It is argued that the nuclear quadrupole -electric field gradi-
, i, j Γ
x, y, z. [1] ent ( EFG ) interaction is, in principle, dependent on the presence of a magnetic field B . A rough estimate of the size of this effect yields 10 04 in fields up to 10 T. However, if the Since this potential is determined by the surroundings of the site symmetry of the nucleus in question includes time-reversal nucleus on a microscopic scale, the EFG provides informasymmetry, the linear dependence of the EFG on B vanishes.
tion about the structure of the surroundings, e.g., about the
In diamagnetic compounds, time-reversal symmetry is violated only by the presence of nuclear spins. In such compounds, the direction and length of a chemical bond. Therefore, the deterdominant dependence of the EFG on B should be quadratic mination of EFGs is one of the most often undertaken tasks and should be described by a fourth-rank tensor. In ferro-and in NQR and in solid-state NMR.
antiferromagnetic compounds time-reversal symmetry is
It is generally assumed that the EFG is independent of strongly violated and a linear dependence of the EFG on B , the presence or absence of a magnetic field B. Intuition, described by a third-rank tensor, is expected. A search for a however, tells us that a dependence of the EFG on a magnetic magnetic field dependence of the EFG was carried out by meafield should exist: The presence of a magnetic field modifies suring the quadrupole coupling constants ( QCCs ) of the 27 Al the electronic wavefunction and thus the charge distribution and 14 N nuclei in corundum and sodium nitroprusside ( SNP ) of a molecule. If it did not, we would not observe chemical by pure NQR, and by NMR in fields of 6.3 and 11 T. These shifts. Because the electronic charge distribution in a molediamagnetic compounds were selected because previous meacule or crystal, together with the nuclear point charges, detersurements, done in different fields, yielded differing results for the QCCs. A new technique for measuring QCCs by NMR is mines the EFG sensed by a nucleus, a magnetic-field-inintroduced that circumvents the necessity of precisely orienting duced modification of the electronic wavefunction must also the sample crystals. For the QCCs of both the 27 Al and 14 N lead to a modified EFG nuclei in corundum and SNP, respectively, a precision of distinctly better than 10 04 is reached. The results obtained in 0,
[2]
6.3, and 11 T fields fully agree with each other which means that, in fields up to 11 T, any possible field dependence of the QCCs is smaller than 10 04 . These results confirm that in dia-Henceforth, a nonzero modification DV ij (B) of the EFG will magnetic compounds a linear dependence of QCCs on B is be called the effect. The search for the effect is the subject largely suppressed. α§ 1997 Academic Press of this paper. We may also argue that there is no general law prohibiting a magnetic field dependence of the EFG and that, therefore, the effect should exist.