Chapter 7 Answers to Problems 7.1 In Ih all d orbitals transform as the species Hg. Therefore they remain degenerate.

All degenerate d orbitals will rise in energy due to electron-electron repulsions between the metal and ligand C60. 7.2 See J. J. Zuckerman, J. Chem Educ., 1965, 42, 315 and R. Krishnamurthy and W. B. Schaap, J. Chem. Educ., 1969, 46, 799. (a) ML2 - linear (D4h) ___ ___ ___ ___ ___ (b) ML3 - trigonal planar (D3h) ___ ___ ___ ___ ___ (c) ML5 - trigonal bipyramid (D3h) ___ ___ ___ ___ ___ (d) ML5 - square pyramid (C4v) ___ ___ ___ ___ ___ (e) ML6 - trigonal antiprism (D3d) ___ ___ ___ ___ ___ eg dxz, dyz eg dx2-y2, dxy a1g dz2 b1 dx2-y2 a1 dz2 b2 dxy e dxz, dyz a1' dz2 e' dx2-y2, dxy e" dxz, dyz e' dx2-y2, dxy a1' dz2 e" dxz, dyz Fg+ dz2 Bg dxz, dyz *g dx2-y2, dxy

-108-

(f) ML7 - capped trigonal prism (C2v) The order of the orbitals is not unambiguous in this case and will depend upon the details of the geometry. The ordering suggested below is based on the following idealized model, in which the central atom lies in the plane of the four ligand atoms from which the capped position emerges.

___ ___ ___ ___ ___ (g) ML7 - pentagonal bipyramid (D5h)

a2 dxy a1 dz2 b1 dxz a1 dx2-y2 b2 dyz

___ ___ ___ ___ ___ (h) ML8 - square antiprism (D4d) ___ ___ ___ ___ ___

a1' dz2 e2'dx2-y2, dxy e1" dxz, dyz

e3 dxz, dyz e2 dx2-y2, dxy a1 dz2

(i) ML8 - cubic (Oh) Same splitting as an octahedron, but with reversed order. ___ ___ ___ t2g dxz, dyz, dxy ___ ___ eg dx2-y2, dz2 (j) ML9 - tricapped trigonal prism (D3h) ___ ___ ___ ___ ___

e" dxz, dyz e' dx2-y2, dxy a1' dz2

-109-

7.3 One way of approaching this to consider an initial tetragonal distortion, involving stretching along z, followed by compression along x to give the orthorhombically distorted structure; i.e., Oh 6 D4h 6 D2h. This would lead to the following CFT orbital splitting diagram:

As a result of the orthorhombic distortion, orbitals with an x component will rise in energy, while those with a z component will fall in energy. Relative to the D4h z-stretch case, the separation between dx2-y2 and dz2 orbitals will increase, as will that between the dxy and dxz orbitals. The degeneracy between dxz and dyz in D4h (eg) will be lifted in D2h as the dyz orbital becomes more stable. This is an allowed Jahn-Teller distortion, because D2h is centrosymmetric, as is Oh. For example, for d 1 the proposed orthorhombic distortion would give a nondegenerate ground state. This distortion would be more likely than a tetragonal z-stretch (which leads to a doubly degenerate ground state), but it is no more likely than a tetragonal z-compression (which leads to a nondegenerate state). As always, we cannot predict the nature of the distortion that will occur on the basis of the Jahn-Teller theorem.

-110-

7.4 Using the C3 axis shown below, clockwise rotations will result in the transformations listed in the following table.

C3 x y z dxy dyz dyz dx2-y2 d2z2-x2-y2 z x y dzx dzy dxy dz2-x2 d2y2-z2-x2

C3 2 y z x dyz dyx dzx dy2-z2 d2x2-y2-z2

The first three rows of d orbital listings show the equivalence of the orbitals comprising the t2 set. The last two lines show the equivalence of the orbitals comprising the e set, when the transformed orbitals are seen to be related to dx2-y2 and d2z2-x2-y2 by Eqs. (7.1a) - (7.1d). 7.5 Taking the functions dx2-y2, dz2-x2, and dz2-y2 as reference functions, C8 rotations about an axis perpendicular to the plane of the cloverleaf in each case (z, y, and x) will effect transformations to dxy, dxz, and dyz, respectively. The interconversion of dx2-y2 and dxy is shown below.

-111-

As noted in the statement of the problem, dz2-x2 and dz2-y2 can be combined into the conventional function for dz2. As a result, the operations of C8(z), C8(y), and C8(x) would interconvert all five conventional d orbitals into each other or the components of the combination that makes up dz2, making them degenerate with each other. Of course, C8 is not an operation in Oh, and the five d orbitals cannot be made completely degenerate by any operation in that point group. 7.6 (a) D3h P(E) = 2l + 1 = 5 P(C3) = {sin (2.5)(120o)}/{sin 60o} = -0.866/0.866 = -1 P(C2) = {sin (2.5)(180o)}/{sin 90o} = 1/1 = 1 P(F) = + sin {(2.5)(180o)} = 1 P(S3) = + {sin (2.5)(1200 + 180o)}/{sin 150o} = 0.5/0.5 = 1 D3h 'd E 5 2C3 -1 3C2 1 Fh 1 2S3 1 3Fv 1

'd = A1' + E' + E" (b) D2h P(E) = 2l + 1 = 5 P(C2) = {sin (2.5)(180o)}/{sin 90o} = 1/1 = 1 P(i) = + {2l + 1) = 5 P(F) = + sin {(2.5)(180o)} = 1 D2h 'd E 5 C2(z) C2(y) 1 1 C2(x) 1 i 5 F(xy) 1 F(xz) F(yz) 1 1

'd = 2Ag + B1g + B2g + B3g

-112-

(c) C4v P(E) = 2l + 1 = 5 P(C4) = {sin (2.5)(90o)}/{sin 45o} = -0.707/0.707 = -1 P(C2) = {sin (2.5)(180o)}/{sin 90o} = 1/1 = 1 P(F) = + sin {(2.5)(180o)} = 1 C4 v 'd E 5 2C4 -1 C2 1 2Fv 1 2Fd 1

'd = A1 + B1 + B2 + E (d) I P(E) = 2l + 1 = 5 P(C5) = {sin (2.5)(72o)}/{sin 36o} = 0/0.588 =0 P(C52 ) = {sin (2.5)(144o)}/{sin 72o} = 0/0.951 =0 P(C3) = {sin (2.5)(120o)}/{sin 60o} = -0.866/0.866 = -1 P(C2) = {sin (2.5)(180o)}/{sin 90o} = 1/1 = 1 Ih 'd E 5 12C5 0 12C52 0 20C3 -1 15C2 1

'd = H Y Hg in Ih 7.7 With l = 3, f orbitals are inherently ungerade, requiring the use of the negative sign with Eqs. (7.4) - (7.6). (a) Td P(E) = 2l + 1 = 7 P(C3) = {sin (3.5)(120o)}/{sin 60o} = 0.866/0.866 = 1 P(C2) = {sin (3.5)(180o)}/{sin 90o} = -1/1 = -1 P(S4) = {sin (3.5)(270o)}/{sin 135o} = (-0.707)/0.707 = 1 P(F) = sin {(3.5)(180o)} = (-1) = 1 Td 'f E 7 8C3 1 3C2 -1 6S4 1 6Fd 1

'f = A1 + T1 + T2

-113-

(b) Oh. By correlation of the Td results obtained in part (a) to the corresponding ungerade species of Oh we obtain Td 6 Oh A1 6 A2u T1 6 T2u T2 6 T1u Alternately, the reducible representation can be generated and decomposed in the usual way. P(E) = 2l + 1 = 7 P(C3) = {sin (3.5)(120o)}/{sin 60o} = 0.866/0.866 = 1 P(C2) = {sin (3.5)(180o)}/{sin 90o} = -1/1 = -1 P(C4) = {sin (3.5)(90o)}/{sin 45o} = -0.707/0.707 = -1 P(i) = (2l + 1) = -7 P(S4) = {sin (3.5)(270o)}/{sin 135o} = (-0.707)/0.707 = 1 P(S6) = {sin (3.5)(240o)}/{sin 120o} = (0.866)/0.866 = -1 P(F) = sin {(3.5)(180o)} = (-1) = 1 Oh 'f E 7 8C3 1 6C2 -1 6C4 -1 3C2 -1 i -7 6S4 1 8S6 -1 3Fh 1 6Fd 1

'f = A2u + T1u + T2u

-114-

(c) D4h. By correlation of the Oh results obtained in part (b) to the corresponding ungerade species D4h of we obtain Oh 6 D4h A2u 6 B1u T2u 6 A2u + Eu T1u 6 B2u + Eu Alternately, the reducible representation can be generated and decomposed in the usual way. P(E) = 2l + 1 = 7 P(C4) = {sin (3.5)(90o)}/{sin 45o} = -0.707/0.707 = -1 P(C2) = {sin (3.5)(180o)}/{sin 90o} = -1/1 = -1 P(i) = (2l + 1) = -7 P(S4) = {sin (3.5)(270o)}/{sin 135o} = (-0.707)/0.707 = 1 P(F) = sin {(3.5)(180o)} = (-1) = 1 D4h 'f E 7 2C4 -1 C2 -1 2C2' -1 2C2" -1 i -7 2S4 1 Fh 1 2Fv 1 2Fd 1

'f = A2u + B1u + B2u + 2Eu

-115-

(d) D3d. By correlation of the Oh results obtained in part (b) to the corresponding ungerade species D4h of we obtain Oh 6 D3d A2u 6 A2u T2u 6 A2u + Eu T1u 6 A1u + Eu Alternately, the reducible representation can be generated and decomposed in the usual way. P(E) = 2l + 1 = 7 P(C3) = {sin (3.5)(120o)}/{sin 60o} = 0.866/0.866 = 1 P(C2) = {sin (3.5)(180o)}/{sin 90o} = -1/1 = -1 P(i) = (2l + 1) = -7 P(S6) = {sin (3.5)(240o)}/{sin 120o} = (0.866)/0.866 = -1 P(F) = sin {(3.5)(180o)} = (-1) = 1 D3d 'f E 7 2C3 1 3C2 -1 i -7 2S6 -1 3Fd 1

'f = A1u + 2A2u + 2Eu 7.8 (a) Only degenerate states are subject to Jahn-Teller distortions. Assuming the ordering of d orbitals shown on the extreme right of Fig. 7.9, only the configurations d 1 and d 3 result in degenerate ground states and would be distorted. If the eg orbitals are above the a1g orbital, the ground state configurations d 3 and d 5 would be degenerate and would be distorted. Of course, any excited states in which one or three electrons occupy the eg orbitals would also be degenerate and result in Jahn-Teller distortions. (b) D4h is a centrosymmetric point group, so any Jahn-Teller distortion would produce a centrosymmetric structure. Disregarding ligand structure, the only centrosymmetric subgroups are D2h and Ci. A D2h structure would result from shortening or lengthening a trans-related pair of bonds relative to the other pair of trans-related bonds. If the LML' bond angle in such a structure were distorted from 90o, the symmetry would descend further to Ci. As always with the Jahn-Teller theorem, the precise distortion cannot be predicted. (c) Almost all ML4 D4h complexes are d 7-8, in which the eg orbitals are filled, regardless of their relative order with the a1g orbital. Consequently, square planar complexes have nondegenerate ground states and are immune to Jahn-Teller distortions.

-116-

7.9 (a) As shown for PF5 (cf. problem 4.13), for the ligand F-SALCs, ' = 2A1' + E' + A2". The symmetries of the d orbitals, taken from the direct product transformations in D3h, are dz2 = A1', (dx2-y2, dxy) = E', (dxz, dyz) = E". Bonding and antibonding combinations can be formed with A1' and E' symmetries, but the E" symmetry (dxz, dyz) pair must be nonbonding in this model. The following qualitative MO can be constructed.

(b) The upper three levels (boxed region, above) have identical splitting to that in the CFT tbp model (cf. problem 7.2 c). Electron pairs contributed by the ligands are sufficient to fill all lower levels. Therefore, any electrons contributed by the metal ion result in filling in the upper three levels, in the same manner as predicted in the CFT model.

-117-

(c) The a1' combination with the dz2 orbital would use opposite signs for the axial and equatorial ligand F AOs. The SALC would be M1 = 1//5{N1 + N2 + N3 N4 N5}. The resulting LCAO would be:

The two e' SALCs, which combine with dx2-y2 and dxy, involve only equatorial functions. They are M2 = 1//6{2N1 N2 N3} and M3 = 1//2{N2 N3}, respectively. The resulting LCAOs, projected in the xy plane, would be:

(d) The a1' and a2" SALCs, which form nonbonding MOs in the d-only scheme, could form bonding and antibonding combinations with s and pz AOs on the metal. The px and py AOs could form bonding and antibonding combinations with the e' SALCs, which in the d-only scheme form bonding and antibonding combinations with dx2-y2 and dxy AOs. This competition would result in six mixed MOs of varying bonding, nonbonding, and antibonding character, depending upon the relative contributions of d and p AOs.

-118-

7.10 (a) The representation of F-SALCs is based on the following model:

C4 v 'F

E 5

2C4 1

C2 1

2Fv 3

2Fd 1

'F = 2A1 + B1 + E The AO symmetries of the central M atom are shown below. (n 1)d: ns: np: dz2 = A1, dx2-y2 = B1, dxy = B2, (dxz, dyz) = E s = A1 pz = A1, (px, py) = E

The dxy orbital (b2) has no symmetry match with a SALC and must remain nonbonding. (b) For simplicity, the following sketches show 2 = 90o. The three kinds of potentially bonding interactions between (n 1)d orbitals and F-SALCs are the following:

The e (xz, yz) combinations are nonbonding at 2 = 90o and become only slightly bonding with 2 > 90o. A better bonding combination with the E SALCs can be formed with px and py AOs (see below).

-119-

The ns and np AOs can form the following bonding combinations:

The ns and np combinations for a1 are probably mixed. With a total of three a1 AOs (dz2, s, pz) and two a1 SALCs, there can be only five MOs. We will assume bonding and antibonding combinations for dz2 (a1) and s (a1), and treat pz (a1) as essentially nonbonding. The two e combinations with px and py have better overlap with the E SALCs than dxz and dyz. Therefore we will form bonding and antibonding combinations with these and assume that the dxz and dyz orbitals are essentially nonbonding.

-120-

(c) The following MO scheme is based on the assumptions noted above.

(d) In the MO scheme shown for part (c), electrons from ligand would fill all levels below those in the box. The MOs in the box have the same symmetries and are in the same relative energy order as the d orbitals in a square pyramidal crystal field [cf. answer to 7.2 (d)]. Thus, the filling of electrons in MOs above the four lowest bonding levels is equivalent to the presumed filling of electrons in d orbitals of the CFT model.

-121-

7.11 (a) The SALCs are identical to the oxygen F- and B-SALCs derived for CO2 in Chapter 4 (pp. 118-126). Thus, ' F = E g+ + E u+ 'B = A u + A g (b) Assuming p orbitals on the ligands, the SALCs would have the following forms: Eg+ Eu +

Au

both xz and yz planes

Ag

both xz and yz planes

The normalized wave function expressions for these SALCS are as follows: Eg+ = 1//2{Nz(1) + Nz(2)} Au(x) = 1//2{Nx(1) + Nx(2)} Ag(x) = 1//2{Nx(1) Nx(2)} (c) Eu+ = 1//2{Nz(1) Nz(2)} Au(y) = 1//2{Ny(1) + Ny(2)} Ag(y) = 1//2{Ny(1) Ny(2)}

s = Eg+ Y bonding and antibonding with Eg+ SALC pz = Eu+ Y bonding and antibonding with Eu+ SALC (px, py) = Au Y bonding and antibonding with Au SALCs dz2 = Eg+ Y bonding and antibonding with Eg+ SALC (dxz, dyz) = Ag Y bonding and antibonding with Ag SALCs (dx2-y2, dxy) = )g Y nonbonding.

-122-

(d) Both ns and (n - 1)dz2 AOs can form bonding and antibonding combinations with the Eg+ SALC, so s-d mixing is possible. Being higher in energy, the ns AO probably contributes more to antibonding, while the lower-energy (n - 1)dz2 AO contributes more to bonding. For simplicity, the following MO scheme assumes the ns AO on the metal forms an essentially nonbonding Fn level, leaving the (n - 1)dz2 AO to form bonding and antibonding MOs without showing possible s-d mixing.

The MOs in the box have the same symmetry and splitting as the presumed splitting of d orbitals in the CFT model (cf. problem 7.2 a). Electron pairs contributed by ligands are sufficient to fill all lower levels. Therefore, any electrons contributed by the metal ion result in filling in the upper three levels in the same manner as predicted in the CFT model.

-123-

7.12 (a) For F-SALCs, use the following four-vector basis set to obtain the reducible representation.

D4h 'F

E 4

2C4 0

C2 0

2C2' 2

2C2" 0

i 0

2S4 0

Fh 4

2Fv 2

2Fd 0

'F = A1g + B1g + Eu For B-SALCs, use the following eight-vector basis set to obtain the reducible representation.

D4h 'B

E 8

2C4 0

C2 0

2C2' -4

2C2" 0

i 0

2S4 0

Fh 0

2Fv 0

2Fd 0

'B = A2g + B2g + Eg + A2u + B2u + Eu

-124-

(b) The matches between SALCs and central-atom AOs is shown below. Among the BSALCs, A2g and B2u have no matching AOs and must be nonbonding (n.b.). F-SALCs AOs B-SALCs AOs (c) A2g n.b. F-SALCs: G(A1g) = (F1 + F2 + F3 + F4) G(B1g) = (F1 F2 + F3 F4) G"(Eu) = 1//2(F1 F3) G$(Eu) = 1//2(F2 F4) B-SALCs: A(A2g) = (B5 + B6 B7 B8) A(B2g) = (B5 B6 + B7 + B8) A"(Eg) = 1//2(B2 B4) A$(Eg) = 1//2(B1 B3) A(A2u) = (B1 + B2 + B3 + B4) A(B2u) = (B1 B2 + B3 B4) A"(Eu) = 1//2(B6 + B8) A$(Eu) = 1//2(B5 + B7) (nonbonding) (nonbonding) B2g dxy A1g s, dz2 Eg (dxz, dyz) B1g dx2-y2 Eu (px, py) A2u pz B2u n.b. Eu (px, py)

-125-

(d)

G(A1g) + s

G(A1g) + dz2

G(B1g) + dx2-y2

G"(Eu) + py; G $(Eu) + px

-126-

A(A2g)

A(B2g) + dxy

A"(Eg) + dxz; A$(Eg) + dyz

A(A2u) + pz

A(B2u)

-127-

A"(Eu) + px; A$(Eu) + py

7.13 We can save labor by doing the work in O and then correlating to the matching gerade species of Oh, simply adding g subscripts to the Mulliken symbols in O. For G, L = 4. P(E) = 2l + 1 = 9 P(C3) = {sin (4.5)(120o)}/{sin 60o} = 0/0.866 = 0 P(C2) = {sin (4.5)(180o)}/{sin 90o} = 1/1 = 1 P(C4) = {sin (4.5)(90o)}/{sin 45o} = 0.707/0.707 = 1 O 'G E 9 8C3 0 3C2 1 6C4 1 6C2' 1

'G = A1g + Eg + T1g + T2g For H, L = 5. P(E) = 2l + 1 = 11 P(C3) = {sin (5.5)(120o)}/{sin 60o} = -0.866/0.866 = -1 P(C2) = {sin (5.5)(180o)}/{sin 90o} = -1/1 = -1 P(C4) = {sin (5.5)(90o)}/{sin 45o} = 0.707/0.707 = 1 O 'H E 11 8C3 -1 3C2 -1 6C4 1 6C2' -1

'H = Eg + 2T1g + T2g For I, L = 6. P(E) = 2l + 1 = 13 P(C3) = {sin (6.5)(120o)}/{sin 60o} = 0.866/0.866 = 1 P(C2) = {sin (6.5)(180o)}/{sin 90o} = 1/1 = 1 P(C4) = {sin (6.5)(90o)}/{sin 45o} = -0.707/0.707 = -1

-128-

O 'I

E 13

8C3 1

3C2 1

6C4 -1

6C2' 1

'I = A1g + A2g + Eg + T1g + 2T2g The splitting of g and i orbitals would be the same as the states, because both are gerade. The h orbitals are ungerade, and contrary to the H state will split as eu + 2t1u + t2u. 7.14 (a) Use the Oh 6 D4h correlation table. Free-Ion Terms S P D F G H I

Terms in Oh A1g T1 g Eg + T2g A2g + T1g + T2g A1g + Eg + T1g + T2g Eg + 2T1g + T2g A1g + A1g + Eg + T1g + 2T2g

Terms in D4h A1g A2g + Eg A1g + B1g + B2g + Eg A2g + B1g + B2g + 2Eg 2A1g + A2g + B1g + B2g + 2Eg A1g + 2A2g + B1g + B2g + 3Eg 2A1g + A2g + 2B1g + 2B2g + 3Eg

(b) The most direct correlation for terms S through F is to use the R3 correlation table showing R3 6 D3. For the I terms, use Oh 6 D3d and then D3d 6 D3. Free-Ion Terms S P D F G H I

Terms in Oh A1g T1 g Eg + T2g A2g + T1g + T2g A1g + Eg + T1g + T2g Eg + 2T1g + T2g A1g + A1g + Eg + T1g + 2T2g -129-

Terms in D3 A1 A2 + E A1 + 2E A1 + 2A2 + 2E 2A1 + A2 + 3E A1 + 2A2 + 4E 3A1 + 2A2 + 4E

(c)

There is no correlation table for Oh 6 D2d. Use results in part (a) with the table for D4h 6 D2d (C2' 6 C2'). Note that for D2d, the ligand-field terms are the same as for D4h without the subscript g. Alternately, one could use the correlation Td 6 D2d. Free-Ion Terms S P D F G H I

Terms in D4h A1g A2g + Eg A1g + B1g + B2g + Eg A2g + B1g + B2g + 2Eg 2A1g + A2g + B1g + B2g + 2Eg A1g + 2A2g + B1g + B2g + 3Eg 2A1g + A2g + 2B1g + 2B2g + 3Eg

Terms in D2d A1 A2 + E A1 + B1 + B2 + E A2 + B1 + B2 + 2E 2A1 + A2 + B1 + B2 + 2E A1 + 2A2 + B1 + B2 + 3E 2A1 + A2 + 2B1 + 2B2 + 3E

(d) One the basis of vector transformations, determine the correlation between D4h and D4h, as follows: D4h A1g A2g B1g + B2g Eg D4h Eg+ Eg )g Ag Basis z2 Rz (x2 y2, xy) (Rx, Ry) (xz, yz)

-130-

Free-Ion Terms S P D F G H I

Terms in D4h A1g A2g + Eg A1g + B1g + B2g + Eg A2g + B1g + B2g + 2Eg 2A1g + A2g + B1g + B2g + 2Eg A1g + 2A2g + B1g + B2g + 3Eg 2A1g + A2g + 2B1g + 2B2g + 3Eg

Terms in D4h E g+ E g + A g E g+ + A g + ) g Eg + 2 A g + )g 2Eg+ + Eg + 2Ag + )g Eg+ + 2Eg + 3Ag + )g 2Eg+ + Eg + 3Ag + 2)g

7.15 (a) There are four microstates: eg t2g (b)

Use Table 7.2, or apply Eqs. (7.2) - (7.6) for L = 4. Either way, the states emerging from 2G are 2A1g, 2Eg, 2T1g, 2T2g. For 2G, Dt = (2)(9) = 18. For 2A1g + 2Eg + 2T1g + 2T2g, Dt = (2)(1) + (2)(2) + (2)(3) + (2)(3) = 18.

(c)

7.16 The free-ion terms of d n and d 10-n are the same because they arise from the same number of microstates, as shown by Eq. (7.14):

The free-ion terms are unique to a given number of microstates. Moreover, each free-ion term splits into a characteristic set of ligand-field terms in any point group. Thus, the ligand-field terms for d n and d 10-n must be the same.

-131-

7.17 The strong-field terms emerging from the d 2 Td configurations in the hypothetical extreme field are the same as for the comparable d 2 Oh configurations, except that the subscript g is omitted for each. The extreme-field configurations, however, will be in the reverse energy order of the d 2 Oh configurations; i.e., e2 < e1t21 < t22. Identifying the triplet terms versus the singlet terms follows the same logic as shown in the d 2 Oh case; i.e., each configuration gives rise to the same multiplicities for comparable Td orbital terms. Making the correlations in the same way gives a diagram identical to that shown for d 8 Oh (Fig. 7.26), except for the gerade designations on the states. 7.18 Weak-Field Terms. In Oh, the free-ion 2D term of a d 9 configuration becomes a 2Eg ground state term and a 2T2g excited state terms. From the correlation tables, it can be seen that these terms split on descent to D4h square planar geometry as follows: 2 2 Eg 6 2A1g + 2B1g T2g 6 2B2g + Eg Extreme Field Configurations and Strong-Field Terms. The configurations for the d 9 D4h square planar geometry in an extremely strong field are shown below in order of increasing energy. The terms in a strong field are obtained by taking the orbital assignment of the unpaired electron as a basis for a representation in D4h. All paired electrons can be ignored, because the representation for any pair must be the totally symmetric representation; e.g., B1g B1g = A1g. With only one unpaired electron in all configurations, all terms must be doublets. b1g b2g a1g eg 2 2 2 2 B1g B2g A1g Eg The resulting qualitative correlation diagram is shown below.

-132-

7.19 (a) FeF63 is a d 5 high-spin complex, because F is a weak-field ligand. Therefore, there are no spin-allowed transitions from the 6A1g ground state. Fe(CN)63 is a d 5 low-spin complex, because CN is a strong-field ligand. The ground state is 2T2g term. From the TanabeSugano diagram for d 5 low-spin cases (right side), it is apparent that there are at least three possible spin allowed transitions: 2T2g 6 [2A2g, 2T1g], 2T2g 6 2Eg, 2T2g 6 2A2g. As the TanabeSugano diagram indicates, the 2A2g and 2T1g states for the first transition are virtually degenerate and would give rise to a single broad band. These transitions account for the color of Fe(CN)63. (b) Tetrahedral complexes are not centrosymmetric, so the LaPorte Rule does not apply. Nonetheless, d orbitals have an inherent gerade character. Therefore, although molar absorptivities are slightly higher for tetrahedral complexes compared to octahedral complexes, they are still relatively low by comparison to charge-transfer transitions. (c) {The charge on the complex should be 3+, not 2+ as shown in the first printing.} [Cr(H2O)3]3+ is an octahedral d 3 case. From either the Tanabe-Sugano diagram or Orgel diagram, three bands are expected from the transitions 4A2g 6 4T2g, 4A2g 6 4T1g(F), 4A2g 6 4 T1g(P). The last of these involves a change in configuration from t2g3 to t2g1eg2; i.e., a twoelectron transition. This is a less probable event and therefore would have a lower absorptivity. Moreover, it is the highest energy transition and may be obscured by the tail of a stronger charge-transfer band in the ultraviolet. (d) A d 6 high-spin complex such as [Fe(H2O)6]2+ is expected to show a single band from the transition 5T2g 6 5Eg. However, both the ground state and excited state are degenerate and subject to Jahn-Teller distortions. The distortion is greater for the doubly degenerate excited state. Therefore, the observed band splitting is primarily due to lifting of the degeneracy of the upper 5Eg state. (e) With d 0 configurations, both complex ions have no possible d-d transitions. The observed intense colors arise from L6M charge transfer transitions, which are LaPorte allowed. Such transitions can have very high molar absorptivities. 7.20 (a) Use direct products for the configurations to determine the orbital term designations. Where two electrons are paired in the same orbital the product is totally symmetric. The ground state eg4a1g2b2g2 consists of all paired electrons, so the term is 1A1g. For the first excited singlet state eg4a1g2b2g1b1g1, take the direct product B2g B1g:

-133-

D4h B2g B1g ' = A2g

E 1 1 1

2C4 -1 -1 1

C2 1 1 1

2C2' -1 1 -1

2C2" 1 -1 -1

i 1 1 1

2S4 -1 -1 1

Fh 1 1 1

2Fv -1 1 -1

2Fv 1 -1 -1

Therefore, the first singlet excited state is 1A2g. For the second excited singlet state eg4a1g1b2g2b1g1 the direct product is A1g B1g = B1g , and the resulting term is 1B1g. For the third excited singlet state eg3a1g2b2g2b1g1, the direct product is Eg B1g: D4h Eg B1g ' = Eg E 2 1 2 2C4 0 -1 0 C2 -2 1 -2 2C2' 0 1 0 2C2" 0 -1 0 i 2 1 2 2S4 0 -1 0 Fh -2 1 -2 2Fv 0 1 0 2Fv 0 -1 0

Therefore the term is 1Eg. (b) The three electronic transitions are 1 A1g 6 1A2g 1 A1g 6 1B1g 1 A1g 6 1Eg (c) For an electronic transition to be vibronically allowed, the transition moment IReRv:Re' Rv' dJ must be totally symmetric to be non-zero. In this specific case Re is also totally symmetric, and the ground vibrational state, Rv is totally symmetric for all molecules except free radicals. Therefore the symmetry of the integral depends upon the symmetries of the direct products : Re' Rv'. The symmetries of the dipole moment components, obtained from the unit vector transformation properties listed in the D4h character table are :z = A2u (:x, :y) = Eu

The symmetries of the excited electronic states, as shown in part (a), are A2g, B1g, and Eg. The symmetries of the components of : are ungerade, and the symmetries of the electronic excited states are gerade, so the products : Re' must all be ungerade. In order for the integral to be totally symmetric and nonvanishing, the Rv' must belong to an identical ungerade species as : Re' . The ungerade normal modes of vibration, which might couple with the electronic transitions, are <3 (A2u) <5 (B2u) <6 (Eu) <7(Eu)

-134-

For the purpose of vibronic coupling, the gerade normal modes can be ignored. By matching symmetries of : Re' with one or more of these Rv' we obtain the following results, which show the normal modes that vibronically allow each electronic transition: Transition
1

: Re' A2u A2g = A1u Eu A2g = Eu

Matching normal modes none <6, <7 (Eu) <5 (B2u) <6, <7 (Eu) <6, <7 (Eu) <3 (A2u), <5 (B2u)

A1g 6 1A2g

A1g 6 1B1g

A2u B1g = B2u Eu B1g = Eu

A1g 6 1Eg

A2u Eg = Eu Eu Eg = A1u + A2u + B1u + B2u

In the first case for 1A1g 6 1A2g, the integral with the dipole moment component :z (A2u) vanishes, because there is no matching normal mode with the resultant A1u symmetry. Nonetheless, the transition 1A1g 6 1A2g is vibronically allowed by coupling with the normal modes <6 and <7 (Eu), because the direct product : Re' with the degenerate pair of dipole moment components (:x, :y) is Eu, making the full integral totally symmetric. In the last case for 1A1g 6 1Eg, the direct product Eu Eg is a reducible representation, only two whose component irreducible representations (A2u and B2u) match symmetries with normal modes. Those matches are sufficient to vibronically allow the 1A1g 6 1Eg electronic transition.

-135-