Talk:Thom spaces (Ex)

From Manifold Atlas
(Difference between revisions)
Jump to: navigation, search
Line 1: Line 1:
<wikitex>;
<wikitex>;
Part 1
+
'''Part 1'''
We define
We define
Line 20: Line 20:
where $\mathrm{arg}(z)\in(0,2\pi)$.
where $\mathrm{arg}(z)\in(0,2\pi)$.
Part 2
+
'''Part 2'''
If $i$: $M\to\mathbb{R}^{n+k}$ is an embedding, we denote by $j$: $M\to\mathbb{R}^{n+k+1}$ the composition
If $i$: $M\to\mathbb{R}^{n+k}$ is an embedding, we denote by $j$: $M\to\mathbb{R}^{n+k+1}$ the composition
Line 44: Line 44:
$$
$$
Part 3
+
'''Part 3'''
Let the embeddings $i$ and $j$ be as in Part 2.
Let the embeddings $i$ and $j$ be as in Part 2.
Line 68: Line 68:
$$
$$
Part 4
+
'''Part 4'''
Of course one can do similar things for non oriented manifolds or spin manifolds.
Of course one can do similar things for non oriented manifolds or spin manifolds.
One only has to modify the definition of $\Omega_n(X)$ and use the corresponding universal bundle instead of $\xi_k$.
One only has to modify the definition of $\Omega_n(X)$ and use the corresponding universal bundle instead of $\xi_k$.
</wikitex>
</wikitex>

Latest revision as of 12:17, 2 April 2012

Part 1

We define

\displaystyle  \mathrm{Th}(\xi_1)\wedge\mathrm{Th}(\xi_2)\to\mathrm{Th}(\xi_1\times\xi_2),\quad [v_1,v_2]\mapsto \left\{ \begin{array}{ll}\infty, & \textrm{if }v_1=\infty\textrm{ or }v_2=\infty \\ v_1\oplus v_2, & \textrm{else}\end{array} \right.

and

\displaystyle  S^1\wedge\mathrm{Th}(\xi)\to\mathrm{Th}(\xi\oplus\underline{\mathbb{R}}),\quad [z,v]\mapsto \left\{ \begin{array}{ll}\infty, & \textrm{if }z=1\textrm{ or }v=\infty \\ v\oplus\cot(\mathrm{arg}(z)/2), & \textrm{else}\end{array} \right.

where \mathrm{arg}(z)\in(0,2\pi).

Part 2

If i: M\to\mathbb{R}^{n+k} is an embedding, we denote by j: M\to\mathbb{R}^{n+k+1} the composition of i with the inclusion \mathbb{R}^{n+k}=\mathbb{R}^{n+k}\times\{0\}\subset\mathbb{R}^{n+k+1}. In particular the normal bundles are related by \nu(M,j)=\nu(M,i)\oplus\underline{\mathbb{R}}. The bundle map (i_k,\overline{i_k}) induces

\displaystyle  \Omega_n(\overline{i_k}): \Omega_n(\gamma_k)\to\Omega_n(\gamma_{k+1}),\quad  [M,i,f,\overline{f}]\mapsto[M,j,i_k\circ f,\overline{i_k}\circ(\overline{f}\oplus\mathrm{id}_{\underline{\mathbb{R}}})].

From the definition

\displaystyle  V_k:\quad\Omega_n(\gamma_k)\to\Omega_n(X),\quad  [M,i,f,\overline{f}]\mapsto[M,\mathrm{pr}_X\circ f]

we find for all k\geq0

\displaystyle  (V_{k+1}\circ\Omega_n(\overline{i_k}))([M,i,f,\overline{f}]) =[M,\mathrm{pr}_X\circ i_k\circ f] =[M,\mathrm{pr}_X\circ(\mathrm{id_X\times j_k})\circ f] =[M,\mathrm{pr}_X\circ f] =V_k([M,i,f,\overline{f}]).

Part 3

Let the embeddings i and j be as in Part 2. Define the collapse maps

\displaystyle  c_k:\quad S^{n+k}\to\mathrm{Th}(\nu(M,i)),\quad c_{k+1}:\quad S^{n+k+1}\to\mathrm{Th}(\nu(M,j))

as in [Lück2001, page 57]. Then we have c_{k+1}=\Sigma c_k. For all k\geq0 we obtain

\displaystyle  (P_n(\gamma_{k+1})\circ\Omega_n(\overline{i_k}))([M,i,f,\overline{f}]) =P_n(\gamma_{k+1})([M,j,i_k\circ f,\overline{i_k}\circ(\overline{f}\oplus\mathrm{id}_{\underline{\mathbb{R}}})]) =[\mathrm{Th}(\overline{i_k}\circ(\overline{f}\oplus\mathrm{id}_{\underline{\mathbb{R}}}))\circ c_{k+1}] =[\mathrm{Th}(\overline{i_k})\circ\mathrm{Th}(\overline{f}\oplus\mathrm{id}_{\underline{\mathbb{R}}})\circ c_{k+1}]

and

\displaystyle  (s_k\circ P_n(\gamma_k))([M,i,f,\overline{f}]) =s_k([\mathrm{Th}(\overline{f})\circ c_k]) =[\mathrm{Th}(\overline{i_k})\circ\Sigma(\mathrm{Th}(\overline{f})\circ c_k)] =[\mathrm{Th}(\overline{i_k})\circ\mathrm{Th}(\overline{f}\oplus\mathrm{id}_{\underline{\mathbb{R}}})\circ c_{k+1}].

Part 4

Of course one can do similar things for non oriented manifolds or spin manifolds. One only has to modify the definition of \Omega_n(X) and use the corresponding universal bundle instead of \xi_k.

Personal tools
Namespaces
Variants
Actions
Navigation
Interaction
Toolbox