110 p. George, J. Beetlestone and J. S. Griffith 



(identical) pyrrole nitrogen atoms bonded on the x- and j-axes. For such 

 complexes the fg^ and Cg orbitals are split further, perhaps as shown in Fig. Ic. 

 The asymmetry in the field splits the Cg much more than the ^2? orbitals 

 because the former but not the latter point towards the nearest neighbour 

 atoms. It appears probable from electron resonance measurements that 

 d^y lies lowest (Gibson and Ingram, 1957; Griffith, 1957), d^^^y^ is almost 

 certainly at the top except possibly in some of the low-spin derivatives, but 



,,„^ ®® ®o oo 



d-electrons ®(J)® ®®® ®®® 



®® ®0 OO 



d-eiectrons (g)®0 ®®® ®®® 



SIX 



a b c 



Fig. 3. Schematic illustration of the three ways in which the Cg and the /jj, orbitals 



can be occupied in regular octahedral complexes of metal ions containing five 



and six c^-electrons (e.g. ferric and ferrous complexes). 



(a) Maximum number of parallel spins: ferric 5, ferrous 4. 



(b) Intermediate number of parallel spins: ferric 3, ferrous 2. 



(c) Minimum number of parallel spins: ferric 1, ferrous 0. 



the position of d^-. is less certain. Similar considerations to those mentioned 

 before determine whether electron pairing occurs. 



For octahedral complexes of ions with five and six ^-electrons, a third 

 electronic configuration in addition to the two characterizing the high- and 

 low-spin complexes has to be considered. The passage of just one electron 

 from the Cg orbital to the ^03 orbital results in a configuration with an inter- 

 mediate number of unpaired electrons, as shown in Fig. 3b. For a ferric 

 complex this configuration with three unpaired electrons corresponds to that 

 envisaged by Coryell, Stitt and Pauling (1937) forferrihaemoglobin hydroxide. 

 Theoretical treatment has shown however that a configuration of this kind is 

 inherently unstable in the case of regular octahedral complexes. If spin 

 pairing occurs to reduce the number of unpaired electrons from five to three, 

 then further pairing is even more favoured energetically, reducing the number 

 from three to one (Griffith, 1956a, b). Because of the lower symmetry this is 

 not necessarily true for haemoprotein derivatives. However a similar type of 

 argument suggested that it is improbable that there should exist a haemo- 

 protein which possesses derivatives of all the three kinds, high-spin, low-spin 

 and intermediate spin (Griffith, 1956c). Because of the definite existence of 

 the first two this casts doubt on the suggestion that ferrihaemoglobin hydroxide 



