Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $X$ be a smooth variety over $\mathbb{C}$. If $\mathscr{F}$ is a coherent sheaf on $X$ with connection, does it follow that $\mathscr{F}$ is locally free? I can't think of any counterexamples. What if $X$ is a complex analytic manifold?

In his article "Regular connections after Deligne," Malgrange begins to sketch the proof: fix a point $x \in X$ and a minimal system of generators for $\mathscr{F}_x$. Consider a relation in which some coefficient has minimal order at $x$ and differentiate it. How do we finish the proof? I do not see how this calculation implies that the coefficients must vanish (and therefore $\mathscr{F}_x$ is free).

share|improve this question

2 Answers 2

See http://mathoverflow.net/questions/81338/coherent-sheaf-with-connection-is-locally-free

share|improve this answer
These proofs use quite a bit of machinery, and I do not understand them... Surely there is a simple proof of this "folklore" result? –  Justin Campbell May 2 '12 at 23:39

More details for the proof of Malgrange.

First we easily reduce to the local case. Fix a point $x$ and a minimal system of generators $v_1,\dots, v_n$ of $F_x$. We want to prove it is free. Let $m_x$ be the maximal ideal of $O_{X,x}$, let $$a_1v_1+\dots+a_nv_n=0, \quad a_i\in O_{X,x}$$ and let $r$ be the smallest non-negative integer such that $a_i\in m_x^r$ for all $i$ and $a_i\notin m_x^{r+1}$ for at least one $i$. Using the connection, we are going to find a new vanishing relation as above, but with smaller $r$.

So apply the connection to the above equality, we get $$\sum_i a_i\nabla(v_i)+\sum_i v_i\otimes da_i=0.$$ Note that $a_i\nabla(v_i)\in M\otimes \Omega_{X}^1$ and $da_i\in m_x^{r-1}\Omega_{X}^1\setminus m_x^r\Omega_{X}^1$ (here we use the characteristic $0$). As $\Omega_X^{1}$ is free, considering the above equality components by components, we see that there exist $b_i\in O_{X,x}$ such that $$\sum_i b_iv_i=0$$ at leat one $b_i\notin m_x^{r}$. So we get a new vanishing relation with smaller $r$. Repeating util we get $r=1$, so one of the coefficients is invertible and the system of generators is not minimal. Contradiction.

share|improve this answer

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.