As a quick reminder, the Gel'fand Yaglom theorem uses the generalized zeta-function approach to compute functional determinants of differential operators. Given a differential operator on $x\in [0,1]$


The functional determinant can be obtained from a solution $y(x)$ of the following differential equation with initial conditions $y(0)=0$ and $y'(0)=1$:


so that the determinant is given by

$$\text{det}O_1 = y(1).$$


Lets now assume that a more complicated differential operator is given on $x \in [0,1]$ by:

$$O_2=\left(\frac{d^2}{dx^2}+V_1(x)\right) \binom{1~~~0}{0~~~1}+V_2 (x)\binom{0~~~1}{1~~~0}+V_3 (x)\binom{0~-i}{i~~~~0}$$ $$=\left(\frac{d^2}{dx^2}+V_1(x)\right) 1_{2 \times 2}+V_2 (x)\sigma_1+V_3 (x)\sigma_2$$

The Pauli matrices $\sigma_1$ and $\sigma_2$ are orthogonal, and lets take $V_2 (x)$ and $V_3 (x)$ such that the operator indeed cannot be diagonalized by any transformation. Is the Gel'fand Yaglom theorem still aplicable here, in the sense of solving

$$O_2~ {\vec y}(x) = {\vec 0} $$

with initial conditions ${\vec y}(0)={\vec 0}$ and ${\vec y}'(0)={\vec 1}$? And obtaining the determinant from:

$$\text{det}O_2 = \prod_i y_i(1) = y_1 (1) y_2 (1)~?$$

Formally, I do not see any problem. Considering the generalized zeta-function origin of the technique I also do not see that any eigenvalues get lost. But maybe I miss some crucial point? A confirmation or falsification would be nice.


Not an expert, but you caught my interest. Looks like you may need to be a little careful in higher dimensions, but there are versions of the theorem which work. See, for example:



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.