OctoInformBeta

OpenAI preprint, September 2026 · §1. Introduction · §1

Theorem 1.1 [Sharp one-dimensional Lieb-Thirring constants]: For one-dimensional Schrödinger forms and operators, optimal and one-state constants equal the sharp constant, with equality at a sech⁡2\operatorname{sech}^2 potential

The sharp Lieb-Thirring bound holds for exponents in (1/2,3/2)(1/2, 3/2)..

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

Consider a one-dimensional Schrödinger operator −d2dx2−W-\frac{d^2}{dx^2} - W on ℝ\mathbb{R} with a spectral exponent γ∈(1/2,3/2)\gamma \in (1/2, 3/2) and an almost everywhere nonnegative potential WW in the Lebesgue space Lγ+1/2(ℝ)L^{\gamma + 1/2}(\mathbb{R}). The question is what sharp constant bounds the γ\gamma-th negative eigenvalue moment in terms of ∫ℝW(x)γ+1/2dx\int_{\mathbb{R}} W(x)^{\gamma + 1/2}\,dx, and whether this constant agrees with the single-state constant and is attained by an explicit potential. One also asks whether the same sharp bound holds for self-adjoint operator realizations as well as in the weak variational setting.

Admissible Potentials and Sharp Constants

A γ\gamma-admissible potential W:ℝ→ℝW : \mathbb{R} \to \mathbb{R} is a real-valued function on ℝ\mathbb{R} that is almost everywhere nonnegative and belongs to Lγ+1/2(ℝ)L^{\gamma + 1/2}(\mathbb{R}).✦ For a function W:ℝ→ℝW : \mathbb{R} \to \mathbb{R}, the positive part W+:ℝ→ℝW_+ : \mathbb{R} \to \mathbb{R} is the function W+(x)=max⁡(W(x),0)W_+(x) = \max(W(x), 0), and the potential mass Mγ(W)∈[0,∞]M_\gamma(W) \in [0, \infty] is the extended integral Mγ(W)=∫ℝW+(x)γ+1/2dx.M_\gamma(W) = \int_{\mathbb{R}} W_+(x)^{\gamma + 1/2}\,dx.(1)✦

For γ>1/2\gamma > 1/2, the one-dimensional semiclassical Lieb-Thirring constant Lγcl∈ℝL_\gamma^{\mathrm{cl}} \in \mathbb{R} is given in terms of the Gamma function Γ\Gamma by Lγcl=Γ(γ+1)2πΓ(γ+3/2).L_\gamma^{\mathrm{cl}} = \frac{\Gamma(\gamma + 1)}{2\sqrt{\pi}\,\Gamma(\gamma + 3/2)}.(2)✦ For γ>1/2\gamma > 1/2, the sharp constant Lγ∈ℝL_\gamma \in \mathbb{R} is defined from (2) by Lγ=2(γ−1/2γ+1/2)γ−1/2Lγcl.L_\gamma = 2\left(\frac{\gamma - 1/2}{\gamma + 1/2}\right)^{\gamma - 1/2} L_\gamma^{\mathrm{cl}}.(3)✦ For γ>1/2\gamma > 1/2 and r=(γ−1/2)−1r = (\gamma - 1/2)^{-1}, the extremizer potential W*:ℝ→ℝW_* : \mathbb{R} \to \mathbb{R} is the function W*(x)=(r+1)sech⁡2(rx).W_*(x) = (r + 1)\operatorname{sech}^2(rx).(4)✦

The Variational Formulation

Let L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}) denote the Hilbert space of complex-valued square-integrable functions on ℝ\mathbb{R}, with inner product ⟨f,g⟩L2=∫ℝf(x)¯g(x)dx\langle f, g \rangle_{L^2} = \int_{\mathbb{R}} \overline{f(x)}\,g(x)\,dx(5)✦ and norm ∥f∥L2=⟨f,f⟩L21/2\|f\|_{L^2} = \langle f, f \rangle_{L^2}^{1/2}. The Sobolev space H1(ℝ;ℂ)H^1(\mathbb{R};\mathbb{C}) consists of pairs (u,u′)(u, u') of functions in L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}) such that for every smooth compactly supported test function φ∈Cc∞(ℝ;ℂ)\varphi \in C_c^\infty(\mathbb{R};\mathbb{C}), ∫ℝu(x)φ′(x)dx=−∫ℝu′(x)φ(x)dx.\int_{\mathbb{R}} u(x)\,\varphi'(x)\,dx = -\int_{\mathbb{R}} u'(x)\,\varphi(x)\,dx.(6)✦ Given a potential W:ℝ→ℝW : \mathbb{R} \to \mathbb{R}, the Schrödinger quadratic form qWq_W on H1(ℝ;ℂ)H^1(\mathbb{R};\mathbb{C}) is the sesquilinear pairing qW(u,v)=∫ℝu′(x)¯v′(x)dx−∫ℝW(x)u(x)¯v(x)dx.q_W(u, v) = \int_{\mathbb{R}} \overline{u'(x)}\,v'(x)\,dx - \int_{\mathbb{R}} W(x)\,\overline{u(x)}\,v(x)\,dx.(7)✦

Given W:ℝ→ℝW : \mathbb{R} \to \mathbb{R} and k>0k > 0, a weak eigenfunction u∈H1(ℝ;ℂ)u \in H^1(\mathbb{R};\mathbb{C}) of qWq_W at eigenvalue −k2-k^2 is an element of H1(ℝ;ℂ)H^1(\mathbb{R};\mathbb{C}) satisfying, in terms of (5) and (7), qW(v,u)=−k2⟨v,u⟩L2for every v∈H1(ℝ;ℂ).q_W(v, u) = -k^2 \langle v, u \rangle_{L^2} \qquad \text{for every } v \in H^1(\mathbb{R};\mathbb{C}).(8)✦ A finite L2L^2-orthonormal family (u1,…,uN)(u_1, \dots, u_N) in H1(ℝ;ℂ)H^1(\mathbb{R};\mathbb{C}) is an NN-tuple of elements of H1(ℝ;ℂ)H^1(\mathbb{R};\mathbb{C}) satisfying ⟨ui,uj⟩L2=δij\langle u_i, u_j \rangle_{L^2} = \delta_{ij} for all 1≤i,j≤N1 \le i, j \le N, where δij\delta_{ij} is the Kronecker delta.✦

For γ∈ℝ\gamma \in \mathbb{R} and W:ℝ→ℝW : \mathbb{R} \to \mathbb{R}, the variational negative moment 𝒯γ(W)∈[0,∞]\mathcal{T}_\gamma(W) \in [0, \infty] is the supremum 𝒯γ(W)=sup⁡∑i=1Nki2γ,\mathcal{T}_\gamma(W) = \sup \sum_{i=1}^N k_i^{2\gamma},(9)✦ taken over all N≥0N \ge 0, all L2L^2-orthonormal families (u1,…,uN)(u_1, \dots, u_N) in H1(ℝ;ℂ)H^1(\mathbb{R};\mathbb{C}), and all k1,…,kN>0k_1, \dots, k_N > 0 such that each uiu_i is a weak eigenfunction of qWq_W at −ki2-k_i^2 as in (8). The variational one-state moment 𝒯γ(1)(W)∈[0,∞]\mathcal{T}_\gamma^{(1)}(W) \in [0, \infty] of a potential W:ℝ→ℝW : \mathbb{R} \to \mathbb{R} at exponent γ∈ℝ\gamma \in \mathbb{R} is the supremum 𝒯γ(1)(W)=sup⁡k2γ,\mathcal{T}_\gamma^{(1)}(W) = \sup k^{2\gamma},(10)✦ taken over all u∈H1(ℝ;ℂ)u \in H^1(\mathbb{R};\mathbb{C}) with ⟨u,u⟩L2=1\langle u, u \rangle_{L^2} = 1 and all k>0k > 0 such that uu is a weak eigenfunction of qWq_W at −k2-k^2 as in (8).

At exponent γ∈ℝ\gamma \in \mathbb{R}, the variational optimal constant Lγopt∈[0,∞]L_\gamma^{\mathrm{opt}} \in [0, \infty] is the supremum Lγopt=sup⁡𝒯γ(W)Mγ(W),L_\gamma^{\mathrm{opt}} = \sup \frac{\mathcal{T}_\gamma(W)}{M_\gamma(W)},(11)✦ taken over all γ\gamma-admissible potentials WW with 0<Mγ(W)0 < M_\gamma(W) in (1) and (9). The variational one-state constant Lγ(1)∈[0,∞]L_\gamma^{(1)} \in [0, \infty] at exponent γ∈ℝ\gamma \in \mathbb{R} is the supremum Lγ(1)=sup⁡𝒯γ(1)(W)Mγ(W),L_\gamma^{(1)} = \sup \frac{\mathcal{T}_\gamma^{(1)}(W)}{M_\gamma(W)},(12)✦ taken over all γ\gamma-admissible potentials WW with 0<Mγ(W)0 < M_\gamma(W) in (1) and (10).

Self-adjoint Operators and Spectral Moments

A linear operator AA on L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}) with domain dom⁡(A)⊆L2(ℝ;ℂ)\operatorname{dom}(A) \subseteq L^2(\mathbb{R};\mathbb{C}) is associated with the Schrödinger form qWq_W of a potential W:ℝ→ℝW : \mathbb{R} \to \mathbb{R} if for all f,g∈L2(ℝ;ℂ)f, g \in L^2(\mathbb{R};\mathbb{C}), one has f∈dom⁡(A)f \in \operatorname{dom}(A) and Af=gAf = g if and only if ff is represented by an element u∈H1(ℝ;ℂ)u \in H^1(\mathbb{R};\mathbb{C}) satisfying qW(v,u)=⟨v,g⟩L2for every v∈H1(ℝ;ℂ).q_W(v, u) = \langle v, g \rangle_{L^2} \qquad \text{for every } v \in H^1(\mathbb{R};\mathbb{C}).(13)✦ For a linear operator AA on L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}) and e∈ℝe \in \mathbb{R}, the eigenspace Ee(A)⊆L2(ℝ;ℂ)E_e(A) \subseteq L^2(\mathbb{R};\mathbb{C}) is the linear subspace Ee(A)={f∈dom⁡(A):Af=ef},E_e(A) = \{f \in \operatorname{dom}(A) : Af = ef\},(14)✦ whose elements are the operator eigenfunctions of AA at energy ee (called a negative energy when e<0e < 0).

The negative eigenvalue sum 𝒮γ(A)∈[0,∞]\mathcal{S}_\gamma(A) \in [0, \infty] of a self-adjoint operator AA on L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}) at exponent γ∈ℝ\gamma \in \mathbb{R} is the sum 𝒮γ(A)=∑e<0∑f∈Be|e|γ,\mathcal{S}_\gamma(A) = \sum_{e < 0} \sum_{f \in B_e} |e|^\gamma,(15)✦ formed by choosing an orthonormal Hilbert basis BeB_e of (14) for each negative energy e<0e < 0. For any linear operator AA on L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}) and γ∈ℝ\gamma \in \mathbb{R}, the operator one-eigenvalue moment 𝒮γ(1)(A)∈[0,∞]\mathcal{S}_\gamma^{(1)}(A) \in [0, \infty] is the supremum 𝒮γ(1)(A)=sup⁡|e|γ,\mathcal{S}_\gamma^{(1)}(A) = \sup |e|^\gamma,(16)✦ taken over all f∈dom⁡(A)f \in \operatorname{dom}(A) with ⟨f,f⟩L2=1\langle f, f \rangle_{L^2} = 1 and all e<0e < 0 such that Af=efAf = ef.

Inside the Hilbert direct sum L2(ℝ;ℂ)⊕L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}) \oplus L^2(\mathbb{R};\mathbb{C}) with inner product ⟨u,v⟩H1=⟨u,v⟩L2+⟨u′,v′⟩L2\langle u, v \rangle_{H^1} = \langle u, v \rangle_{L^2} + \langle u', v' \rangle_{L^2} and norm ∥u∥H1=⟨u,u⟩H11/2\|u\|_{H^1} = \langle u, u \rangle_{H^1}^{1/2}, the space H1(ℝ;ℂ)H^1(\mathbb{R};\mathbb{C}) is a closed subspace carrying bounded linear projections ι(u,u′)=u\iota(u, u') = u and ∂(u,u′)=u′\partial(u, u') = u' onto L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}).

For an almost everywhere nonnegative measurable potential W:ℝ→ℝW : \mathbb{R} \to \mathbb{R}, a potential decomposition of WW is a pair (b,V)(b, V) consisting of a constant b≥0b \ge 0 and an everywhere nonnegative integrable function V:ℝ→ℝV : \mathbb{R} \to \mathbb{R} with ∫ℝV(x)dx≤1/32\int_{\mathbb{R}} V(x)\,dx \le 1/32 such that |W(x)|≤b+V(x)|W(x)| \le b + V(x) for every x∈ℝx \in \mathbb{R}.✦

Given a potential decomposition (b,V)(b, V) of WW, the bounded linear weighted multiplication operator MW:H1(ℝ;ℂ)→L2(ℝ;ℂ)M_W : H^1(\mathbb{R};\mathbb{C}) \to L^2(\mathbb{R};\mathbb{C}) acts by (MWu)(x)=|W(x)|1/2u(x)(M_W u)(x) = |W(x)|^{1/2} u(x)✦, and the shifted form operator T:H1(ℝ;ℂ)→H1(ℝ;ℂ)T : H^1(\mathbb{R};\mathbb{C}) \to H^1(\mathbb{R};\mathbb{C}) is T=∂*∂−MW*MW+(b+1)ι*ι.T = \partial^*\partial - M_W^* M_W + (b + 1)\iota^*\iota.(17)✦ The operator TT in (17) has a continuous linear inverse T−1:H1(ℝ;ℂ)→H1(ℝ;ℂ)T^{-1} : H^1(\mathbb{R};\mathbb{C}) \to H^1(\mathbb{R};\mathbb{C}), giving the variational solution operator T−1ι*:L2(ℝ;ℂ)→H1(ℝ;ℂ)T^{-1}\iota^* : L^2(\mathbb{R};\mathbb{C}) \to H^1(\mathbb{R};\mathbb{C}) and the shifted resolvent R=ιT−1ι*:L2(ℝ;ℂ)→L2(ℝ;ℂ)R = \iota T^{-1}\iota^* : L^2(\mathbb{R};\mathbb{C}) \to L^2(\mathbb{R};\mathbb{C}). For an injective bounded linear operator R0R_0 on a complex Hilbert space EE and c∈ℝc \in \mathbb{R}, the shifted inverse R0−1−cIR_0^{-1} - cI is the linear operator on EE with domain R0(E)R_0(E) mapping R0fR_0 f to f−cR0ff - c R_0 f.✦ The Schrödinger operator HWH_W on L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}) for an exponent γ>1/2\gamma > 1/2 and a γ\gamma-admissible potential WW is the linear operator R−1−(b+1)IR^{-1} - (b + 1)I on L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}) obtained from the shifted resolvent RR of a potential decomposition (b,V)(b, V) of WW.

For γ>1/2\gamma > 1/2 and a γ\gamma-admissible potential WW, the operator spectral moment 𝒮γ(HW)∈[0,∞]\mathcal{S}_\gamma(H_W) \in [0, \infty] is the negative eigenvalue sum (15) of HWH_W.✦ At exponent γ>1/2\gamma > 1/2, the spectral optimal constant Lγ,opopt∈[0,∞]L_{\gamma,\mathrm{op}}^{\mathrm{opt}} \in [0, \infty] is the supremum Lγ,opopt=sup⁡𝒮γ(HW)Mγ(W),L_{\gamma,\mathrm{op}}^{\mathrm{opt}} = \sup \frac{\mathcal{S}_\gamma(H_W)}{M_\gamma(W)},(18)✦ taken over all γ\gamma-admissible potentials WW with 0<Mγ(W)0 < M_\gamma(W) in (1). The spectral one-state constant Lγ,op(1)∈[0,∞]L_{\gamma,\mathrm{op}}^{(1)} \in [0, \infty] for γ>1/2\gamma > 1/2 is the supremum Lγ,op(1)=sup⁡𝒮γ(1)(HW)Mγ(W),L_{\gamma,\mathrm{op}}^{(1)} = \sup \frac{\mathcal{S}_\gamma^{(1)}(H_W)}{M_\gamma(W)},(19)✦ taken over all γ\gamma-admissible potentials WW with 0<Mγ(W)0 < M_\gamma(W) in (1) and (16).

The Claim

Theorem 1. Suppose 1/2<γ<3/21/2 < \gamma < 3/2. In the variational formulation, 𝒯γ(W)≤LγMγ(W)for every γ-admissible W,Lγopt=Lγ,Lγ(1)=Lγ,W* is γ-admissible,0<Mγ(W*),𝒯γ(W*)=LγMγ(W*).\begin{aligned} &\mathcal{T}_\gamma(W) \le L_\gamma M_\gamma(W) \quad \text{for every $\gamma$-admissible } W, \\ &L_\gamma^{\mathrm{opt}} = L_\gamma, \qquad L_\gamma^{(1)} = L_\gamma, \\ &W_* \text{ is $\gamma$-admissible}, \quad 0 < M_\gamma(W_*), \quad \mathcal{T}_\gamma(W_*) = L_\gamma M_\gamma(W_*). \end{aligned}(20)✦ For 1/2<γ<3/21/2 < \gamma < 3/2, every γ\gamma-admissible potential WW satisfies 𝒯γ(W)≤LγMγ(W),\mathcal{T}_\gamma(W) \le L_\gamma M_\gamma(W),(21)✦ and the variational optimal constant satisfies Lγopt≤Lγ.L_\gamma^{\mathrm{opt}} \le L_\gamma.(22)✦ In the operator formulation for 1/2<γ<3/21/2 < \gamma < 3/2, 𝒮γ(HW)≤LγMγ(W)for every γ-admissible W,Lγ,opopt=Lγ,Lγ,op(1)=Lγ,W* is γ-admissible,0<Mγ(W*),𝒮γ(HW*)=LγMγ(W*).\begin{aligned} &\mathcal{S}_\gamma(H_W) \le L_\gamma M_\gamma(W) \quad \text{for every $\gamma$-admissible } W, \\ &L_{\gamma,\mathrm{op}}^{\mathrm{opt}} = L_\gamma, \qquad L_{\gamma,\mathrm{op}}^{(1)} = L_\gamma, \\ &W_* \text{ is $\gamma$-admissible}, \quad 0 < M_\gamma(W_*), \quad \mathcal{S}_\gamma(H_{W_*}) = L_\gamma M_\gamma(W_*). \end{aligned}(23)✦ If 1/2<γ<3/21/2 < \gamma < 3/2, WW is a γ\gamma-admissible potential, and AA is a self-adjoint operator on L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}) associated with qWq_W, then 𝒮γ(A)≤LγMγ(W)anddim⁡ℂEe(A)<∞ for every e<0.\mathcal{S}_\gamma(A) \le L_\gamma M_\gamma(W) \quad \text{and} \quad \dim_{\mathbb{C}} E_e(A) < \infty \text{ for every } e < 0.(24)✦ For every γ>1/2\gamma > 1/2, the variational one-state constant satisfies Lγ≤Lγ(1),L_\gamma \le L_\gamma^{(1)},(25)✦ the extremizer potential W*W_* satisfies LγMγ(W*)=1,L_\gamma M_\gamma(W_*) = 1,(26)✦ and its real Lebesgue integral satisfies Lγ∫ℝW*(x)γ+1/2dx=1.L_\gamma \int_{\mathbb{R}} W_*(x)^{\gamma + 1/2}\,dx = 1.(27)✦

In the paper, the result reads: for every 1/2<γ<3/21/2 < \gamma < 3/2 and every nonnegative W∈Lγ+1/2(ℝ)W \in L^{\gamma + 1/2}(\mathbb{R}), Tr⁡(H−W)−γ≤2((γ−1/2)/(γ+1/2))γ−1/2LγclMγ(W)\operatorname{Tr}(H_{-W})_-^\gamma \le 2((\gamma - 1/2)/(\gamma + 1/2))^{\gamma - 1/2} L_\gamma^{\mathrm{cl}} M_\gamma(W), the constant is optimal and equals Lγ,1(1)L_{\gamma,1}^{(1)}, and equality is attained by W*(x)=(r+1)sech⁡2(rx)W_*(x) = (r + 1)\operatorname{sech}^2(rx) with r=(γ−1/2)−1r = (\gamma - 1/2)^{-1} (here LγclL_\gamma^{\mathrm{cl}} stands for Lγ,1clL_{\gamma,1}^{\mathrm{cl}}, Mγ(W)M_\gamma(W) for ∫ℝWγ+1/2\int_{\mathbb{R}} W^{\gamma + 1/2}, W*W_* for the paper’s extremizing WW, 𝒯γ(W)\mathcal{T}_\gamma(W) and 𝒮γ(HW)\mathcal{S}_\gamma(H_W) for Tr⁡(H−W)−γ\operatorname{Tr}(H_{-W})_-^\gamma, and Lγ(1)L_\gamma^{(1)} and Lγ,op(1)L_{\gamma,\mathrm{op}}^{(1)} for Lγ,1(1)L_{\gamma,1}^{(1)}).

What Is Checked

The variational claims (20), (22), (25), (26), (21), and (27) are stated over real exponents γ\gamma (in (1/2,3/2)(1/2, 3/2) for (20), (22), and (21), and γ>1/2\gamma > 1/2 for (25), (26), and (27)), γ\gamma-admissible potentials WW, the Sobolev space H1(ℝ;ℂ)H^1(\mathbb{R};\mathbb{C}), weak eigenfunctions of qWq_W, the constants (3), (11), and (12), and the extremizer (4). The spectral claims (23) and (24) are stated over exponents γ∈(1/2,3/2)\gamma \in (1/2, 3/2), γ\gamma-admissible potentials WW, L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}), self-adjoint operators associated with qWq_W as in (13), the spectral constants (18) and (19), and the constructed Schrödinger operator HWH_W.

For those constructions, the library proves that for any potential decomposition (b,V)(b, V) of an almost everywhere nonnegative measurable potential WW and any u∈H1(ℝ;ℂ)u \in H^1(\mathbb{R};\mathbb{C}), the weighted product |W|1/2u|W|^{1/2} u belongs to L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C})✦ and satisfies ∥MWu∥L2≤b+1∥u∥H1\|M_W u\|_{L^2} \le \sqrt{b + 1}\,\|u\|_{H^1}✦, and that the shifted resolvent RR is injective✦. It also proves that for every γ>1/2\gamma > 1/2 and every γ\gamma-admissible potential WW, the operator HWH_W is self-adjoint✦, and that for any self-adjoint operator AA on L2(ℝ;ℂ)L^2(\mathbb{R};\mathbb{C}) and any e∈ℝe \in \mathbb{R}, the eigenspace Ee(A)E_e(A) admits an orthonormal Hilbert basis✦.