It looks to me that the bordism group $$\Omega_3^{SO} (B(O(2) \times SO(3))) \tag{1}$$ (whose Pontryagin dual for the manifold generator) contains at least a nontrivial invariant: $$ w_1(O(2))\big(w_1(O(2))^2 +w_2(SO(3))\big). \tag{2} $$

Question: Is it true that if we lift this invariant respect to a new bordism group $$ \Omega^{SO}_3 (B((Pin^-(2) \times Spin(3))/ \mathbb{Z}_2)), \tag{3} $$ the invariant eq.(2) becomes zero in bordism group (2)?

My trial/attempt: The lifting is simply that the $O(2)$ bundle in eq(1) is lifted to $Pin^-(2)$ bundle in eq(3), and the $SO(3)$ bundle in eq(1) is lifted to $Spin(3)$ bundle in eq(3). There is an overall constraint: $$ w_2(O(2))+w_1(O(2))^2=w_2(SO(3)). \tag{4} $$

It also looks that, after I look it up, the $w_3(O(2))=w_1(O(2))w_2(O(2))+\frac{d w_2(O(2))}{2}=0$. But I am not sure that $\frac{d w_2(O(2))}{2}$ is well-defined. By using the fact that the O(2) bundle can be lifted to $Pin^-(2)$ bundle, we shall have this constraint $w_2(O(2))+w_1(O(2))^2=0$. But here we may need to apply eq.(4) instead.


1 Answer 1


Since you have the constraint $$w_1(O(2))^2+w_2(O(2))=w_2(SO(3)),$$ then $$w_1(O(2))(w_1(O(2))^2+w_2(SO(3)))=w_1(O(2))w_2(O(2))=Sq^1(w_2(O(2)))=w_1(TM)w_2(O(2))=0$$ by Wu formula. ($w_1(TM)=0$ since you are considering oriented bordism.)

$\frac{dw_2(O(2))}{2}=Sq^1(w_2(O(2)))$ by the definition of Bockstein homomorphism.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.