The electronic properties of high-temperature superconducting cuprates are encoded in NMR data. Without microscopic theory, reliable NMR phenomenologies are in demand. Here we make use of the extensive literature data to develop a different understanding of the cuprates from the shifts of the CuO2 plane. The Cu shift analysis is based only on the symmetry of the two Cu hyperfine couplings, without assumptions about their size. We use an anisotropic Aα and isotropic B, as from atomic Cu orbitals, and find two spin components (A- and B-spins) that explain all the shift data. The components differ in size and temperature dependence according to simple rules. Upon doping the cuprates, metallic B-spin appears above a pseudogap temperature, which is shared with the A-spin. Further doping decreases the pseudogap temperature and increases the B-spin, but less so the A-spin. The apparent linear rate of increase in the density of states of the B-spin with doping is nearly threefold above x=0.20, where the pseudogap disappears. The pseudogap temperature is a measure of the coupling between A and B, which suppresses the shifts but not nuclear relaxation. Spin-singlet pairing involves A and B according to three simple condensation rates, which will be discussed. The optimal Tc demands a special match between A and B. However, the shifts do not simply predict the highest Tc of all cuprates, in contrast to nuclear relaxation anisotropy and charge sharing between planar Cu and O. Relations to other probes are discussed.
Lee et al. (Sat,) studied this question.