Higher-order non-local gradient theory of phase-transitions

Margherita Solci · 2024

Read Higher-order non-local gradient theory of phase-transitions by Margherita Solci (2024) 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-2411-01586.txt.

Description

Abstract We study the asymptotic behaviour of double-well energies perturbed by a higher-order fractional term, which, in the one-dimensional case, take the form 1 ε ​ ∫ I W ⁡ ( u ⁡ ( x ) ) ​ 𝑑 x + ε 2 ​ ( k + s ) − 1 ​ s ⁡ ( 1 − s ) 2

Catalog record

CREATORMargherita Solci
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE88 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2411-01586

Browse the full library · About the Tower of Babel

Full text

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

Higher-order non-local gradient theory of phase-transitions Margherita Solci Abstract We study the asymptotic behaviour of double-well energies perturbed by a higher-order fractional term, which, in the one-dimensional case, take the form 1 ε ​ ∫ I W ⁡ ( u ⁡ ( x ) ) ​ 𝑑 x + ε 2 ​ ( k + s ) − 1 ​ s ⁡ ( 1 − s ) 2 1 − s ​ ∫ I × I | u ( k ) ​ ( x ) − u ( k ) ​ ( y ) | 2 | x − y | 1 + 2 ​ s ​ 𝑑 x ​ 𝑑 y \frac{1}{\varepsilon}\int_{I}W(u(x))dx+\varepsilon^{2(k+s)-1}\frac{s(1-s)}{2^{1-s}}\int_{I\times I}\frac{|u^{(k)}(x)-u^{(k)}(y)|^{2}}{|x-y|^{1+2s}}dx\,dy defined on the higher-order fractional Sobolev space H k + s ​ ( I ) H^{k+s}(I) , where W W is a double-well potential, k ∈ ℕ k\in\mathbb{N} and s ∈ ( 0 , 1 ) s\in(0,1) with k + s > 1 2 k+s>\frac{1}{2} . We show that these functionals Γ \Gamma -converge to a sharp-interface functional with domain B ​ V ​ ( I , { − 1 , 1 } ) BV(I;\{-1,1\}) of the form m k + s ​ # ​ ( S ⁡ ( u ) ) m_{k+s}\#(S(u)) , with m k + s m_{k+s} given by the optimal-profile problem m k + s = inf { ∫ ℝ W ( v ) d x + s ⁡ ( 1 − s ) 2 1 − s ∫ ℝ 2 | v ( k ) ​ ( x ) − v ( k ) ​ ( y ) | 2 | x − y | 1 + 2 ​ s d x d y : \displaystyle m_{k+s}=\inf\Big\{\int_{\mathbb{R}}W(v)dx+\frac{s(1-s)}{2^{1-s}}\int_{\mathbb{R}^{2}}\frac{|v^{(k)}(x)-v^{(k)}(y)|^{2}}{|x-y|^{1+2s}}dx\,dy: v ∈ H loc k + s ( ℝ ) , lim x → ± ∞ v ( x ) = ± 1 } . \displaystyle\hskip 142.26378ptv\in H^{k+s}_{\rm loc}(\mathbb{R}),\lim_{x\to\pm\infty}v(x)=\pm 1\Big\}. The normalization coefficient s ⁡ ( 1 − s ) 2 1 − s \frac{s(1-s)}{2^{1-s}} is such that m k + s m_{k+s} interpolates continuously the corresponding m k m_{k} defined on standard higher-order Sobolev space H k ​ ( I ) H^{k}(I) , obtained by Modica–Mortola in the case k = 1 k=1 [ 25 ] , Fonseca–Mantegazza in the case k = 2 k=2 [ 18 ] and Brusca, Donati and Solci for k ≥ 3 k\geq 3 [ 12 ] . The results also extends previous works by Alberti, Bouchitté and Seppecher [ 3 ] , Savin–Valdinoci [ 29 ] and Palatucci–Vincini [ 27 ] in the case k = 0 k=0 and s ∈ ( 1 2 , 1 ) s\in(\frac{1}{2},1) . MSC codes: 49J45, 35B25, 82B26, 35R11. Keywords: Γ \Gamma -convergence, non-convex energies, higher-order fractional Sobolev spaces, phase-transition problems. 1 Introduction Variational models of phase transitions consider integrals depending on double-well potentials W with two (or more) minimizers. These simple models, as such admit highly discontinuous unphysical solutions; a selection of minimizers can be ensured by a singular-perturbation approach in which a higher-order term is added, the form of which is sometimes dictated by modeling considerations. In the Cahn–Hilliard theory of phase separations [ 13 ] (but earlier models leading to similar formulations date back to Landau [ 21 ] , and even to van der Waals [ 31 ] ) the corresponding energy is of the form \int_{\Omega}W(u)dx+\varepsilon^{2}\int_{\Omega}|\nabla u|^{2}dx, with a first-order perturbation depending on the small parameter \varepsilon . Analytically, the description of the solutions for \varepsilon small translates in an asymptotic analysis, which leads to a sharp-interface model in which the interaction of the two integral terms gives rise to an interfacial energy separating two phases characterized by the minimization of W , and minimum problems for \varepsilon>0 