1: \documentclass[12pt]{article}
2: \usepackage{latexsym,amsmath,amssymb}
3:
4: \addtolength{\textwidth}{2cm} \addtolength{\hoffset}{-1cm}
5: \addtolength{\textheight}{3cm} \addtolength{\voffset}{-1.5cm}
6: \newif\ifdviwin
7:
8: \usepackage{enumerate}
9: \usepackage[latin1]{inputenc}
10: \usepackage[english]{babel}
11: \usepackage{indentfirst}
12: \usepackage[mathscr]{eucal}
13: \usepackage{amssymb,amsmath,amsfonts}
14: %\usepackage{fancybox,fancyhdr}
15: \usepackage{graphicx}
16:
17: \newif\ifdviwin
18:
19: \dviwintrue
20:
21: \def\cA{\mathcal{A}}
22: \def\cP{\mathcal{P}}
23: \def\cS{\mathcal{S}}
24: \def\cC{\mathcal{C}}
25: \def\cR{\mathcal{R}}
26: \def\cB{\mathcal{B}}
27: \def\cD{\mathcal{D}}
28: \def\cL{\mathcal{L}}
29: \def\cU{\mathcal{U}}
30: \def\cG{\mathcal{G}}
31: \def\cE{\mathcal{E}}
32: \def\cN{\mathcal{N}}
33: \def\cT{\mathcal{T}}
34: \def\cK{\mathcal{K}}
35:
36: \def\N{\mathbb{N}}
37: \def\L{\mathbb{L}}
38: \def\E{\mathbb{E}}
39: \def\B{\mathbb{B}}
40: \def\Q{\mathbb{Q}}
41: \def\R{\mathbb{R}}
42: \def\Z{\mathbb{Z}}
43: \def\C{\mathbb{C}}
44: \def\D{\mathbb{D}}
45: \def\CP{\mathbb{C}\mathbb{P}}
46: \def\H{\mathbb{H}}
47: \def\S{\mathbb{S}}
48: \def\X{\widetilde{X}}
49: \def\SL{\mathbf{SL}(2,\mathbb{C})}
50:
51: \def\Om{\Omega}
52: \def\om{\omega}
53: \def\ve{\varepsilon}
54: \def\Sg{\Sigma}
55:
56: \newcommand{\longui}{\operatorname{length}}
57: \newcommand{\dist}{\operatorname{dist}}
58:
59: \renewcommand{\thefootnote}{\fnsymbol{footnote}} % Para poner símbolos en los footnotes
60:
61: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
62: %%%% Grupos tipo teorema
63: \newtheorem{definition}{Definition}
64: \newtheorem{theorem}{Theorem}
65: \newtheorem{proposition}{Proposition}
66: \newtheorem{corollary}{Corollary}
67: \newtheorem{lemma}{Lemma}
68: \newtheorem{claim}{Claim}
69: \newtheorem{remark}{Remark}
70: \newtheorem{note}{Note}
71:
72:
73: \newenvironment{proof}{\rm \trivlist \item[\hskip \labelsep{\it
74: Proof.}]}{\nopagebreak \hfill $\Box$ \endtrivlist}
75:
76: \newenvironment{proof1}{\rm \trivlist \item[\hskip \labelsep{\it
77: Proof of Theorem \ref{th:main}}:]}{\par\nopagebreak \hfill $\Box$ \endtrivlist}
78:
79: \newcommand\esc[2]{\langle{#1},{#2}\rangle}
80:
81:
82: %\numberwithin{equation}{section}
83:
84: \begin{document}
85: \mbox{}\vspace{0.4cm}\mbox{}
86:
87: \begin{center}
88: \rule{14cm}{1.5pt}\vspace{0.5cm}
89:
90: {\Large \bf Complete minimal surfaces in $\R^3$} \\ [0.3cm]{\Large \bf with a prescribed coordinate function }\\
91: \vspace{0.5cm} {\large Antonio Alarc\'{o}n\footnote{Research
92: partially supported by MEC-FEDER grant number MTM2007-61775.} and
93: Isabel Fern\'{a}ndez\footnote{Research partially supported by
94: MEC-FEDER grant number MTM2007-64504.} }
95:
96: \vspace{0.3cm} \rule{14cm}{1.5pt}
97: \end{center}
98:
99:
100: \vspace{0.5cm}
101:
102: \begin{abstract}
103: In this paper we construct complete simply connected minimal
104: surfaces with a prescribed coordinate function. Moreover, we prove
105: that these surfaces are dense in the space of all minimal surfaces
106: with this coordinate function (with the topology of the smooth
107: convergence on compact sets).
108: \\
109:
110: \noindent 2000 Mathematics Subject Classification: Primary 53A10, Secondary 53C42, 49Q05, 30F15 \\
111:
112: \noindent Keywords: Complete minimal surfaces, harmonic functions.
113: \end{abstract}
114:
115: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
116: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
117: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
118:
119: \section{Introduction}\label{sec:intro}
120:
121: An isometric immersion $X:M\to\R^3$ of a Riemann surface into
122: space is said to be minimal if its coordinate functions are
123: harmonic on $M$. In 1980, Jorge and Xavier \cite{JX} constructed a
124: complete minimal surface contained in a slab of $\R^3$, disproving
125: a conjecture by Calabi \cite{C}. They looked for a minimal
126: immersion of the disk with third coordinate $x_3(z)=\mbox{Re}(z)$
127: and complete metric.\\
128:
129:
130: In light of the above, it appears as a natural question whether
131: any harmonic function can be realized as a coordinate of a
132: complete minimal surface. The present paper is devoted to answer
133: this question in the simply connected case. More specifically, we
134: extend Jorge-Xavier's result to prove that any harmonic function
135: defined on a simply connected domain is a coordinate function of a
136: conformal complete minimal immersion (see Theorem
137: \ref{th:corolario}). Moreover, we show that complete surfaces are
138: dense in the space of (simply connected) minimal surfaces with a
139: prescribed coordinate (Corollary \ref{co:dens}). These results
140: come as a consequence of the following one.
141:
142: \begin{theorem}\label{th:main}
143: Let $X=(X_1,X_2,X_3):\Sg\to\R^3$ be a conformal minimal immersion
144: on $\Sg=\C,\D$, with $X_3$ being non-constant. Consider $K\subset
145: \Sg$ a compact set and $\ve>0$.
146:
147: Then, there exists a complete conformal minimal immersion $Y=(Y_1,Y_2,Y_3):\Sg\to\R^3$ such that
148: \begin{enumerate}[\rm (a)]
149: \item $\|X-Y\|<\ve$ in $K$.
150:
151: \item $X_3=Y_3$.
152: \end{enumerate}
153: \end{theorem}
154:
155: We also derive from Theorem \ref{th:main} some results concerning
156: existence of complete null curves in $\C^3$ and complete maximal
157: surfaces in $\L^3$ with a prescribed coordinate (Section \ref{sec:cons}).\\
158:
159:
160: The construction of the Jorge-Xavier's surface relies on a clever
161: use of the Runge's classical theorem with a suitable labyrinth
162: close to the boundary of the disk. A refinement of Jorge and
163: Xavier's ideas led to Nadirashvili \cite{N} to construct conformal
164: complete bounded minimal disks. Nadirashvili's arguments have
165: given rise to the construction of complete bounded minimal
166: surfaces with other additional properties (see for instance
167: \cite{LMM,MN,AFM,A}). However, all the coordinate functions of
168: these examples are implicit.
169:
170: Despite we use some ideas related to Nadirashvili's technique in
171: the proof of the above theorem, it is not possible in general to
172: construct complete bounded minimal surfaces with a prescribed
173: (bounded) coordinate function. We show a requirement for a
174: harmonic function on the disk to be the coordinate function of a
175: complete bounded minimal surface in Proposition \ref{prop} (see
176: also \cite{N2,AN}).\\
177:
178: Finally, we would like to point out that the only complete simply
179: connected embedded minimal surfaces are the plane and the
180: helicoid \cite{MR,CM}. Therefore, our surfaces are not embedded,
181: except for the aforementioned cases.
182:
183: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
184: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
185: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
186:
187: \section{Preliminaries}\label{sec:prelim}
188:
189: This section is devoted to briefly summarize the notation and
190: results that we use in the paper.
191:
192: From now on, we denote by $\Sg$ an open simply-connected Riemann surface. By the Uniformization Theorem we can assume that $\Sg$ is eihter the complex plane $\C$ or the open unit disk $\D$. Furthermore, for any $r>0$, we denote by $\D_r=\{z\in\C\;|\; |z|<r\}$ and $\S^1_r=\{z\in\C\;|\; |z|=r\}$.
193:
194: Consider a Riemannian metric $d\tau^2$ in $\Sg$. Given a curve $\alpha$ in $\Sg$, by $\longui_{d\tau^2}(\alpha)$ we mean the length of $\alpha$ with respect to the metric $d\tau^2$. Moreover, we define:
195: \begin{itemize}
196: \item $\dist_{d\tau^2}(p,q)=\inf \{\longui_{d\tau^2}(\alpha) \: | \: \alpha:[0,1]\rightarrow \Sg, \; \alpha(0)=p,\alpha(1)=q \}$, for any $p,q\in \Sg$.
197: \item $\dist_{d\tau^2}(T_1,T_2)=\inf \{\dist_{d\tau^2}(p,q) \;|\;p \in T_1, \;q \in T_2 \}$, for any $T_1, T_2 \subset \Sg$.
198: \end{itemize}
199: Throughout the paper, we work with metrics induced by conformal minimal
200: immersions $X:\Sg\to\R^3.$ Then, by $\lambda_{X}^2|dz|^2$ we
201: mean the Riemannian metric induced by $X$ in $\Sg.$ We
202: also write $\dist_{X}(T_1,T_2)$ instead of
203: $\dist_{\lambda_X^2|dz|^2}(T_1,T_2),$ for any sets $T_1$ and $T_2$ in $\Sg.$
204:
205: \subsection{Minimal surfaces background}
206:
207: The theory of complete minimal surfaces is closely related to the theory of Riemann surfaces. This is due to the fact that any such surface is given by a triple $\Phi=(\Phi_1, \Phi_2, \Phi_3)$ of holomorphic 1-forms defined on some Riemann surface $M$ such that
208: \begin{equation} \label{eq:conforme}
209: \Phi_1^2+\Phi_2^2+\Phi_3^2=0,
210: \end{equation}
211: \begin{equation}\label{eq:inmersion}
212: \|\Phi_1\|^2+\|\Phi_2\|^2+\|\Phi_3\|^2 \neq 0,
213: \end{equation}
214: and all periods of the $\Phi_j$ are purely imaginary.
215: %Here we consider $\Phi_i$ to be a holomorphic function $\phi_i$ times $dz$ in a local parameter $z$.
216: Then the minimal immersion $X:M
217: \rightarrow \R^3$ can be parameterized by $z \mapsto \mbox{Re}
218: \int^z \Phi.$ The above triple is called the Weierstrass
219: representation of the immersion $X$. Usually, the first
220: requirement \eqref{eq:conforme} (which ensures the conformality
221: of $X$) is guaranteed by introducing the formulas
222: \[
223: \Phi_1 =\frac12 \left( \frac1{g}-g\right) \, \Phi_3, \quad \Phi_2 =\frac{\rm i}2 \left( \frac1{g}+g\right) \, \Phi_3,
224: \]
225: with a meromorphic function $g$ (the stereographic projection of
226: the Gauss map) and a holomorphic 1-form $\Phi_3$. The pair
227: $(g,\Phi_3)$ is called the Weierstrass data of the minimal
228: immersion $X$. In this article all the minimal immersions are
229: defined on the simply connected Riemann surface $\Sg$. Then, the
230: Weierstrass data have no periods and so the only requirements are
231: \eqref{eq:conforme} and \eqref{eq:inmersion}. The metric of $X$
232: can be expressed as
233: \begin{equation}\label{eq:metric}
234: \lambda^2_X|dz|^2=\frac12
235: \|\Phi\|^2=\left(\frac12\left(\frac1{|g|}+|g|\right) \|\Phi_3
236: \|\right)^2.
237: \end{equation}
238:
239: \subsubsection{The L\'{o}pez-Ros transformation}
240: The proof of Lemma \ref{lem:lemma} exploits what has come to be called the L\'{o}pez-Ros transformation. If $M$ is a Riemann surface and $(g,\Phi_3)$ are the Weierstrass data of a minimal immersion $X:M \rightarrow \R^3$, we define on $M$ the data
241: \[
242: \widetilde g= \frac{g}{h}, \qquad \widetilde \Phi_3= \Phi_3,
243: \]
244: where $h:M \rightarrow \C$ is a holomorphic function without
245: zeros. If the periods of this new Weierstrass representation are purely imaginary, then it defines a minimal immersion $\widetilde X: M \rightarrow \R^3$. This method provides us with a powerful and natural tool for deforming minimal surfaces. From our point of view, the most important property of the resulting surface is that the third coordinate function is preserved. Note that the intrinsic metric is given by \eqref{eq:metric} as
246: \[
247: \lambda^2_{\widetilde X}|dz|^2=\left(\frac12\left(\frac{|h|}{|g|}+\frac{|g|}{|h|}\right)\, \|\Phi_3 \|
248: \right)^2.
249: \]
250: This means that we can increase the intrinsic distance in a prescribed compact of $M$, by using suitable functions $h$.
251:
252: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
253: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
254: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
255:
256: \section{Proof of Theorem \ref{th:main}}\label{sec:main}
257:
258: In order to prove the main theorem we need the following technical lemma. It will be proved later in subsection \ref{sub:lemma}.
259:
260: \begin{lemma}\label{lem:lemma}
261: Let $X=(X_1,X_2,X_3):\Sg\to\R^3$ be a conformal minimal immersion
262: being $X_3$ non-constant. Consider two positive constants $0<r<R$
263: (with $R<1$ if $\Sg=\D$).
264:
265: Then, for any $\ve,s>0$ there exists a conformal minimal
266: immersion $\X=(\X_1,\X_2,\X_3):\Sg\to\R^3$ such that
267: \begin{enumerate}[\rm ({L}1)]
268: \item $\dist_{\X}(0,\S^1_R)>s$.
269:
270: \item $\|\X-X\|<\ve$ in $\overline\D_r$.
271:
272: \item $\X_3=X_3$.
273: \end{enumerate}
274: \end{lemma}
275:
276: Assuming the above lemma, the proof of Theorem \ref{th:main} goes as follows. First of all, consider $r_0>0$ ($0<r_0<1$ in case $\Sg=\D$) such that $K\subset \D_{r_0}\subset \Sg$. Let $\{r_n\}_{n\in\N}$ be an increasing sequence of positives, with $r_1=r_0$, and such that $\{r_n\}\nearrow +\infty$ in case $\Sg=\C$, and $\{r_n\}\nearrow 1$ in case $\Sg=\D$. Finally, take any sequence $\{\sigma_n\}_{n\in\N}$ with $0<\sigma_n<1$ and so that $\prod_{k=1}^\infty\sigma_k = 1/2$.
277:
278: We will obtain the immersion $Y$ as a limit of a sequence of immersions $\{X_n\}_{n\in\N}$. For any $n\in\N$, we will construct a family $\chi_n=\{X_n,\ve_n\}$ where $X_n:\Sg\to\R^3$ is a conformal minimal immersion and $\ve_n<6\ve/(n^2\pi^2)$ is a positive number. Furthermore, the sequence $\{\chi_n\}_{n\geq 2}$ will satisfy the following properties.
279: \begin{enumerate}[\rm (A$_{n}$)]
280: \item $\|X_n-X_{n-1}\|<\ve_n$ in $\overline{\D}_{r_{n-1}}$.
281:
282: \item $\dist_{X_n}(0,\S^1_{r_n})>n$.
283:
284: \item $\lambda_{X_n}\geq \sigma_n\cdot \lambda_{X_{n-1}}$ in $\overline{\D}_{r_{n-1}}$.
285:
286: \item $(X_n)_3=X_3$.
287: \end{enumerate}
288:
289: The sequence is constructed in a recursive way. The first element of the sequence is the immersion $X_1=X$ and any positive $\ve_1<6\ve/\pi^2$. Assume we have defined $\chi_1,\ldots,\chi_n$. Let us show how to construct the family $\chi_{n+1}$. Consider a sequence $\{\xi_m\}_{m\in\N}$ decreasing to zero and such that
290: \[
291: \xi_m<\min \left\{\ve_n\,,\,\frac{6\ve}{\pi^2(n+1)^2}\right\},\quad \forall m\in\N.
292: \]
293: Let $F_m:\Sg\to\R^3$ be the immersion obtained from Lemma \ref{lem:lemma} for the data
294: \[
295: X=X_n,\quad r=r_n,\quad R=r_{n+1},\quad \ve=\xi_m,\quad s=n+1.
296: \]
297: The sequence $\{F_m\}_{m\in\N}$ converges to $X_n$ uniformly on $\overline{\D}_{r_n}$. Therefore, there exists $m_0\in\N$ large enough so that
298: \begin{equation}\label{equ:metricas}
299: \lambda_{F_{m_0}}\geq \sigma_{n+1}\cdot \lambda_{X_n}\quad \text{in }\overline{\D}_{r_{n-1}}.
300: \end{equation}
301: Recall that $0<\sigma_{n+1}<1$. Label $X_{n+1}:=F_{m_0}$ and $\ve_{n+1}:=\xi_{m_0}$. Properties (L2), (L1) and (L3) imply (A$_{n+1}$), (B$_{n+1}$) and (D$_{n+1}$), respectively. Finally, inequality \eqref{equ:metricas} gives (C$_{n+1}$). This concludes the construction of the sequence $\{\chi_n\}_{n\in\N}$.
302:
303: Since $\Sg=\cup_{n\in\N}\overline{\D}_{r_n}$, we infer from
304: properties (A$_n$), $n\in\N$, that $\{X_n\}_{n\in\N}$ converges
305: to a smooth map $Y:\Sg\to\R^3$ uniformly on compact sets of
306: $\Sg$. Let us check that $Y$ satisfies the conclusion of the
307: theorem.
308:
309: \noindent{$\bullet$} $Y$ is an immersion. Indeed, consider $p\in\Sg$. Fix $n_0\in\N$ so that $p\in \overline{\D}_{r_{n_0}}$. From properties (C$_n$), $n> n_0$, we obtain that
310: \[
311: \lambda_{X_n}(p)\geq \frac12 \cdot\lambda_{X_{n_0}}(p),
312: \]
313: where we have used that $\prod_{k=1}^\infty \sigma_k=1/2$. If we take limits in the above inequality as $n\to\infty$ we infer that $\lambda_Y(p)\geq \frac12\lambda_{X_{n_0}}(p)>0$, and therefore, $Y$ is an immersion.
314:
315: \noindent{$\bullet$} Since $Y$ is harmonic (Harnack's Theorem), it is minimal and conformal.
316:
317: \noindent$\bullet$ $Y$ is complete. In order to check it, let $\alpha$ be a divergent curve in $\Sg$ starting at $0$. Then, for any $k\in\N$, we have
318: \[
319: \longui_{Y}(\alpha)\geq \longui_{Y}(\alpha\cap
320: \overline{\D}_{r_k})\geq \frac12 \longui_{X_k}(\alpha\cap
321: \overline{\D}_{r_k})>\frac{k}2,
322: \]
323: where we have used properties (B$_n$) and (C$_n$), $n\geq k$.
324: Hence, $\longui_{Y}(\alpha)=\infty$, which proves the
325: completeness of $Y$.
326:
327: \noindent$\bullet$ Statement (a) follows from properties (A$_n$), $n\in\N$, and the facts that $\sum_{n\geq 1}\ve_n<\ve$ and $K\subset \overline{\D}_{r_n}$, $\forall n\in\N$.
328:
329: \noindent$\bullet$ Statement (b) is a trivial consequence of properties (D$_n$), $n\in\N$.
330:
331: The proof is done.
332:
333: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
334: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
335: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
336:
337: \subsection{Proof of Lemma \ref{lem:lemma}}\label{sub:lemma}
338:
339: Let $(g,\Phi_3)$ be the Weierstrass data of the immersion $X$.
340: Since $X_3$ is non-constant and $\Sg$ is simply connected we can
341: write $\Phi_3=\phi_3(z)dz$ with $\phi_3$ non identically zero.
342: Therefore, there exist a constant $\delta>0$ and two real numbers
343: $r'$ and $R'$ with $r<r'<R'<R$, satisfying
344: \begin{equation}\label{eq:nocero}
345: |\phi_3|>\delta \quad \text{in }\D_{R'}\setminus
346: \overline{\D}_{r'}.
347: \end{equation}
348:
349: Fix a natural $N$ (which will be specified later) such that
350: $2/N<R'-r'$. The immersion $\X$ will be obtained from $X$ by
351: using L\'{o}pez-Ros transformation. The effect of this deformation
352: will be concentrated on a labyrinth of compact sets contained in
353: $\D_{R'}\setminus \overline{\D}_{r'}$. On the other hand, the
354: deformation hardly acts on $\overline\D_r$. The shape of the
355: labyrinth is inspired in those used by Jorge and Xavier
356: \cite{JX}. Let us describe it.
357:
358: For any $n\in\N$, $n=1,\ldots,2N^2$, define $s_n=R'-n/N^3$ and
359: label $s_0=R'$. Now, consider the set (see Figure
360: \ref{fig:laberinto})
361: \[
362: \cK_n=\left\{ z\in\C\; \left|\; s_n+\frac1{4N^3}\leq |z|\leq
363: s_{n-1}-\frac1{4N^3},\quad \frac1{N^2}\leq {\rm arg}((-1)^{n}z)
364: \leq 2\pi-\frac1{N^2}\right.\right\}.
365: \]
366: \begin{figure}[ht]
367: \begin{center}
368: \scalebox{0.5}{\includegraphics{laberinto}}
369: \end{center}
370: %\vspace{-2cm}
371: \caption{The labyrinth of compact sets.}\label{fig:laberinto}
372: \end{figure}
373:
374: Then, consider
375: \[
376: \cK=\bigcup_{n=1}^{2N^2}\cK_n.
377: \]
378:
379: From the definition of the compact set $\cK$ follows that any
380: curve joining $\S^1_{r'}$ and $\S^1_{R'}$ without going through
381: $\cK$ must have large Euclidean length. This fact is stated in the
382: following claim.
383:
384: \begin{claim}\label{claim:lemma}
385: Let $\lambda^2|dz|^2$ be a conformal metric on $\Sg$ satisfying
386: \[
387: \lambda\geq
388: \begin{cases}
389: c & \text{in }\D_{R'}\setminus\overline{\D}_{r'}
390: \\
391: c\, N^4 & \text{in }\cK,
392: \end{cases}
393: \]
394: for some $c>0$.
395:
396: Then, there exists a positive constant $\rho$ not depending on $c$ nor $N$ and such that
397: \[
398: \longui_{\lambda^2|dz|^2}(\alpha)>\rho\cdot c\cdot N
399: \]
400: holds, for any $\alpha$ curve in $\Sg$ joining $\S^1_{r'}$ and $\S^1_{R'}$.
401: \end{claim}
402:
403: Now, we define the function we use as parameter of the L\'{o}pez-Ros transformation.
404: In order to do it, for any $\beta>0$ we consider a holomorphic function $h_\beta:\Sg\to \C^*$ with the following two properties:
405: \begin{itemize}
406: \item $|h_\beta-1|<1/\beta$ in $\overline{\D}_r$.
407:
408: \item $|h_\beta-\beta|<1/\beta$ in $\cK$.
409: \end{itemize}
410: The set $\cK\cup \overline{\D}_r$ is a compact set and
411: $\Sg\setminus (\cK\cup \overline{\D}_r)$ is connected. Hence, the
412: existence of the above functions is guaranteed by Runge's Theorem.
413:
414: Define the Weierstrass data on $\Sg$
415: \[
416: g^\beta:=\frac{g}{h_\beta},\qquad \Phi_3^\beta :=\Phi_3.
417: \]
418: These Weierstrass data give rise to a minimal immersion
419: $X^\beta:\Sg\to\R^3$. Notice that $h_\beta$ converges to $1$
420: (resp. $\infty$) uniformly on $\overline{\D}_r$ (resp. $\cK$).
421: Hence, there exists a large enough $\beta_0$ such that
422: \begin{enumerate}[\rm (i)]
423: \item $\|X^{\beta_0}-X\|<\ve$ in $\overline{\D}_r$.
424:
425: \item $\lambda_{X^{\beta_0}}\geq \delta\cdot N^4$ in $\cK$, where $\delta$ was defined in equation
426: \eqref{eq:nocero}, and $\lambda_{X^{\beta_0}}^2$ is the conformal
427: factor of the metric induced by $X^{\beta_0}$ (see
428: \eqref{eq:metric}).
429: \end{enumerate}
430: Label $\X:=X^{\beta_0}$. Let us check that $\X$ has the desired
431: properties, provided that $N$ was chosen to be large enough.
432: From the very definition of $\X$, statements (L2) and (L3) trivially hold. In order to check (L1) notice first that
433: \begin{equation}\label{equ:iii}
434: \lambda_{\X}\geq |\phi_3|>\delta\quad \text{in }
435: \D_{R'}\setminus\overline{\D}_{r'},
436: \end{equation}
437: where we have taken into account \eqref{eq:nocero}. Then,
438: properties (ii), \eqref{equ:iii} and Claim \ref{claim:lemma}
439: guarantee that
440: \[
441: \longui_{\X}(\alpha)\geq \delta\cdot\rho\cdot N,
442: \]
443: for any curve $\alpha$ in $\Sg$ joining $0$ with $\S^1_R$. Assume that we chose $N$ large enough so that $\delta\cdot\rho\cdot N>s$ (recall that neither $\rho$ nor $\delta$ depend on $N$). This finishes the proof.
444:
445: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
446: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
447: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
448:
449: \section{Some consequences of Theorem \ref{th:main}}\label{sec:cons}
450:
451: It may be followed from our main theorem some interesting results concerning not only minimal surfaces.
452:
453: As we deal with simply connected surfaces, it is not hard to realize any harmonic function as a coordinate function of a minimal immersion. Hence, Theorem \ref{th:main} implies that any harmonic function on a simply connected domain is a coordinate function of a complete minimal immersion, as stated in the next theorem.
454:
455: \begin{theorem}\label{th:corolario}
456: Let $\Sg=\C,$ $\D$ and $u:\Sigma\to\R$ be a harmonic function.
457: Assume $u$ is non constant in case $\Sg=\D$.
458:
459: Then, there exists a complete conformal minimal immersion
460: $Y=(Y_1,Y_2,Y_3):\Sg\to\R^3$ such that $Y_3=u$.
461: \end{theorem}
462:
463: \begin{proof}
464: Fist of all notice that in case $\Sg=\C$ and $u$ is constant, the
465: plane $x_3=u$ satisfies the conclusion of the theorem.
466:
467: Thus, let us assume that $u$ is non constant, and define the
468: holomorphic $1$-form $$\Phi_3=du+{\rm i} (\star du),$$ where
469: $\star$ denotes de Hodge operator. Since $\Sg$ is simply
470: connected we can write $\Phi_3=\phi_3(z) dz$ on $\Sg$. Then the
471: pair $(\phi_3,\Phi_3)$ are Weierstrass data on $\Sg$ and so they
472: define a conformal minimal immersion
473: $X=(X_1,X_2,X_3):\Sg\to\R^3$. Moreover, it is straightforward to
474: check that $X_3=u$. Now, applying Theorem \ref{th:main} to the
475: immersion $X$, any compact set $K\subset\Sg$ and any positive
476: $\ve$, we obtain an immersion fulfilling the statement of the
477: theorem.
478: \end{proof}
479:
480:
481: Other interesting (and immediate) consequence of Theorem
482: \ref{th:main} is a density type result. Let $u:\Sg\to\R$ be a non
483: constant harmonic function, and label $\cA_u=\{
484: X=(X_1,X_2,X_3):\Sg\to\R^3 \text{ conformal minimal
485: immersion}\;|\; X_3=u \}$. With this notation the following holds.
486:
487: \begin{corollary}\label{co:dens}
488: Let $u:\Sg\to\R$ be a non constant harmonic function on
489: $\Sg=\C,\D$.
490:
491: Then, complete immersions in $\cA_u$ are dense in $\cA_u$,
492: endowed with the topology of the uniform convergence on compact
493: sets.
494: \end{corollary}
495:
496: Theorem \ref{th:main} can be applied to other geometric theories.
497: Maximal surfaces in the $3$-dimensional Lorentz-Minkowski space
498: $\L^3=(\R^3,dx_1^2+dx_2^2-dx_3^2)$ are spacelike surfaces with
499: vanishing mean curvature. There is a close connection between
500: minimal and maximal surfaces. Indeed, if
501: $X=\mbox{Re}\int(\Phi_1,\Phi_2,\Phi_3):\Sg\to\R^3$ is a conformal
502: minimal immersion defined on a simply connected surface $\Sg$,
503: then $$\widehat X=\mbox{Re}\int({\rm i}\Phi_1,{\rm
504: i}\Phi_2,\Phi_3):\Sg\to\L^3$$ is a conformal maximal immersion
505: (possibly with lightlike singularities), with the same third
506: coordinate function. See \cite{K,FLS} for more details on maximal
507: surfaces. Using this connection we can translate Theorem
508: \ref{th:corolario} to the Lorentzian setting.
509:
510: \begin{corollary}
511: Let $\Sg=\C,$ $\D$ and $u:\Sigma\to\R$ be a harmonic function.
512: Assume $u$ is non constant in case $\Sg=\D$.
513:
514: Then, there exists a conformal maximal immersion (possibly with
515: lightlike singularities), $Y=(Y_1,Y_2,Y_3):\Sg\to\L^3$, such that
516: $Y_3=u$ and $Y$ is weakly complete in the sense of Umehara and
517: Yamada \cite{UY}.
518: \end{corollary}
519:
520: Other geometrical objects related with minimal surfaces are null
521: curves (see for instance \cite{MUY}). By definition, a complex
522: curve $F=(F_1,F_2,F_3):\Sg\to\C^3$ is said to be a holomorphic
523: null curve if its coordinate functions are holomorphic and they
524: satisfy $$(F_1')^2+(F_2')^2+(F_3')^2=0,$$ where $'$ denotes the
525: complex derivative. Using Weierstrass representation, simply
526: connected minimal surfaces in $\R^3$ can be seen as the real part
527: of holomorphic null curves in $\C^3$, and conversely. Moreover,
528: the minimal surface and the associated holomorphic null curve
529: have the same metric. This allows us to prove the next result.
530:
531: \begin{corollary}
532: Let $f:\Sg\to\C$ be a holomorphic function on $\Sg=\C,\D$. Assume
533: $f$ is non constant if $\Sg=\D$.
534:
535: Then, there exists a complete holomorphic null curve
536: $F=(F_1,F_2,F_3):\Sg\to\C^3$ with $F_3=f$.
537: \end{corollary}
538:
539: Finally, we would like to remark that our results are sharp in
540: the following sense. Recall that Nadirashvili's techniques give
541: complete conformal bounded minimal disks. Since our arguments are
542: inspired in his techniques, it could be expected that, starting
543: from a bounded harmonic function on the disk, one could obtain a
544: complete bounded minimal immersion having this function as a
545: coordinate function. However, the following proposition shows that
546: this is not possible in general.
547:
548: \begin{proposition}\label{prop}
549: Let $X=(X_1,X_2,X_3):\D\to\R^3$ be a complete conformal minimal
550: immersion. Assume that $X_3$ can be extended smoothly to the
551: closed disk $\overline\D$.
552:
553: Then, $X_1$ and $X_2$ are unbounded on $\D$.
554: \end{proposition}
555:
556: \begin{proof}
557: Let $\Phi=(\Phi_1,\Phi_2,\Phi_3)$ be the Weierstrass data of $X$
558: and write $\Phi_j=\phi_j(z)dz$, $j=1,2,3$.
559:
560: Reasoning by contradiction, let us assume that $X_2$ is bounded.
561: Then, Bourgain's Theorem \cite{B} gives the existence of a real
562: number $0<\theta<2\pi$ such that $$\int_{0}^1 |\phi_2(r
563: e^{i\theta})| dr <\infty.$$
564:
565: On the other hand, the assumption on $X_3$ guarantees that
566: $$\int_{0}^1 |\phi_3(r e^{i\theta})| dr <\infty.$$ Since $X$ is a
567: conformal map, it follows that $|\phi_1|^2\leq |\phi_2|^2 +
568: |\phi_3|^2$ and so, from the above inequalities we get
569: $$\int_{0}^1 \|(\phi_1,\phi_2,\phi_3)(r e^{i\theta})\| dr
570: <\infty,$$ which contradicts the completeness of $X$.
571: \end{proof}
572:
573: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
574: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
575: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
576:
577: \def\refname{References}
578: \begin{thebibliography}{99999}
579:
580: \bibitem[A]{A} A. Alarc\'{o}n, {\it Compact complete minimal
581: immersions in $\R^3$.} Trans. Amer. Math. Soc. (To appear).
582:
583: \bibitem[AFM]{AFM} A. Alarc\'{o}n, L. Ferrer, F. Mart\'{\i}n, {\it Density
584: theorems for complete minimal surfaces.} Geom. Funct. Anal. {\bf 18} (1), 1--49 (2008).
585:
586: \bibitem[AN]{AN} A. Alarc\'{o}n, N. Nadirashvili, {\it Limit sets for complete minimal
587: immersions.}
588: Math. Z. {\bf 258} (1), 107--113 (2008).
589:
590: \bibitem[B]{B} J. Bourgain, {\it On the radial variation of bounded analytic functions on the disc.}
591: Duke Math. J. {\bf 69} (3), 671--682 (1993).
592:
593: \bibitem[C]{C} E. Calabi, {\it Problems in differential geometry.} Ed. S. Kobayashi and J. Ells, Jr., Proceedings
594: of the United States-Japan Seminar in Differential Geometry,
595: Kyoto, Japan, 1965. Nippon Hyoronsha Co., Ltd., Tokyo 170 (1966).
596:
597: \bibitem[CM]{CM} T. H. Colding and W. P. Minicozzi, {\em The Calabi-Yau conjectures for embedded
598: surfaces.} Ann. of Math. {\bf 167} (1), 211--243 (2008).
599:
600: \bibitem[FLS]{FLS} I. Fern\'{a}ndez, F. J. L\'{o}pez, R. Souam, {\it The space of complete embedded maximal
601: surfaces with isolated singularities in the 3-dimensional
602: Lorentz-Minkowski space.} Math. Ann. {\bf 332} (3), 605--643
603: (2005).
604:
605: \bibitem[JX]{JX}
606: L. P. Jorge, F. Xavier, {\it A complete minimal surface in
607: $\R^{3}$ between two parallel planes.} Ann. of Math. {\bf 112}
608: (1), 203--206 (1980).
609:
610: \bibitem[K]{K} O. Kobayashi, {\it Maximal surfaces in the $3$-dimensional Minkowski space $\L^{3}$.}
611: Tokyo J. Math. {\bf 6} (2), 297--309 (1983).
612:
613: \bibitem[LMM]{LMM} F. J. L\'{o}pez, F. Mart\'{\i}n, S. Morales, {\it Adding handles to Nadirashvili's surfaces.}
614: J. Differential Geom. {\bf 60} (1), 155--175 (2002).
615:
616: \bibitem[MN]{MN} F. Mart\'{\i}n, N. Nadirashvili, {\it A Jordan curve spanned by a complete minimal surface.}
617: Arch. Ration. Mech. Anal. {\bf 184} (2), 285--301 (2007).
618:
619: \bibitem[MUY]{MUY} F. Mart\'{\i}n, M. Umehara, K. Yamada, {\it Complete
620: bounded null curves immersed in $\C^3$ and ${\rm SL}(2,\C)$.}
621: Preprint.
622:
623: \bibitem[MR]{MR} W. H. Meeks III, H. Rosenberg, {\it The uniqueness of the helicoid.} Ann. of Math. {\bf 161} (2), 727--758 (2005).
624:
625: \bibitem[N1]{N} N. Nadirashvili, {\it Hadamard's and Calabi-Yau's conjectures on negatively curved and minimal surfaces.}
626: Invent. Math. {\bf 126} (3), 457--465 (1996).
627:
628: \bibitem[N2]{N2} N. Nadirashvili, {\it An application of potential analysis
629: to minimal surfaces.} Dedicated to the memory of I. G. Petrovskii
630: on the occasion of his 100th anniversary. Mosc. Math. J. {\bf 1}
631: (4), 601--604 (2001).
632:
633: \bibitem[UY]{UY} M. Umehara, K. Yamada, {\it Maximal surfaces with singularities in Minkowski
634: space.} Hokkaido Math. J. {\bf 35} (1), 13--40 (2006).
635:
636: \end{thebibliography}
637:
638: \vspace*{1cm}
639:
640: \noindent
641: {\bf Antonio Alarc\'{o}n}\\
642: Departamento de Geometr\'{\i}a y Topolog\'{\i}a, \\
643: Universidad de Granada, E-18071 Granada, Spain. \\
644: e-mail: alarcon@ugr.es.
645:
646: \vspace*{0.5cm}
647:
648: \noindent
649: {\bf Isabel Fern\'{a}ndez}\\ Departamento de Matem\'{a}tica Aplicada I,\\
650: Universidad de Sevilla, E-41012 Sevilla, Spain.\\ e-mail:
651: isafer@us.es
652:
653:
654:
655:
656: \end{document}
657: