Citation: Donatella Danielli, Rohit Jain. Regularity results for a penalized boundary obstacle problem[J]. Mathematics in Engineering, 2021, 3(1): 1-23. doi: 10.3934/mine.2021007
[1] | Morteza Fotouhi, Andreas Minne, Henrik Shahgholian, Georg S. Weiss . Remarks on the decay/growth rate of solutions to elliptic free boundary problems of obstacle type. Mathematics in Engineering, 2020, 2(4): 698-708. doi: 10.3934/mine.2020032 |
[2] | Daniela De Silva, Giorgio Tortone . Improvement of flatness for vector valued free boundary problems. Mathematics in Engineering, 2020, 2(4): 598-613. doi: 10.3934/mine.2020027 |
[3] | Hugo Tavares, Alessandro Zilio . Regularity of all minimizers of a class of spectral partition problems. Mathematics in Engineering, 2021, 3(1): 1-31. doi: 10.3934/mine.2021002 |
[4] | Tatsuya Miura . Polar tangential angles and free elasticae. Mathematics in Engineering, 2021, 3(4): 1-12. doi: 10.3934/mine.2021034 |
[5] | Camilla Nobili . The role of boundary conditions in scaling laws for turbulent heat transport. Mathematics in Engineering, 2023, 5(1): 1-41. doi: 10.3934/mine.2023013 |
[6] | Filippo Gazzola, Gianmarco Sperone . Remarks on radial symmetry and monotonicity for solutions of semilinear higher order elliptic equations. Mathematics in Engineering, 2022, 4(5): 1-24. doi: 10.3934/mine.2022040 |
[7] | Catharine W. K. Lo, José Francisco Rodrigues . On the obstacle problem in fractional generalised Orlicz spaces. Mathematics in Engineering, 2024, 6(5): 676-704. doi: 10.3934/mine.2024026 |
[8] | Xavier Fernández-Real, Alessio Figalli . On the obstacle problem for the 1D wave equation. Mathematics in Engineering, 2020, 2(4): 584-597. doi: 10.3934/mine.2020026 |
[9] | Luca Spolaor, Bozhidar Velichkov . On the logarithmic epiperimetric inequality for the obstacle problem. Mathematics in Engineering, 2021, 3(1): 1-42. doi: 10.3934/mine.2021004 |
[10] | Aleksandr Dzhugan, Fausto Ferrari . Domain variation solutions for degenerate two phase free boundary problems. Mathematics in Engineering, 2021, 3(6): 1-29. doi: 10.3934/mine.2021043 |
Al carissimo amico Sandro Salsa, con tanto affetto, ammirazione e gratitudine.
In this paper we study a penalized boundary obstacle problem of interest in thermics, fluid mechanics, and electricity. Given a domain Ω in Rn, n≥2, with sufficiently regular boundary ∂Ω=Γ1∪Γ2 and unit outer normal ν, we consider the following stationary problem:
{Δu=f in Ω,u=g on Γ1∂u∂ν=−k+((u−h)+)p−1+k−((u−h)−)p−1 on Γ2. | (1.1) |
Here f:Ω→R, g:Γ1→R and h:Γ2→R are given functions, u+=max{u,0}, u−=−min{u,0}≥0, k+ and k− are non-negative constants, and p>1. Our goal is to establish optimal regularity of the solutions, and to study properties of the free boundary (∂{u>h}∪∂{u<h})∩Γ. We begin by observing that in the limiting case k+=k−=0, u is clearly the solution of a classical Neumann problem. The other limiting case, when k+=0 and k−=+∞ (or equivalently k+=+∞ and k−=0), is more interesting. The boundary condition, in fact, becomes
u≥h,∂u∂ν≥0,(u−h)∂u∂ν=0, |
and u is a solution of the Signorini problem, also known as the thin obstacle problem. The Signorini problem has received a resurgence of attention in the last decade, due to the discovery of several families of powerful monotonicity formulas, which in turn have allowed to establish the optimal regularity of the solution, a full classification of free boundary points, smoothness of the free boundary at regular points, and the structure of the free boundary at singular points. We refer the interested reader to [1,3,6,11,14,16], see also the survey [10] and the references therein.
The general scheme of a solution to the Signorini problem provides a road map for the solution of problem (1.1), but there are two new substantial difficulties. The first one is due to the non-homogeneous nature of the boundary condition in (1.1), which in particular implies that this problem does not admit global homogeneous solutions of any degree. This is in stark contrast with the Signorini problem, where the existence and classification of such solutions play a pivotal role. Moreover, in the thin obstacle problem it is readily seen that continuity arguments force u to be always above h (hence the nomenclature), whereas the case h(x)>u(x) is no longer ruled out in (1.1). Allowing for both constants k+, k− to be finite (even when one of the two vanishes) de facto destroys the one-phase character of the problem. In order to focus the attention on these new aspects, it is useful to understand first a simplified local version of (1.1), posed in the upper half ball
B+1={x∈B1∣xn>0}, |
with f=h=0. In this setting problem (1.1) becomes
{Δu=0 in B+1u=g on (∂B1)+∂u∂xn=k+(u+)p−1−k−(u−)p−1 on Γ. | (1.2) |
Here
(∂B1)+={x∈∂B1∣xn>0},Γ={x∈B1∣xn=0}. |
An alternate perspective is given by the associated energy. We seek to minimize
J(v)=12(∫B1|∇v|2 dx+∫Γ(˜k−(v−)p+˜k+(v+)p) dx′) | (1.3) |
over all v∈W1,q(B1) with q=max{2,p} and v−g∈W1,q0(B1) for given boundary data g. Here ˜k±=2k±/p, and x=(x′,xn). In this context we think of the data in (1.2) as extended to all of B1 by even reflection. A minimizer to this energy will be symmetric about Γ and u will correspond to the restriction to B+1.
Our first main result is the following:
Theorem 1.1. Let g∈W1,q(Ω), 0≤k±<∞, k+≠k−, and p>1. Then there exists a unique minimizer u∈W1,q(B1) of the energy J(v) in (1.3). If p is an integer, then u∈Cp−1,α(¯B+1/2) for every α<min{1,p−1}, and there exists a constant C=C(n)>0 such that
‖u‖Cp−1,α(B+1/2)≤C(‖u‖L2(B+1)+‖u‖Lp(Γ)). | (1.4) |
If instead p is not an integer, then u∈C⌊p−1⌋,α(¯B+1/2) for every α<p−1−⌊p−1⌋, and there exists a constant C=C(n)>0 such that
‖u‖C⌊p−1⌋,α(B+1/2)≤C(‖u‖L2(B+1)+‖u‖Lp(Γ)). | (1.5) |
Additionally, if p is a positive integer and k−=k+, or if g does not change sign, then u∈C∞(¯B+1/2).
