Tải bản đầy đủ - 0 (trang)
PARAMAGNETIC, HIGH-OXIDATION-STATE, AND HIGH-COORDINATION-NUMBER COMPLEXES

PARAMAGNETIC, HIGH-OXIDATION-STATE, AND HIGH-COORDINATION-NUMBER COMPLEXES

Tải bản đầy đủ - 0trang

464



PARAMAGNETIC COMPLEXES



in group 7 and Os in group 8 are the last elements that are able to attain their

theoretical maximum oxidation states (e.g., ReF7 and OsO4 ); Ir and Pt only reach

M(VI) in MF6 , and gold shows its highest oxidation state, Au(V), in [AuF6 ]− .

It is therefore not surprising that most of the organometallic complexes having

an oxidation state in excess of 4 come from the elements Ta, W, Re, Os, and

Ir. While high oxidation states are usual for the earlier elements [e.g., Ti(IV),

Ta(V)], high oxidation states are rare for the later elements, and it is here that

we might expect to see interesting oxidizing properties. Just as the study of

low-valent organotransition metal complexes led to the development of methods

for the selective reduction of organic compounds, we can anticipate that highoxidation-state chemistry will lead to better methods of oxidation. We already

looked at OsO4 in Section 14.2. The higher oxidation states in general are more

stable for the third-row transition metals (Section 2.7). We will see that this is

also true for organometallic compounds.

As we saw in Section 2.2, the 18e rule is most likely to be obeyed by lowvalent diamagnetic complexes. In this chapter, we will find many examples of

stable species with electron counts less than 18e, but this is especially true of

polyalkyls, some of which are paramagnetic. One reason is that an alkyl ligand

occupies much space around the metal in exchange for a modest contribution

to the electron count. Second, the high ∂ + character of the metal leads to a

contraction in its covalent radius because the metal electrons are contracted by

the positive charge. Note that this only leads to a slight decrease in the M−L

bond lengths because the ligands acquire ∂ − character and so their covalent radii

increase. An increase in the ligand size and a decrease in the metal size makes

it more difficult to fit a given number of ligands around a metal in the highvalent case. The low apparent electron count in such species as MeReO3 may

be augmented somewhat by contributions from the ligand (O, Cl, NR, etc.) lone

pairs. Agostic interactions with the alkyl C−H bonds are probably not widespread

in d 0 and high-valent complexes because this interaction needs back donation

from the metal (Chapter 3). This means that electron counting in these species

is not completely unambiguous. High-valent Cp complexes are more likely to

be conventional 18e species because Cp contributes many more electrons to the

metal in proportion to the space it occupies than do alkyl groups. Polyhydrides

are almost always 18e, as we might expect for what is one of the smallest, and

one of the least electronegative, ligands present in the complexes discussed in

this section.

ž



Paramagnetic and high-oxidation-state organometallics have been rela­

tively neglected because they are harder to study.



15.1 MAGNETISM AND SPIN STATES

Diamagnetic materials are weakly repelled by a magnetic field gradient while

paramagnetic ones are attracted. From the weight change of a sample in the



465



MAGNETISM AND SPIN STATES



presence or absence of a magnetic field gradient, or by an NMR method (Evans

method; Section 10.11), one can measure the magnetic moment of a complex.

This gives the number of unpaired electrons on the central metal. Specialist

texts1 cover a number of possible factors that can complicate the interpretation,

such as spin coupling in metal clusters and orbital contributions in third-row

(5d) transition metals. Table 15.1 shows the situation in the absence of such

complications, where the measured magnetic moment in Bohr magnetons gives

the number of unpaired electrons. This number is often indicated by the spin

quantum number, S, which is simply half the number of unpaired electrons. The

multiplicity (singlet, doublet, triplet, etc.) is also used as shown in the table.

