Loading [MathJax]/jax/output/SVG/jax.js
Research article Special Issues

The shape of a free boundary driven by a line of fast diffusion

  • We complete the description, initiated in [6], of a free boundary traveling at constant speed in a half plane, where the propagation is controlled by a line having a large diffusion on its own. The main result of this work is that the free boundary is asymptotic to a line at infinity, whose angle to the horizontal is dictated by the velocity of the wave on the line. This helps understanding some counter-intuitive simulations of [7].

    Citation: Luis A. Caffarelli, Jean-Michel Roquejoffre. The shape of a free boundary driven by a line of fast diffusion[J]. Mathematics in Engineering, 2021, 3(1): 1-25. doi: 10.3934/mine.2021010

    Related Papers:

    [1] Anne-Charline Chalmin, Jean-Michel Roquejoffre . Improved bounds for reaction-diffusion propagation driven by a line of nonlocal diffusion. Mathematics in Engineering, 2021, 3(1): 1-16. doi: 10.3934/mine.2021006
    [2] La-Su Mai, Suriguga . Local well-posedness of 1D degenerate drift diffusion equation. Mathematics in Engineering, 2024, 6(1): 155-172. doi: 10.3934/mine.2024007
    [3] Glen Wheeler . Convergence for global curve diffusion flows. Mathematics in Engineering, 2022, 4(1): 1-13. doi: 10.3934/mine.2022001
    [4] Raúl Ferreira, Arturo de Pablo . A nonlinear diffusion equation with reaction localized in the half-line. Mathematics in Engineering, 2022, 4(3): 1-24. doi: 10.3934/mine.2022024
    [5] Mario Pulvirenti . On the particle approximation to stationary solutions of the Boltzmann equation. Mathematics in Engineering, 2019, 1(4): 699-714. doi: 10.3934/mine.2019.4.699
    [6] Yihong Du, Wenjie Ni . The Fisher-KPP nonlocal diffusion equation with free boundary and radial symmetry in R3. Mathematics in Engineering, 2023, 5(2): 1-26. doi: 10.3934/mine.2023041
    [7] 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
    [8] Giuseppe Maria Coclite, Lorenzo di Ruvo . On the initial-boundary value problem for a Kuramoto-Sinelshchikov type equation. Mathematics in Engineering, 2021, 3(4): 1-43. doi: 10.3934/mine.2021036
    [9] Stefano Berrone, Stefano Scialò, Gioana Teora . The mixed virtual element discretization for highly-anisotropic problems: the role of the boundary degrees of freedom. Mathematics in Engineering, 2023, 5(6): 1-32. doi: 10.3934/mine.2023099
    [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
  • We complete the description, initiated in [6], of a free boundary traveling at constant speed in a half plane, where the propagation is controlled by a line having a large diffusion on its own. The main result of this work is that the free boundary is asymptotic to a line at infinity, whose angle to the horizontal is dictated by the velocity of the wave on the line. This helps understanding some counter-intuitive simulations of [7].


    We would like to dedicate this article to Sandro Salsa. He is a wonderful person and an exceptional, generous mathematician. We have greatly enjoyed his work and his friendship.

    The system under study involves an unknown real c>0, an unknown function u(x,y) defined in R2:=R×R, and an unknown curve ΓR2 satisfying

    {dΔu+cxu=0(x,y){u>0}|u|=1((x,y)Γ:={u>0} Duxx+cxu+1/μuy=0for xRy=0u(,y)=1,uniformly in yRu(+,y)=0pointwise in yR (1.1)

    Note that the convergence of u to 1 to the left, and 0 to the right, are not requested to hold in the same sense. This is not entirely innocent, we will explain why in more detail below. We will also be interested in a (seemingly) more complex version of (1.1). We look for a real c>0, a function u(x), defined for xR, a function v(x,y) defined in R2:=R+×R, and a curve ΓR2 such that

    {dΔv+cxv=0(x,y){v>0}|v|=1((x,y)Γ:={v>0} Duxx+cxu+1/μuv=0for xRy=0vy=μuvfor xRy=0 and v(x,0)>0u()=1/μ, u(+)=0v(,y)=1uniformly with respect to y0v(+,y)=0pointwise with respect to y0. (1.2)

    We ask for the global shape of the free boundary Γ. Before that, we ask about the existence of a solution (c,Γ,u) to (1.1), and of a solution (c,Γ,u,v) to (1.2), this indeed deserves some thought, as the condition at is rather strong.

    Systems (1.1) and (1.2) arise from a class of models proposed by H. Berestycki, L. Rossi and the second author to model the speed-up of biological invasions by lines of fast diffusion, see for instance [3] or [4]. The two-dimensional lower half-plane (that was called "the field" in the afore-mentionned references), in which reaction-diffusion phenomena occur, interacts with the x axis ("the road") which has a much faster diffusion D of its own. In Model (1.2), u(x) the density of individuals on the road, and v(t,x,y) the density of individuals in the field. The road yields the fraction μu to the field, and retrieves the fraction νv in exchange; the converse process occurs for the field. Model (1.1) is obtained from (1.2) by forcing μu=v on the road, so that the sole unknown is v(x,y), and the exchange term is replaced by vyμ (and v has been renamed u). In the sequel, Model (1.2) will be called the "model with two species" (that is, the density on the road and in the field may be different), while Model (1.1) will be called "Model with one species". Also note that, in both models, the unknown functions are travelling waves of an evolution problem where the term cxu (resp. cxv) is replaced by tu (resp. tv). This is explained in more detail in [6], where the study of (1.1) and (1.2) was initiated.

    Theorem 1.1. (Existence for the model with one species) Assume Dd. System (1.1) has a solution (c>0,Γ,u) with u globally Lipschitz, and we have xu<0. Moreover

    Γ is an analytic curve, as well as a locally Lipschitz graph in the y variable:

    Γ={(φ(y),y), y<0}. (1.3)

    Γ{y=0} is nonempty, assume (0,0)Γ. There is ε0>0 such that

    ΓBε0(0)={y=ϕ(x), ε0x0},ϕ(x)=x22D+ox0(x2). (1.4)

    Theorem 1.2. (Existence for the model with two species) System (1.2) has a solution (c>0,Γ,u,v) such that the function v is globally Lipschitz, and we have xv<0. Moreover

    Γ is an analytic curve, as well as a locally Lipschitz graph in the y variable:

    Γ={(φ(y),y), y<0}. (1.5)

    Γ{y=0} is nonempty, assume (0,0)Γ. There is λ>0 and ε0>0 such that

    ΓBε0(0)={y=ϕ(x), ε0x0},ϕ(x)=λx+ox0(|x|). (1.6)

    The main question of this work, namely, how Γ looks like, is addressed in the following theorem. Both models with one species and model with two species are concerned. Let c0 be the speed of the basic travelling wave ϕ0(x):

    ϕ0+c0ϕ0=0 (xR)limxϕ0(x)=1,ϕ0(x)=0 (xR+)[ϕ0]x=0=0, [ϕ0]x=0=1. (1.7)

    In what follows, we stress the dependence on D of the velocity, free boundary and solution of the PDE by denoting them cC, ΓD, uD, vD.