In the case p=2, we can in fact establish that the regularity is optimal at points where the gradient does not vanish.
Theorem 1.2. Let u be the unique solution to (1.2) (see Definition 3.2) when p=2. If ∇x′u(x′,0)≠0, then u is not in C1,1 at (x′,0).
As an immediate consequence of the regularity of the solution and of the implicit function theorem, we obtain the following result on the regularity of the free boundary.
Definition 1.3. The regular set of the free boundary is defined as
R(u)={(x′,0)∈Γ | u(x′,0)=0, ∇x′u(x′,0)≠0} |
Theorem 1.4. Let u be the unique solution to (1.2), with p>1. If x0∈R, then in a neighborhood of x0 the free boundary {u(x′,0)=0} is a C1,α− graph for all α<1.
We next turn our attention to the study of the singular set. To this end, in what follows we assume p≥2.
Definition 1.5. For
Nx0(r;u)=r ∫B+r(x0)|∇u|2 dx∫(∂Br(x0))+u2 dσ(x) |
and
μ=Nx0(0+;u)=limr→0Nx0(r;u), |
we define the set of singular points with frequency μ as
Σμ(u)={x0∈Γ | u(x0)=0, ∇x′u(x0)=0,and Nx0(0+;u)=μ}. |
The dimension of Σμ(u) at a point x∈Σ(u) is
dx0μ=dim{ζ∈Rn | ⟨ζ,∇x′px0μ(x′,0)⟩=0 for all x′∈Rn−1}, |
where px0μ is a homogeneous polynomial of degree μ as in Theorem 7.1. Finally, we introduce
Σdμ(u)={x0∈Σμ | dx0μ=d}. |
We observe here that the existence of the limit in Definition 1.5 is guaranteed by Corollary 4.3, and that it follows from the proof of Theorem 5.1 below that μ is necessarily a positive integer. The structure of the singular set is described in the following result.
Theorem 1.6. Let u be the unique solution to (1.2), with p≥2. Then for every μ∈N and d=0,1,…,n−2, the set Σdμ(u) is contained in the countable union of d-dimensional C1-submanifolds of Γ.
The proof of Theorem 1.6 follows the ideas of the corresponding result in [14] for the Signorini problem. It hinges on the monotonicity (or almost-monotonicity) of a perturbed Almgren functional and a Monneau-type functional (see Theorem 4.1 and Corollary 6.2). From these results we infer the growth rate and nondegeneracy of the solution near the free boundary. In turn, these properties allow to prove uniqueness and continuous dependance on the singular point of the blow-up limits. The rest of the proof is based on Whitney's extension and the implicit function theorem.
To conclude, we remark that considering a more general situation as in (1.1) introduces significant technical difficulties. A standard approach, under suitable smoothness assumptions, consists in flattening Γ2, which in turn leads to the study of a variable-coefficient operator and flat portion of the boundary. This problem, also with non-vanishing h, is the object of the recent paper [9].
The paper is organized as follows. In Section 2 we describe some applications to problems of semi-permeable membranes and of temperature control, which motivate the study of (1.1). In Section 3 we establish existence and uniqueness of solutions, and prove Theorems 1.1 and 1.2. In Section 4 we prove the monotonicity of the perturbed functional of Almgren type, and infer some properties of the solution as a consequence. In Section 5 we introduce the Almgren rescalings, and discuss their blow-up limits. In Section 6 we prove the almost-monotonicity of a Monneau-type functional, and establish nondegeneracy of solutions. Finally, Section 7 is devoted to the proof of Theorem 1.6.
Following [12,Section 2.2.2], we briefly describe the process of osmosis through semi-permeable walls. By Ω we denote a domain in Rn, n≥2, with sufficiently regular boundary ∂Ω. The region Ω consists of a porous medium occupied by a viscous fluid which is only slightly compressible, and we denote its pressure field by u(x). We assume that a portion Γ of ∂Ω consists of a semi-permeable membrane of finite thickness, i.e., the fluid can freely enter in Ω, but the outflow of fluid is prevented. Combining the law of conservation of mass with Darcy's law, one finds that u satisfies the equation
Δu−∂u∂t=f in Ω, |
where f=f(x,t) is a given function. When a fluid pressure h(x), for x∈Γ, is applied to Γ on the outside of Ω, one of two cases holds:
h(x)<u(x,t) or h(x)≥u(x,t). |
In the former, the semi-permeable wall prevents the fluid from leaving Ω, so that the flux is null. If we let ν denote the outer unit normal to Γ, we then have
∂u∂ν=0. | (2.1) |
In the latter case, the fluid enters Ω. It is reasonable to assume the outflow to be proportional to the difference in pressure, so that
−∂u∂ν=k(u−h), | (2.2) |
where k>0 measures the conductivity of the wall. Combining (2.1) and (2.2), we obtain the boundary condition
∂u∂ν=k(u−h)− on Γ. | (2.3) |
In our model (1.1), we allow for fluid flow to occur both into and out of Ω with different permeability constants, under the assumption that the flux in each direction is proportional to a power of the pressure.
An alternative interpretation of the model is as a boundary temperature control problem, which we only briefly outline here. We assume that a continuous medium occupies a region Ω in Rn, with boundary Γ and outer unit normal ν. Given a reference temperature h(x), for x∈Γ, it is required that the temperature at the boundary u(x,t) deviates as little as possible from h(x). To this end, thermostatic controls are placed on the boundary to inject an appropriate heat flux when necessary. The controls are regulated as follows:
(i) If u(x,t)=h(x), no correction is needed and therefore the heat flux is null.
(ii) If u(x,t)≠h(x), a quantity of heat proportional to the difference between u(x,t) and h(x) is injected.
We can thus write the boundary condition as
−∂u∂ν=Φ(u), |
where
Φ(u)={k−(u−h) if u<h0 if u=hk+(u−h) if u>h |
More in general, one can assume that Φ(u) is a continuous and increasing function of u. For further details, we refer to [12,Section 2.3.1], see also [1] for the limiting case k−=0 and k+=+∞, [4] for the case p=1 in (1.3), and [2] for different boundary conditions Φ(u).
We begin this section by proving existence and uniqueness of minimizers to (1.3). We let K={v∈W1,2(B1)∣v−g∈W1,20(B1)}.
Lemma 3.1. There exists a unique minimizer u∈K for the energy J(v) given by (1.3).
Proof. Throughout this proof we will pass to subsequences whenever necessary without comment. Let ul be a minimizing sequence. Then ‖∇ul‖2 is clearly bounded owing to the form of the energy itself. By using the Poincaré inequality on ul−g we deduce that the sequence ul is bounded in the W1,2(B1) norm. Thus there exists a weak limit u which is necessarily in K. We may assume that ul→u in L2 and a.e. The weak convergence of ul to u in W1,2 and the strong convergence in L2 imply that
∫B1|∇u|2 dx≤lim infl→∞∫B1|∇ul|2 dx. |
This clearly follows from the property of weak convergence
‖u‖W1,2(B1)≤lim infl→∞‖ul‖W1,2(B1) |
and, because of the strong L2 convergence, the inequality must fall on the gradient part of the norm.
To prove that u is a minimizer we must show then that
∫Γ(u±)p dx′≤lim infl→∞∫Γ(u±l)p dx′. |
It will suffice to demonstrate this for u−; the result for u+ is proved in an analogous fashion. The trace operator T:W1,2(B+1)→L2(∂B+1) is a bounded linear operator, since the half ball is a Lipschitz domain. Furthermore, in this setting it is a compact operator, and thus takes weakly convergent sequences to strongly convergent ones. Suppressing the Tul notation and simply writing ul we then have that
ul→uin L2(Γ). |
From this we may assume that ul→u a.e. on Γ. But then clearly (u−l)p→(u−)p a.e. and applying Fatou's Lemma we have
∫Γ(u−)p dx′≤lim inf∫Γ(u−l)p dx′ |
which completes the proof of existence.
Uniqueness follows by observing that (f+g)±≤f±+g±, and then applying standard arguments.
Next, we recall the definition of a weak solution (see [17]):
Definition 3.2. We say that u is a weak solution to
{Δu=0 in B+1uxn=f on Γ |
if for every ξ∈C∞(B+1) vanishing on (∂B1)+ we have
∫B+1∇u∇ξ dx=−∫Γfξ dx′ |
It is easy to show that the minimizer u is a weak solution to our problem.
Lemma 3.3. The minimizer u obtained in Lemma 3.1 is a weak solution to (1.2). That is,
∫B+1∇u∇ξ dx=−∫Γ(−k−(u−)p−1+k+(u+)p−1)ξ dx′ | (3.1) |
for all ξ∈C∞(B+1) vanishing on (∂B1)+.
Proof. This is a standard variational fact. See for example the proof of Lemma 4.1 in [4].
Remark 3.4. The −k− term appears since u−=−min{u,0}.
We now turn to the regularity of the solution. Our strategy will be to first prove an initial Hölder regularity which will improve afterwards. The first step is an energy estimate for u.
Lemma 3.5. Let u be the minimizer of (1.3). Then we have for any B2r⊂B1
∫Br|∇u|2 dx≤cr2∫B2ru2 dx. |
Proof. We first prove the corresponding estimate for u−=−min{u,0}. Let η∈C∞0(B2r) with
η≡1 in Br,|∇η|≤cr2. |
Taking ξ=u−η2 and using (3.1) we have
∫B1∇u∇(u−η2) dx=−∫Γ(−k−(u−)p−1+k+(u+)p−1)u−η2 dx′=−∫Γ(−k−(u−)p−1)u−η2 dx′≥0. |
Expanding yields
∫B1(η2∇u∇u−+2u−η∇u∇η) dx=∫B1−|∇u−|2η2−2u−η∇u−∇η dx≥0 |
or
∫B1|∇u−|2η2 dx≤−∫B12u−η∇u−∇η dx. |
At this point standard energy arguments imply
∫Br|∇u−|2 dx≤cr2∫B2r(u−)2 dx. |
A similar argument implies the same inequality with u+; together they yield the energy estimate for u.
Next, we use the energy estimate to prove an initial Hölder modulus of continuity for u. This regularity is much lower than optimal, but it will allows us to bootstrap to obtain higher regularity.
Lemma 3.6. The solution to (1.2) is in C0,1/2(¯B1/2).
Proof. Let Br:=Br(x) for r<1/4 and x∈B1, and let v be the harmonic replacement of u in Br. Set Γr=Br∩Γ. By minimality we have
∫Br(|∇u|2−|∇v|2) dx≤∫Γr(k−((v−)p−(u−)p)+k+((v+)p−(u+)p)) dx′. | (3.2) |
However, since v is harmonic we have
∫Br∇v⋅∇(v−u) dx=0, |
and thus
∫Br|∇u−∇v|2 dx=∫Br(|∇u|2−|∇v|2) dx. | (3.3) |
Next, since v is the harmonic lifting of u, |v|≤|u| in Br. In turn, the computation used in Lemma 3.5 demonstrated that u± are subharmonic, and therefore |u|=u++u− is as well. Thus, by the maximum principle, supB1|u|≤sup∂B1|u|=supg, the given boundary data in (1.2). In particular,
∫Γr(k−((v−)p−(u−)p)+k+((v+)p−(u+)p)) dx′≤Crn−1, |
with C independent of x and v. From this fact, combined with (3.2) and (3.3), we infer
∫Br|∇u−∇v|2 dx≤Crn−1. |
At this point, we can mimic the derivation in [5,Theorem 3.1] to deduce that
∫Br|∇u|2 dx≤Crn−1. |
In turn Morrey's Dirichlet Growth Theorem (see for instance [15,Corollary 9.1.6]) implies the desired Hölder-1/2 regularity inside B1/2.
We have reached the proof of our main result:
Proof of Theorem 1.1. Existence and uniqueness follow from Lemma 3.1. Thus, we need only to show the desired regularity. From Lemma 3.3 we know that u is a weak solution to our problem on B+1. Moreover, u is C0,1/2(¯B1/2) by Lemma 3.6. The ′±′ operation preserves Hölder regularity (with the same Hölder norm) so u±∈C0,1/2(¯B1/2) and in particular on the thin region Γ. This implies that
−k−(u−)p−1+k+(u+)p−1 | (3.4) |
is Hölder continuous of order γ, although γ will in general not be 1/2.
Nevertheless, this implies that u is a weak solution to an oblique derivative problem with Hölder continuous boundary data, namely −k−(u−)p−1+k+(u+)p−1. Regularity theory for such a problem (see e.g., [17,Proposition 5.53]) then yields that u must be C1,γ up to the boundary, with
|u|1+γ≤C(sup|u|+|u|γ). |
But in turn this implies that u is Lipschitz up to the boundary, in which case (3.4) is Hölder continuous of order p−1 when p≤2; if p>2 this is to be interpreted as differentiablity with a Hölder modulus of continuity. Applying the regularity theory once again we have the result of the theorem.
