“…2 the PWGGA functional in this work, while the LDA was used in the latter. Replacing the PWGGA by the LDA in our calculation, the valence-band maximum shifts from −10.27 eV ͑see Table III͒ to −9.23 eV, while the conduction bands were nearly unaffected, leading to lowering of the band gap from 9.51 eV to 8.55 eV, which agrees reasonably well with value of 8.09 eV obtained by Bacalis et al 2 When DFT and quasiparticle energies for Ne at X and L points are compared ͑Table IV͒, we find a widening of the valence bands by approximately 30%. Our results for valence band energies and widths are in very good agreement with those reported earlier by Bacalis et al 2 A similar pattern of valence band widening for GW valence bands in Ne is found in Ar, and our results are again in good agreement with those of Bacalis et al 2 The energy difference of the first and second conduction bands at the ⌫ point ͑⌫ 25c Ј − ⌫ 1c ͒ in our GW calculation is ϳ2 eV smaller than in the projector augmented wave ͑PAW͒ 2 calculation for Ne, while the value of 8.21 eV for Ar agrees well with the value of 8.44 eV obtained by Bacalis et al This energy difference is sensitive to completeness of Gaussian orbital basis sets ͑see the Appendix͒ as the ⌫ 25c Ј state has significant amplitude in octahedral interstitial regions.…”