    Theorem 1.3. In both models (1.1) and (1.2), there is D00 such that, for every DD0 we have cD>c0. Let αD be given by sinαD=c0cD. If φD is given by (1.4) we have

    limyφD(y)=cotanαD. (1.8)

    In other words, ΓD has an asymptotic direction, which is a line making the angle αD with the horizontal. We have a more precise version of Theorem 1.3:

    Theorem 1.4. Assume that α<π2. For every ω(0,12cDsinαD), we have

    φD(y)=cotanαD+O(eωy). (1.9)

    Thus there is a straight line making the angle αD with the horizontal that is asymptotic to ΓD at infinity, and the distance between the two shrinks exponentially fast.

    Let us be more specific about the question that we wish to explore here. We want to account for a loss of monotonicity phenomenon for u (Model (1.1)) or v (Model (1.2)) in the y variable, a study that was initiated in [6]. This phenomenon was discovered numerically by A.-C. Coulon in her PhD thesis [7]. There, she provided simuations for the evolution system (notice that the propagation takes place for y>0):

    tuDxxu=νv(t,x,0)μu  xRtvdΔv=f(v)  (x,y)R×R+yv(t,x,0)=μu(t,x,t)νv(t,x,0)  xR. (1.10)

    The figures below account for some of her results; the parameters are

    f(v)=vv2,  D=10,  u(0,x)=1[1,1](x),  v(0,x,y)0.

    The top figure represents the levels set 0.5 of v at times 10, 20, 30, 40, while the bottom figure represents the shape of v(40,x,y).

    The surprise is the location of the leading edge of the invasion front: rather than being located on the road, as one would have expected (especially for large D), it sits a little further in the field. This entails a counter-intuitive loss of monotonicity. As working directly with the reaction-diffusion (1.10) has not resulted in a significant outcome so far, the idea was to replace the reaction-diffusion model by a free boundary problem, that may be seen as an extreme instance of reaction-diffusion. Using this idea, a first explanation of the location of the leading edge is provided in [6]. The conclusions are included in Theorems 1.1 and 1.2.

    The goal of the paper is, as already announced, to account for the global shape of the free boundary. We claim that it will provide a good theoretical explanation of A.-C. Coulon's numerics. Indeed it may be expected (although it is not a totally trivial statement) that the solution of (1.10) will converge to a travelling wave. As the underlying situation is that of an invasion, it is reasonable to assume that no individuals (if we refer, as was the initial motivation, to a biological invasion) are present ahead of the front. This is why we impose the seemingly stringent, but reasonable from the modelling point of view, condition at . Theorem 1.3 shows that it entails a very specific behaviour.

    Let us explain the consequences of our results on the understanding of the model. Theorems 1.1 to 1.3 put together depict a free boundary whose leading edge is in the lower half plane and which, after a possibly nonempty but finite set of turns, becomes asymptotic to a line that goes to the right of the lower half plane. This is in good qualitative agreement with the upper two-thirds of the picture presented here, the lower third accounting for the fact (still to be described in mathematically rigorous terms) that the free boundary bends in order to connect to a front propagating downwards, which is logical as we start from a solution that is compactly supported. Let us, however, point out that the analogy should not be pushed further than what is reasonable, as the logistic nonlinearity displays – and this is especially true for the model (1.10) – more counter-intuitive oddities of its own, see [5].

    Some additional mathematical comments are in order. The first one is that we have put some effort in proving existence theorems. The reason is that we could not entirely rely on our previous study [6] for that: Although the leading edge was analysed, we had chosen to wipe out the additional difficulties coming from the study in the whole lower half-plane, by studying a model in a strip of finite length with Neumann conditions at the bottom. While the study of the leading edge is purely local, and will not need any more development, the condition limxu(x,y)=1 (resp. limxu(x,y)=1) uniformly in yR requires some additional care, that is presented in Section 3.

    Theorem 1.3 is hardly incidental. It is in fact a general feature in reaction-diffusion equations in the plane. The heuristic explanation is the following: looking down very far in the lower half plane, we may think that the free boundary ΓD propagates like the 1D wave in its normal direction, that is, Vn=c0. On the other hand, it propagates with speed cD horizontally: this imposes the angle αD. For a rigorous proof of that, we take the inspiration from previous works on conical-shaped waves for reaction-diffusion equations in the plane. A first systematic study may be found in [2], while the stability of these objects is studied in [9]. Further qualitative properties are derived in [10]. Solutions of the one phase free boundary problem for cxuΔu=0 are classified by Hamel and Monneau in [8]. One of their results will play an important role in the proof of Theorem 1.3, this will be explained in detail in Section 4.

    Although this work is clearly aimed at understanding the situation for large D (the case Dd poses interesting technical questions in for the one species model) we have not really provided a systematic study of the limit D+. This will be the object of a forthcoming paper. Another interesting question is whether the free boundary has turning points. While the simulations cleary point at convexity properties of the sub-level sets, we do not, at this stage, have real hints of what may be true.

    The paper is organised as follows. In Section 2, we provide some universal bounds on the velocities, in therms of the diffusion on the road D. In Section 3, we construct the wave for the one species model and prove Theorem 1.2. In Section 4, we indicate the necessary modifications for the two-species model. In Section 5, we prove the exponential convergence of the level sets.

    Solutions to (1.1) will be constructed through a suitable approximation in a finitely wide cylinder; we set

    ΣL=R×(L,0),

    from (a trivial modification of) [6], for L>1, there is a solution (cLD,ΓLD,uLD) to the auxiliary problem

    {dΔu+cxu=0(x,y)ΣL{u>0}|u|=1((x,y)Γ:={u>0}ΣL Duxx+cxu+1/μuy=0for xRy=0u(,y)=1,u(+,y)=(1yL)+u(x,L)=1. (2.1)

    In biological terms, this means that the boundary {y=L} is lethal for the individuals. The limit at x=+, namely the function (1yL)+, is of course not chosen at random, it solves the one-dimensional free boundary problem

    {u=0in (L,0){u>0}|u|=1on ((L,0){u>0}) (that is, at y=1L)u(L)=1.

    To ensure the maximum chance to retrieve, in the end, a solution that converges to 1 uniformly in y as x, we have imposed the Dirichlet condition u=1 at the bottom of the cylinder.

