8a79c64ddb5c0286.tex
1: \begin{abstract}
2: We consider the differential sum rule for the  effective scattering rate $%
3: 1/\tau (\omega)$ and optical conductivity $\sigma _{1}(\omega ) $ in a 
4: dirty BCS superconductor, for arbitrary ratio of the superconducting gap $%
5: \Delta$ and the normal state constant damping rate $1/\tau$. We show that if 
6: $\tau$ is independent of $T$,  the area under $1/\tau (\omega)$ does not
7: change between the normal and the superconducting states, i.e., there exists
8: an exact  differential sum rule for the scattering rate.  For \textit{any}
9: value of the dimensionless parameter $\Delta\tau $, the sum rule is
10: exhausted at frequencies controlled by $\Delta$. 
11: %but the numerical convergence is weak. 
12: We show that in the dirty limit the convergence of the differential 
13: sum rule for the scattering rate is much faster then the convergence of 
14: the $f-$sum rule, but slower then the convergence of the differential 
15: sum rule for conductivity.
16: \end{abstract}
17: