Processing math: 100%
Research article Special Issues

Local Calderón-Zygmund estimates for parabolic equations in weighted Lebesgue spaces

  • We prove local Calderón-Zygmund type estimates for the gradient of weak solutions to degenerate or singular parabolic equations of p-Laplacian type with p>2nn+2 in weighted Lebesgue spaces Lqw. We introduce a new condition on the weight w which depends on the intrinsic geometry concerned with the parabolic p-Laplace problems. Our condition is weaker than the one in [13], where similar estimates were obtained. In particular, in the case p=2, it is the same as the condition of the usual parabolic Aq weight.

    Citation: Mikyoung Lee, Jihoon Ok. Local Calderón-Zygmund estimates for parabolic equations in weighted Lebesgue spaces[J]. Mathematics in Engineering, 2023, 5(3): 1-20. doi: 10.3934/mine.2023062

    Related Papers:

    [1] Cristiana De Filippis . Optimal gradient estimates for multi-phase integrals. Mathematics in Engineering, 2022, 4(5): 1-36. doi: 10.3934/mine.2022043
    [2] María Ángeles García-Ferrero, Angkana Rüland . Strong unique continuation for the higher order fractional Laplacian. Mathematics in Engineering, 2019, 1(4): 715-774. doi: 10.3934/mine.2019.4.715
    [3] Lucas C. F. Ferreira . On the uniqueness of mild solutions for the parabolic-elliptic Keller-Segel system in the critical $ L^{p} $-space. Mathematics in Engineering, 2022, 4(6): 1-14. doi: 10.3934/mine.2022048
    [4] Nicolai Krylov . On parabolic Adams's, the Chiarenza-Frasca theorems, and some other results related to parabolic Morrey spaces. Mathematics in Engineering, 2023, 5(2): 1-20. doi: 10.3934/mine.2023038
    [5] 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
    [6] Márcio Batista, Giovanni Molica Bisci, Henrique de Lima . Spacelike translating solitons of the mean curvature flow in Lorentzian product spaces with density. Mathematics in Engineering, 2023, 5(3): 1-18. doi: 10.3934/mine.2023054
    [7] David Cruz-Uribe, Michael Penrod, Scott Rodney . Poincaré inequalities and Neumann problems for the variable exponent setting. Mathematics in Engineering, 2022, 4(5): 1-22. doi: 10.3934/mine.2022036
    [8] Arthur. J. Vromans, Fons van de Ven, Adrian Muntean . Homogenization of a pseudo-parabolic system via a spatial-temporal decoupling: Upscaling and corrector estimates for perforated domains. Mathematics in Engineering, 2019, 1(3): 548-582. doi: 10.3934/mine.2019.3.548
    [9] François Murat, Alessio Porretta . The ergodic limit for weak solutions of elliptic equations with Neumann boundary condition. Mathematics in Engineering, 2021, 3(4): 1-20. doi: 10.3934/mine.2021031
    [10] Pengfei Guan . A weighted gradient estimate for solutions of $ L^p $ Christoffel-Minkowski problem. Mathematics in Engineering, 2023, 5(3): 1-14. doi: 10.3934/mine.2023067
  • We prove local Calderón-Zygmund type estimates for the gradient of weak solutions to degenerate or singular parabolic equations of p-Laplacian type with p>2nn+2 in weighted Lebesgue spaces Lqw. We introduce a new condition on the weight w which depends on the intrinsic geometry concerned with the parabolic p-Laplace problems. Our condition is weaker than the one in [13], where similar estimates were obtained. In particular, in the case p=2, it is the same as the condition of the usual parabolic Aq weight.



    Dedicated to Giuseppe Rosario Mingione, on the occasion of his 50th birthday.

    We study local regularity theory for weak solutions to the following parabolic equations of p-Laplacian type:

    utdiva(Du)=div(|F|p2F)in  ΩT:=Ω×(0,T], (1.1)

    where ΩRn (n2) is an open set, T is a positive constant, u=u(x,t) is a real valued function with (x,t)Ω×(0,T]=ΩT, ut is the partial derivative of u with respect to the time variable t, and DuRn is the gradient of u with respect to the space variable x (i.e., Du=Dxu). For given p(1,), we assume that a:RnRn satisfies the following p-growth and p-ellipticity conditions:

    |a(ξ)|+|Dξa(ξ)||ξ|L|ξ|p1 (1.2)

    and

    Dξa(ξ)ηην|ξ|p2|η|2, (1.3)

    for every ξ,ηRn{0}, and for some constants ν and L with 0<ν1L. The prototype of a is

    a(ξ)=|ξ|p2ξ.

    Lq-regularity theory with Calderón-Zygmund estimates for partial differential equations is originated from the classical result of Calderón and Zygmund [16] about the boundedness of linear operators including the Laplace operator. For the following p-Laplacian type equation

    div(|Du|p2Du)=div(|F|p2F),

    a fundamental Lq-regularity theory, which is also called (nonlinear) Calderón-Zygmund theory, is to show the following implication:

    |F|pLq    |Du|pLq,q>1, (1.4)

    and obtain corresponding estimates, so-called Calderón-Zygmund estimates. In this regard, Iwaniec [26] first obtained Calderón-Zygmund estimates in the whole space Rn when p2, and DiBenedetto and Manfredi [21] extended this result to the corresponding system with 1<p<. Thereafter, Caffarelli and Peral [15] considered general elliptic equations with p-growth, for instance, the stationary case of (1.1), applying a new approach by means of maximal functions and a covering argument obtained by Krylov and Safonov based on the Calderón-Zygmund decomposition. We further refer to e.g., [1,11,14,17,28] for Lq-regularity theory with corresponding Calderón-Zygmund estimates for elliptic problems.

    Difficulty of the study on regularity theory for parabolic problems of p-Laplacian type is originated from the absence of the scaling invariant property: a constant multiple of a solution of the parabolic p-Laplace equation utdiv(|Du|p2Du)=0 does not become a solution. It could be overcome by considering intrinsic parabolic cylinders depending on solutions, instead of the usual parabolic cylinders. This idea was introduced by DiBenedetto and Friedman in [19,20], see also the monograph [18], where Hölder regularity for parabolic p-Laplace systems had been established.

