Published on

Geometric Topology 10

Authors

Geometric Topology 10

Manufacturing Surfaces

We shall be concerned with connected compact surfaces initally without boundary. These come in two families: orientable and non-orientable. The former ones can be embedded into R3\mathbb{R}^3, the others not (unless they have a boundary like the Mobius band). We shall classify both types.

Definition: A surface without boundary (or 22-manifold) is a Hausdorff topological space, each point of which lies in an open set that is homeomorphic to a disk in R2\mathbb{R}^2.

Examples:

  • A torus
  • A 'fat' trefoil knot

Cartesian Equations

Exercise: The equations all describe subsets of R3\mathbb{R}^3. Which are surfaces? Which are connected and which are compact? What do these look like in space?

  1. x+y+z=1|x| + |y|+|z|=1,
  2. yz+zx+xy=0yz+zx+xy=0,
  3. xyz=1xyz=1,
  4. x8+y8+z8=1000x^8+y^8+z^8=1000,
  5. 1x+1y+1z=0\frac{1}{x} + \frac{1}{y} + \frac{1}{z} = 0

Models for Surfaces

If enough cuts are made on a surface, so it is subdivided by a graph of vertices and edges, the pieces can be deformed and laid out as plane figures. The surface can then be reconstructed by 'sewing' along the cuts. In four simple cases, a single square suffices with the cuts defining pairs of edges.

We can define the models of four surfaces as follows:

  1. Torus: aba1b1aba^{-1}b^{-1}
  2. Klein Bottle: aba1baba^{-1}b
  3. Projective Plane: abababab
  4. Sphere: $abb^-1a^-1

By labelling the edges, and reading clockwise from the top left, these models can be described by the 'words' above. The last two can be simplified AAAA and AA1AA^{-1}.

Embedded Surfaces in R3\mathbb{R}^3

A compact surface M\mathscr{M} is embedded in R3\mathbb{R}^3 if there is a continuous injective map MR3\mathscr{M} \rightarrow \mathbb{R}^3. It will then be homeomorphic to its image.

As a topological space, the unit sphere:

S2(1)={(x,y,z):x2+y2+z2=1}R3S^2(1) = \{(x,y,z) : x^2 + y^2 + z^2 = 1\} \subset \mathbb{R}^3

is what we get when we take a blue disk, push it into a bag and fasten its top boundary up with a 'zip'.

The symbol SS stands for any surface homeomorphic to S2(1)S^2(1), for example the octahedron x+y+z=1|x|+|y|+|z|=1. We have already seen the torus parametrized by a rectangle R\mathscr{R} (equivalently square). All 44 vertices of R\mathscr{R} map to common point pp on the torus, and the 44 edges of mathscrRmathscr{R} are identified in pairs and mapped to 22 circles intersecting in pp.

Immersed Surfaces in R3\mathbb{R}^3

  • The Klein bottle KK is formed by passing a cylinder through itself. It can be mapped almost injectively into R3\mathbb{R}^3, but there is a circle of self-intersection that actually represents two disjoint circles on KK.
  • The projective plane PP can be immersed in R3\mathbb{R}^3 as a sphere with crosscap, parametrized by {(cosusin2v,sinusin2v,cos2vcos2usin2v)}\{(\cos u \sin 2v, \sin u \sin 2v, \cos^2 v - \cos^2 u \sin^2 v)\}. It has a segment of self-intersection, where the image passes through itself.

These immersed surfaces in R3\mathbb{R}^3 are one-sided in the sense that the image of a continuous path in PP or KK can travel to all sides.

Classification

Our next aim is to prove this main theorem:

Main theorem (v1): Any connected compact surface without boundary is homeomorphic to one of the following:

  1. A sphere SS, or a sphere with g1g \geq 1 handles, equivalently a torus with gg holes.
  2. A sphere with h1h \geq 1 crosscaps.

In case 1, the surface M\mathscr{M} can be embedded in R3\mathbb{R}^3 and is orientable. The integer gg is called its genus, and SS itself corresponds to g=0g=0. One can regard a surface of genus gg as the 'connected sum' (in analogy of knots) of gg tori, and one often writes M=T#...#Tg=gT\mathscr{M} = \overbrace{T \# ... \# T}^g = gT, where "==" really means "homeomorphic to".

In (2), the surface must contain a Mobius band and is non-orientable. Any such surface can be regarded as the connected sum of hh copies of the projective plane, so M=hP\mathscr{M} = hP. We shall see that the Klein bottle has h=2h=2. so we write K=P#P=2PK = P \# P = 2P. A key step in the proof of the main theorem is to show that K#P=T#PK \# P = T \# P.