The S value of a complex depends first on the d n configuration. The d 0

and d 10 cases are necessarily diamagnetic (S = 0), having no possibility for

unpaired electrons. In contrast, d 1 and d 9 are necessarily paramagnetic with one

unpaired electron (↑, S = 12 ). The d 3 , d 5 , and d 7 odd-electron configurations are

necessarily paramagnetic but may have different accessible spin states depending

on how the spins are paired [e.g., (↑↑↑, S = 32 ) or (↑↑↓, S = 12 ) for d 3 ]. Evenelectron d 2 , d 4 , d 6 , and d 8 may be diamagnetic or paramagnetic with the spin

states depending on spin pairing [e.g., (↑↑, S = 1) or (↑↓, S = 0) for d 2 ].

Spin States

Spin states are isomeric forms with distinct energies, structures, and reactivities.

Which spin state is stablest for a given metal and oxidation state depends on the

geometry and ligand set that lead to a splitting pattern for the d orbitals. As we

fill these orbitals, we have alternative spin states whenever we have choices in

the electron filling pattern. Instead of the idealized octahedral splitting pattern of

three dπ orbitals below two dσ orbitals that we considered in Chapter 1, which

gives the high-spin/low-spin alternative spin states of Fig. 1.2, we deal instead

with more realistic splitting patterns of low-symmetry organometallic complexes.



TABLE 15.1 Terms Used in Discussing Magnetism

Spin Quantum

Number, S

0

1

2



1

3

2



2

5

2

a



Number of

Magnetic

Unpaired

Moment (bohr

Electrons Multiplicity magnetons)a

0

1

2

3

4

5



Singlet

Doublet

Triplet

Quartet

Pentet

Sextet



0

1.73

2.83

3.87

4.90

5.92



The magnetic moment can also be affected by orbital contri­

butions and magnetic coupling in metal clusters, effects that we

ignore here.



466



PARAMAGNETIC COMPLEXES



As discussed by Poli,2a a simple picture, based on the ionic model, starts from

the coordination number, represented in what follows by the symbol m, a value

determined by Eq. 15.1 for the complex [MXa Lb ]c+ . Of the nine valence orbitals

of the metal, we expect to find m orbitals in the M−L σ ∗ group (Fig. 15.1a). Of

these m orbitals, four are the single s and the three p orbitals, so (m − 4) is the

number of d orbitals in this M−L σ ∗ group. For the octahedral case, we have

(6 − 4), or two d orbitals, in agreement with the presence of just two dσ orbitals

in the familiar octahedral crystal field pattern. We can usually avoid further

consideration of these (m − 4) orbitals because electrons rarely go into M−L σ ∗

antibonding orbitals in organometallic complexes, although this is not uncommon

in Werner complexes with their generally lower

values. In the middle set of

orbitals, in a dotted box in Fig. 15.1, we find (9 − m) d orbitals, which are either

nonbonding or involved in π back bonding. For the familiar octahedral case, we

have (9 − 6) or three orbitals, corresponding with the familiar dπ set. Below these

orbitals, we have m M−L σ -bonding levels. The electron count of the complex

will be (2m + n); for the familiar d 6 octahedral case, this will be (2 × 6 + 6), or

18 electrons.

CN = m = a + b



(15.1)



Number of M−L antibonding d orbitals = (m − 4)



(15.2)



Number of M−L nonbonding d orbitals = (9 − m)



(15.3)



To find the possible spin states for any system, we first find the d n config­

uration, then we see what choices are available to distribute n electrons among

(9 − m) orbitals. To take the d 2 case, typical coordination numbers seen in real

complexes are 6 and 7. The examples of Fig. 15.2b and 15.2c show how the

LX2 -type Cp ligand contributes three to the coordination number. Small changes

in the ligand set can be sufficient to alter the energies of the d orbitals so that the

magnetism changes from one spin state to the other. If the energies of the two

states are close enough together, there can even be a spin equilibrium between

the two forms, as for S = 0 and S = 1 spin isomers of [(C5 H4 Me)NbCl2 (PEt3 )2 ].

The relative energies of the spin states in such a case is decided by the relative

magnitudes of the electron pairing energy and the HOMO–LUMO splitting, .

A large electron pairing energy (PE) favors the S = 1 state because it makes it

difficult to put two electrons in the same orbital where they repel each other more

strongly than when they are in different orbitals. A large favors the S = 0 state

because it makes it difficult to promote the electron because this now requires

more energy. In Fig. 15.2b and 15.2c, 1 is larger than 2 and 3 is larger than

4 , as expected on the basis of this argument.

The value of

depends on the geometry, ligands, and metal. The geometry

therefore often changes to some extent when the spin state changes. An example

where a large change occurs is d 8 16e NiX2 (PR3 )2 : the S = 0 complexes are

square planar and the S = 1 species are tetrahedral. The often increases as we

move from 3d to 4d and 5d metals; for example, PdX2 (PR3 )2 and PtX2 (PR3 )2



467



MAGNETISM AND SPIN STATES



m empty M L s∗

antibonding levels:

one s, three p, and

(m − 4) d orbitals.



(a)



p

s

(9 − m)



nonbonding



M d levels





d



(m = coord. no.

= single orbital

= set of orbitals)



m filled

M L s-bonding

levels

(b)

∆1



d 2, CN = 7





∆2



d 2, CN = 7



16e, S = 0





16e, S = 1

Cp∗NbCl 2(PMe3)2



CpNbCl2(dppe)



(c)

∆3



∆4



d 2, CN = 6



d 2, CN = 6





14e, S = 0



14e, S = 1





TiMe 2(dppe)2



TiCl 2(dppe)2





FIGURE 15.1 Poli model2 for discussion of open-shell organometallic compounds

(dppe = Ph2 CH2 CH2 PPh2 ). (a) The number of nonbonding levels (dotted box) depends

on the coordination number, m. The number of electrons, n, available to fill these levels

depends on the d n configuration. (b, c) For 6- and 7-coordinate species, such as the ones

shown, two spin states are possible, S = 0 and S = 1. Thick lines denote sets of orbitals.



468



PARAMAGNETIC COMPLEXES



(a)



(b)



p lone

pair



••



P



Ph



Mo



L

L



p lone

pairs



Mo



••



Cl•





L

L



Ph



FIGURE 15.2 (a) The single π-donor lone pair of PPh2 splits the d orbitals so that the

four d electrons prefer to occupy the two lower levels leading to an S = 0 state. (a) The

pair of π-donor lone pairs of Cl split the d orbitals so that the four d electrons now prefer

to occupy the three lower levels as shown, leading to an S = 1 state. The two unpaired

electrons are parallel according to Hund’s rule.



are always square planar with S = 0 as a result of Pd and Pt having higher

values than Ni.

The π bonding also strongly alters by the mechanism of Figs. 1.7 and 1.8 if

different orbitals are differently affected. In [Cp∗ Mo(PMe3 )2 (PPh2 )], for example

(Fig. 15.2), there is one π-bonding lone pair on the phosphide ligand that raises

one of the three nonbonding d levels appropriate for this 6-coordinate system.

The result is a diamagnetic S = 0 state for this d 4 case. If the ligand has two

π-bonding lone pairs, as in the chloro analog [Cp∗ Mo(PMe3 )2 Cl], however, the

two d-orbitals now affected by π bonding are both raised in energy, resulting in

an S = 1 state.

Influence of Spin State Changes on Kinetics and Thermodynamics

Often, one spin state may be very reactive, the other not. Where alternate spin

states are possible, there may be a change of spin state in a reaction, as has been

discussed by Shaik et al.2b and by Harvey et al.2c . A molecule in one spin state

could undergo a spin change to give a reactive form if the latter is close enough

in energy; the energy cost of the spin state change would merely contribute to

the reaction barrier. Such a case is illustrated in Fig. 15.3a for the reaction of

A to give B in a case where we have a ground spin state with a high reaction



469



MAGNETISM AND SPIN STATES



barrier and an excited spin state with a low barrier. If the spin state change were

very fast, the system could take the path A → 1 → 2 → 3 → B. If the spin

change could not occur rapidly enough to happen during the reaction, however,

we would have to go via the pathway A → A∗ → 2 → B∗ → B (where A∗

and B∗ are the excited spin states of reactant and product). In either case, the

reaction would still be faster than going via point 4, which would be the case if

there were no alternate spin states available (as is often the case in conventional

low-valent organometallic chemistry). This implies that organometallic species

with alternate spin states could be more kinetically labile than other cases, but

good data are still lacking.

Another situation, discussed by Poli,2a also involves a system with alternate

spin states but with a change of spin state occurring during the reaction. As

shown in Fig. 15.3b, this can alter the thermodynamics of the reaction. Assume



(a)



4



Energy

1

A



2







3



B∗



B



A



Reaction coordinate

(b)

Energy

A∗



1



B∗

2



A

B



Reaction coordinate



FIGURE 15.3 Reactivity patterns for species with alternate spin states. (a) The kinetics

of a reaction can be accelerated if a more reactive accessible excited spin state exists

with a lower net barrier for the reaction. We assume that spin change is fast. (b) The

thermodynamics of a reaction can be affected if the product has a spin state different

from that of the reagent. In this case, a reaction is unfavorable in the starting spin state

but favorable if the system crosses to the other spin state. The star refers to the excited

(less stable) spin state in each case.



470



PARAMAGNETIC COMPLEXES



the reagent spin state, A, leads to an excited spin state of the product, B∗ ; this

can even be an endothermic, unfavorable process, as shown here. If this reaction

pathway intersects the corresponding curve for the other spin state, crossover is

expected to give not B∗ but B. The path is now A → 1 → 2 → B and the

reaction now becomes thermodynamically favorable thanks to the accessibility

of the alternate spin state.

If the unsaturated product of ligand loss is stabilized by this mechanism, the

M−L bond strength will be lower than if no such stabilization occurred. This is

because the bond strength is defined as the difference in energy between Ln M−L

and ground state Ln M + L. Indeed, exceptionally low M−CO bond energies of

10–15 kcal/mol have been reported for a series of compounds where spin state

changes of this sort occur.3

Examples of spin state control of reaction rates have been given

by Harvey et al.2c For example, the slow addition of H2 to Schrock’s

[W{N(CH2 CH2 NSiMe3 )3 }H] is “spin-blocked” with a high barrier due to the

crossing between reactant triplet and product singlet surfaces. In contrast, addition

of CO to Theopold’s [TpCo(CO)] is fast because the triplet and singlet surfaces

cross at an early stage of reaction and therefore at low energy.

3d Versus 4d and 5d Metals

First-row (3d) transition metals are the most likely to be paramagnetic with a

<18e structure. Later metal analogs often adopt a different, often 18e, structure.

For example, the CpMCl2 series (M = Cr, Mo, W), shown below, starts with

15.1 without M−M bonds, where each Cr is S = 32 15e Cr. In contrast, the Mo

and W analogs 15.2 and 15.3 are both 18e, S = 0 with M−M bonds. Similarly,

the 3d metals may have a lower coordination number in their compounds. For

example, 15.1 reacts with dppe to give S = 32 , 15e 15.4 having a monodentate

dppe, but with 15.2 to give S = 12 , 17e 15.5.4,5

Cl

Cp

Cr

Cl



Cp



Cl



Cp



Cr



Cl



Mo



Mo



Cp



Cl



Cl



Cr

Cl



PPh2

15.4



PPh2



W



W

Cl

PPh2

Cl Ph2P

15.5



Cl



Cl

15.3



Cp



Cp

Cl



W

Cl

Cl



Cl

15.2



15.1



Cp



Cp



Cl



471



POLYALKYLS



ž

ž



Simple models are available to predict the magnetism of organometallics.

The reactions may involve crossing between one potential energy surface

and another, which can lead to faster reaction (Fig. 15.3)



15.2 POLYALKYLS

Group 4

We saw in Section 14.1 how MeTiCl3 is used in organic synthesis. The homolep­

tic TiMe4 (a homoleptic complex contains only one type of ligand) was reported

as early as 1959.6 The bright yellow crystalline material decomposes above ∼ 0◦ C

to methane and a black powder containing Ti, C, and H. Adducts with such lig­

ands as NMe3 , tmeda, or PMe3 are thermally more stable. Note the hard character

of the ligands that bind to TiMe4 ; this suggests that the high formal oxidation state

is real and that the electrophilic metal requires good σ -donor ligands but is inca­

pable of significant back donation. Another clue that points in the same direction

is the Grignard-like reactivity of the Ti(IV) alkyls (Section 14.1), which implies

the presence of a ∂ − carbon. Since the electronegativity difference between C

(2.5) and Ti (1.5) is considerable, the real charge on Ti must be quite positive. As

we go to the right and down in the periodic table from Ti, we find that the elec­

tronegativity increases from 1.5 to about 2.2 for the heavy platinum metals, and

so the M−C bond becomes less polar for these elements. This means the metal

will be less positive and the alkyl groups less negatively charged in homoleptic

alkyls of the later metals in a given oxidation state.

The red Ti(CH2 Ph)4 has been studied crystallographically,7 and it has been

found that the Ti−Ca −Cb angle is only 84◦ –86◦ (Fig. 15.4). The Cb carbon

of the aromatic ring interacts to some extent with the metal and the structure

is reminiscent of the η2 -allyl (Section 5.2). The soft ligand CO does react with

Ti(CH2 Ph)4 , but although initial formation of a CO adduct has been proposed, the



85°



Ti



CH2Ph



PhCH2

CH2Ph



FIGURE 15.4 The structure of Ti(CH2 Ph)4 showing the unusual distortion of the

T−Cα −Cβ bond.



472



PARAMAGNETIC COMPLEXES



final product is Ti(COCH2 Ph)2 (CH2 Ph)2 .8 In contrast to the low thermal stability

and high air and acid sensitivity of these alkyls, the bulky complexes 15.6 and

15.7 are unusually stable, thanks to steric protection of the metal. Complex 15.6

decomposes only over several days at 100◦ C, is stable to air even in solution, and

decomposes only slowly in dilute H2 SO4 ,9 and 15.7 is stable enough to melt at

234◦ C.10 The Zr and Hf alkyls are less well studied but behave rather similarly

to their Ti analogs.



Ti



Ti



15.6



15.7



Group 5

Even though vanadium has a stable (V) oxidation state, the only alkyls so far

discovered are the dark paramagnetic d 1 VR4 species, such as the green-black

benzyl complex. The 1-norbornyl is the most stable, decomposing only slowly

at 100◦ . Tantalum, the third-row element gives stable alkyls, such as TaMe5 ,

which forms a dmpe adduct.11 As we go to the right in the transition series,

the differences between the first-, second-, and third-row elements become more

marked. An example is the increasing reluctance of the first- and even secondrow elements to give alkyls having the highest possible oxidation state, a feature

that first appears in group 5 and becomes dominant in groups 6 and 7. TaMe5 is

trigonal bipyramidal, but attempts to make bulkier TaR5 complexes always lead

to α elimination to carbenes.

Group 6

A dark red Cr(IV) alkyl [Cr(CH2 SiMe3 )4 ] is known, but Cr(III) is the common

oxidation state, as in the orange Li3 [CrPh6 ]. WMe6 was the first homoleptic

alkyl of group 6 having maximum oxidation state allowed for the group. It can

decompose explosively at room temperature, but the reactions shown below have

been identified.12

O2



WMe6 −−−→ W(OMe)6

CO



WMe6 −−−→ W(CO)6 + Me2 CO



473



POLYALKYLS

heat



WMe6 −−−→ 3MeH + C2 H6 (traces)

Hal2



WMe6 −−−→ WHal6 + MeHal

N

WMe6



NO



(15.4)

O



O



N



Me4W

N



(15.5)

N



O



O



The reaction with CO may go by migratory insertion, then reductive elimination

of species containing the W(COMe)Me unit. The reaction with NO may go via

insertion to give W−O−Nž−Me, the N-centered radical center may then bind a

further NO to give the final product.

Schrock and co-workers13 have found that the hydrolysis of some of their

alkylidyne complexes lead to oxoalkyls, such as neopentyl tungsten trioxide,

which is air stable and is hydrolyzed further only by strong acid or base. The

S(TMS)2 reagent (Eq. 15.6)13 is useful for replacing oxygen with sulfur because

the formation of Si−O bonds provides a strong driving force. The mechanistic

scheme proposed for the hydrolysis is also shown (Eq. 15.7). Note in Eq. 15.8

how the alkyl groups resist hydrolysis under conditions that would lead to cleav­

age of Ti−C bonds, a sign of the greater electronegativity of W compared to Ti.

OH−



S(TMS)2



t-BuC≡W(Ot-Bu)3 −−−→ t-BuCH2 −WO3 −−−→ t-BuCH2 −WS3 (15.6)

OH−



t-BuC≡W(Ot-Bu)3 −−−→ {t-BuC≡W(OH)(Ot-Bu)3 }−

−t-BuOH



−−−−→ {t-BuCH=W(=O)(Ot-Bu)2 }−

H2 O



−−−→ {t-BuCH2 −W(=O)(OH)(Ot-Bu)2 }−

H2 O



−−−→ t-BuCH2 −WO3 + 3t-BuOH

H2 O



t-BuC≡W(CH2 t-Bu)3 −−−→ {(CH2 t-Bu)3 W(=O)}2 (µ-O)



(15.7)

(15.8)



Wilkinson et al.14 have made an analogous series of M(VI) complexes of the

type M(=Nt-Bu)2 (2,4,6-Me3 C6 H2 )2 for Cr, Mo, and W. The Cr complex is deep

red and air stable.

Group 7

Only one Mn(IV) alkyl is known, the green Mn(1-norbornyl)4 , but rhenium

has one of the most extensive series of high-oxidation-state alkyls, some of which



474



PARAMAGNETIC COMPLEXES



are illustrated in Eq. 15.9.15

MeLi



ReOCl4



AlMe3



ReOMe4



ReMe6

green



carmine

O2



Cl4Re



ReCl4



MeLi



Me4Re

red



heat



Re

Cl



Me





Me



Cl

Cl



ReMe4



Cl



Cl



Re



Cl



Cl



Re



Re

Cl



Cl



Re



Me

Me



Me



MeLi



Cl



Me



Cl

Cl

MeLi



Re



Me



Re

Cl



Me



Me



Re



Cl



Cl



Re

Me



Me



Me



(15.9)

In contrast to the reactions of O2 and NO with WMe6 (Eqs. 15.4 and 15.5),

interesting oxo-alkyls can be obtained by oxidation of ReMe6 with these oxidants.

The higher electronegativity of Re compared to W may make the Re alkyls gen­

erally more stable to air, acids, and attack by nucleophiles. ReOMe4 fails to

react with the Lewis bases that usually give complexes with the polyalkyls of the

earlier metals. The dirhenium alkyls probably have the eclipsed structure charac­

teristic of quadruply bonded metals (Section 13.1), and the trirhenium complexes

are triangular clusters with Re−Re bonds and bridging halide or alkyl groups.15

O2



NO



O2



ReMe6 −−−→ ReOMe4 −−−→ cis-ReO2 Me3 −−−→ ReO3 Me



(15.10)



The NO reactions are said to go as follows:

Me

N

LnReMe



NO



LnRe



O



O



LnRe

•N



Me



ReLn

O



N

Me

−LnRe



(15.11)

O



Me

N



N

Me



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

PARAMAGNETIC, HIGH-OXIDATION-STATE, AND HIGH-COORDINATION-NUMBER COMPLEXES

Tải bản đầy đủ ngay(0 tr)

×