The peculiar field-dependent magnetism of N a0.82CoO2 has been investigated through an analysis of its DC and AC spin susceptibilities. To account for the easily activated narrow b2g-a1g gap of the crystal field for Co in the cobalt oxide layer, the spin-state transition of Co 3+ (3d 6 ) between the low spin (LS) N a x CoO 2 has been shown to have an intriguing and rich electronic phase diagram as a function of Na ions that are sandwiched between CoO 2 layers, including a superconducting state for x∼ 1/3 after hydration,[1] a metal-to-insulator transition for x∼1/2, and a CurieWeiss metal phase with A-type antiferromagnetic ordering below T N ∼ 22K for x∼0.82-0.86. [2,3] In addition, several large hexagonal superlattices were identified for x∼0.71 ( √ 12a) and 0.82-0.86 ( √ 13a) due to Na vacancy/ion ordering, which has also be been shown to have a direct impact on the Fermi surface reconstruction. [4] In addition, N a x CoO 2 has also been introduced as a promising thermoelectric material, [5] being used either at low-temperature for its large enhancement of thermopower near 50K for x near 0.85, [6] or at high temperature for its thermoelectric figure-of merit (ZT ) higher than 1.[7] The findings of large enhancement of thermopower and magnetic field suppression have been interpreted as being important evidence that spin entropy may play a key role in the Heikes form of the Seebeck coefficient (the thermopower), [8] i.e., the spin entropy contribution to the Seebeck coefficient of a semiconductor can be suppressed by the magnetic field. Alternative explanations such as strong electron correlations, [9], combined disorder and electronic correlations, [10], and the narrow manifold of t 2g bands in this system [11] have also been proposed. [12] However, ab initio LDA calculations have yielded a reversed t 2g splitting of a 1g -e ′ g relative to the classical crystal electric field theory (CEF) prediction, in addition, the predicted e ′ g hole pockets along Γ− K at k z = 0 have never been observed by the ARPES experiments. [13][14][15][16] We have resolved this issue by revising the view of CEF that the narrow d-orbital splitting should be assigned to be b 2g -a 1g rigorously following the