are approximated by corresponding minimum problems for interfacial energies. This minimal-interface criterion was conjectured e.g. by Gurtin [ 20 ] . Its rigorous statement was made possible by the development of the theory of sets of finite perimeter and of functions of bounded variation, which allowed to rigorously define the limit interfacial functionals. The proof of the minimal-interface criterion was then achieved by Modica [ 24 ] , following earlier results in a seminal paper by Modica and Mortola [ 25 ] , showing the \Gamma -convergence of the scaled energies \frac{1}{\varepsilon}\int_{\Omega}W(u)dx+\varepsilon\int_{\Omega}|\nabla u|^{2}dx to an interfacial energy, which is simply given by m_{W}{\rm Per}(E,\Omega), where E is a set of finite perimeter, and {\rm Per}(E,\Omega) denotes the perimeter of E in \Omega . In the case where the minimizers of W are -1 and 1 (which can be assumed up to a change of variables), minimal functions u_{\varepsilon} can be described as approximating u=-1+2\chi_{E} , where E is a minimal set. In this process, an interesting observation is that functions u_{\varepsilon} approximating \chi_{E} tend to follow an optimal profile in the direction orthogonal to the interface. This remark is highlighted by the fact that the constant m_{W} can be expressed as a one-dimensional optimal-profile problem m_{W}=\inf\Big\{\int_{-\infty}^{+\infty}(W(v)+|v^{\prime}|^{2})dt:v\in H^{1}_{\rm loc}(\mathbb{R}),\lim_{t\to\pm\infty}v(t)=\pm 1\Big\}. The Modica result has been generalized and revisited in many ways and in different contexts, leading also to more complex non one-dimensional descriptions (see e.g. [ 6 , 7 , 8 , 26 ] ). However, even when dealing with higher-dimensional problems, a reduction to the one-dimensional case is often useful to ensure compactness by a slicing argument (see e.g. [ 1 , 10 ] ). We note moreover that the analysis of perturbed double-well energies is the building block for many “phase-field” approximation procedures of free-discontinuity problems starting from the seminal work by Ambrosio and Tortorelli [ 5 ] (see [ 10 ] for a presentation of different approximations). The higher-order perturbation term of the double-well energy can be taken of different forms involving higher-order derivatives , justified either by the necessity of higher regularity (useful for the approximation procedures mentioned above), or by modeling assumptions (e.g. from atomistic theories with long-range interactions as in [ 11 ] ). In this case the energies take the form \int_{\Omega}W(u)dx+\varepsilon^{2k}\int_{\Omega}\|\nabla^{(k)}u\|^{2}dx, the second term denoting a norm of the k -th order tensor of derivatives of order k . Such functionals have been considered by Fonseca and Mantegazza [ 18 ] for k=2 and by Brusca, Donati, and Solci [ 12 ] for arbitrary k . They showed that, taking \|\cdot\| the operator norm, the scaled functionals \frac{1}{\varepsilon}\int_{\Omega}W(u)dx+\varepsilon^{2k-1}\int_{\Omega}\|\nabla^{(k)}u\|^{2}dx, still \Gamma -converge to a sharp-interface energy m_{k}{\rm Per}(E,\Omega) , where now the (one-dimensional) optimal-profile problem giving m_{k} is m_{k}=\inf\Big\{\int_{-\infty}^{+\infty}(W(v)+|v^{(k)}|^{2})dt:v\in H^{k}_{\rm loc}(\mathbb{R}),\lim_{t\to\pm\infty}v(t)=\pm 1\Big\}. (1) We note that in this case even the proof of the inequality m_{k}>0 is not trivial if k\neq 1 . In a different direction, it is also natural in some problems to consider perturbations of the form \int_{\Omega}W(u)dx+\varepsilon^{2s}\int_{\Omega\times\Omega}\frac{|u(x)-u(y)|^{2}}{|x-y|^{d+2s}}dx\,dy, where s\in(0,1) and \Omega\subset\mathbb{R}^{d} . The additional term is a Gagliardo fractional s -seminorm , which can appear as a boundary term from a higher-dimensional problem (see e.g. [ 4 ] ). Under different scaling, these problems have been studied by Savin and Valdinoci [ 29 ] (see also Palatucci and Vincini [ 27 ] , earlier results by Alberti, Bouchitté and Seppecher [ 3 ] and Alberti and Bellettini [ 2 ] , and a recent review paper by Dipierro

Read the complete text