Now suppose that g does not change sign. We aim to show that u does not change sign either, in which case u±=u (and thus u± is as smooth as u is) and the regularity result above can be bootstrapped to prove that u is smooth. To this end, suppose that g≥0, but u attains a minimum value which is negative, say u(z)=m<0. Then z must lie on Γ. In particular, z∈ΓR=Γ∩BR for some 0<R<1. Now, trivial modifications to the above arguments allow to show u∈C1,α(¯BR), and therefore we can assume that the restriction of u to ΓR is C1,α. Next, we apply the Hopf Lemma. Since u is harmonic in the interior we must have
∂u∂ν(z)<0. |
Here ν is the outer normal vector, which at the point z is −en. Thus
∂u∂xn(z)>0. |
However, the boundary condition along Γ is given by
∂u∂xn=k+(u+)p−1−k−(u−)p−1, |
which holds in a classical sense since u is C1,α in a neighborhood of z. But u(z)<0, and therefore the boundary condition at z is uxn=−k−u−(z)<0, a contradiction. We have thus shown that, if g≥0, u cannot be negative along Γ. As a consequence, u is non-negative everywhere, so that u±=u and higher regularity follows by bootstrapping.
A similar argument shows that if g≤0 then u≤0 everywhere, which again implies higher regularity. Finally, the case p integer and k+=k− follows immediately from a repeated application of [17,Proposition 5.53].
We now show that, at least in the case p=2, the regularity obtained in Theorem 1.1 is optimal at points where the gradient of u is non-vanishing.
Proof of Theorem 1.2. We argue by contradiction, and assume that u∈C1,1(0), with ∇u(0)≠0. Thanks to Theorem 1.1, we know that u has a unique differential P=∇u(0). Without loss of generality, we may assume that P is also a superdifferential for u− (if not, consider u+). We refer, for instance, to [8,Chapter 3] for the definition and properties of superdifferentials. We begin by observing that we can write
∂u∂xn=k+u+−k+u−+k+u–k−u−=k+u+(k+−k−)u−. |
Thus,
(k+−k−)u−=∂u∂xn−k+u. |
From this, applying the extension theorem in [7] (with a slight abuse of notation, u(x′) denotes the restriction of u(x)=u(x′,xn) to xn=0) and the semigroup property of (−Δ)s, we deduce
(k+−k−)[−(−Δx′)1/2u−(x′)]=[−(−Δx′)1/2]∘[−(−Δx′)1/2]u(x′)−k+[−(−Δx′)1/2]u(x′)=Δx′u(x′)−k+∂u∂xn(x′,0). | (3.5) |
Because of our C1,1 assumption, we have that C0≤uττ(0)≤C1 for some constants C0,C1>0 and for any tangential direction τ. Hence, keeping also Theorem 1.1 in mind, it follows from (3.5)
|−(−Δx′)1/2u−(0)|≤C2 |
for some C2>0. We now consider
ψ(x)=[u−(0)+min{P⋅x,0}+C12|x|2]χB1. |
A straightforward computation yields
−(−Δx′)1/2ψ(0)=−∞. |
In addition, u−(x)≤ψ(x), with equality at x=0. From the definition of (−Δ)1/2, we infer
−(−Δx′)1/2u−(0)≤−(−Δx′)1/2ψ(0)=−∞. |
But we showed above that −(−Δx′)1/2u−(0)≥−C2. We have thus reached a contradiction.
In this section we establish some properties of the solution around free boundary points in the case p≥2. For u solution to (1.2), we define the coincidence set Λ(u)={(x′,0) | u(x′,0)=0}, and the free boundary F(u)=∂Λ(u). In the Signorini problem, the monotonicity of the Almgren's Frequency Functional
N(r;u)=N(r)=r ∫B+r|∇u|2 dx∫(∂Br)+u2 dσ(x) | (4.1) |
plays a fundamental role in the study of both the solution and the free boundary. In our setting, N(r) may fail to be monotone, but a suitable perturbation is. We thus introduce the perturbed Almgren Frequency Functional at the point x0=0 as
˜N(r;u)=˜N(r)=r ∫B+r|∇u|2 dx+2p∫ΓrF(u) dx′∫(∂Br)+u2 dσ(x), | (4.2) |
with F(u)=k−(u−)p+k+(u+)p and Br=Br(0).
Theorem 4.1. Let u be a solution to (1.2), with p≥2. Then ˜N(r;u) is monotone increasing in r∈(0,1).
Proof. Let
H(r)=∫(∂Br)+u2 dσ(x),D(r)=∫B+r|∇u|2 dx. |
We begin by observing
H′(r)=n−1rH(r)+2∫(∂Br)+uuν dσ(x). | (4.3) |
We also have
D(r):=∫B+r|∇u|2 dx=∫B+r(|∇u|2+uΔu) dx=∫B+rΔ(u22) dx=∫(∂Br)+uuν dσ(x)+∫Γruuν dx′=∫(∂Br)+uuν dσ(x)+∫Γr[k−(u−)p−1−k+(u+)p−1]u dx′=∫(∂Br)+uuν dσ(x)−∫Γr[k+(u+)p+k−(u−)p] dx′. |
By Rellich's Identity
D′(r)=∫(∂Br)+|∇u|2 dσ(x)=n−2r∫B+r|∇u|2 dx+2∫(∂Br)+u2ν dσ(x)−2r∫Γr⟨x,∇u⟩uxn dx′=n−2r∫B+r|∇u|2 dx+2∫(∂Br)+u2ν dσ(x)−2r∫Γr⟨x,∇u⟩(−k−(u−)p−1+k+(u+)p−1) dx′, |
which we can rewrite as
D′(r)=n−2rD(r)+2∫(∂Br)+u2ν dσ(x)−2pr∫Γr[k−⟨x,∇(u−)p⟩+k+⟨x,∇(u+)p⟩] dx′. | (4.4) |
Using integration by parts we note that
∫Γr⟨x,∇(u±)p⟩ dx′=∫∂Γrr(u±)p dσ(x′)−(n−1)∫Γr(u±)p dx′. |
Applying this fact in (4.4) we obtain
D′(r)=n−2rD(r)+2∫(∂Br)+u2ν dσ(x)−2pr[∫∂Γrr(k−(u−)p+k+(u+)p) dσ(x′)−(n−1)∫Γr(k−(u−)p+k+(u+)p)] dx′. | (4.5) |
For the sake of brevity we will define
˜D(r)=D(r)+2p∫ΓrF(u) dx′. |
A direct computation, together with (4.4) and (4.5), yields
˜N′(r)˜N(r)=1r+˜D′(r)˜D(r)−H′(r)H(r)=1r+D′(r)+2p∫∂ΓrF(u) dσ(x′)D(r)+2p∫ΓrF(u) dx′−n−1r−2∫(∂Br)+uuν dσ(x)∫(∂Br)+u2 dσ(x)=1r+n−2rD(r)D(r)+2p∫ΓrF(u) dx′+2∫(∂Br)+u2ν dσ(x)D(r)+2p∫ΓrF(u) dx′+2(n−1)pr∫ΓrF(u) dx′D(r)+2p∫ΓrF(u) dx′+1−nr−2∫(∂Br)+uuν dσ(x)∫(∂Br)+u2 dσ(x). |
Collecting terms we have
˜N′(r)˜N(r)=1r[1−D(r)D(r)+2p∫ΓrF(u) dx′]+2[∫(∂Br)+u2ν dσ(x)D(r)+2p∫ΓrF(u) dx′−∫(∂Br)+uuν dσ(x)∫(∂Br)+u2 dσ(x)]+n−1r[D(r)D(r)+2p∫∂ΓrF(u) dx′+2p∫ΓrF(u) dx′D(r)+2p∫ΓrF(u) dx′−1] | (4.6) |
Clearly the first term in (4.6) is non-negative, whereas the last one vanishes. On the other hand, from (1.2) we know
∫(∂Br)+uuν dσ(x) =D(r)+∫ΓrF(u) dx′≥D(r)+2p∫ΓrF(u) dx′, | (4.7) |
since p≥2. In turn this implies
∫(∂Br)+u2ν dσ(x)D(r)+2p∫ΓrF(u) dx′−∫(∂Br)+uuν dσ(x)∫(∂Br)+u2 dσ(x)≥∫(∂Br)+u2ν dσ(x)∫(∂Br)+uuν dσ(x)−∫(∂Br)+uuν dσ(x)∫(∂Br)+u2 dσ(x)≥0 |
by the Cauchy-Schwartz inequality. Hence ˜N′(r)N(r)≥0, and the proof is complete.