    For the Lq-regularity theory, on the other hand, the approaches used in [26] and [15] are not directly applicable to the parabolic p-Laplacian type problems when p2, since the intrinsic geometry prevents the use of maximal functions. Finally, Acerbi and Mingione [2] established local Calderón-Zygmund estimates when p>2nn+2 with a new approach, hence proved the implication (1.4) in the local sense. We also refer to the higher integrability result of Kinnunen and Lewis in [27], where the implication (1.4) is obtained when q is sufficiently close to 1. Note that the condition p>2nn+2 is essential since there exists an unbounded weak solution to the parabolic p-Laplace system in this case, see [18]. It is worth pointing out that the approach in [2] does not employ maximal functions and the covering argument by Krylov and Safonov but a new covering argument used in [31] which is based on the Vitali covering lemma. It has led to the development of the Calderón- Zygmund theory for parabolic problems. For instance, we refer to [5,7,10] for global Calderón Zygmund theory in bounded domains, [6,33] for parabolic obstacle problems, [4,9] for parabolic problems with variable exponent, [25,32] for parabolic problems with growth and [35] for parabolic variational problems.

    Research on Calderón-Zygmund estimates in general function spaces such as weighted Lebesgue spaces, Orlicz spaces, variable exponent Lebesgue spaces, Lorentz spaces have been actively conducted for the last decade, e.g., [3,8,12,29,36]. In particular, estimates in weighted Lebesgue spaces are crucial since these imply estimates in various function spaces by extrapolation argument, see [24, Section 5]. For parabolic problems with p-growth as in (1.1), Byun and Ryu [13] obtained global Calderón-Zygmund estimates in the weight Lebesgue spaces Lqw hence proved the following implication:

    |F|pLqw    |Du|pLqw, (1.5)

    with q>1 and the weight w satisfying the following Muckenhoupt type condition:

    supQC(Qwdz)(Qw1q1dz)q1<, (1.6)

    where C is the set of all cylinders of the form Br(x0)×(t1,t2)Rn×R. On the other hand, if p=2, the same implication can be obtained for every usual parabolic Aq weight w, that is, w satisfies (1.6) with C the set of all parabolic cylinders of the form Br(x0)×(t0r2,t0+r2). Therefore, there is a drastic change between the conditions of weights when p=2 and p2.

    In this paper, we introduce a new parabolic Muckenhoupt type condition depending on the intrinsic geometry concerned with the parabolic p-Laplacian setting, see Definition 2.1. We emphasize that our condition on weights depends on p, and is weaker than the one in [13] and exactly the same as the parabolic Aq condition when p=2. With this condition we prove the implication (1.5) in the local sense by obtaining corresponding Calderón-Zymund estimates.

    Now, we state our main result. Notation and the definition of the p-intrinsic Aq weight are introduced in next section. We say that uC0(0,T;L2(Ω))Lp(0,T;W1,p(Ω)) is a weak solution to (1.1) if

    ΩTuζtdz+ΩTa(Du)Dζdz=ΩT|F|p2FDζdz

    holds for every ζC0(ΩT).

    Theorem 1.1. Let p>2nn+2 and uC0(0,T;L2(Ω))Lp(0,T;W1,p(Ω)) be a weak solution to (1.1) with FLp(ΩT,Rn). If w is a p-intrinsic Aq weight with q>1 and |F|pLqw,loc(ΩT), then |Du|pLqw,loc(ΩT).

    Furthermore, there exists R0=R0(n,ν,L,p,q,[w]q,Du,F)>0 such that for every Q2rΩT with 2r<R0,

    (1w(Qr)Qr|Du|pqwdz)1qc(Q2r[|Du|p+|F|p+1]dz)d+c(1w(Q2r)Q2r|F|pqwdz)1q (1.7)

    for some c=c(n,ν,L,p,q,[w]q)>0, where

    d:={2pp(n+2)2n,if2nn+2<p<2,p2,ifp2. (1.8)

    Remark 1.1. In the above theorem, R0 will be chosen as in (3.1). Furthermore, when p=2 we may put R0=.

    Remark 1.2. (Possible extensions) In this paper, we deal with only scalar problems without coefficients for simplicity. We can consider more general problems such as general non-autonomous parabolic equations with p-growth

    utdiva(x,t,Du)=div(|F|p2F),

    where a(x,t,ξ) satisfies (1.2), (1.3) and a VMO condition (see [10]), and parabolic p-Laplace systems with coefficients

    utdiv((A(x,t)Du:Du)p22A(x,t)Du)=div(|F|p2F),

    where u:ΩTRN and A(x,t):Rn+1Rn2N2 satisfies a VMO condition (see [2]). Moreover, as in [13], we can also consider global Calderón-Zygmund estimates in Reifenberg flat domains.

    The remaining part of the paper is organized as follows. In Section 2, we introduce notation, weights with their main assumption, and comparison and regularity estimates for corresponding homogeneous problems. In Section 3, we prove our main theorem, Theorem 1.1.

    Let z0=(x0,t0)Rn×R with x0=(x10,,xn0), r,α,λ>0 and 1<p<. We define an α-parabolic cylinder by Qr,α(z0)=Br(x0)×(t0αr2,t0+αr2) and an α-parabolic cube by ˜Qr,α(z0)=Cr(x0)×(t0αr2,t0+αr2), where Br(x0):={xRn:|xx0|<r} and Cr(x0):={x=(x1,,xn)Rn:max{|x1x10|,,|xnxn0|}<r}. Note that Qr(z0):=Qr,1(z0) and ˜Qr(z0):=˜Qr,1(z0) is the usual parabolic cylinder and cube, respectively, and we denote pQr(z0):=(Br(x0)×{t=t0r2})(Br(x0)×[t0r2,t0+r2)) Furthermore, when α=λ2p we write Qλr(z0)=Qr,λ2p(z0)=Br(x0)×(t0λ2pr2,t0+λ2pr2) which is usually called a p-intrinsic parabolic cylinder since we will consider λ related to the weak solution to (1.1).

    For an integrable function f:URm with URn+1 and 0<|U|<, we write (f)U=Ufdz:=1|U|Ufdz, where |U| is the Lebesgue measure of U in Rn+1.

    We say that w:Rn+1R is a weight if it is nonnegative and locally integrable. For a weight w and a bounded open set URn+1, we write

    w(U):=Uwdz,

    and define by weighted Lebesgue space Lqw(U), 1q<, the set of all measurable function f on U such that

    fLqw(U):=(U|f|qwdz)1q<.

