MAGNETIC RESONANCE. II 395 



Two or more electrons — ■ N electrons, let me say — may form, in 

 effect, a rigid unit having a total spin S = N/2 and a total magnetic 

 moment Nue . Such a unit will have (iV + 1) allowed orientations in the 

 big magnetic field. These will engender N resonance-peaks. In the ideal 

 case, all of these would have the same frequency 2neH/h, and would 

 therefore coalesce into a single peak at the position appropriate to g = 

 2.0023. But in these crystals Ave are likely to find cases far from ideal, 

 because of the conjoined influence of two factors. These are the presence 

 of orbital motions of the electrons, and the presence of a big electric 

 field within the crystal. 



Were the atoms in question free, we could allow for the orbital mo- 

 tions. There would be a single resonance-peak, corresponding to a value 

 of g which could be computed by a formula well known and much used 

 in optical spectroscopy. Incidentally, this formula was used in interpreting 

 the earliest molecular-beam experiments (not here described) that were 

 the first to show that g in the ideal case is not exactly equal to 2. 



Now, however, we are dealing with resonating electrons that are in a 

 strong electric field, and moreover, an electric field which is usually 

 unsymmetrical. If the asymmetry is sufficiently great, the orbital mo- 

 tions suffer a singular effect. This effect is known as "quenching." It is 

 impossible to explain and difficult even to describe without invoking 

 quantum mechanics. One may say that the orbital angular momentum 

 is no longer constant in time, and the associated magnetic moment 

 abnost but not quite disappears. 



The spin survives the quenching: but it would not be right to say that 

 the quenching restores the ideal case. The resonance is affected by what 

 have been called the ''remains" of the orbital magnetic moment. These 

 have the following consequences : 



(a) The N resonance-peaks, which coincide in the ideal case, may be 

 drawn apart. They then form a group of N separate peaks, which is 

 known as a ''fine-structure pattern." The number N tells us the number 

 of electrons coupled parallel in the atom, for these two numbers are the 

 same. Often the number of electrons coupled parallel is known from 

 independent evidence, and in such cases it is confirmed by the number of 

 lines in the fine-structure pattern. Sometimes it is not otherwise known, 

 and in such cases it is identified with the number N. 



(b) The value of g corresponding to the centre of the fine-structure 

 pattern may be altered considerably from 2.0023, falling as low as 1.35 

 or rising as high as 6.5. This is as though a part of the orbital magnetic 

 moment were added to or subtracted from the magnetic moment of the 

 spin. 