We now state some consequences of Theorem 4.1. The first result shows that, even if the Almgren's Frequency Functional N(r) in (4.1) fails to be monotone, it still has a limit as r→0+, and in fact its limit coincides with the one of ˜N(r). In order to prove this result, we will need the following trace-type inequality (see, for instance, [13,Lemma 2.5]).
Lemma 4.2. Let u∈W1,2(B+r). Then there is a bounded linear function T:W1,2(B+r)→L2(∂B+r) such that T(u) is the restriction of u to ∂B+r for any u∈C1(¯B+r). Moreover, there exists a constant C>0 such that
∫Γru2 dx′≤C(r∫B+r|∇u|2+∫(∂Br)+u2 dσ(x)). | (4.8) |
Corollary 4.3. Let N(r) and ˜N(r) be given by (4.1) and (4.2), respectively. Define μ=limr→0+˜N(r). Then there exists N(0+):=limr→0+N(r), and N(0+)=μ.
Proof. We begin by observing that, since ˜N(r)≥0, Theorem 4.1 guarantees that μ exists, and that μ∈[0,∞). Since F(u)≥0, trivially
N(r)≤˜N(r). | (4.9) |
On the other hand, if we let k=max{k+,k−} and 0<r<1/2,
∫ΓrF(u) dx′≤k∫Γr|u|p dx′≤ksupB+1/2|u|p−2∫Γr|u|2 dx′. |
Applying (4.8) we get
∫ΓrF(u) dx′≤C(r∫B+r|∇u|2+∫(∂Br)+u2 dσ(x)). | (4.10) |
Using the notations introduced in the proof of Theorem 4.1, we thus obtain
˜N(r)≤N(r)+Cr2D(r)H(r)+Cr=(1+Cr)N(r)+Cr. |
Hence,
N(r)≥˜N(r)−Cr1+Cr, | (4.11) |
and the desired conclusion follows from (4.9) and (4.11).
Next, we introduce the quantity
φ(r)=φ(r;u)=−∫(∂Br)+u2. |
Corollary 4.4. Let μ=limr→0+˜N(r)∈[0,∞). The following hold:
(a) The function r↦r−2μφ(r) is nondecreasing for 0<r<1/2. In particular,
φ(r)≤(r/2)2μφ(1/2)≤Cn(r/2)2μsupB+1/2|u|2, |
where Cn>0 is a dimensional constant.
(b) Let 0<r<1/2. Then for any δ>0 there exists R0=R0(δ)>0 such that for all r<R≤R0
φ(R)≤e2C(1−2p)(μ+δ+1)(R−r)(Rr)2(μ+δ)φ(r). |
Here C is the constant appearing in (4.10).
Proof. We begin the proof of (a) by computing
φ′(r)=ddr−∫(∂Br)+u2=2−∫(∂Br)+uuν. |
Hence, taking (4.7) into account, we have
ddr(r−2μφ(r))=−2μr−2μ−1φ(r)+2r−2μ|(∂Br)+|(∫B+r|∇u|2+∫ΓrF(u) dx′)=2r−2μ−1|(∂Br)+|(−μ∫(∂Br)+u2+r∫B+r|∇u|2+r∫ΓrF(u) dx′)=2r−2μ−1|(∂Br)+|(−μ∫(∂Br)+u2+r∫B+r|∇u|2+2pr∫ΓrF(u) dx′)+2r−2μ|(∂Br)+|(1−2p)∫ΓrF(u) dx′≥0. |
In the last inequality we have used Theorem 4.1 and the fact that p≥2.
For the proof of (b), we compute
r2φ′(r)φ(r)=r∫(∂Br)+uuν dσ(x)∫(∂Br)+u2 dσ(x)(by (4.7))=rD(r)+∫ΓrF(u) dx′∫(∂Br)+u2 dσ(x)=˜N(r)+r(1−2p)∫ΓrF(u) dx′∫(∂Br)+u2 dσ(x)(by (4.10))≤˜N(r)+Cr(1−2p)(N(r)+1). |
Thanks to Corollary 4.3, there exists R0=R0(δ)>0 such that N(r)≤˜N(r)≤μ+δ for r<R≤R0. We then have
ddrlogφ(u)≤2r(μ+δ)+2C(1−2p)(μ+δ+1). |
To conclude we integrate the inequality over (r,R).
Corollary 4.5. Let u be a solution to (1.2). Then, for all x∈B+r, 0<r<1/2,
|u(x)|≤Cn(r/2)μsupB+1/2|u|, |
where Cn>0 is a dimensional constant.
Proof. We note that (u+)2 is a positive subharmonic function in the domain. Hence,
(u+)2(0)≤−∫(∂Br)+(u+)2≤Cn(r/2)2μsupB+1/2|u|2, |
by Corollary 4.4(a). A similar estimate holds for (u−)2.
The next step in our analysis is to study blow-up sequences around a free boundary point x0∈F(u). Without loss of generality, we may assume x0=0. We define, for 0<r<1, the Almgren rescalings
vr(x)=u(rx)[φ(r;u)]1/2. | (5.1) |
We note that ‖vr‖L2((∂B1)+)=1. Moreover, for R0=R0(δ) as in Corollary 4.4(b), a fixed R>1, and every r>0 such that rR≤R0, we have, thanks to Corollaries 4.3 and 4.4(b),
∫B+R|∇vr|2 dx=Rn−2N(rR;u)φ(rR;u)φ(r,u)≤C(μ+δ)Rn−2+2(μ+δ). |
Hence, after an even reflection across {xn=0}, any sequence {vrj}, with rj→0+ as j→∞, is equibounded in H1loc(Rn), and by Theorem 1.1, it is also bounded in C1,αloc(Rn). Thus, there exists a subsequence, denoted by vj, and a function v∗ (which we will refer to as the Almgren blow-up), such that
vj→v∗and∇vj→∇v∗as j→∞, |
uniformly on every compact subset of Rn. We note that the fact ‖vj‖L2((∂B1)+)=1 in particular implies that the blow-up is nontrivial. In addition, by rescaling,
μ=limj→∞N(rj;u)=limj→∞N(1;vj)=limj→∞∫B+1|∇vj|2 dx=∫B+1|∇v∗|2 dx, |
and therefore we have (keeping in mind that u(0)=0) μ>0. A similar rescaling argument, in fact, shows that for any ρ>0
N(ρ;v∗)=limj→∞N(ρ;vj)=limj→∞N(rjρ;u)=μ | (5.2) |
Next, for a function ξ∈C∞0(B1) we compute
∫B+1∇vr(x)∇ξ(x) dx=r[φ(r;u)]1/2∫B+1∇u(rx)∇ξ(x) dx(by (3.1))=r[φ(r;u)]1/2∫Γ(−k−(u−)p−1(rx′,0)+k+(u+)p−1(rx′,0))ξ(x,0) dx′. | (5.3) |
We now assume p≥3. An application of (4.8) yields
|∫B+1∇vr(x)∇ξ(x) dx|≤Cr[φ(r;u)]1/2supB+r|u|p−3∫Γu2(rx′,0) dx′≤Cr2−n[φ(r;u)]1/2supB+r|u|p−3∫Γru2(x′,0) dx′≤Cr2−n[φ(r;u)]1/2supB+r|u|p−3(r∫B+r|∇u(x)|2 dx+∫(∂Br)+u2(x) dσ(x))≤Cr[φ(r;u)]1/2supB+r|u|p−3(N(r;u)+1). | (5.4) |
Since [φ(r;u)]1/2≤Crμ by Corollary 4.4(a) and supB+r|u|≤Crμ by Corollary 4.5, we conclude that the last term in (5.4) goes to zero as r→0+. Hence, if we extend v∗ by even reflection across {xn=0}, then v∗ is harmonic in B1. The same conclusion can be reached in the case 2≤p<3 by applying Hölder's inequality in the last integral in (5.3). Now, it is well known (see, for instance [18,Section 9.3.1]) that a function harmonic in B1 and satisfying (5.2) is necessarily an homogeneous harmonic polynomial of degree μ∈N (since we have already ruled out the possibility μ=0). If we also assume ∇x′u(0)=0, from the uniform convergence of ∇vj to ∇v∗ we deduce μ≥2. We have thus proved the following.
Theorem 5.1. Let u be a solution to (1.2), with u(0)=0 and ∇x′u(0)=0. If vr is as in (5.1), then for any sequence rj→0+ there exists a subsequence {vj} of {vrj} and a function v∗ such that
vj→v∗in H1(B+1) and in C1(B+1). |
Furthermore, the even reflection of v∗ across {xn=0} is an homogeneous harmonic polynomial of degree μ=N(0+;u)∈N, μ≥2.
Our next step consists in establishing almost-monotonicity of a functional of Monneau type. Using the notations introduced in the proof of Theorem 4.1, we define the Weiss functional
Wμ(r;u)=H(r;u)rn−1+2μ(N(r;u)−μ). |
Theorem 6.1. Let u be as in Theorem 5.1, and let pμ be an harmonic polynomial, homogeneous of degree μ and even in xn. If we define the Monneau functional as
Mμ(r;u,pμ)=1rn−1+2μ∫(∂Br)+(u−pμ)2 dσ(x), |
then there exists C>0 such that
ddr(Mμ(r;u,pμ)+Cr)≥2rWμ(r;u). | (6.1) |
Proof. Let pμ be an harmonic polynomial, homogeneous of degree μ and even in xn. Since N(r;pμ)=μ, we have Wμ(r;pμ)=0. We now rewrite
Wμ(r;u)=1rn−2+2μD(r;u)−μrn−1+2μH(r;u), |
and let w=u−pμ. Then
Wμ(r;u)=Wμ(r;u)−Wμ(r;pμ)=1rn−2+2μ∫B+r(|∇w|2+2∇w⋅∇pμ) dx−μrn−1+2μ∫(∂Br)+(w2+2pμw) dσ(x). |
Integrating by parts in the first integral, keeping in mind that pμ is harmonic and ∂pμ∂xn=0 on Γr, we obtain
Wμ(r;u)=1rn−2+2μ∫(∂Br)+w∇w⋅xr dσ(x)−1rn−2+2μ∫Γr∂u∂xn(u−pμ) dx′+2rn−2+2μ∫(∂Br)+w∇pμ⋅xr dσ(x)−μrn−1+2μ∫(∂Br)+(w2+2pμw) dσ(x). |
Noting that ∇pμ⋅x=μpμ, we infer
Wμ(r;u)=1rn−1+2μ∫(∂Br)+w∇w⋅x dσ(x)−1rn−2+2μ∫Γr∂u∂xn(u−pμ) dx′−μrn−1+2μ∫(∂Br)+w2 dσ(x). | (6.2) |
We now observe the following facts:
1rn−1+2μ∫(∂Br)+w(∇w⋅x−μw) dσ(x)=r2ddr(1rn−1+2μ∫(∂Br)+w2 dσ(x)), | (6.3) |
∫Γru∂u∂xn dx′=∫ΓrF(u) dx′≥0, | (6.4) |
∫Γrpμ∂u∂xn dx′≤Crn−1+pμ. | (6.5) |
In (6.5) we have used the boundary condition in (1.2), Corollary 4.5, and the fact that pμ is homogeneous of degree μ. As a consequence, the constant in (6.5) will depend on supB+1/2|u| and ‖pμ‖L1(Γ). Using (6.3)–(6.5) in (6.2), we obtain
Wμ(r;u)≤r2ddr(1rn−1+2μ∫(∂Br)+w2 dσ(x))+Cr1+μ(p−2). | (6.6) |
An application of (6.6) yields
ddr(1rn−1+2μ∫(∂Br)+w2)≥2rWμ(r,u)−Crμ(p−2)≥2rWμ(r;u)−C, |
thus concluding the proof.
Corollary 6.2. Under the assumption of Theorem 6.1, there exists C>0 such that
ddr(Mμ(r;u,pμ)+Cr)≥0. |
In particular, there exists limr→0+Mμ(r;u,pμ).
Proof. Thanks to the inequality (4.11) and Theorem 4.1, we have for 0<r<1/2
N(r;u)≥˜N(r;u)−Cr1+Cr≥μ−Cr1+Cr |
and therefore
N(r;u)−μ≥−C(μ+1)r1+Cr. | (6.7) |
Thus, using (6.7) and Corollary 4.4(a), we obtain
Wμ(r;u)≥−C(μ+1)r1+CrH(r;u)rn−1+2μ≥−Cr. | (6.8) |
Inserting this information in (6.1) gives the desired conclusion.
With Corollary 6.2 at our disposal, we can prove nondegeneracy of the solution at free boundary points.
Lemma 6.3. Let u be as in Theorem 5.1. There exists C>0 and 0<R0<1, depending possibly on u, such that
sup(∂Br)+|u|≥Crμ | (6.9) |
for all 0<r<R0.
Proof. Arguing by contradiction, assume that (6.9) does not hold. We thus have, for a sequence r=rj→0+,
φ(r)=o(r2μ). | (6.10) |
Possibly passing to a subsequence, Theorem 5.1 guarantees that the Almgren rescalings ur(x) introduced in (5.1) converge uniformly, as r→0+, to a nontrivial harmonic polynomial pμ, homogeneous of degree μ and even in xn. We now compute Mμ(0+;u,pμ)=limr→0Mμ(r;u,pμ), whose existence follows from Corollary 6.2. We have
Mμ(r;u,pμ)=1rn−1+2μ∫(∂Br)+u2 dσ(x)+1rn−1+2μ∫(∂Br)+(−2upμ+p2μ) dσ(x). | (6.11) |
We observe that the first integral in (6.11) goes to 0 as r→0+ because of (6.10). Moreover, the homogeneity of pμ implies
1rn−1+2μ∫(∂Br)+p2μ dσ(x)=∫(∂B1)+p2μ dσ(y), | (6.12) |
and therefore
1rn−1+2μ∫(∂Br)+|upμ| dσ(x)≤(1rn−1+2μ∫(∂Br)+u2 dσ(x))1/2(1rn−1+2μ∫(∂Br)+p2μ dσ(x))1/2→0. | (6.13) |
Combining (6.11)–(6.13) we infer
Mμ(0+;u,pμ)=∫(∂B1)+p2μ dσ(y)=1rn−1+2μ∫(∂Br)+p2μ dσ(x) |
for all 0<r<1/2. An application of Corollary 6.2 then yields
1rn−1+2μ∫(∂Br)+(u−pμ)2 dσ(x)+Cr≥Mμ(0+;u,pμ)=1rn−1+2μ∫(∂Br)+p2μ dσ(x), |
which we can rewrite as
1rn−1+2μ∫(∂Br)+(u2−2upμ) dσ(x)≥−Cr. |
Rescaling according to (5.1), we obtain
1r2μ∫(∂B1)+(φ(r)v2r−2[φ(r)]1/2rμvrpμ) dσ(x)≥−Cr, |
or equivalently
∫(∂B1)+([φ(r)]1/2rμv2r−2vrpμ) dσ(x)≥−Crμ+1[φ(r)]1/2. | (6.14) |
At this point we observe that, thanks to Corollary 4.4(b), for each 0<δ<1 there exist C=C(p, μ, δ), and R0=R0(δ)>0 such that φ(r)≥C1r2(μ+δ) for all 0<r<R0. Hence, letting r→0+ in (6.14) we conclude
−∫(∂B1)+p2μ≥0, |
which is a contradiction since pμ is nonzero.
Corollary 6.4. Let u and μ be as in Theorem 5.1. The set Σμ(u) (see Definition 1.5) is of type Fσ, i.e., it is the union of countably many closed sets.
Proof. The proof follows the lines of the one of Lemma 1.5.3 in [14], and it is omitted.
To continue with our analysis, we introduce the homogeneous rescalings
v(μ)r(x)=u(rx)rμ,0<r<1, | (7.1) |
and show existence and uniqueness of the blow-ups with respect to this family of rescalings.
Theorem 7.1. Let u and μ be as in Theorem 5.1. If v(μ)r is as in (7.1), then there exists a unique function v0 such that
v(μ)r(x)→v0in H1(B+1) and in C1(B+1). |
Furthermore, the even reflection of v0 across {xn=0} is an homogeneous harmonic polynomial of degree μ.
Proof. By Corollary 4.4(a) and Lemma 6.3, there exist constants C1, C2>0 such that
C1rμ≤[φ(r)]1/2≤C2rμ,0<r<1. |
From this it follows that any limit of the rescalings v(μ)r over any sequence rj→0+ is a positive multiple of the Almgrens rescalings vr as in (5.1). By Theorem 5.1, we know that the even reflection of v0 across {xn=0} is an homogeneous harmonic polynomial of degree μ, and that the convergence is both in H1(B+1) and in C1(B+1). At this point, we only need to show uniqueness. To this end, we apply Corollary 6.2 with pμ=v0. We thus have
Mμ(0+,u,v0)=limrj→0+Mμ(rj,u,v0)=limrj→0+∫(∂B1)+(u(μ)rj−u0)2 dσ(x)=0, |
where the last equality follows from the first part of the proof. In particular, we have that
Mμ(r,u,v0)=∫(∂B1)+(u(μ)r−u0)2 dσ(x)→0 |
as r→0+, and not only over rj→0+. If v′0 is a limit of v(μ)r over another sequence r′j→0+, we infer that
∫(∂B1)+(u′0−u0)2 dσ(x)=0. |
Hence, u′0=u0 and the proof is complete.
The next step consists in showing the continuous dependance of the blow-ups. In what follows, we denote by Pμ, with μ∈N, the class of harmonic polynomials homogeneous of degree μ and even in xn.
Theorem 7.2. Let u be a solution to (1.2), μ∈N with μ≥2, and x0∈Σμ(u). Denote by px0μ the blow-up of u at x0 as in Theorem 7.1, so that
u(x)=px0μ(x−x0)+o(|x−x0|μ). |
Then the mapping x0↦px0μ from Σμ(u) to Pμ is continuous. Moreover, for any compact set K⊂Σμ(u)∩B1 there exists a modulus of continuity ωμ, with ωμ(0+)=0, such that
|u(x)−px0μ(x−x0)|≤ωμ(|x−x0|)|x−x0|μ |
for any x0∈K.
Proof. We begin by observing that Pμ is a convex subset of the finite-dimensional vector space of all polynomials homogeneous of degree μ, and therefore all norms are equivalent. We choose to endow it with the norm in L2((∂B1)+). We begin by fixing x0∈Σμ(u) and ε>0 sufficiently small. Then there is rε=rε(x0)>0 such that
Mx0μ(rε;u,px0μ):=1rn−1+2με∫(∂Brε)+(u(x+x0)−px0μ)2dσ(x)<ε. |
In turn, there exists δε=δε(x0)>0 such that if x1∈Σμ(u)∩Bδε(x0), then
Mx1μ(rε;u,px0μ)=1rn−1+2με∫(∂Brε)+(u(x+x1)−px0μ)2dσ(x)<2ε. | (7.2) |
Corollary 6.2 yields that Mx1μ(r;u,px0μ)<3ε, provided that 0<r<rε and rε is small enough. Rescaling and passing to the limit as r→0+, we obtain
∫(∂B1)+(px1μ−px0μ)2dσ(x)=Mx1μ(0+;u,px0μ)≤3ε | (7.3) |
and the first part of the theorem is proved. In order to establish the second part, we observe that, for |x1−x0|<δε and 0<r<rε, combining (7.2) and (7.3) we obtain
‖u(⋅+x1)−px1μ‖L2((∂Br)+)≤‖u(⋅+x1)−px0μ‖L2((∂Br)+)+‖px0μ−px1μ‖L2((∂Br)+)≤2(3ε)1/2rn−12+μ. |
Integrating in r we obtain
‖u(⋅+x1)−px1μ‖L2((Br)+)≤Cε1/2rn/2+μ, | (7.4) |
with C=C(n,μ)>0. Letting
v(μ)r,x1(x)=u(rx+x1)rμ, |
we infer from (7.4)
‖v(μ)r,x1(x)−px1μ‖L2(B+1)≤Cε1/2. | (7.5) |
At this point we observe that the difference wr=v(μ)r,x1(x)−px1μ is a weak solution to
{Δwr=0 in B+1,∂wr∂ν=r[−k+((v(μ)r,x1)+)p−1+k−((v(μ)r,x1)−)p−1] on Γ. |
By the L∞−L2 interior estimates (see, for instance [17,Theorem 5.36]), there exists a positive constant C=C(n,k+,k−) such that, for some q>n−1,
‖v(μ)r,x1(x)−px1μ‖L∞((B1/2)+)≤C(‖v(μ)r,x1(x)−px1μ‖L2((B1)+)+r‖|v(μ)r,x1|p−1‖Lq(Γ)). | (7.6) |
To estimate the right-hand side in (7.6), we recall that |v(μ)r,x1|≤C by Corollary 4.5, and thus
r‖|v(μ)r,x1|p−1‖Lq(Γ)≤Cr1+n−1q. | (7.7) |
Combining (7.6) with (7.5) and (7.7), we obtain
‖v(μ)r,x1(x)−px1μ‖L∞((B1/2)+)≤Cε | (7.8) |
for 0<r<rε sufficiently small, and Cε→0 as ε→0. To conclude, we cover the compact set K⊂Σμ(u)∩B1 with a finite number of balls Bδε(xi0)(xi0) for some choice of xi0∈K, i=1,…,N. Hence, for r<rKε:=min{rε(xi0)|i=1,…,N}, we have that (7.8) holds for all x1∈K. The desired conclusion readily follows.
Proof of Theorem 1.6. The proof of the structure of Σdμ is centered on Corollary 6.4, Theorem 7.2, Whitney's extension theorem, and the implicit function theorem. Since the arguments are essentially identical to the ones in the proof of Theorem 1.3.8 in [14], we omit the details and refer the interested reader to that source.
The authors wish to thank the anonymous referee, whose comments and suggestions helped to improve the readability of the paper.
The authors declare no conflict of interest.
[1] |
Athanasopoulos I, Caffarelli LA (2006) Optimal regularity of lower dimensional obstacle problems. J Math Sci 132: 274-284. doi: 10.1007/s10958-005-0496-1
![]() |
[2] |
Athanasopoulos I, Caffarelli LA (2010) Continuity of the temperature in boundary heat control problems. Adv Math 224: 293-315. doi: 10.1016/j.aim.2009.11.010
![]() |
[3] |
Athanasopoulos I, Caffarelli LA, Salsa S (2008) The structure of the free boundary for lower dimensional obstacle problems. Am J Math 130: 485-498. doi: 10.1353/ajm.2008.0016
![]() |
[4] |
Allen M, Lindgren E, Petrosyan A (2015) The two-phase fractional obstacle problem. SIAM J Math Anal 47: 1879-1905. doi: 10.1137/140974195
![]() |
[5] |
Allen M, Petrosyan A (2012) A two-phase problem with a lower-dimensional free boundary. Interface Free Bound 14: 307-342. doi: 10.4171/IFB/283
![]() |
[6] |
Caffarelli LA, Salsa S, Silvestre L (2008) Regularity estimates for the solution and the free boundary of the obstacle problem for the fractional Laplacian. Invent Math 171: 425-461. doi: 10.1007/s00222-007-0086-6
![]() |
[7] |
Caffarelli L, Silvestre L (2007) An extension problem related to the fractional Laplacian. Commun Part Diff Eq 32: 1245-1260. doi: 10.1080/03605300600987306
![]() |
[8] | Cannarsa P, Sinestrari C (2004) Semiconcave Functions, Hamilton-Jacobi Equations, and Optimal Control, Boston: Birkhäuser Boston, Inc. |
[9] | Danielli D, Krummel B (2020) Existence and regularity results for the penalized thin obstacle problem with variable coefficients. arXiv:2005.05524. |
[10] | Danielli D, Salsa S (2017) Obstacle Problems Involving the Fractional Laplacian, Berlin: De Gruyter. |
[11] |
De Silva D, Savin O (2016) Boundary Harnack estimates in slit domains and applications to thin free boundary problems. Rev Mat Iberoam 32: 891-912. doi: 10.4171/RMI/902
![]() |
[12] | Duvaut G, Lions JL (1976) Inequalities in Mechanics and Physics, Berlin-New York: SpringerVerlag. |
[13] | Fall MM, Felli V (2014) Unique continuation property and local asymptotics of solutions to fractional elliptic equations, Commun Part Diff Eq 39: 354-397. |
[14] |
Garofalo N, Petrosyan A (2009) Some new monotonicity formulas and the singular set in the lower dimensional obstacle problem. Invent Math 177: 415-461. doi: 10.1007/s00222-009-0188-4
![]() |
[15] | Jost J (2007) Partial Differential Equations, 2 Eds., New York: Springer. |
[16] |
Koch H, Petrosyan A, Shi W (2015) Higher regularity of the free boundary in the elliptic Signorini problem. Nonlinear Anal 126: 3-44. doi: 10.1016/j.na.2015.01.007
![]() |
[17] | Lieberman GM (2013) Oblique Derivative Problems for Elliptic Equations, Hackensack: World Scientific Publishing Co. Pte. Ltd. |
[18] | Petrosyan A, Shahgholian H, Uraltseva N (2012) Regularity of Free Boundaries in Obstacle-Type Problems, Providence: American Mathematical Society. |
1. | Donatella Danielli, Brian Krummel, Existence and regularity results for the penalized thin obstacle problem with variable coefficients, 2025, 432, 00220396, 113213, 10.1016/j.jde.2025.02.084 |