7
$\begingroup$

Let $E$ be a spin$^c$ bundle and $spin^c(E)$ the corresponding $spin^c(n)$-principial bundle. Let $g_{U,V}: U \cap V \to spin^c(n)$ denote transition functions for this principial bundle and consider the map $\nu:spin^c(n) \to \mathbb{T}$ defined by $\nu(w)=w^!w$ where $(v_1 \cdot ... \cdot v_r)^!=v_r \cdot ... \cdot v_1$. We can form the composition $\nu \circ g_{U,V}:U \cap V \to \mathbb{T}$. As this satisfies cocycle property we can form the line bundle $L_E$.

Let us assume that $E$ is $spin^c$ but is not spin.

How to prove that the first Chern class of $L_E$ is odd (in the sense that $j^*(c_1(L_E)) \neq 0 $ where $j^*$ is mod 2 reduction of coefficients)?

$\endgroup$
3
  • 1
    $\begingroup$ Since $E$ is a spin^c bundle, $w_2(E)$ is the mod 2 reduction of $c_1(L_E)$ (here $w_2$ is the second Stiefel-Whitney class). $w_2$ is precisely the obstruction for a spin structure, so since $E$ is not spin, $w_2(E)\ne 0$, and hence $c_1(L_E)$ is odd. $\endgroup$ Commented Feb 17, 2018 at 3:54
  • 1
    $\begingroup$ This is exactly my point: I would like to prove that $w_2(E)$ is precisely mod 2 reduction of $c_1(L_E)$.In order to prove this I managed to show that both sides are natural wih respect to pullbacks and both side vanish on spin bundles. The remaning part is that this mod 2 reduction does not vanish if the bundle is not spin and this boils down to my original question. So how you can show that $w_2(E)$ is mod 2 reduction of $c_1(L_E)$? $\endgroup$
    – truebaran
    Commented Feb 17, 2018 at 13:58
  • $\begingroup$ Ah, of course; sorry about that. One way to do that would be to prove it for the universal bundle over $\mathit{BSpin}^c$, but there ought to be an easier proof. $\endgroup$ Commented Feb 17, 2018 at 17:46

1 Answer 1

6
+100
$\begingroup$

One way to prove this is to first check that it is true when $E$ is the spin$^c$ bundle canonically associated to a complex vector bundle $F$ and then argue that complex and spin$^c$ bundles have the essentially same characteristic classes in degree up to 4.

To associate a spin$^c$-bundle to a complex vector bundle, consider the inclusion $\rho: U \hookrightarrow Spin^c$ obtained by lifting $i \times \det : U \hookrightarrow SO \times U(1)$ by the double cover $Spin^c \to SO \times U(1)$ (where $i : U \hookrightarrow SO$ is the obvious inclusion). Given a principal $U$-bundle $F$, we can thus consider $E = F \times_\rho Spin^c$, i.e. the quotient $(F \times Spin^c)/U$ where $h \in U$ acts by $(f,g) \mapsto (f\cdot h, \,\rho(h^{-1})g)$. Then we get a well-defined action of $Spin^c$ on $E$ by setting $[(f,g)]\cdot h \mapsto [(f, g\rho(h))]$ for $h \in Spin^c$, and this makes $E$ a principal $Spin^c$-bundle.

In a similar way, the fundamental line bundle $L_E$ of any spin$^c$ bundle $E$ is the $U(1)$-bundle associated to $E$ by the homomorphism $\psi : Spin^c \to U(1)$ defined by composing the double cover of $SO \times U(1)$ with projection to the right factor. Since $\psi \circ \rho : U \to U(1)$ is precisely $\det$, the fundamental line bundle of the spin$^c$ bundle associated to a complex vector bundle is simply $\det F$. Hence $c_1(L_E) = c_1(\det F) = c_1(F)$ in this case, and it is well-known that $w_2(E) = c_1(F)$ mod 2.

For the final part, note that the classifying map $f : BU \to BSpin^c$ for the universal complex bundle $EU$ as a spin$^c$ bundle can be represented as a fibre bundle with fibres $Spin^c/U$. For if $ESpin^c$ is a contractible space with a free $Spin^c$ action, then the subgroup $U$ also acts freely. Hence we can take $ESpin^c/U$ as a representative of $BU$, and then $f$ is simply the projection $ESpin^c/U \to ESpin^c/Spin^c$, whose fibres are $Spin^c/U$.

This fibre has a transitive action by $Spin$ with stabiliser $SU$, so it's homeomorphic to $Spin/SU$. In degree $0 < i \leq 5$, the homotopy groups are $$ \pi_i(Spin/SU) = \pi_i(Spin(6)/SU(3)) = \pi_i(SU(4)/SU(3) = \pi_i(S^7) = 0, $$ implying that $f$ is a 5-equivalence (i.e. $f_* : \pi_i(BU) \to \pi_i(BSpin^c)$ is an isomorphism for $i < 5$, and surjective for $i = 5$). Therefore $f^* : H^i(BSpin^c) \to H^i(BU)$ is an isomorphism for $i \leq 4$. Hence any relation between spin$^c$ characteristic classes of degree up to $4$ that is valid for complex vector bundles is valid for any spin$^c$ bundle.

$\endgroup$
2
  • $\begingroup$ Thank you for your answer. I don't know what to do since I would be happy to accept your answer and award the boutny for your help. However still I'm missing many details and I'm unable to fill them. Ley me list some of them: 1. I don't fully understand how you associate spin^c bundle to given complex bundle. 2. I would also like to see more details what's going on in part one. 3. Finally what do you mean by saying that the classifying map can be represented as a fiber bundle? $\endgroup$
    – truebaran
    Commented Mar 2, 2018 at 0:31
  • $\begingroup$ I've added some additional explanation, hope that helps. $\endgroup$ Commented Mar 2, 2018 at 9:02

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .