I don't think there is a consensus of norm conserving PP being more accurate. There are some references I am aware of which have calculated dielectric tensor using NC-PP, but without justification though:
- Yu, E. K.; Stewart, D. A.; Tiwari, S. Ab Initio Study of Polarizability and Induced Charge Densities in Multilayer Graphene Films. Phys. Rev. B - Condens. Matter Mater. Phys. 2008, 77 (19), 1–8.
- Kumar, P.; Bhadoria, B. S.; Kumar, S.; Bhowmick, S.; Chauhan, Y. S.; Agarwal, A. Thickness and Electric-Field-Dependent Polarizability and Dielectric Constant in Phosphorene. Phys. Rev. B 2016, 93 (19), 1–8.
But then I have seen often (and performed calculations myself) PAW-based methods in VASP used for calculating the dielectric tensor. Here is a reference for you to dig deeper:
- Laturia, A.; Van de Put, M. L.; Vandenberghe, W. G. Dielectric Properties of Hexagonal Boron Nitride and Transition Metal Dichalcogenides: From Monolayer to Bulk. npj 2D Mater. Appl. 2018, 2 (1), 6.
One way to validate your method is to compare polarizabilities (related to dielectric constant) with more accurate methods such as polarizability calculations using maximally localized Wannier functions - see below for some insights:
- Pan, D.; Govoni, M.; Galli, G. Dielectric Properties of Condensed Systems Composed of Fragments. J. Chem. Phys. 2018, 149 (5), 051101.