Search the corpus

    Closed surfaces with one $0$-cell and two $1$-cells

    P-3WQES

    Area
    Topology
    Topics
    Cell Complexes, Classification, Surfaces
    Seen at
    UGA
    Years
    2014
    Status
    draft

    Identify (with proof, but of course you can appeal to the classification of surfaces) all of the compact surfaces without boundary that have a cell decomposition having exactly one 0-cell and exactly two 1-cells (with no restriction on the number of cells of dimension larger than 1).

    Solution

    1. Any such cell decomposition is connected and has a finite number, say \(r\), of \(2\)-cells.

    The unique \(0\)-cell lies in the \(0\)-skeleton. Each \(1\)-cell is attached to that vertex at both endpoints, so the \(1\)-skeleton is connected. Every \(2\)-cell is attached to the \(1\)-skeleton, hence the whole surface is connected.

    Since the surface is compact and the given cell decomposition is a CW decomposition, the surface is contained in a finite subcomplex; because the whole space is the surface, only finitely many cells occur. As the ambient space is a surface, the cells have dimensions at most \(2\). Let \(r\) denote the number of \(2\)-cells.

    2. One necessarily has\[ r\ge1. \]

    If \(r=0\), the surface would equal its \(1\)-skeleton, a graph with one vertex and two edges. Its second homology with \(\mathbb Z/2\) coefficients would therefore vanish.

    On the other hand, every connected compact surface without boundary has a mod-\(2\) fundamental class, so \[ H_2(S;\mathbb Z/2)\cong\mathbb Z/2. \] This contradiction shows that at least one \(2\)-cell is present.

    3. The Euler characteristic satisfies\[ \chi(S)=1-2+r=r-1\ge0. \]

    For a finite CW decomposition, \[ \chi(S)=c_0-c_1+c_2. \] Here \[ c_0=1, \qquad c_1=2, \qquad c_2=r. \] Thus \[ \chi(S)=1-2+r=r-1. \] By 2, \(r\ge1\), so \(\chi(S)\ge0\).

    4. By the classification of compact connected surfaces, the only possible surfaces are\[ S^2, \qquad \RP^2, \qquad T^2, \qquad K, \]where \(K\) is the Klein bottle.

    For the closed orientable surface \(\Sigma_g\) of genus \(g\), \[ \chi(\Sigma_g)=2-2g. \] The condition \(\chi\ge0\) gives \[ g=0\quad\text{or}\quad g=1, \] namely \(S^2\) and \(T^2\).

    For the closed nonorientable surface \(N_k\) of nonorientable genus \(k\ge1\), \[ \chi(N_k)=2-k. \] The condition \(\chi\ge0\) gives \[ k=1\quad\text{or}\quad k=2, \] namely \(\RP^2\) and the Klein bottle. No other closed connected surface has nonnegative Euler characteristic.

    5. The sphere \(S^2\) admits a decomposition with one \(0\)-cell and two \(1\)-cells.

    Embed a figure-eight graph \[ S^1\vee S^1 \] in \(S^2\) so that the two circles meet at a single point. Use the common point as the unique \(0\)-cell and the two circle interiors as the two \(1\)-cells. The complement of this figure-eight in \(S^2\) has three connected components, each an open disk. Taking these three components as \(2\)-cells gives a CW decomposition with cell counts \[ (c_0,c_1,c_2)=(1,2,3). \] Indeed, its Euler characteristic is \[ 1-2+3=2=\chi(S^2). \]

    6. The projective plane \(\RP^2\) admits such a decomposition with two \(2\)-cells.

    Start with the standard polygon model of \(\RP^2\) as a \(2\)-gon whose boundary word is \[ aa. \] Its two polygon vertices are identified to one vertex, and the two boundary sides are identified to one \(1\)-cell \(a\).

    Inside the polygon add an arc \(b\) joining the two vertices. After the vertex identification, the interior of this arc is a second \(1\)-cell whose two endpoints are the unique vertex. The arc divides the polygon into two disks, which become two \(2\)-cells after the quotient. This is merely a subdivision of the standard polygon model, so the quotient is still \(\RP^2\). Thus \[ (c_0,c_1,c_2)=(1,2,2). \]

    7. The torus \(T^2\) admits such a decomposition with one \(2\)-cell.

    Use the standard square model with boundary word \[ aba^{-1}b^{-1}. \] All four vertices are identified to one \(0\)-cell, the edge classes \(a\) and \(b\) give exactly two \(1\)-cells, and the square interior is one \(2\)-cell. Hence \[ (c_0,c_1,c_2)=(1,2,1). \]

    8. The Klein bottle \(K\) also admits such a decomposition with one \(2\)-cell.

    Use the standard square model of the Klein bottle with boundary word \[ aba^{-1}b. \] Again all four vertices become one \(0\)-cell, the two edge classes \(a,b\) give exactly two \(1\)-cells, and the square interior gives one \(2\)-cell. Thus \[ (c_0,c_1,c_2)=(1,2,1). \]

    9. Therefore the complete list is\[ \boxed{S^2,\ \RP^2,\ T^2,\ \text{and the Klein bottle}.} \]

    Necessity follows from <1>3–<1>4, and <1>5–<1>8 explicitly realize every surface on the list.