    We introduce the assumption of the weight w in Theorem 1.1.

    Definition 2.1. Let p,q(1,). We say that weight w:Rn+1R is a p-intrinsic parabolic Aq weight if it satisfies that

    [w]q:=supQCp(Qwdz)(Qw1q1dz)q1<, (2.1)

    where

    Cp:={Qr,α(z0):z0Rn+1,  r>0,  α=λ2p,  1λmax{1,rn+22}}.

    Here, the α-parabolic cylinders Qr,α(z0) can be replaced by the α-parabolic cubes ˜Qr,α(z0).

    Note that in the definition of the class Cp, the range of α with respect to p is following:

    {1αmax{1,r(p2)(n+2)2}if  p<2,α=1if  p=2,min{1,r(p2)(n+2)2}α1if  p>2.

    Hence, Cp contains all the parabolic cylinders Qr(z0) and, in particular, C2 (i.e., p=2) consists of only the parabolic cylinders. Moreover, since rn+22ρn+22 for ρ(0,r], we have

    Qr,αCp    Qρ,αCp  for every  ρ(0,r].

    From this fact, we can obtain the following properties for p-intrinsic parabolic Aq weights, which are well known properties of the usual Aq weights, see e.g., [23, Section 7.2]. Proofs are exactly the same as the ones in there with just replacing cubes and the dimension n by the α-parabolic cubes and n+2, respectively. Therefore, we omit their proofs.

    Proposition 2.1. Let p,q(1,) and w:Rn+1R be a p-intrinsic parabolic Aq weight.

    (1) For every fLqw(Q) with QCp,

    (Q|f|dz)q[w]qw(Q)Q|f|qwdz. (2.2)

    (2) There exist γ,c>0 depending on n, q and [w]q such that

    (Qw1+γdz)11+γcQwdz.

    (3) There exist γ1>0 and c1,c21 depending on n, q and [w]q such that for every QCp and EQ

    1c1(|E||Q|)qw(E)w(Q)c2(|E||Q|)γ1. (2.3)

    (4) w is a p-intrinsic parabolic Aq1 weight for every q1>q. Moreover, w is a p-intrinsic parabolic Aq weight for some q(1,q), where q and [w]q depend on n, q and [w]q.

    Example. On Rn+1, the function w(x,t)=max{|x|,|t|}γ is a p-intrinsic Aq weight for p2, q(1,) when n<γ<n(q1). Indeed, we first write

    I[Q]:=(Qmax{|x|,|t|}γdz)(Qmax{|x|,|t|}γq1dz)q1,QCp.

    We divide the α-parabolic cylinders Qr,λ2p(z0) in Cp for z0=(x0,t0)Rn+1 into three cases:

    (i) min{|x0|,|t0|}3r,

    (ii) |x0|<3r and 3λ2p2r|t0|<3r,

    (iii) |x0|<3r and |t0|<3λ2p2r.

    Note that if xBr(x0) and |x0|3r,

    23|x0||x0||xx0||x||xx0|+|x0|43|x0|,

    and if t0λ2pr2tt0+λ2pr2 and |t0|>3λ2p2r,

    83|t0||t0||tt0||t||tt0|+|t0|103|t0|.

    In Case (i), if (x,t)Qr,λ2p(z0), we have |x||x0| and |t||t0|, hence

    I[Qr,λ2p(z0)]c(Qr,λ2p(z0)max{|x0|,|t0|}γdz)(Qr,λ2p(z0)max{|x0|,|t0|}γq1dz)q1=1=c.

    In Case (ii), if (x,t)Qr,λ2p(z0), then |x|<4r and |t||t0|<9r2, and hence

    I[Qr,λ2p(z0)]c(t0+λ2pr2t0λ2pr2B4r(0)max{|x|,|t0|}γdxdt)×(t0+λ2pr2t0λ2pr2B4r(0)max{|x|,|t0|}γq1dx)q1c[1rn(|t0|0|t0|γ2ρn1dρ+4r|t0|ργ+n1dρ)]×[1rn(|t0|0|t0|γ2(q1)ρn1dρ+4r|t0|ργq1+n1dρ)]q1crnq[γ|t0|γ+n2n(n+γ)+(4r)γ+nγ+n][γ|t0|nqnγ2(q1)n(nqnγ)+(q1)(4r)nqnγq1nqnγ]q1crnqrγ+nrnqnγ=c,

    where we have used n<γ<n(q1).

    In Case (iii), if (x,t)Qr,λ2p(z0), then |x|4r and |t|10λ2pr2, and hence, by the similar computation as in Case (ii), we have

    I[Qr,λ2p(z0)]crnq[γn(n+γ)10λ2pr210λ2pr2|t|γ+n2dt+(4r)γ+nγ+n]×[γn(nqnγ)10λ2pr210λ2pr2|t|n(q1)γ2(q1)dt+(q1)(4r)nqnγq1nqnγ]q1.

    Finally, using the facts that n<γ<n(q1) and λ1,

    I[Qr,λ2p(z0)]crnq[γ(λ2pr2)γ+n2+rγ+n][γ(λ2pr2)n(q1)γ2(q1)+rnqnγq1]q1crnq(γλ(2p)(γ+n)2rγ+n+rγ+n)(γλ(2p)(nqnγ)2(q1)rnqnγq1+rnqnγq1)q1crnqrγ+nrnqnγ=c.

    We consider the following homogeneous problem in simple parabolic cylinder Q2=Q2(0):

    {htdiva(Dh)=0 inQ2,h=u onpQ2, (2.4)

    where uC0(22,22;L2(B2))Lp(22,22;W1,p(B2)) is a weak solution to (1.1) with replacing ΩT by Q2. For the existence and the uniqueness of the weak solution hC0(22,22;L2(B2))Lp(22,22; W1,p(B2)) to the above equation, we refer to e.g., [34, Section III.4]. Then, we obtain the following regularity estimates for h and comparison estimate between u and h.

    Lemma 2.1. Let u be a weak solution to (1.1) in Q2 with

    Q2|Du|pdz1andQ2|F|pdzδp (2.5)

    for some δ(0,1), and let h be the weak solution to (2.4). Then

    DhL(Q1,Rn)c(Q2|Dh|pdz+1)dpcLip (2.6)

    for some c,cLip1 depending on n,ν,L and p, where d1 is from (1.8).

    Moreover, for any ε(0,1), there exists small δ=δ(n,ν,L,p,ε)(0,1) such that

    Q2|DuDh|pdzε. (2.7)

    Proof. In view of [18, Section VIII.5], we have the first inequality in (2.6). We note that the Lipschitz regularity estimates in [18] are obtained for the parabolic p-Laplace systems. However, the same argument can apply to equations of p-Laplacian type such as (1.1) with the nonlinearity a satisfying (1.2) and (1.3).

    Regarding (2.7) and the second inequality in (2.6), similar comparison estimates can be found in numerous papers, see e.g., [2,5,10]. But, we shall prove them in details for completeness.

    We take ζ=uh as a test function in (1.1) and (2.4) to obtain

    Q2ut(uh)dz+Q2a(Du)(DuDh)dz=Q2|F|p2F(DuDh)dz

    and

    Q2ht(uh)dz+Q2a(Dh)(DuDh)dz=0.

    We notice that u and h are not differentiable for t. However, by considering their Steklov averages (see e.g., [18, Section I.3] and [6]), we may assume that they are differentiable for t. Then we have

    Q2(uh)t(uh)dz+Q2(a(Du)a(Dh))(DuDh)dz=Q2|F|p2F(DuDh)dz.

    Note that

    Q2(uh)t(uh)dz=Q212t(uh)2dz=12Br(uh)2|t=4dx12Br(uh)2|t=40dx0.

    We remark that the condition (1.3) implies the monotonicity condition:

    (a(ξ)a(η))(ξη)c(p,ν)(|ξ|2+|η|2)p22|ξη|2

    for every ξ,ηRn{0}. Then we see

    (|Du|2+|Dh|2)p22|DuDh|2c(a(Du)a(Dh))(DuDh).

    Therefore, by the above estimates, Young's inequality and the second inequality in (2.5), we have that for any κ1(0,1),

    Q2(|Du|2+|Dh|2)p22|DuDh|2dzcκ1Q2|DuDh|pdz+cκ1p11Q2|F|pdzcκ1Q2|DuDh|pdz+cκ1p11δp.

    If p2, since |DuDh|p(|Du|2+|Dh|2)p22|DuDh|2, by taking sufficiently small κ1=κ1(n,ν,L,p)>0 we have

    Q2|DuDh|pdzcδp.

    The second inequality in (2.6) follows from the first inequality in (2.5) together with δ1. Moreover, by choosing small δ depending on ε, we get (2.7).

    If 2nn+2<p<2, on the other hand, applying Young's inequality, we have that for any κ2(0,1),

    Q2|DuDh|pdz=Q2(|Du|2+|Dh|2)p(2p)4(|Du|2+|Dh|2)p(p2)4|DuDh|pdzcκ2Q2(|Du|2+|Dh|2)p2dz+cκ2pp2Q2(|Du|2+|Dh|2)p22|DuDh|2dzcκ2Q2[|Du|p+|Dh|p]dz+cκ2pp2κ1Q2|DuDh|pdz+cκ2pp2κ1p11δp.

    Hence by choosing κ1 sufficiently small depending on κ2 we have

    Q2|DuDh|pdzcκ2(Q2|Dh|pdz+Q2|Du|pdz)+c(κ2)δp. (2.8)

    We first note that

    Q2|Dh|pdzcκ2Q2|Dh|pdz+cQ2|Du|pdz++c(κ2)δp.

    Then by choosing κ2 sufficiently small and using the first inequality in (2.5) and δ1 we have the second inequality in (2.6). Finally, applying the second inequalities of (2.5) and (2.6) to (2.8), we have

    Q2|DuDh|pdzcκ2+c(κ2)δp.

    Finally, choosing κ2 and δ sufficiently small depending on ε we get (2.7).

    Now we start with the proof of the main theorem, Theorem 1.1. As we mentioned in the introduction, we follow the approach introduced in [2], see also [13] for the case of the weighted Lebesgue space. We divide the proof into five steps.

    Step 1. (Setting and stopping time argument)

    Let δ(0,1), which will be determined as a small constant depending only on n, ν, L, p, q and [w]q in below (3.17). Then there exists R0>0 satisfying that

    QR0(z0)ΩT[|Du|p+|Fδ|p]dz2|B1|5n+2for all  z0ΩT. (3.1)

    We fix any Q2r=Q2r(z0)ΩT with 2r<R0. For simplicity, we write Qρ=Qρ(z0), ρ(0,2r]. In addition, for ρ>0 and λ>0, we define the super level set

    E(ρ,λ):={zQρ:|Du(z)|>λ},

    and

    λpd0:=Q2r[|Du|p+|Fδ|p+1]dz1, (3.2)

    where d1 is from (1.8).

    Let rr1<r22r and consider any λ satisfying the following:

    λBλ0with  B:=(20rr2r1)d(n+2)p. (3.3)

    We notice that Qλρ(˜z)Qr2Q2r for any ˜z=(˜x,˜t)E(r1,λ) and all ρ<ρ0 where

    ρ0:={λp22(r2r1) if  2nn+2<p<2,r2r1 if  p2.

    Then we obtain the following Vitali type covering result for the super-level set E(r1,λ).

    Lemma 3.1. For each rr1<r22r and λBλ0, there exist ziE(r1,λ) and ρi(0,ρ010), i=1,2,3,, such that the intrinsic parabolic cylinders Qλρi(zi) are mutually disjoint,

    E(r1,λ)N  i=1Qλ5ρi(zi)

    for some Lebesgue measure zero set N,

    Qλρi(zi)[|Du|p+|Fδ|p]dz=λp (3.4)

    and

    Qλρ(zi)[|Du|p+|Fδ|p]dz<λp   for allρ(ρi,r2r1]. (3.5)

    Proof. For ˜zE(r1,λ) and ρ[ρ010,ρ0), by (3.2) and (3.3), we derive

    Qλρ(˜z)[|Du|p+|Fδ|p]dz|Q2r||Qλρ(˜z)|Q2r[|Du|p+|Fδ|p+1]dz=|Q2r|λpd0|Qλρ(˜z)|λp.

    To attain the last bound, we consider two cases p<2 and p2. When p2, we see pd=2 and so

    |Q2r|λpd0|Qλρ(˜z)|=(2r)n+2λ20λ2pρn+2(20rr2r1)n+2λp2λ20(20rr2r1)n+2λp(Bλ0)2λ20=λp.

    When p<2, we see pd=(p2)(n+2)2+2 and ρλp22(r2r1)10 and so

    |Q2r|λpd0|Qλρ(˜z)|=(2r)n+2λpd0λ2pρn+2(20rλp22(r2r1))n+2λp2λpd0=(20rr2r1)n+2(λ0λ)pdλp(20rr2r1)n+2(λ0Bλ0)pdλp=λp.

    Moreover, from the parabolic Lebesgue differentiation theorem, we deduce that, for almost every ˜zE(r1,λ),

    limρ0+Qλρ(˜z)[|Du|p+|Fδ|p]dz|Dw(˜z)|p>λp.

    Since the map ρQλρ(˜z)[|Du|p+|Fδ|p]dz is continuous, there exists ρ˜z(0,r2r110) such that

    Qλρ˜z(˜z)[|Du|p+|Fδ|p]dz=λp

    and

    Qλρ(˜z)[|Du|p+|Fδ|p]dz<λpfor all  ρ(ρ˜z,r2r1].

    Hence we apply Vitali's covering lemma for {Qλρ˜z(˜z):˜zE(r1,λ)} to complete the proof.

    From now on, let us set for i=1,2,3,,

    Q(0)i:=Qλρi(zi)andQ(j)i:=Qλ5jρi(zi),  j=1,2.

    Step 2. (Estimates of super-level sets)

    With the result in Lemma 3.1, we first estimates the Lebesgue measure of super-level set

    |{zQ(1)i:|Du(z)|>Aλ}|with  λBλ0,

    where A1 will be determined below in (3.9), by using estimates in Lemma 2.1. Note from (3.5) that

    Q(2)i[|Du|p+|Fδ|p]dz<λp. (3.6)

    We consider the following rescaled functions:

    aλ(ξ):=a(λξ)λp1for ξRn,
    uλ,i(z):=u(Zi)5ρiλandFλ,i(z):=F(Zi)λfor  Zi=zi+(5ρix,λ2p(5ρi)2t)

    with z=(x,t)Q2. Then it is obvious that aλ(ξ) satisfies (1.2) and (1.3) with ΩT=Q2(0)=Q2. Then we see that uλ,i is a weak solution to

    (uλ,i)tdivaλ(Duλ,i)=div(|Fλ,i|p2Fλ,i)  in Q2.

    Moreover we have from (3.6) that

    Q2[|Duλ,i|p+|Fλ,iδ|p]dz=1λpQ(2)i[|Du|p+|Fδ|p]dz<1,

    which implies

    Q2|Duλ,i|pdz1andQ2|Fλ,i|pdzδp. (3.7)

    In addition, let ˜hλ,i be a weak solution to

    (˜hλ,i)tdivaλ(D˜hλ,i)=0  in Q2,and˜hλ,i=uλ,i  on pQ2.

    Now, we consider sufficiently small constant ε>0 which will be determined below in (3.17). Then by applying Lemma 2.1, one can find δ=δ(n,ν,L,p,ε)>0 satisfying (3.7) such that

    Q1|Duλ,iD˜hλ,i|pdzεandD˜hλ,iL(Q1)cLip.

    Remark that both δ and cLip are independent of λ and i. Therefore setting

    hλ,i(z)=hλ,i(x,t):=5ρiλ˜hλ,i(xyi5ρi,tτiλ2p(5ρi)2)

    where zi=(yi,τi), we obtain

    Q(1)i|DuDhλ,i|pdzελp,andDhλ,iL(Q(1)i)cLipλ. (3.8)

    We set

    A:=2cLip>1. (3.9)

    Then since

    {zQ(1)i:|Du(z)|>Aλ}{zQ(1)i:|Du(z)Dhλ,i(z)|>Aλ2}{zQ(1)i:|Dhλ,i(z)|>Aλ2},

    we have from the estimates in (3.8) that

    |{zQ(1)i:|Du(z)|>Aλ}||{zQ(1)i:|Du(z)Dhλ,i(z)|>cLipλ}|+|{zQ(1)i:|Dhλ,i(z)|>cLipλ}|=01λpQ(1)i|DuDhλ,i|pdzε|Q(1)i|

    which implies

    |{zQ(1)i:|Du(z)|>Aλ}||Q(1)i|ε. (3.10)

    Step 3. (weighted estimates of supper-level sets)

    In this step, we estimate the weighted measure of super-level set

    w(E(r1,Aλ))with  λBλ0.

    We first observe from (3.1) and (3.4) that

    λp1|Q(0)i|Q2r[|Du|p+|Fδ|p]dz2|B1|5n22|B1|ρn+2iλ2p,

    hence

    λ(5ρi)n+22.

    This and the fact λ1 from (3.3) imply Q(1)iCp for every i=1,2,3,. Then we obtain from (2.3) and (3.10) that

    w({zQ(1)i:|Du(z)|>Aλ})w(|Q(1)i|)c2εγ1. (3.11)

    By Proposition 2.1, w is a p-intrinsic Aq for some q(1,q). Now we suppose that

    Q2r|Du|pqwdz<. (3.12)

    Then by (3.4) and (2.2) with q replaced by q we have

    λpq2q1(Q(0)i|Du|pdz)q+2q1(Q(0)i|Fδ|pdz)q2q1[w]qw(Q(0)i)(Q(0)i|Du|pqwdz+Q(0)i|Fδ|pqwdz)2q1[w]qw(Q(0)i)(Q(0)i{|Du|>λc0}|Du|pqwdz+Q(0)i{|F|δ>λc0}|Fδ|pqwdz+2cpq0λpqw(Q(0)i)),

    where c0:=(2q+1[w]q)1pq. Note that the right hand side is finite by the assumptions FLpqw,loc(ΩT) and (3.12). The above estimate means

    w(Q(0)i)2q[w]qλpq(Q(0)i{|Du|>λc0}|Du|pqwdz+Q(0)i{|F|δ>λc0}|Fδ|pqwdz). (3.13)

    Therefore, using Lemma 3.1, (3.11), (2.3) and (3.13), we obtain

    w(E(r1,Aλ))=w({zQr1:|Du(z)|>Aλ})i=1w({zQ(1)i:|Du(z)|>Aλ})cεγ1i=1w(Q(1)i)cεγ1i=1(|Q(1)i||Q(0)i|)qw(Q(0)i)cεγ1λpqi=1(Q(0)i{|Du|>λc0}|Du|pqwdz+Q(0)i{|F|δ>λc0}|Fδ|pqwdz)cεγ1λpq(Qr2{|Du|>λc0}|Du|pqwdz+Qr2{|F|δ>λc0}|Fδ|pqwdz). (3.14)

    Step 4. (A priori estimates)

    We prove the estimate (1.7) under the additional assumption

    Q2r|Du|pqwdz<, (3.15)

    where 2r<R0 and R0 satisfies (3.1). Note that (3.15) implies (3.12).

    Fix any rr1<r22r. Observe that

    Qr1|Du|pqwdz=pqApq0w(E(r1,Aλ))λpq1dλ=pqApqBλ00w(E(r1,Aλ))λpq1dλ+pqApqBλ0w(E(r1,Aλ))λpq1dλ=:I=(ABλ0)pqw(Q2r)+pqApqI, (3.16)

    where A and B are from (3.9) and (3.3). We estimate the second term I. Applying (3.14), we derive

    Icεγ1Bλ0(Qr2{|Du|>λc0}|Du|pqwdz+Qr2{|F|δ>λc0}|Fδ|pqwdz)λpqpq1dλcεγ10(Qr2{|Du|>λc0}(c0|Du|)pqwdz+Qr2{|F|δ>λc0}|c0Fδ|pqwdz)λpqpq1dλcεγ1(Qr2|Du|pqwdz+Qr2|Fδ|pqwdz).

    In the last inequality we apply the following elementary identity with g=c0|Du| or c0|F|δ, β2=pq, β1=pq and U=Qr2:

    Ugβ2wdz=(β2β1)0λβ2β11{zU:g(z)>λ}gβ1wdzdλ,β2>β1>1.

    Inserting the estimate for I into (3.16) and recalling the definitions of A and B and the fact that ε(0,1), we have

    Qr1|Du|pqwdzcεγ1Qr2|Du|pqwdz+cw(Q2r)λpq0rd(n+2)q(r2r1)d(n+2)q+cQ2r|Fδ|pqwdz,

    where the constants c, γ1 and c depend on n,p,ν,L,q and [w]q. At this stage, we choose ε=ε(n,p,ν,L,q,[w]q) such that

    cεγ112, (3.17)

    hence δ is also determined as a small constant depending on n,p,ν,L,q and [w]q. Therefore we obtain

    Qr1|Du|pqwdz12Qr2|Du|pqwdz+cλpq0w(Q2r)rd(n+2)q(r2r1)d(n+2)q+cQ2r|F|pqwdz

    for every rr1<r22r. Finally, applying Lemma 3.2 below with Ψ(ρ)=Qρ|Du|pqwdz with R1=r and R2=2r and recalling (3.2), we have that

    Qr|Du|pqwdzcw(Q2r)λpq0+cQ2r|F|pqwdzcw(Q2r)(Q2r[|Du|p+|F|p+1]dz)dq+cQ2r|F|pqwdz.

    This together with (2.3) implies (1.7).

    Lemma 3.2 (Lemma 6.1 in [22]). Let Ψ:[R1,R2][0,) be a bounded function. Suppose that for any r1 and r2 with 0<R1r1<r2R2,

    Ψ(r1)ϑΨ(r2)+C(r2r1)κ+D

    where C>0 and D0, κ>0 and ϑ[0,1). Then there exists c=(ϑ,κ)>0 such that

    Ψ(R1)c(ϑ,κ)[A(R2R1)κ+B].

    Step 5. (Approximation)

    Finally, we remove the a priori assumption (3.15) by a standard approximation argument. Suppose w is a p-intrinsic Aq weight and FLpqw,loc(ΩT,Rn). Fix any Q2r=Q2r(z0)ΩT with 2r<R0. Then there exists QR=QR(z0) such that Q2rQRΩT. Note that by Proposition 2.1 (4), w is a p-intrinsic Apq weight hence a usual parabolic Apq weight. Therefore, Cc(Rn+1) is dense in Lpqw(Rn+1), see e.g., [30, Lemma 2.1]. Therefore there exist FkCc(Rn+1,Rn), k=1,2,3,, such that

    FkFin  Lpqw(QR,Rn) as k,

    hence by (2.2),

    FkFin  Lp(QR,Rn) as k. (3.18)

    We further assume that

    QR|Fk|pdz2QR|F|pdz. (3.19)

    Let ukC0(t0R2,t0+R2;L2(BR(x0))Lp(t0R2,t0+R2;W1,p(BR(x0)) be the unique weak solution to

    {(uk)tdiva(Duk)=div(|Fk|p2Fk) inQR,uk=u onpQR, (3.20)

    see e.g., [34, Section III.4] for the existence of such uk. In view of [2], we have at least |Duk|pLγloc(QR) for every γ>1 since |Fk|pLγ(QR). In particular, by Hölder's inequality with Proposition 2.1 (2),

    Q2r|Duk|pqwdz(Q2r|Duk|pq(1+γ)γdz)γ1+γ(Q2rw1+γdz)11+γ<,

    which implies the a priori assumption in (3.15) for uk hence it follows from the previous results in Step 4 that

    (1w(Qr)Qr|Duk|pqwdz)1qc(Q2r[|Duk|p+|Fk|p+1]dz)d+c(1w(Q2r)Q2r|Fk|pqwdz)1q. (3.21)

    Now we take uuk as a test function in the weak forms of (1.1) and (3.20) to get

    QR(uuk)t(uuk)dz+QR(a(Du)a(Duk))(DuDuk)dz=QR(|F|p2F|Fk|p2Fk)(DuDuk)dz.

    Then, in a similar way as in the proof of Lemma 2.1, we derive

    QR(|Duk|2+|Du|2)p22|DukDu|2dzcQR||Fk|p2Fk|F|p2F||DukDu|dzcτ1p11QR||Fk|p2Fk|F|p2F|pp1dz+τ1QR|Duk|p+|Du|pdz (3.22)

    for any τ1(0,1), by applying Young's inequality.

    If p2, since |DukDu|p(|Duk|2+|Du|2)p22|DukDu|2, we infer

    QR|Duk|pdzcQR|F|p+|Du|pdz

    by taking sufficiently small τ1>0 and (3.19). If 2nn+2<p<2, applying Young's inequality, we have that for any τ2(0,1),

    QR|DukDu|pdzτ2QR(|Duk|2+|Du|2)p2dz+cτ2pp2QR(|Duk|2+|Du|2)p22|DukDu|2dzc(τ2+τ1τ2pp2)QR[|Duk|p+|Du|p]dz+cτ1p11τ2pp2QR||Fk|p2Fk|F|p2F|pp1dz

    and then by taking sufficiently small τ1,τ2>0 and (3.19),

    QR|Duk|pdzcQR|F|p+|Du|pdz.

    Eventually, for any p>2nn+2, we obtain that

    QR|Duk|pdzcQR|F|p+|Du|pdz<for all  k=1,2,3,. (3.23)

    Moreover, from (3.18) we see

    QR||Fk|p2Fk|F|p2F|pp1dz0 as k.

    Then taking into account (3.22) with (3.23),

    lim supkQR(|Duk|2+|Du|2)p22|DukDu|2dzcτ1QR|F|p+|Du|pdz.

    Since τ1(0,1) is arbitrary, we have that

    QR(|Duk|2+|Du|2)p22|DukDu|2dz0 as  k.

    Now, if 2nn+2<p<2, the Hölder inequality yields

    QR|DukDu|pdz=QR(|Duk|2+|Du|2)p(2p)4(|Duk|2+|Du|2)p(p2)4|DukDu|pdz(QR(|Duk|2+|Du|2)p2dz)2p2(QR(|Duk|2+|Du|2)p22|DukDu|2dz)p2

    and therefore by virtue of (3.23), we obtain

    QR|DukDu|pdz0 as k.

    This also holds in case p2 from (3.22), because

    QR|DukDu|pdzQR(|Duk|2+|Du|2)p22|DukDu|2dz.

    In turn, we obtain that for every p>2nn+2

    DukDuin  Lp(QR,Rn)Lp(Q2r,Rn) as k.

    In particular, we also have that DukDu a.e. in Lp(Q2r,Rn) as k, up to subsequence.

    Finally by passing k from (3.21) and applying the above convergence results for Fk and Duk with Fatou's lemma, we obtain

    (1w(Qr)Qr|Du|pqwdz)1qlim infk(1w(Qr)Qr|Duk|pqwdz)1qlim infk[c(Q2r[|Duk|p+|Fk|p+1]dz)d+c(1w(Q2r)Q2r|Fk|pqwdz)1q]=c(Q2r[|Du|p+|F|p+1]dz)d+c(1w(Q2r)Q2r|F|pqwdz)1q.

    Therefore, we complete the proof.

    We would like to thank the referees for many helpful comments. M. Lee was supported by the National Research Foundation of Korea by the Korean Government (NRF-2021R1A4A1032418). J. Ok was supported by the National Research Foundation of Korea by the Korean Government (NRF-2022R1C1C1004523).

    The authors declare no conflict of interest.



    [1] E. Acerbi, G. Mingione, Gradient estimates for the p(x)-Laplacean system, J. Reine Angew. Math., 2005 (2005), 117–148. http://doi.org/10.1515/crll.2005.2005.584.117 doi: 10.1515/crll.2005.2005.584.117
    [2] E. Acerbi, G. Mingione, Gradient estimates for a class of parabolic systems, Duke Math. J., 136 (2007), 285–320. http://doi.org/10.1215/S0012-7094-07-13623-8 doi: 10.1215/S0012-7094-07-13623-8
    [3] P. Baroni, Lorentz estimates for degenerate and singular evolutionary systems, J. Differ. Equations, 255 (2013), 2927–2951. http://doi.org/10.1016/j.jde.2013.07.024 doi: 10.1016/j.jde.2013.07.024
    [4] P. Baroni, V. Bögelein, Calderón-Zygmund estimates for parabolic p(x,t)-Laplacian systems, Rev. Mat. Iberoam., 30 (2014), 1355–1386. http://doi.org/10.4171/RMI/817 doi: 10.4171/RMI/817
    [5] V. Bögelein, Global gradient bounds for the parabolic p-Laplacian system, Proc. Lond. Math. Soc., 111 (2015), 633–680. http://doi.org/10.1112/plms/pdv027 doi: 10.1112/plms/pdv027
    [6] V. Bögelein, F. Duzaar, G. Mingione, Degenerate problems with irregular obstacles, J. Reine Angew. Math., 2011 (2011), 107–160. http://doi.org/10.1515/crelle.2011.006 doi: 10.1515/crelle.2011.006
    [7] S.-S. Byun, W. Kim, Global Calderón-Zygmund estimate for p-Laplacian parabolic system, Math. Ann., 383 (2022), 77–118. http://doi.org/10.1007/s00208-020-02089-z doi: 10.1007/s00208-020-02089-z
    [8] S.-S. Byun, J. Ok, On W1,q()-estimates for elliptic equations of p(x)-Laplacian type, J. Math. Pure. Appl., 106 (2016), 512–545. http://doi.org/10.1016/j.matpur.2016.03.002 doi: 10.1016/j.matpur.2016.03.002
    [9] S.-S. Byun, J. Ok, Nonlinear parabolic equations with variable exponent growth in nonsmooth domains, SIAM J. Math. Anal., 48 (2016), 3148–3190. http://doi.org/10.1137/16M1056298 doi: 10.1137/16M1056298
    [10] S.-S. Byun, J. Ok, S. Ryu, Global gradient estimates for general nonlinear parabolic equations in nonsmooth domains, J. Differ. Equations, 254 (2013), 4290–4326. http://doi.org/10.1016/j.jde.2013.03.004 doi: 10.1016/j.jde.2013.03.004
    [11] S.-S. Byun, J. Ok, S. Ryu, Global gradient estimates for elliptic equations of p(x)-Laplacian type with BMO nonlinearity, J. Reine Angew. Math., 2016 (2016), 1–38. http://doi.org/10.1515/crelle-2014-0004 doi: 10.1515/crelle-2014-0004
    [12] S.-S. Byun, S. Ryu, Global weighted estimates for the gradient of solutions to nonlinear elliptic equations, Ann. Inst. H. Poincaré C Anal. Non Linéaire, 30 (2013), 291–313. http://doi.org/10.1016/j.anihpc.2012.08.001 doi: 10.1016/j.anihpc.2012.08.001
    [13] S.-S. Byun, S. Ryu, Weighted Orlicz estimates for general nonlinear parabolic equations over nonsmooth domains, J. Funct. Anal., 272 (2017), 4103–4121. http://doi.org/10.1016/j.jfa.2017.01.024 doi: 10.1016/j.jfa.2017.01.024
    [14] S.-S. Byun, L. Wang, Elliptic equations with BMO coefficients in Reifenberg domains, Commun. Pure Appl. Math., 57 (2004), 1283–1310. http://doi.org/10.1002/cpa.20037 doi: 10.1002/cpa.20037
    [15] L. A. Caffarelli, I. Peral, On W1,p estimates for elliptic equations in divergence form, Commun. Pure Appl. Math., 51 (1998), 1–21. http://doi.org/10.1002/(SICI)1097-0312(199801)51:1<1::AID-CPA1>3.0.CO;2-G doi: 10.1002/(SICI)1097-0312(199801)51:1<1::AID-CPA1>3.0.CO;2-G
    [16] A. P. Calderon, A. Zygmund, On the existence of certain singular integrals, Acta Math., 88 (1952), 85–139. http://doi.org/10.1007/BF02392130 doi: 10.1007/BF02392130
    [17] M. Colombo, G. Mingione, Calderón-Zygmund estimates and non-uniformly elliptic operators, J. Funct. Anal., 270 (2016), 1416–1478. http://doi.org/10.1016/j.jfa.2015.06.022 doi: 10.1016/j.jfa.2015.06.022
    [18] E. DiBenedetto, Degenerate parabolic equations, New York: Springer, 1993. http://doi.org/10.1007/978-1-4612-0895-2
    [19] E. DiBenedetto, A. Friedman, Regularity of solutions of nonlinear degenerate parabolic systems, J. Reine Angew. Math., 1984 (1984), 83–128. http://doi.org/10.1515/crll.1984.349.83 doi: 10.1515/crll.1984.349.83
    [20] E. DiBenedetto, A. Friedman, Hölder estimates for nonlinear degenerate parabolic systems, J. Reine Angew. Math., 1985 (1985), 1–22. http://doi.org/10.1515/crll.1985.357.1 doi: 10.1515/crll.1985.357.1
    [21] E. DiBenedetto, J. Manfredi, On the higher integrability of the gradient of weak solutions of certain degenerate elliptic systems, Amer. J. Math., 115 (1993), 1107–1134.
    [22] E. Giusti, Direct methods in the calculus of variations, River Edge, NJ: World Scientific Publishing Co., Inc., 2003. http://doi.org/10.1142/5002
    [23] L. Grafakos, Classical Fourier analysis, 3 Eds., New York: Springer, 2014. http://doi.org/10.1007/978-1-4939-1194-3
    [24] P. Harjulehto, P. Hästö, Orlicz spaces and generalized Orlicz spaces, Cham: Springer, 2019. http://doi.org/10.1007/978-3-030-15100-3
    [25] P. Hästö, J. Ok, Higher integrability for parabolic systems with Orlicz growth, J. Differ. Equations, 300 (2021), 925–948. http://doi.org/10.1016/j.jde.2021.08.012 doi: 10.1016/j.jde.2021.08.012
    [26] T. Iwaniec, Projections onto gradient fields and Lp-estimates for degenerated elliptic operators, Stud. Math., 75 (1983), 293–312. http://doi.org/10.4064/sm-75-3-293-312 doi: 10.4064/sm-75-3-293-312
    [27] J. Kinnunen, J. Lewis, Higher integrability for parabolic systems of p-Laplacian type, Duke Math. J., 102 (2000), 253–271. http://doi.org/10.1215/S0012-7094-00-10223-2 doi: 10.1215/S0012-7094-00-10223-2
    [28] J. Kinnunen, S. Zhou, A local estimate for nonlinear equations with discontinuous coefficients, Commun. Part. Diff. Eq., 24 (1999), 2043–2068. http://doi.org/10.1080/03605309908821494 doi: 10.1080/03605309908821494
    [29] T. Mengesha, N. C. Phuc, Global estimates for quasilinear elliptic equations on Reifenberg flat domains, Arch. Rational Mech. Anal., 203 (2012), 189–216. http://doi.org/10.1007/s00205-011-0446-7 doi: 10.1007/s00205-011-0446-7
    [30] N. Miller, Weighted Sobolev spaces and pseudodifferential operators with smooth symbols, Trans. Amer. Math. Soc., 269 (1982), 91–109. http://doi.org/10.1090/S0002-9947-1982-0637030-4 doi: 10.1090/S0002-9947-1982-0637030-4
    [31] G. Mingione, The Calderón–Zygmund theory for elliptic problems with measure data, Ann. Scuola Norm. Sup. Pisa Cl. Sci. (5), 6 (2007), 195–261. http://doi.org/10.2422/2036-2145.2007.2.01 doi: 10.2422/2036-2145.2007.2.01
    [32] J. Oh, J. OK, Gradient estimates for parabolic problems with Orlicz growth and discontinuous coefficients, Math. Method. Appl. Sci., 45 (2022), 8718–8736. http://doi.org/10.1002/mma.7845 doi: 10.1002/mma.7845
    [33] C. Scheven, Existence of localizable solutions to nonlinear parabolic problems with irregular obstacles, Manuscripta Math., 146 (2015), 7–63. http://doi.org/10.1007/s00229-014-0684-8 doi: 10.1007/s00229-014-0684-8
    [34] R. E. Showalter, Monotone operators in Banach space and nonlinear partial differential equations, Providence, RI: American Mathematical Society, 1997.
    [35] D. Signoriello, T. Singer, Local Calderón-Zygmund estimates for parabolic minimizers, Nonlinear Anal., 125 (2015), 561–581. http://doi.org/10.1016/j.na.2015.06.005 doi: 10.1016/j.na.2015.06.005
    [36] A. Verde, Calderón–Zygmund estimates for systems of φ-growth, J. Convex Anal., 18 (2011), 67–84.
  • Reader Comments
  • © 2023 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(2127) PDF downloads(141) Cited by(0)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog