0\) inside the range of the potential. Outside the range of V, we use (192). We find
We conclude that
Finally, we show the bounds in (170). For \(r\in [R_0, R]\), from (192) we have
As for the derivative of \(w_R\) we use (193) to compute
Moreover \(\partial _r f_R (x)= 0\) if \(|x| = R\), by construction.
On the other hand, if \(|x| \le R_0\), we have \(w_R (x)= 1- f_R(x)\le 1\). As for the derivative, we define \(\widetilde{f}_R\) on \({\mathbb {R}}_+\) through \(\widetilde{f}_R (r) = f_R (x)\), if \(|x| = r\), and we use the representation
With (167), we have (with \(\widetilde{V}\) defined on \({\mathbb {R}}_+\) through \(V(x) = \widetilde{V}(r)\), if \(|x| = r\))
By (200), we can estimate \(\widetilde{f}_R (R_0) \le C / \log (R/{{\mathfrak {a}}})\). From (196), we also recall that
for any \(r < R_0\). We conclude therefore that
\(\square \)
Rights and permissions
This article is published under an open access license. Please check the 'Copyright Information' section either on this page or in the PDF for details of this license and what re-use is permitted. If your intended use exceeds what is permitted by the license or if you are unable to locate the licence and re-use information, please contact the Rights and Permissions team.
About this article
Cite this article
Caraci, C., Cenatiempo, S. & Schlein, B. Bose–Einstein Condensation for Two Dimensional Bosons in the Gross–Pitaevskii Regime. J Stat Phys 183, 39 (2021). https://doi.org/10.1007/s10955-021-02766-6
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10955-021-02766-6