MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let M be a closed spin manifold of dimension $d$. One form of the elliptic genus of $M$ is $$ F(q)=q^{-d/8} \hat A(M) {\rm ch} \otimes_{k=1/2,3/2,\cdots} \Lambda_{q^k}T \otimes_{\ell=1}^\infty S_{q^\ell}T [M] $$ where the notation follows that of E. Witten, ``The Index of the Dirac Operator in Loop Space." The coefficient of $q^{n/2-d/8}$ is the index of a Dirac operator $D_n$ which acts on sections of $S \otimes T_{R_n}$ where $S$ is the spinor bundle and $T_{R_n}$ is the bundle associated to a representation $R_n$ of $Spin(d)$ with the first few representations being $$ R_0=1, \qquad R_1=T, \qquad R_2=\Lambda^2 T \oplus T $$ where $T$ is the fundamental (vector) representation.

I'm interested in the generalization of the elliptic genus to manifolds with boundary. In the actual application I'm interested in one eventually takes the boundary to infinity to obtain a noncompact manifold, but I'd be happy to understand the situation for a compact manifold with boundary first. The index of the Dirac operator in such a situation acquires boundary corrections of the form $$ CS[ \partial M] - \frac{1}{2}(\eta(0)+h) $$ where $h$ is the number of zero modes of the Dirac operator on $\partial M$ and $\eta(0)$ is the $\eta$ invariant. In the examples I'm interested in I believe the Chern-Simons contributions $CS[\partial M]$ vanish.

Summing up these boundary contributions to the index of $D_n$ weighted by $q^{n/2-d/8}$ leads to a "boundary" contribution to the elliptic genus on manifolds with boundary with the "bulk" contribution given by $F(q)$. My questions are whether this variant of the elliptic genus has been studied and if so where, whether this leads to interesting invariants of manifolds with boundary, and whether the modular properties of the bulk and boundary contributions are known.

share|cite|improve this question
up vote 8 down vote accepted

This, or a similar variant has been studied in Secondary Invariants for String Bordism and tmf and The f-invariant and index theory. The deviation from modularity caused by the boundary gives the interesting invariant of the boundary.

share|cite|improve this answer

I studied such modular invariants in my thesis, Modular invariants for manifolds with boundary, , , , and never find the time to get back to them, they are very interesting.

share|cite|improve this answer
Thank you for the reference. – Jeff Harvey Jan 19 '12 at 13:46

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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