    At this point, we need to make a short recollection of what the exponential solutions of the linear problem are. System (2.1), linearised around 0, reads

    {dΔu+cxu=0(x,y)ΣL Duxx+cxu+1/μuy=0for xRy=0u(x,L)=0u(,y)=0. (2.2)

    It has exponential solutions that decay to 0 as x, i.e., solutions of the form ψLD(x,y)=eαxh(y) (we have to put here the dependence on L and D. If L=+ we may choose the function h as an exponential in y. We have

    ψLD(x,y)=eαLDxsh(βLD(y+L)),

    so that the exponents αLD and βLD satisfy

    {d(α2+β2)+cα=0Dα2+cα+βμcotanh(βL)=0. (2.3)

    Three types of limits will be considered.

    Case 1. The limit DL1, c bounded. We expect βLD to go to 0 as D+, so that

    βLDcotanh(βLDL)1L.

    Then (2.3) yields estimates of the form

    αLD1μLD,βLDcdμDL. (2.4)

    It is to be noted that these equivalents may be pushed up to c=o(D).

    Case 2. The limit DL1, cD. This time we expect βLD to go to infinity as D+, so that

    cotanh(βLD)1.

    We have

    αLD(cdD2)1/3,βLDμ(cD)2/3. (2.5)

    We have here a first occurrence of the critical order of magnitude cD.

    Case 3. The limit LD1, c1. We expect that LβLD will go to infinity, so that cotanh(βLD)1. In this setting, we have the estimates (2.5).

    In the construction of a travelling wave for (1.1), the first task is to bound the velocity from above. We will prove straight away the upper bound that will serve us in the later section, namely that cLD cannot exceed a (possibly large) multiple of D.

    Theorem 2.1. There is K>0, independent of D and L, such that

    cLDKD. (2.6)

    Proof. Assume cLDDL1, so that (2.5) hold for αLD and βLD. Set

    u_LD(x,y)=1ψ2LD(x,y).

    The function u_LD vanishes on the curve Γ_2LD whose equation is

    x=1α2LDln shβ2LD(y+2L).

    In particular, for y=L we have

    x=1α2LDln sh(β2LDL). (2.7)

    Translate the solution uLD(x,y) of (1.1) so that the leftmost point of ΓLD is located at the vertical of the origin. Then, from (2.7), Γ_2LD lies entirely to the left of ΓLD. By the maximum principle we have

    uLD(x,y)u_LD(x,y).

    Now, translate u_LD as much as this allows it to remain under uLD. If λLD is the maximum amount by which one may translate u_LD, there is a contact point (xLD,yLD) between uLD(x,y) and u_LD(xλLD,y).

    Still from (2.7), we have yLD>L. So, it now depends on whether we have yLD=0 or yLD(L,0). In the first case, let ΓLD intersect the line {y=0} at the point (˜xLD,0). If xLD=˜xLD, we have xuLD(xLD,0)=0 but we also have

    xuLD(xLD,0)<xu_LD(xLD,0)<0,

    an impossibility. If not, we have xLD<˜xLD and this time we have

    xuLD(xLD,0)=xu_LD(xLD,0),xxuLD(xLD,0)xxu_LD(xLD,0),yuLD(xLD,0)<yu_LD(xLD,0),

    the last inequality because of the Hopf lemma. This prevents the Wentzell relation from holding at (xLD,0). So, the only possibility left is yLD<0. By the strong maximum principle, the point (xLD,yLD) is on the free boundary ΓLD. We have

    yuLD(xLD,yLD)=β2LDeα2LD(xLDλLD)chβ2LD(yLD+2L),

    that is,

    |yuLD(xLD,yLD)|β2LDeα2LD(xLDλLD)shβ2LD(yLD+2L)=β2LD,

    because u_LD(xLD,yLD)=0. Thus, if D is sufficiently large, (2.5) implies

    |u_LD(xLD,yLD)|>1,

    contradicting the free boundary relation for uLD. This proves the theorem.

    Theorem 2.2. There is L0>0 such that the velocity cLD in Model (2.1) satisfies

    limD+cLD=+,

    uniformly with respect to LL0.

    The proof of Theorem 2.2 will be by contradiction. From now on, and until this has been proved wrong, we assume that, cLD is bounded, both with respect to L and D. Recall that ΓLD, the free boundary, is an analytic graph {x=φLD(y)}. Moreover, from [6], we may always assume that it intersects the line {x=0}, so that we may always assume φLD(0)=0. Define ˉxLD as the last x such that, for all y(L,0), then (x,y)ΓLD. Our main step is to prove that the front goes far to the left of the domain, this is expressed by the following lemma.

    Lemma 2.3. There is q>0 universal such that

    ˉxLDqD. (2.8)

    Proof. Recall that we have

    dφLDdy(0)>0.

    For every x<0, let yLD(x) be the first y such that x=φLD(y). We will prove that

    limε0yLD(εD)=0, uniformly in D2 and L2, (2.9)

    which implies Lemma 2.3. Assume (2.9) does not hold, and consider δ0>0 such that, for a sequence (Dn,Ln)n, going to +, and a sequence (εn)n going to 0 as n+, we have

    yLnDn(εnDn)δ0.

    Obviously, we must assume the boundedness from below of the sequence (εnDn)n. For every n set

    Xn=(εnDn,δ0/2), ˉXn=(εnDn,0).

    For ε0, to be chosen small in due time, the segment (ˉXn,Xn) is at distance at least δ0/2 from the free boundary. By nondegeneracy (recall the boundedness of (cLnDn)n), there is q0>0 universal such that, for all n:

    uLnDn(X)q0 on [ˉXn,Xn]. (2.10)

    On the other hand, recall that yuLnDn(x,0)1, this follows from [6]. So, the equation for uLnDn(x,0) reads, simply

    xxuLnDn(x,0)O(D1), u(0,0)=ux(0,0)=0.

    Thus we have

    uLnDn(εnDn,0)O(εn).

    This yields, for n large enough, the existence of δ0>0, universal, such that

    yuLnDn(x,0)δ0, x(ε0Dn,εnDn). (2.11)

    This comes from the Hopf boundary lemma. So now, we now write the equation for uLnDn(x,0) as

    xxuLnDn(x,0)o(1)DxuLnDn(x,0)=yuLnDn(x,0)Dδ0D,uLnDn(0,0)=xuLnDn(0,0)=0.

    Integrating this equation on (ε0Dn,0) and invoking (2.11) allows us to find a small constant δ0>0 such that

    uLnDn(ε0Dn,0)δ0,

    a contradiction.

    Proof of Theorem 2.2. Recall that we have still assumed the boundedness of (cLnDn)n. In order to allevite the notations a little, we omit the index n. Integration of (2.1) over ΣL yields

    (1/μ+L)cLD=ΓLDνuLD+|yuLD(x,L)|dx. (2.12)

    This expression has to be handled with care, because each integral, taken separately, diverges. We will see, however, that there is much less nonsense in (2.12) than it carries at first sight. The curve ΓLD has an upper branch, that we call ΓLD,+, and that connects (0,0) to XLD; the latter point being a turning point. The lower branch, called ΓLD,, connects XLD to (+,L+1), in other words it is asymptotic to the line {y=L+1} as x goes to +. We decompose (2.12) into

    (1/μ+L)cLD=ΓLD,+νuLDxLD|yuLD(x,L)|dx+(ΓLD,+νuLD+xLD|yuLD(x,L)|dx):=III+III.

    Let us remark that yuLD(x,L)1 for all xR. Indeed, the function uLD(x,y) being decreasing in x, it is larger than (1Ly)+. So, we have uLD(x,L)1. Now, the graph ΓLD has a discrete set of turning points, due to its analyticity. Away from these turning points, for xxLD, there is a finite number of ys: y1>y2>...ynD(x) such that (x,yi)ΓLD for 1inD. The point (x,y1) has already been counted in the integration over ΓLD,+, and so does not need to be counted again. Notice then that, if x is not the abscissa of a turning point, then nD (or nD2) is even, because of the configuration of ΓLD. So, uν(x,yi)=uν(x,yi+1)=1 for 1inD(x)2: it suffices to consider the sole point (x,ynD(x)) in the computation of III, however at this point we also have νuLD=1. Therefore we end up with III0. Of course the integral giving III converges, but we do not even have to bother to prove it.

    From Lemma 2.3, we have

    ICD,

    for some universal C>0. We already saw that III was nonnegative, so let us deal with II. By nondegeneracy, there is A>0 universal such that

    uLD(x,y)1AψLD(x,y),xxLD1.

    Notice that we do not change anything if, in II, we integrate up to xLD1 instead of xLD. We deduce, because ψLD(x,L)=0 and uLD(x,L)=1:

    yuLD(x,L)AyψLD(x,L),xxLD1.

    This implies

    IIO(1)+O(cDL)1/2.

    This yields

    cLDDLqLO(cDL)1/2,

    for a universal q, as soon as we choose DL1. This is an obvious contradiction. Now, note that, because of the Dirichlet condition at y=L, the sequence (cLD)L, for fixed D, is increasing. This ends the proof of the theorem.

    In this short section we recall a result of Hamel and Monneau that we will use to analyse the behaviour of the free boundary at infinity. Consider a solution (c>0,Γ,u) of the free boundary problem in the whole plane

    Δu+cxu=0in Ω:={u>0}|u|=1on Γ:=Ω. (3.1)

    The result is a classification of the solutions of (3.1) having certain additional properties. We rephrase it here to avoid any confusion, since the function u in [8] corresponds to 1u in our notations.

    Theorem 3.1. (Hamel-Monneau [8], Theorem 1.6) Assume that

    1). Γ is a C1,1 curve with globally bounded curvature,

    2). R2Ω has no bounded connected component,

    3). we have

    lim infd(X,Γ)+,XΩu(X)=1. (3.2)

    Then cc0 and, if we set

    sinα=c0c, (3.3)

    either u is the tilted one-dimensional solution ϕ0(ycosα±xsinα), or u is a conical front with angle α to the horizontal, that is, the unique solution uα(x,y) of (3.1) such that Γ is asymptotic to the cone Cα, with

    Cα={(x,y)R×R: x|y|cotanα}. (3.4)

    Note that the fact that uα is unique is not exactly trivial, it is given by Theorem 1.3 of [8]. Let us already notice that Properties 1 and 2 of this theorem are satisfied by the solution (cLD,ΓLD,uLD) of (2.1): Property 1 is clear, and Property 2 is readily granted by the monotonicity in x. Property 3 will be a little more involved to check.

    From then on, we fix D>0 large enough so that Theorem 2.2 holds. Pick L0 such that cL0D>c0, this implies cLDcL0D>c0. Notice that we have almost all the elements for the proof of Theorem 1.1, we just need, in addition, to control where the free boundary meets the line {y=0}. Since we have this freedom, we assume ΓLD to intersect the line {y=0} at the origin. Let ˉXLD=(ˉxLD,ˉyLD) the point of ΓLD that is furthest to the left, our sole real task will be to prove that XLD cannot escape too far as L+. Indeed, we notice that the property

    limxuLD(x,y)=1,uniformly with respect to y<0 and LL0 (3.5)

    holds easily. Indeed, from the maximum principle we have

    1uLD(x,y)(1eγx)+, (3.6)

    for all γ<cL0DD, as it is a subsolution to the equation for u in the plane and on the line {y=0}, and below uLD on the bottom line and also the vertical segment {x=0,Ly0}. Hence, at that point, we have almost everything for the construction of the wave in the whole half-plane, except the attachment property.

    Lemma 3.2. There is a constant KD>0 independent of L such that |ˉXLD|KD.

    Proof. Let us first assume that

    limL+ˉyLD=.

    Translate ΓLD and uLD so that ˉXLD becomes the new origin, the free boundary ΓLD meets therefore the horizontal line at the point (ˉxLD,ˉyLD). Up to a subsequence the triple (cLD,ΓLD,uLD) converges to a solution (cD,ΓD,uD) of the free boundary problem (3.1) in the whole plane. Moreover, the origin is its leftmost point. So, we may slide ϕ0 from x= to the first point where it touches uD, this can only be at ΓD, thus at the origin. But then we have

    ϕ0(0)=xuD(0,0)=1,

    a contradiction with the Hopf Lemma. So this scenario is impossible, and the family (ˉyLD)LL0 is bounded as L.

    To prove that the family (ˉxLD)LL0 is bounded, we consider

    ˉL2lim supL+(ˉyLD),

    and integrate the equation for uLD on ΣˉL=R×(ˉL,0), we obtain

    (1μ+1ˉL)cLD=ΓLDΣˉLνuLD+{u(x,ˉL)>0}yuLD(x,ˉL)dx:=I+II.

    From (3.6) and elliptic estimates, we have

    |yuLD(x,ˉL)|Ceγx,

    moreover the choice of ˉL implies that the rightmost point of {u(x,ˉL)>0 is at the left of the origin. Hence II is uniformly bounded with respect to L. On the other hand we have

    IˉxLD,

    which, from the uniform boundedness of cLD in Theorem 2.1, yields the boundedness of the family (ˉxLD)LL0.

    Proof of Theorem 1.1. We send L to infinity, a sequence (cLnD,ΓLnD,uLnD)n will converge to a solution (cD,ΓD,uD) of (1.1). Because of Lemma 3.2, the free boundary ΓD meets the line {y=0} at a point (ˉxD,0) and the expansion (1.4) is granted by Theorem 1.4 of [6]. Notice also that the uniform limit at is also granted because of (3.6). So, to finish the proof of the theorem, it remains to prove that, for all y<0, the positivity set of u only extends to a finite range. Such were it not the case, the limit

    uD(x)=limx+uD(x,y)

    would exist and be nonzero. It would solve the free boundary problem

    (uD)=0 on the positivity set,|(uD)|=1 at the free boundary points.

    This only allows for uD(y)=(y+a), a contradiction to the bounedness of uD.

    Let (cD,ΓD,uD) the solution constructed in Theorem 1.1 as one of the limits L+ of (cLD,ΓLD,uLD), with

    ΓD={(φD(y),y), y0},

    where φD is a smooth, locally Lipschitz function. Theorem 2.2 readily implies the first part of Theorem 1.3, that is

    limD+cD=+,

    simply because cLDcD. The only use of this result that we are going to make in this section is that there exists D0>0 such that cD>c0 for all DD0, which is the first part of Theorem 1.3. And so, we drop the indexes D in the rest of this section, for the simple reason that the dependence with respect to D will not appear anymore. To prove the second part of Theorem 1.3, we apply Theorem 3.1 to any sequence of translates of u:

    un(x,y)=u(φ(yn)+x,yn+y), (3.7)

    to infer that any possible limit of un is the one-dimensional wave, tilted in the correct direction. Properties 1 and 2 of the theorem being readily true, we concentrate on Property 3. The main step will be to prove that ϕ is in fact globally Lipschitz in (any sub-plane of) the half plane, once this is done a suitably designed Hamel-Monneau type [8] subsolution, placed under u, will give the property.

    Proposition 3.3. The function φ is globally Lipschitz in (,y0], for any y0<0.

    Proof. Note that the lemma is trivially false if we inisist in making y vary on the whole half-line. Also, as will be clear from the proof, the value of y0 will play no role as soon as it remains a little away from 0. So, we will assume for definiteness y0=1. The main step of the proposition consists in proving that no point of Γ{y1} may have a horizontal tangent. Assume that there is such a point X0=(φ(y0),0), y01. Translate u and Γ so that it becomes the origin, still denoting them by Γ and u. In a neighbourhood of 0 of size, say, ρ>0, Γ may be written (recall that it is an analytic curve) as

    ΓBρ={(x,ψ(x),|x|ρ}.

    We have ψ(0)=ψ(0)=0. Two cases have to be distinguished. The first one is ψ0 in [ρ,ρ]. By analyticity and Cauchy-Kovalevskaya's Theorem, this implies that ψ is defined and equal to 0 on the whole line; actually this case may happen only if X0 is a point at infinity, that is, u is a limit of translations of infinite size. But then we have

    u(x,y)=y+ or u(x,y)=y,

    depending on whether the positivity set of u is above or below Γ. This is in contradiction with the boundedness of u. The other case is ψ nonconstant in [ρ,ρ], so that u is nonconstant either in Bρ. By analyticity again, u has an expansion of the following type, in a neighbourhood of the origin:

    u(x,y)=xn+yP(x)+y2R(x,y), (3.8)

    with n2, the functions P and R being smooth in their arguments. Because xu it maximal at the origin, we have uxy(0,0)0 from the Hopf Lemma. Consider the situation where we have, for instance uy(0,0)=1, uxy(0,0)=β>0. We have therefore

    P(x)=1+βx+O(x2),

    and

    ψ(x)=xn1+βx+O(x2).

    Inside Bρ we also have

    u(x,y)=(nxn1+y(2β+O(x)+yxQ(x,y)1+2βx+O(x2)+y(y2R)(x,y))

    so that

    u(x,ψ(x)))=(nxn1+O(xn)1+2βx+O(x2))

    This implies, because n2:

    |u(x,ψ(x))|2=(12βx+O(x2))2+O(x2n2)=14βx+O(x2).

    Thus the free boundary relation cannot be satisfied in Bρ, except at the origin.

    We note that these situations exhaust what can happen on Γ. Indeed, if there is a sequence (Xn)n of Γ, going to infinity, such that the tangent to Γ at Xn makes an angle αn with the horizontal, with limn+αn=0, the usual translation and compactness argument yields a solution of the free boundary problem (3.1), where the tangent at the origin is horizontal. Once again we are in one of he above two cases, that are impossible.

    The last step is to check a property that will imply Property 3 for any limiting translation of (Γ,u) of the form (3.7). This is the goal of the next proposition.

    Proposition 3.4. We have

    lim infu(x,y)>0,y,d((x,y),Γ)+u(x,y)=1. (3.9)

    Proof. We use the notations of Theorem 3.1. Assume for definiteness that the leftmost point of Γ is located at Xl=(0,yl) with yl<0. Pick ˉXΩ, call M>0 its distance to Γ, and also set X0=(x0,N), both M and N will be assumed to be large, independently of one another. From Proposition 3.3 we claim the existence of a cone Cβ (the notation is given by (3.4)), the angle β>0 depending on the Lipschitz constant of φ, but independent of M and N, and a point ˉXM,NΩ, such that

    1). we have d(ˉXM,N,Γ)inf(M,N),

    2). we have ˉXˉXM,N+Cβ,

    3). the upper branch of ˉXM,N+Cβ meets the line of fast diffusion {y=0} at a point (xM,N,0) with

    xM,nγinf(M,N), (3.10)

    γ>0 independent of M and N.

    Two cases have to be distinguished, depending on whether ˉX is below (see Figure 1) or above (see Figure 2) its projection onto Γ (or, if the projection is not unique, the projection that we have selected). Note that, in the first case, there is C>0, once again independent of M and N, such that

    NCM,
    Figure 1.  ˉX below Γ.
    Figure 2.  ˉX above Γ.

    in such a case we have inf(M,N)=M. We now translate the origin to ˉXM,N. Note that the line of fast diffusion becomes the line {y=N}. Consider a smooth, nonpositive, concave, even function φM,N(y) such that

    1). we have φM,N(0)=0,

    2). we have φM,N(y)1inf(M,N),

    3). we have |φM,N(y)|=2cotanβ if |y|inf(M,N).

    Let ΓM,N be the graph

    ΓM,N={(x,y)R×R:x=inf(M,N)+φM,N(y)}.

    It meets the line of fast diffusion at a point ˜xM,N that satisfies an estimate of the form

    ˜xM,NxM,Nγinf(M,N), (3.11)

    with a universal γ>0, possibly different from that of (3.10).

    Pick now any small ε>0 and consider the function

    u_(x,y)=(1ε)(1eλ(xφM,N(y)))+.

    We claim that u_u in ΓM,N{yN}. Indeed, recall that we have

    limxxM,Nu(x,N)=1,

    this is a simple consequence of the uniform convergence of u(x,.) to 1 as x in the original variables, the fact that (xM,N,N) is at the left of the leftmost point of Γ, and (3.11), and the fact that u_1ε. Inside ΓM,N{yN}, we have

    Δu+cxu=(φM,N+λφM,N2λc)λeλ(xφM,N(y)(1inf(M,N)4λcotan2β+λc)λeλ(xφM,N(y)0 as soon as λ is small enough and M,N large enough.

    Therefore, u_u, so that we have, if we still call ˉX the original point ˉX translated by ˉXM,N:

    u(ˉX)1eλ(Minf(M,N)),

    which is the sought for estimate.

    Proof of Theorem 1.2. Proposition 3.3 proves that any limit of a sequence of translations (un)n in (3.4) satisfies the assumptions of Theorem 3.1. Therefore it is either a conical front uα (Case 1), a tilted one-dimensional wave ϕ0(ycosαxsinα) (Case 2), or a tilted wave ϕ0(ycosα+xsinα) (Case 3). We wish to prove that only Case 3 survives. Let us consider the set T of turning points of Γ, that is, the set of all points (φ(y),y) such that φ(y)=0. From the analyticity of Γ, T is discrete:

    T={(φ(yn),yn), nN}.

    If we manage to prove that it is finite we are done because this excludes Case 1 trivially, and Case 2 because u(x,y)1 as x, uniformly in y. So, assume that T is infinite. We claim the existence of two sequences of T, decreasing to , (yin)n, i{1,2}, such that

    ● we have y1n+1y2ny1n for all n,

    ● we have limn+(y2ny1n+1)=limn+(y1ny2n)=+,

    such that, if we set xin=φ(yin) we have

    limn+u(xin+x,yin,y)=uα(x,y),

    uniformly on compact sets in (x,y). Indeed, there is N0N such that T may be organised in disjoint clusters

    T=nNΛn, yz if (φ(y),y)Λn+1(φ(z),z)Λn,

    each Λn having at most N0 elements, and

    limn+(min(y,φ(y))Λnymax(z,φ(z))Λnz)=+.

    This is because of the fact that uα has a finite number of turning points and that the convergence of u to its limits implies the convergence of the free boundaries in C1 norms. We claim that (x1n,y1n) and (x2n,y2n) cannot be in two consecutive clusters, because of the orientation of Cα. So, if Λin is the cluster of (xin,yin), let Λ3n be a cluster in-between. Let (x3n,y3n) be the leftmost point of Λ3n. From the definition of Λ3n, there is Rn+ such that (x3n,y3n) is the leftmost point of Γ in BRn(x3n,y3n). Let u(x,y) be a limit of translations of u with the sequence of points (x3n,y3n), it does not converge to any of the functions prescribed by Theorem 3.1, which is a contradiction. This finishes the proof of Theorem 1.2.

    The first thing one must understand is (1.2) in the truncated cylinder ΣL, namely

    {dΔv+cxv=0(x,y)ΣL{v>0}|v|=1((x,y)Γ:={v>0}ΣL Du+cu+1/μvy=0for xRy=0dvy(x,0)=μu(x)v(x,0)for xR v(,y)=1,v(+,y)=(1yL)+v(x,L)=1. (4.1)

    System (4.1), linearised around 0, reads

    {dΔv+cxv=0(x,y)ΣL Du+cu+1/μvy=0for xRy=0dvy(x,0)=μu(x)v(x,0)for xRv(x,L)=0v(,y)=0. (4.2)

    The solutions that decay to 0 as x are looked for under the form

    (ϕLD(x),ψLD(x,y))=eαLDx(1,γsh(βLD(y+L)),

    so that the exponents αLD and βLD satisfy

    {d(α2+β2)+cα=0Dα2+cα+dβμch(βL)sh+dβch(βL)=0. (4.3)

    Once again we consider types of limits.

    Case 1. The limit DL1, c bounded. We expect βLD to go to 0 as D+, so that

    dβch(βL)sh(βL)+dβch(βL)1L.

    Then (4.3) yields estimates of the form

    αLDμLD,βLDcμdDL. (4.4)

    And, once again, the estimate may be pushed up to c=o(D).

    Case 2. The limit DL1, cD. This time we expect βLD to go to infinity as D+, so that

    dβch(βL)sh(βL)+dβch(βL)1.

    We have

    αLD(cdD2)1/3,βLD1μ(cD)2/3. (4.5)

    Case 3. The limit LD1, c1. We expect that LβLD will go to infinity, so that dβμch(βL)sh(βL)+dβch(βL)1. In this setting, we have the estimates (4.5).

    Let (cLD,ΓLD,vLD) a solution to (4.1), we may infer its existence from a once again slight modification of Theorem 1.1 of [6]. We assume that it meets the line {y=0} at the point (0,0).

    Theorem 4.1. There is a constant K>0, independent of L and D, such that

    cLDKD. (4.6)

    Moreover there is L0>0 such that we have

    limD+cLD=+, (4.7)

    uniformly in LL0.

    Proof. That of (4.6) is similar to that of Theorems 2.1, one compares (uLD,vLD) to

    (u_LD(x),v_LD(x,y))=(1μϕLD(x),1ψLD(x,y)).

    As for (4.7), the only point to prove is that, under the assumption that (cL0D)D is bounded, the width L0 being large but fixed, then the leftmost point of ΓLD, denoted by (ˉxL0D,ˉyL0D) satisfies

    ˉxL0Dδ0D. (4.8)

    Once this is proved, the proof of the theorem proceeds much as that of Theorem 2.2, using in particular estimates (4.4) for the linear exponentials. So, assume (4.8) to be false, that. is, there are two diverging positive sequences (Dn)n and (xn)n, and a positive constant d0 such that

     limn+xnDn=0,minxxn{y<0: (x,y)ΓL0Dn}d0. (4.9)

    By nondegeneracy, there is γ>0 such that

    v(x,d02)γ for xxn. (4.10)

    Choose another sequence (xn)n such that

    limn+xnxn=limn+xnDn=0.

    Then we have

    limn+uL0DnL(xn,xn)=0,

    just by integrating the ODE for uL0Dn. But then, the Robin condition yvL0Dn+vL0Dn=o(1) on (xn,xn)×{0}, together with (4.9) and the Hopf Lemma, yields the existence of γ>0 such that

    vL0Dn(x,0)γ for xnxxn.

    The equation for uL0Dn becomes

    (uL0Dn)+cDnL0(uL0Dn)μ1(xn,0)(x)γ21(xn,xn),

    as soon as n is large enough. This implies uDnL0(xn)<0, contradiction.

    From then on, the rest of the study of the two species model parallels exactly that of the one species model.

    In this section, we assume that all the requirements on the coefficients are fulfilled, and we drop the index D for the velocity c, the free boundary Γ and the solution u. Theorem 1.4 is proved by deriving a differential ineqaulity for φ, exploiting Theorem 1.3 and the fact that φ as a limit at infinity. This allows indeed to write the free boundary problem for u in a suitable perturbative form, and translate the double Dirichlet and Neumann boundary condition into the sought for differential inequality.

    In the whole section, the considerations will be rigorously identical for the one species model or the two species model, except the global estimate in Proposition 5.2 below, where the computations are slightly different - but left to the reader. Thus we will concentrate on the one species model (1.1).

    Recall from [1] that Γ is an analytic graph. Theorem 1.3 readily implies

    Proposition 5.1. We have

    limyφ(y)=limyφ(y)=0.

    This entails the following improvement of Proposition 3.4.

    Proposition 5.2. There is ρ>0 such that, if u(x,y)>0 we have

    u(x,y)1eρdist((x,y),Γ).

    Proof. For ε>0 consider ψε(y) smooth whose derivative ψε satisfies

    {ψε(y)=φ(y) if y1ε4ψε(y)=ε if 1ε2yy0εψε(y)0 everywhere,|ψε|ε everywhere.

    Note that this function exists due to Theorem 1.3 and Proposition 5.1. For ρ>0 consider

    ˉu(x,y)=eρ(xψε(y)),

    it can be estimated by eδερdist((x,y),Γ) for a suitable δε>0, because of Theorem 1.3 again. In the lower half plane we have

    dΔˉu+cxˉu=(d(ρψε+ρ(ψε)2)+c)ρeρ(xψε(y))0 if ρ(0,c) and if ε small enough.

    On the line we have, on the same pattern:

    (Dxx+cx+1μy)ˉu=(ρD+cεμ)ρeρx0 if ρ(0,cD) and if ε small enough.

    And so, as soon as ρ(0,cD) and ε>0 is small enough, the function u_:=(1ˉu)+ is a sub-solution to the equations for u in the region {u>0}, moreover {ˉu>0} coincides Γ sufficiently far in the lower half plane. The maximum principle implies the proposition.

    Proof of Theorem 1.4. From now on, translate the origin so that we are in the following situation: ΓR2 is the graph

    {(x,ϕ(x)), x>0},

    with ϕ(x)<0 and limx+ϕ(x)=tanα. Actually, the translation may be adjusted so that ϕ(x) is as colse as we wish to tanα, this will be quantified later. The following chain of transformations is then made.

    1). Rotate the coordinates by the angle α, so as to obtain the new set (x,Y) given by

    X=xcosαysinα,  Y=xsinα+ycosα.

    In this new system the free boundary may be written as Y=ψ(X), with, by Proposition 5.1:

    limX+ψ(X)=limX+ψ(X)=limX+ψ(X)=0.

    2). Straighten the free boundary by setting

    X=X, Y=ψ(X).

    In this new coordinate system, u solves the over-determined problem

    XXu(1+ψ2)YYu+ψYu2ψXYu+ccosαXu+(c0cψcosα)Yu=0(X>0,Y<0)u(X,0)=0,  Yu(X,0)=11+ψ2(X) (5.1)

    3). A standard compactness/uniqueness argument shows that

    limX+u(X,Y)=ϕ0(Y),

    uniformly in Y<0. As we will not need any additional change of coordinates, let us, for notational simplicity, revert to the initial notation

    x:=X, y:=Y.

    The function u(x,y) is this looked for under a perturbation of ϕ0(y): u(x,y)=ϕ0(y)+˜u(x,y). The free boundary condition writes

    yu(x,0)=ψ2(x)1+ψ2(x).

    The full PDE for ˜u will not be written, as we need another change of unknowns.

    4). In order to transform (5.1) into an over-determined problem with fixed boundary, we look for ˜u under the form

    ˜u(x,y)=v(x,y)ψ2(x)1+ψ2(x)+1+ψ2(x)γ0(y),

    with γ0 smooth, compactly supported, γ(0)=0, γ(0)=1. As ψ appears in the equations only in the form of ψ, we set h(x)=ψ(x).

    The system for (h,v) is thus

    xxv(1+h2)yyv+hyv2hxyv+ccosαxv+(c0chcosα)yv=(h+hccosα)ϕ0+R[h](x,y)(x>0,y<0)v(x,0)=0, yv(x,0)=0, (5.2)

    the function R[h](x,y) being:

    R[h](x,y)=(h21+h2+1+h2)+h2γ0(y)1+h21+1+h2h2hγ0(y)1+h2+1+h2+2h(h21+h2+1+h2)γ0(y)ccosα(h21+h2+1+h2)γ0(y)(c0hccosα)h2γ0(y)1+h2+1+h2+h2ϕ0(y)

    While the expression of R[h](x,y) is utterly unpleasant, its structure is quite simple: It is quadratic in h and its derivatives up to order 2, which are known to vanish at infinity. Finally, set

    S[h,v](x,y)=h2vyyhvy+2hvxy+hccosαvy.

    By elliptic regularity, all derivatives of v go to 0 as x+, uniformly in y. Problem (5.3) now reads

    Δv+ccosαvx+c0vy=(h+hccosα)ϕ0+R[h](x,y)+S[h,v](x,y)(x>0,y<0)v(x,0)=0, yv(x,0)=0, (5.3)

    We will use the estimate

    |R[h](x,y)|+|S[h,v](x,y)|θ(x,y)(|h(x)|+|h(x)|)1(0,1)(y), (5.4)

    with θ(x,y) a positive function that tends to 0, uniformly in y, as x+. It only remains to apply Proposition A1 of the appendix together with (5.4)), with the data

    f(x)=h(x)h(x)ccosα,  g(x,y)=R[h](x,y)+S[h,v](x,y).

    From Proposition 5.2, estimate (A1) holds, so that the function h solves the differential inequality

    |h(x)h(x)ccosα|˜θ(x)(|h(x)|+|h(x)|),

    the function ˜θ(x) going to 0 as x+. This implies the exponential estimate.

    L. A. Caffarelli is supported by NSF grant DMS-1160802. The research of J.-M. Roquejoffre has received funding from the ERC under the European Union's Seventh Frame work Programme (FP/2007-2013) / ERC Grant Agreement 321186 - ReaDi. He also acknowledges a long term visit at UT Austin during the academic year 2018-19, made possible by a delegation within the CNRS for that year, and a J. T. Oden fellowship.

    All authors declare no conflicts of interest in this paper.

    The goal of this section is to prove that the right handside of the solution of an elliptic equation in the lower half plane, both with Dirichlet and Neumann condition, has to satisfy an integral equation.

    Proposition A1. Consider f(x) and g(x,y) two smooth, bounded functions defined on R+ (resp. R+×R). Set e0(y)=ec0y. Assume the existence of ρ>0 such that

    |g(x,y)|=O(eρy), y0, uniformly in x R+. (A1)

    Let u(x,y) solve

    {Δu+ccosαux+c0uy=f(x)e0(y)+k(x,y)  (x>0, y0)u(x,0)=uy(x,0)=0, (A2)

    in the classical sense. Also assume that u and its derivatives satisfy the estimate (A1). There is a continuous function K(x,y), defined on [2,+)×R, such that, for all ε>0 we have

    |K(x,y)|=O(e(βρε)yρx),  x2, y0,

    and such that

    f(x)=0K(x,y)g(x,y)dy,  x2.

    Proof. First, transform (A2) into a problem on the whole line by setting

    v(x,y)=γ(x)u(x,y),

    where γ is a smooth function, supported on (1,+), and equal to 1 on [2,+). the equation for v is thus

    {Δv+ccosαvx+c0vy=γ(x0f(x)e0(y)+k(x,y)  (xR, y0)v(x,0)=vy(x,0)=0, (A3)

    with

    k(x,y)=γ(x)g(x,y)+γ(x)u(x,y)+2γ(x)ux(x,y).

    In other words, the right handside of (A3) equals that of (A2) as soon as x2.

    A second observation is that the general Dirichlet problem

    {Δu+ccosαux+c0uy=F(x,y)  (x>0, y0)u(x,0)=0u(0,y)=u0(y) (A3)

    is well-posed as soon as the Dirichlet datum u0(y) and the right handside F(x,y) satisfy the estimate (A1), with any ρ<c0. Indeed, the classical change of unknown u(x,y)=eρyv(x,y) changes (A3) into an elliptic equation the the zero order term ρc0ρ2. Thus it is enough to assume tht g and k are compactly supported in y, in order to perform Fourier transforms. An easy density argument then allows to treat general data satisfying (A1).

    Set w(x,y)=e(ccosαx+c0y)/2v(x,y), the equation for w is now

    {Δw+β2w=F(x)e0(y)+G(x,y)  (xR, y0)w(x,0)=wy(x,0)=0. (A4)

    The new data are

    F(x)=eccosα/x/2γ(x)f(x), G(x,y)=e(ccosαxc0y)/2k(x,y), E0(y)=ec0y/2,

    and we have set

     β2=c20+c2cos2α4.

    Thanks to both Dirichlet and Neumann conditions, the function w(x,y) may be extended evenly over the whole plane R2, where it solves the same equation as (A4), with G and E0 also extended evenly. The new unknown and data are still denoted by w, E0 and G. Let ˆw(ξ,ζ) be the Fourier transform of w, as well as ˆF(ξ), ˆE0(ζ) and ˆG(ξ,ζ) the Fourier transforms of the data. Thus we have

    ˆw(ξ,ζ)=ˆF(ξ)ˆE0(ζ)+ˆG(ξ,ζ)β2+ξ2+ζ2.

    Now, the Dirichlet condition for w entails Rˆw(ξ,ζ)dζ=0, that is

    ˆF(ξ)=(RˆE0(ζ)β2+ξ2+ζ2dζ)1RˆG(ξ,ζ)β2+ξ2+ζ2dζ.

    From Plancherel's Theorem, we have

    RˆE0(ζ)β2+ξ2+ζ2dζ1c0/2+β2+ξ2,

    and, if we still denote by ˆG(ξ,y) the partial Fourier transform of G with respect to x, we have, again by Plancherel's theorem

    RˆE0(ζ)β2+ξ2+ζ2dζ+ˆG(ξ,y)e|y|β2+ξ2dy.

    Inverting the Fourier transform in x yields, by the residue theorem:

    F(x)=∝eccosα|x|/2R2G(x,y)eixξ|y|β2+ξ2dξdy.

    Shifting the integration line in ξ from R to R+iρ yields, for x>0:

    Reixξ|y|β2+ξ2dξ=eρxReixξ|y|β2+ξ2ρ2+2iρξdξ

    and β2+ξ2ρ2+2iρξ has nonzero real part. This entails the proposition.



    [1] Alt HW, Caffarelli LA (1981) Existence and regularity for a minimum problem with free boundary. J Reine Angew Math 325: 105-144.
    [2] Bonnet A, Hamel F (1999) Existence of nonplanar solutions for a simple model of premixed Bunsen flames. SIAM J Math Anal 31: 80-118.
    [3] Berestycki H, Coulon AC, Roquejoffre JM, et al. (2014) Speed-up of reaction fronts by a line of fast diffusion, In: Séminaire Laurent Schwartz, Equations aux Dérivés Partielles et Applications, 2013-2014, Palaiseau: Ecole Polytechnique, exposé XIX.
    [4] Berestycki H, Roquejoffre JM, Rossi L (2013) The influence of a line with fast diffusion on FisherKPP propagation. J Math Biol 66: 743-766.
    [5] Berestycki H, Roquejoffre JM, Rossi L (2013) Fisher-KPP propagation in the presence of a line: Further effects. Nonlinearity 26: 2623-2640.
    [6] Caffarelli LA, Roquejoffre JM, The leading edge of a free boundary interacting with a line of fast diffusion. arXiv: 1903.05867.
    [7] Coulon Chalmin AC (2014) Fast propagation in reaction-diffusion equations with fractional diffusion, PhD thesis. Available from: thesesups.ups-tlse.fr/2427/.
    [8] Hamel F, Monneau R (2002) Existence and uniqueness of solutions of a conical-shaped free boundary problem in R.2. Interface Free Bound 4: 167-210.
    [9] Hamel F, Monneau R, Roquejoffre JM (2004) Stability of travelling waves in a model for conical flames in two space dimensions. Ann Sci Ecole Norm Sup 37: 469-506.
    [10] Hamel F, Monneau R, Roquejoffre JM (2006) Uniqueness and classification of travelling waves with Lipschitz level lines in bistable reaction-diffusion equations. Discrete Contin Dyn Syst A 14: 75-92.
  • Reader Comments
  • © 2021 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(4075) PDF downloads(554) Cited by(0)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog