International Journal of Mathematics and Mathematical Sciences

Volume 2008, Article ID 350326, 20 pages

http://dx.doi.org/10.1155/2008/350326

## The Schwarz-Christoffel Conformal Mapping for “Polygons” with Infinitely Many Sides

Departamento de Matemáticas, Pontificia Universidad Católica de Chile, Avenue Vicuña Makenna 4860, 7820436 Macul, Santiago, Chile

Received 22 January 2008; Revised 15 May 2008; Accepted 1 July 2008

Academic Editor: Vladimir Mityushev

Copyright © 2008 Gonzalo Riera et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

#### Abstract

The classical Schwarz-Christoffel formula gives conformal mappings of the upper half-plane onto domains whose boundaries consist of a finite number of line segments. In this paper, we explore extensions to boundary curves which in one sense or another are made up of infinitely many line segments, with specific attention to the “infinite staircase” and to the Koch snowflake, for both of which we develop explicit formulas for the mapping function and explain how one can use standard mathematical software to generate corresponding graphics. We also discuss a number of open questions suggested by these considerations, some of which are related to differentials on hyperelliptic surfaces of infinite genus.

#### 1. Introduction

Let be a polygon in the complex plane with vertices and interior angles ; the exterior angles are given by , where . The Schwarz-Christoffel mapping of the upper-half plane onto is effected by a function of the form where are real and .

Although there is ample literature on the subject (see, e.g., [1–3], as well as the complete overview contained in Driscoll and Threfethen [4]), a few remarks would not be out of place here. First of all, even though (1.1) gives the appearance of being an explicit formula, there is actually no known relation between the values of the and the lengths of the sides of . For this reason, when considering an infinite-sided polygon, one needs to justify passage to a corresponding infinite product, which is what we do in what follows. Also, formula (1.1) is necessary but not sufficient in the sense that for some values of the 's, the mapping need not to be one-to-one (see [5]). To establish the necessity, one needs to consider the behavior of at , a possibility also clearly excluded in the case of infinitely many .

Before proceeding, we review some variations of (1.1). One of the points may be located at , the formula remaining the same, and one of the vertices of the polygon may be at , where we use the relation . The formula for mapping the unit disc conformally onto is essentially the same, but with . The formula for the mapping onto the exterior of the polygon is with .

In this paper, we explore the case in which there are infinitely many points on the real line (or the unit circle) and obtain results for two kinds of “polygons."

##### 1.1. Polygons with an Infinite Number of Sides

This is the case, for example, of an “infinite stairway” with interior angles alternately equal to and . The formula in this case is

The zeros of the denominator correspond to and those of the numerator to .

##### 1.2. Fractals

This is the case, for example, of the interior of the Koch snowflake, for which the formula on the interior of the unit disc turns out to be

A similar generalization of (1.2) gives the conformal mapping onto the exterior of the Koch snowflake.

To reproduce the figures in the text, a program such as Derive, Maple, or Mathematica is required; we explain in each case how to obtain these figures.

To our knowledge, formulas of this kind have not previously been considered in the literature, and the initial nature of the present investigation precludes exhaustive results. It is our hope that others will be motivated to seek general theorems along these lines, or at any rate to find explicit mappings of the upper-half-plane onto other infinitely sided polygons. We mention here the open problem of a half-strip slit along infinite sequences of equal-length line segments emanating from the infinite sides. (1.5) The mapping of the half-plane onto this domain is of importance in the study of differentials on hyperelliptic surfaces of infinite genus as considered in the pioneering work by Myrberg, see [5]. (The formula for the case of the analogously slitted full strip is obtained below.)

#### 2. Polygons with an Infinite Number of Sides

##### 2.1. The Infinite Staircase

The first example we consider is the mapping defined on the upper half-plane by the formula

The image is given in Figure 1, where the curves shown are the images of horizontal and vertical segments lying in the half-plane; they can be defined directly by for a vertical segment, and by for a horizontal segment.

The image of the real axis is the infinite staircase with steps of equal length

In order to prove that this formula actually defines a one-to-one mapping, we proceed as follows.

Consider the triangle (i.e., half-strip) of Figure 2.

The image of the positive imaginary axis (labeled (1) in the figure) is given by for , so that it is the half-line making an angle of with the positively oriented real axis. The image of (2) is the real segment , and that image of (3) is a half-line parallel to the first one and beginning at as in Figure 3.

Consider the triangle of Figure 3.

The general theory of conformal mappings applies to these two triangles, so that from the bijectivity on the boundaries we can conclude bijectivity in the interior. With the Schwarz reflection principle as applied to the vertical segments (1), (3), and their reflections, the mapping can now be extended to the entire upper half-plane.

The function with , generalizes this to zigzag patterns.

##### 2.2. The Hairy Half-Plane

We next consider interior angles at

The mapping function is therefore and the image is shown in Figure 4.

To establish that the given mapping is indeed one-to-one, it is first necessary to see that the length of the segment from an angle to the tip of the hair is the same on both sides, that is, that an equality easily obtained by the change of variable .

The proof can now be completed in the same way as for the infinite staircase by first considering the image of a triangle with vertical sides at and , and then extending the mapping by the reflection principle.

In a similar fashion, we obtain a formula for the hairy plane, which can be regarded as a polygon with infinitely many sides and interior angles alternately equal to and .

This means taking as in Figure 5.

##### 2.3. The Infinite Half-Staircase

In order to consider only one half of the stair, we have to take analytic functions in (1.1) having half as many zeros as before, which leads to the function. Indeed, the formula defines such a mapping (see Figure 6).

However, to prove that the mapping given in (2.10) is one-to-one, we cannot use the arguments of examples 1 and 2 since there is no symmetry with respect to a line. We proceed as follows. Since it follows that

The classical Schwarz-Christoffel formula can be applied to each finite product on the right-hand side, which gives a staircase with a finite number of steps. These conformal mappings converge uniformly to the required mapping since for each positive integer, the integral is finite.

To obtain a half-stair with the steps moving to the left one would naturally use

Finally, to obtain half a hairy plane, we take as in Figure 7, or half a hairy half-plane via and so on.

#### 3. Fractals

##### 3.1. The Koch Snowflake

Driscoll and Trefethen ([4, Chapter 4.5]) remark that the conformal mapping into the interior or exterior of complicated fractals, such as the Koch snowflake, ought to be treatable in a manner related to the multiple method. In the next theorem, we give an explicit geometric formula which does this.

Numerical schemes have been studied before using iteration procedures, see [6].

The Koch snowflake is obtained as the limit of a sequence of polygons defined as in Steps 1, 2, and 3, and so on.

*Step 1. *(3.1)

*Step 2. *(3.2)

*Step 3. *(3.3)

At each step, we consider the Schwarz-Christoffel mapping of the unit disc into the interior of the polygon; we use the notation of the introduction throughout.

*Step 1. *

This conformal mapping is portrayed in
Figure 8.

*Step 2. * The formula for the conformal mapping is and the image of the unit disc appears in
Figure 9.

*Step 3. *

Although the combinatorics is not automatic, the formula for the conformal mapping is the image of the unit disc is given in Figure 10.

We may generalize the following.

*Step 3.1 . * with an integrand of the form

Theorem 3.1. *The formula for the
conformal map from the interior of the unit disc to the interior of a Koch
snowflake is*

*Proof. *The wording “a Koch snowflake” means that it is
the limit of polygons with the same angles as the polygons in the Koch pattern,
but with unequal sides.

First, we rewrite the integrand for
Step as
(modulo a cube
root of 1), so that for the integrand converges to the one indicated
in the statement of the theorem.

That the lengths are not equal to the lengths in true
Koch snowflake can already be seen at Step 3:

Define so that is one vertex
of the polygon. But then

The image of a circular sector from to to is therefore a polygon of the form of
Figure 11.

If the lengths
were equal to those of the regular snowflake then the vertex would be the vertex of the isosceles triangle
with base and base angle .

The side would equal .
But
showing that
the second length in the small triangle is necessarily shorter than the first
one, leading to a location of slightly “higher” than it should be.

Observe also that the location of this last vertex
implies (the angles being as required) that the map is one-to-one on the
boundary and therefore one-to-one throughout.

In general, the image in Step differs somewhat, but not by too much, from a
regular snowflake with all sides equal.

If we analyze a sufficiently general case as in Step
3, say that we have the mapping

It is easy to obtain the points

If we compare them to a regular snowflake with
vertices ,
we obtain points at a distance smaller than .

What is to be
noticed is that for the image not be one-to-one on the boundary, the point should differ sufficiently from ,
to be below ,
for example; the position of thus forces the actual polygon to be free of
self-intersections, the angles being or .
We draw one such possible polygon in Figure 12.

The sequence of conformal mappings therefore converges
to a nonconstant conformal mapping (since ).

We check that the image is not all of .

We claim that

Indeed, with , we have and

Thus the integral is bounded from above by

We also check that the image is not a disc.

Set

Then and .

But has modulus
smaller than .

Finally, we have to explain a subtle point that arises
in producing a drawing such as the one in Figure 10.

One cannot write, for example, since the
branch we implicitly use by analytic continuation may not coincide with the
fixed branches of the mathematical software.

With the numerator
of the integrand is and the
denominator is

At any rate,
the formula of the theorem with terms of up to is sufficiently precise (and simpler to use)
and gives Figure 13.

##### 3.2. The Exterior of the Snowflake

At Step 1, we use the formula of the introduction with .

In this case, so that the formula is whose image is in Figure 14.

In the same way, the formulas for Steps 2 and 3 are corresponding to Figures 15 and 16.

The mapping from the unit disc to the exterior of the Koch snowflake is given by

##### 3.3. Periodic Koch Fractals

In a paper on diffusive transport by Brady and Pozrikidis (cf. [6]) the authors consider conformal mappings of a rectangle onto regions as in Figure 16.

The mappings can then be extended by reflections on the vertical sides to mappings from an infinite strip to regions bounded below by the given contours. In the limit, a fractal region is obtained.

The mapping is given by an iteration procedure which we summarize as follows.

(i)Guess initial values .(ii)Locate the vertices on the triangular contours by and compare them to the known exact values of the vertices . If they are equal the calculation stops; otherwise compute improved values for using the corrector formula:

This procedure works well in the first few iterations but, as the authors explain, is computationally prohibitive already for (cf. [6] for details).

Using the Schwarz-Christoffel formula in a similar way as in number 2 above, we may define a conformal mapping to a region bounded below by a periodic fractal of Koch's type obtained from Figure 16. For consider the function

It maps the upper half plane to a region bounded by the periodic contour as in (extended by reflection on the vertical sides).

The actual lengths of the sides in the image are so that they do not coincide with preassigned value 1 for all sides, but the image has the same geometrical shape.

For , the function is defined by and in the limit, for ,

The infinite products converge uniformly in compact sets on the upper half plane and the integral defines a conformal map onto a region bounded below by an infinite periodic fractal of Koch type. As usual with these maps, we cannot so far control precisely the lengths of the sides except numerically for the first few steps.

##### 3.4. A Tree

We define a sequence of “trees” as follows.

*Step 1. *(3.36)

*Step 2. *(3.37)

*Step 3. *(3.38)

We want to define a sequence of Schwarz-Christoffel mappings from the unit disc to the complements of these trees.

The interior angles at Step 1 give with at the points and at the corners.

Therefore is the required formula at Step 1.

At Step 2, we obtain with branches at the 32 roots of unity.

Similarly, at Step 3

The image is shown in Figure 17.

In the limit, we obtain the formula

for the conformal mapping of the unit disc onto the exterior of this fractal.

There are of course many more trees to be considered but the combinatorics arising from angles and points is not straightforward, so that it may be hard to obtain a general expression for the conformal mapping. We leave these as questions for further study.

#### 5. Conclusion

The classical Schwarz-Christoffel formula maps the upper half plane to a polygon with prescribed angles, but with sides of unknown lengths. Notwithstanding this, we have shown that the formula can be extended to a variety of shapes with an infinite number of vertices, namely, stairs, slitted planes, fractals of Koch's type and trees. It would be very interesting in the future to find exactly the branch points necessary in Koch snowflake as well as in specific slitted regions.

#### Acknowledgment

This research was partially suported by Fondecyt 1050904.

#### References

- D. Bonciani and F. Vlacci, “Some remarks on Schwarz-Christoffel transformations from the unit disk to a regular polygon and their numerical computation,”
*Complex Variables and Elliptic Equations*, vol. 49, no. 4, pp. 271–284, 2004. View at Publisher · View at Google Scholar · View at Zentralblatt MATH · View at MathSciNet - P. Henrici,
*Applied and Computational Complex Analysis. Vol. 1: Power Series, Integration, Conformal Mapping, Location of Zeros*, Pure and Applied Mathematics, John Wiley & Sons, New York, NY, USA, 1974. View at Zentralblatt MATH · View at MathSciNet - E. Johnston, “A “counter example for the Schwarz-Christoffel transform,”
*The American Mathematical Monthly*, vol. 90, no. 10, pp. 701–703, 1983. View at Publisher · View at Google Scholar · View at Zentralblatt MATH · View at MathSciNet - T. A. Driscoll and L. N. Trefethen,
*Schwarz-Christoffel Mapping*, vol. 8 of*Cambridge Monographs on Applied and Computational Mathematics*, Cambridge University Press, Cambridge, UK, 2002. View at Zentralblatt MATH · View at MathSciNet - P. J. Myrberg, “Über analytische Funktionen auf transzendenten zweiblättrigen Riemannschen Flächen mit reellen Verzweigungspunkten,”
*Acta Mathematica*, vol. 76, no. 3-4, pp. 185–224, 1945. View at Publisher · View at Google Scholar · View at Zentralblatt MATH · View at MathSciNet - M. Brady and C. Pozrikidis, “Diffusive transport across irregular and fractal walls,”
*Proceedings of the Royal Society of London. Series A*, vol. 442, no. 1916, pp. 571–583, 1993. View at Google Scholar