The weighted L 2 L^{2} -Caffarelli-Kohn-Nirenberg inequalities for the curl-free vector fields and second order derivatives: The sharp constants and stability estimates.

Anh Tuan Duong, Van Hoang Nguyen · 2026

Read The weighted L 2 L^{2} -Caffarelli-Kohn-Nirenberg inequalities for the curl-free vector fields and second order derivatives: The sharp constants and stability estimates. by Anh Tuan Duong, Van Hoang Nguyen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-25692.txt.

Introduction

Interpolation inequalities bound a function's middle behaviour by its extremes, and the Caffarelli Kohn Nirenberg family, weighted by powers of distance, anchors much of modern analysis. Anh Tuan Duong and Van Hoang Nguyen extend that family in 2026 for curl free vector fields and second derivatives, territory where earlier sharp results covered only part of the parameter range. Sharp means the best possible constant in the inequality is identified, not just estimated. Their proofs rest on splitting functions into spherical harmonics and sharpening one dimensional integral bounds.

Two stability results follow: when a field nearly achieves the best constant, it must sit quantifiably close to the extremal shapes that achieve it. One covers the curl free inequality proved by Cazacu, Flynn, and Lam; the other covers the new second order family itself. Stability turns a sharp inequality from a static bound into a measuring tool for near optimal fields.

Catalog record

CREATORAnh Tuan Duong, Van Hoang Nguyen
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE90 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.25692

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The weighted L 2 L^{2} -Caffarelli-Kohn-Nirenberg inequalities for the curl-free vector fields and second order derivatives: The sharp constants and stability estimates. Anh Tuan Duong, Van Hoang Nguyen Abstract In this paper, we study the weighted L 2 L^{2} -Caffarelli-Kohn-Nirenberg inequalities for curl-free vector fields and second order derivatives. Firstly, we prove a family of the sharp weighted second order L 2 L^{2} -Caffarelli-Kohn-Nirenberg inequalities that complements the results in [ C. Cazacu, J. Flynn and N. Lam, Calc. Var. Partial Differential Equations 62 (2023), no. 4, Paper No. 118, 26 pp. ] and [ A. T. Duong and V. H. Nguyen, On the sharp second order Caffarelli-Kohn-Nirenberg inequality. Ann. Fenn. Math., 50(1):275–286, 2025 ]. Secondly, we establish a stability version of the sharp weighted L 2 L^{2} -Caffarelli-Kohn-Nirenberg inequalities for curl-free vector fields proved by Cazacu, Flynn and Lam. Finally, we prove a stability estimate for the sharp weighted second order L 2 L^{2} -Caffarelli-Kohn-Nirenberg inequalities established in this paper. Our approach is based on the spherical harmonic decomposition method, the one dimensional integral inequalities and their improvements. † † Email: [email protected] ; [email protected] . † † 2010 Mathematics Subject Classification : 26D10, 46E35, 26D15 † † Key words and phrases : Caffarelli-Kohn-Nirenberg inequality, sharp constants, extremal functions, stability version, spherical harmonic decomposition. 1 Introduction Let us start by recalling the first order interpolation inequalities introduced by Caffarelli, Kohn and Nirenberg in 1stCKN known as the Caffarelli-Kohn-Nirenberg (CKN for short) inequalities. Theorem A. (CKN inequalities) Let n\geq 1 and let p,q,r,\alpha,\beta,\delta and \sigma be real numbers satisfying p,q\geq 1,\qquad r>0,\qquad\delta\in[0,1], and \frac{1}{p}+\frac{\alpha}{n}>0,\quad\frac{1}{q}+\frac{\beta}{n}>0,\quad\frac{1}{r}+\frac{\gamma}{n}>0, where \gamma=\delta\sigma+(1-\delta)\beta . Then there exists C>0 such that the inequality \Big(\int_{\mathbb{R}^{n}}|u|^{r}|x|^{r\gamma}dx\Big)^{\frac{1}{r}}\leq C\Big(\int_{\mathbb{R}^{n}}|\nabla u|^{p}|x|^{\alpha p}dx\Big)^{\frac{\delta}{p}}\Big(\int_{\mathbb{R}^{n}}|u|^{q}|x|^{\beta q}dx\Big)^{\frac{1-\delta}{q}} (1.1) holds for any function u\in C_{0}^{\infty}(\mathbb{R}^{n}) if and only if the following relations occur \frac{1}{r}+\frac{\gamma}{n}=\delta\Big(\frac{1}{p}+\frac{\alpha-1}{n}\Big)+(1-\delta)\Big(\frac{1}{q}+\frac{\beta}{n}\Big), \alpha-\sigma\geq 0\qquad\text{ if }\qquad a>0, and \alpha-\sigma\leq 1\qquad\text{ if }a>0\quad\text{ and }\quad\frac{1}{p}+\frac{\alpha-1}{n}=\frac{1}{r}+\frac{\gamma}{n}. Furthermore, on any compact set in the space of parameters, the constant C above is bounded. Interestingly, the CKN inequalities ( 1.1 ) contain many well-known and important inequalities in analysis such as Sobolev inequalities, Gagliardo-Nirenberg inequalities, Hardy inequalities, Hardy-Sobolev inequalities, Nash’s inequalities, Heisenberg-Pauli-Weyl uncertainty principle, etc. as special cases. The CKN inequalities play the important roles and have many applications in analysis and theory of partial differential equations (see, e.g, CKNappl ). They have been extensively studied by many mathematicians, especially on the questions about the sharp constant and the classification of extremal functions for the CKN inequalities. In this research direction, we can list some celebrated results of Aubin and Talenti Aubin ; Talenti , of Del Pino and Dolbeault DD1 ; DD2 , and of Carlen and Loss CarlenLoss for the sharp Sobolev inequalities, the sharp Gagliardo-Nirenberg inequalities, and the sharp Nash’s inequalities, respectively. The interest reader may consult the papers NguyenPLMS ; NguyenHC1 ; NguyenHC2 ; BrezisMironescu ; BrezisJeanYung ; Xia ; Ruzhansky1 ; Ruzhansky2 ; Ruzhansky3 ; CatrinaWang ; ChenLuZhang ; Dong ; DongLamLu ; Flynn ; FlynnLamLu ; LamLu for other results concerning the CKN type inequalities. In Costa ; CatrinaCosta , the following important subfamily of the sharp L^{2}- CKN inequalities was proved \int_{\mathbb{R}^{n}}\frac{|u|^{2}}{|x|^{2b}}dx\int_{\mathbb{R}^{n}}\frac{|\nabla u|^{2}}{|x|^{2a}}dx\geq C(n,a,b)^{2}\left(\int_{\mathbb{R}^{n}}\frac{|u|^{2}}{|x|^{a+b+1}}dx\right)^{2},\quad u\in C^{\infty}_{0}(\mathbb{R}^{n}\setminus\{0\}). (1.2) The sharp constant C(n,a,b) and the extremal functions in ( 1.2 ) are explicitly computed and depend on the parameter regions of a,b as follows \begin{cases}\mathcal{A}_{1}:=\{(a,b)\,|\,a+1>b,a\leq(n-2)/2\}\\ \mathcal{A}_{2}:=\{(a,b)\,|\,a+1<b,a\geq(n-2)/2\}\\ \mathcal{A}:=\mathcal{A}_{1}\cup\mathcal{A}_{2}\\ \mathcal{B}_{1}:=\{(a,b)\,|\,a+1<b,a\leq(n-2)/2\}\\ \mathcal{B}_{2}:=\{(a,b)\,|\,a+1>b,a\geq(n-2)/2\}\\ \mathcal{B}=\mathcal{B}_{1}\cup\mathcal{B}_{2}.\end{cases} More precisely, it was shown in CatrinaCosta that \bullet If (a,b)\in\mathcal{A} , then the sharp constant C(n,a,b)=|n-(a+b+1)|/2 and the extremal functions are given by u(x)=c\exp\Big(-\lambda\frac{|x|^{1+a-b}}{1+a-b}\Big), where c,\lambda are constants with \lambda(1+a-b)>0 . \bullet If (a,b)\in\mathcal{B} , then the sharp constant C(n,a,b)=|n-(3a-b+3)|/2 and the extremal functions are given by u(x)=c|x|^{2(a+1)-n}\exp\Big(-\lambda\frac{|x|^{1+a-b}}{1+a-b}\Big) where c,\lambda are constants with \lambda(1+a-b)>0 . In the case 1+a=b , the CKN inequality ( 1.2 ) becomes the weighted Hardy inequality with the sharp constant given by C(n,a,a+1)=|n-2a-2|/2 . In this case, the sharp constant C(n,a,a+1) is not attained by non-zero functions. The inequality ( 1.2 ) was first proved by Costa in Costa for a particular range of parameters by using the e xpanding-the-square method, and then was established by Catrina and Costa in CatrinaCosta for full range of parameters by using the spherical harmonic decomposition method and the Kelvin transformation. A simple and direct proof of ( 1.2 ) was given in CaFlyLa . Moreover, in CaFlyLa , the authors obtained a stronger version of ( 1.2 ) in which the full gradient \nabla u is replaced by the radial derivative \nabla u\cdot\frac{x}{|x|} . It is remarkable that in the case a=0 and b=-1 , the CKN inequality ( 1.2 ) is exactly the famous Heisenberg-Pauli-Weyl uncertainty principle (see Heisenberg ; Weyl ), which asserts that \int_{\mathbb{R}^{n}}|\nabla u|^{2}dx\int_{\mathbb{R}^{n}}|u|^{2}|x|^{2}dx\geq\frac{n^{2}}{4}\left(\int_{\mathbb{R}^{n}}|u|^{2}dx\right)^{2},\quad u\in C_{0}^{\infty}(\mathbb{R}^{n}). (1.3) The constant n^{2}/4 in ( 1.3 ) is sharp and is attained by Gaussian functions u(x)=ce^{-\lambda|x|^{2}} with c\in\mathbb{R} and \gamma>0 . Motivated by questions and applications in hydrodynamics and harmonic analysis, Maz’ya in ( Mazya , Section 3.9 ) raised the question on the sharp constant in ( 1.3 ) when the scalar function u is replaced by a divergence-free vector field U , here a vector field U=(U_{1},U_{2},\ldots,U_{n}):\mathbb{R}^{n}\to\mathbb{R}^{n} is called divergence-free if \text{\rm div}U=\partial_{1}U_{1}+\partial_{2}U_{2}+\cdots+\partial_{n}U_{n}=0. This question was solved by Cazacu, Flynn and Lam (see CFL22 ) in two dimensional case, i.e, n=2 . More precisely, in CFL22 the authors established the following inequality \int_{\mathbb{R}^{n}}|\Delta u|^{2}dx\int_{\mathbb{R}^{n}}|\nabla u|^{2}|x|^{2}dx\geq\frac{(n+2)^{2}}{4}\left(\int_{\mathbb{R}^{n}}|\nabla u|^{2}dx\right)^{2},\quad u\in C_{0}^{\infty}(\mathbb{R}^{n}). (1.4) The constant (n+2)^{2}/4 is sharp and is achieved by Gaussian functions u(x)=ce^{-\lambda|x|^{2}} where c\in\mathbb{R} and \lambda>0 . A vector field U is curl-free if U=\nabla u for some scalar function u:\mathbb{R}^{n}\to\mathbb{R} . Therefore, the inequality ( 1.4 ) is equivalent to following inequality \int_{\mathbb{R}^{n}}|\nabla U|^{2}dx\int_{\mathbb{R}^{n}}|U|^{2}|x|^{2}dx\geq\frac{(n+2)^{2}}{4}\left(\int_{\mathbb{R}^{n}}|U|^{2}dx\right)^{2} (1.5) for any curl-free vector field U\in(C_{0}^{\infty}(\mathbb{R}^{n}))^{n} . The constant (n+2)^{2}/4 is again the best possible and it is attained by the curl-free vector fields of the form U(x)=cxe^{-\lambda|x|^{2}} with c\in\mathbb{R} and \lambda>0 . The inequality ( 1.5 ) provides an uncertainty principle of Heisenberg-Pauli-Weyl type for curl-free vector fields. By an isometrically isomorphic between divergence-free vector fields and curl-free vector fields in \mathbb{R}^{2} (in fact, if U is a divergence-free vector field in \mathbb{R}^{2} then U(x)=(-\partial_{x_{2}}u,\partial_{x_{1}}u) for some scalar function u ), the inequality ( 1.5 ) is equivalent to the following inequality in \mathbb{R}^{2} \int_{\mathbb{R}^{2}}|\nabla U|^{2}dx\int_{\mathbb{R}^{2}}|U|^{2}|x|^{2}dx\geq 4\left(\int_{\mathbb{R}^{2}}|U|^{2}dx\right)^{2} (1.6) for any divergence-free vector fields U\in(C^{\infty}_{0}(\mathbb{R}^{2}))^{2} . The constant 4 is sharp and is attained by the vector fields of the form U(x)=c(-x_{2},x_{1})e^{-\lambda|x|^{2}} , with c\in\mathbb{R} and \lambda>0 . The inequality ( 1.6 ) answers affirmatively the question of Maz’ya in the case n=2 . Recently, this question was completely solved by Hamamoto HamaMaz (see also Hamasimple for a simpler proof). In fact, by using the poloidal-toroidal decomposition for divergence-free vector fields and an one-dimensional variational problem, he proved that \int_{\mathbb{R}^{n}}|\nabla U|^{2}dx\int_{\mathbb{R}^{n}}|U|^{2}|x|^{2}dx\geq\frac{1}{4}(\sqrt{n^{2}-4(n-3)}+2)^{2}\left(\int_{\mathbb{R}^{n}}|U|^{2}dx\right)^{2} (1.7) for any divergence-free vector field U\in(C_{0}^{\infty}(\mathbb{R}^{n}))^{n} . The constant (\sqrt{n^{2}-4(n-3)}+2)^{2}/4 is sharp in ( 1.7 ) and is achieved by some divergence-free vector fields. In CFL23 , Cazacu, Flynn and Lam extended the inequality ( 1.5 ) to the weighted case and proved a family of sharp CKN inequalities for curl-free vector fields. In order to state their results, let us define X_{a,b}(\mathbb{R}^{n}) , with a,b\in\mathbb{R} , to be the set of vector fields U\in C^{\infty}(\mathbb{R}^{n}\setminus\{0\})^{n} such that \int_{\mathbb{R}^{n}}\frac{|\nabla U|^{2}}{|x|^{2a}}dx<\infty,\quad\int_{\mathbb{R}^{n}}\frac{|U|^{2}}{|x|^{2b}}dx<\infty,\quad\int_{\mathbb{R}^{n}}\frac{|U|^{2}}{|x|^{a+b+1}}dx<\infty, \lim_{|x|\to 0,\infty}|x|^{-1+\frac{n}{2}-a}U(x)=0 and

Read the complete text