This is from Hugh Osborn's 'Advanced Quantum Field Theory' notes, Lent 2013, page 15.

I want to evaluate the expression

$$ Z = \exp\Big(\frac{1}{2} \frac{\partial}{\partial \underline{x}} . A^{-1} \frac{\partial}{\partial \underline{x}} \Big) \exp\Big(-V(x) + \underline{b}. \underline{x}\Big) \bigg\vert_{\underline{x} = \underline{0}}$$

assuming that $\underline{b} = \underline{0}$.

We use the notation

$$ V_{i_{1} i_{2} \dots i_{k}} = \frac{\partial}{\partial x_{i_{1}}} \frac{\partial}{\partial x_{i_{2}}} \dots \frac{\partial}{\partial x_{i_{k}}} V(\underline{x})\Big\vert_{\underline{x} = \underline{0}}$$


$\frac{\partial}{\partial \underline{x}} \equiv \Big( \frac{\partial}{\partial x_{1}}, \dots, \frac{\partial}{\partial x_{n}}\Big)$

We also assume that $V(\underline{0}) = V_{i}(\underline{0}) = 0$.

And $A^{-1}$ is an $n \times n$ matrix. I expanded the exponential with the derivative to get:

$$ Z = \Bigg(1 + \frac{1}{2} \frac{\partial}{\partial x_{i}} A^{-1}_{ij} \frac{\partial}{\partial x_{j}} + \frac{1}{2} \frac{1}{2}\frac{\partial}{\partial x_{i}} A^{-1}_{ij} \frac{\partial}{\partial x_{j}} \frac{1}{2}\frac{\partial}{\partial x_{k}} A^{-1}_{kl} \frac{\partial}{\partial x_{l}} + \frac{1}{6}\frac{1}{2} \frac{\partial}{\partial x_{i}} A^{-1}_{ij} \frac{\partial}{\partial x_{j}}\frac{1}{2} \frac{\partial}{\partial x_{k}} A^{-1}_{kl} \frac{\partial}{\partial x_{l}}\frac{1}{2} \frac{\partial}{\partial x_{m}} A^{-1}_{mn} \frac{\partial}{\partial x_{n}} + \dots\Bigg) \exp(-V(x))\biggr\vert_{\underline{x} = \underline{0}}$$

Then this series acts on $\exp(-V(x))$, where summation notation is implied. Am I proceeding correctly?

The answer is enter image description here

  • $\begingroup$ It is not clear if what is expected is an expansion in $V$ or in $A^{-1}$... though I would assume the former. $\endgroup$ – Adam Jul 17 '19 at 10:36
  • $\begingroup$ @Adam I don't understand. How would you take derivatives without expanding in $A^{-1}$? $\endgroup$ – saad Jul 17 '19 at 11:18
  • $\begingroup$ For instance, $e^{a\partial_x}e^{f(x)}=e^{f(x+2)}$, but that is beside the point. In any case, you didn't expand the first exponential correctly $\endgroup$ – Adam Jul 17 '19 at 11:42
  • $\begingroup$ @Adam does the expansion look correct now? $\endgroup$ – saad Jul 18 '19 at 10:35
  • $\begingroup$ Yes, this should give you the expected result $\endgroup$ – Adam Jul 18 '19 at 11:17

The corrected expansion does indeed give you the right answer. For example,

$$ \frac{1}{2} \frac{\partial}{\partial x_{i}} A^{-1}_{ij} \frac{\partial}{\partial x_{j}} \exp(-V(x)) \big\vert_{x = 0} = \frac{1}{2} \frac{\partial}{\partial x_{i}} A^{-1}_{ij} (-\frac{\partial}{\partial x_{j}}V(x)) \exp(-V(x)) \big\vert_{x = 0} = -\frac{1}{2} A^{-1}_{ij} \Big[\frac{\partial}{\partial x_{i}} \frac{\partial}{\partial x_{j}} V(x) - \frac{\partial}{\partial x_{i}} V(x)\Big] \exp(-V(x))\Big|_{x = 0} = -\frac{1}{2}A_{ij}^{-1}V_{ij}$$

And so on and so forth for higher derivatives.

| cite | improve this answer | |
  • 1
    $\begingroup$ The second line is not correct. $\endgroup$ – Adam Jul 19 '19 at 11:22

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.