cơ_merged
cơ_merged
(NR) are biradicals that form a double bond with the metal, and will therefore be
considered as X2 ligands, whereas the singlet carbenes give a pair of electrons to the
Chapter 1 metal, and they are thus considered as L ligands (see Chap. 9).
The number of electrons usually given to the metal by the most common ligands are
the following:
MONOMETALLIC 1-electron, radical-type X ligands
TRANSITION-METAL COMPLEXES H, F, Cl, Br, I
OH, OR, SR, NH2, NR2, PR2, AsR2
CH3, CR3 (alkyl), Ph, Ar (aryl), CH=CR2 (vinyl), C CR (alkynyl), COR (acyl),
The transition metals are, by definition, the elements with an incomplete d shell, and SiR3 (silyl), SnR3, (stannyl)
an empty last p shell (the valence one). These elements will need to complete more Metal-carbonyl radicals: M(CO)n (M being a metal with an odd number of
or less these subshells with electrons given or shared by the ligands in order to give valence electrons), MCp(CO)n (M being a metal with an even number of valence
rise to stable compounds. These electrons provided by the ligands allow the metal to electrons)
reach more or less exactly the electronic structure of the rare gas following them on
the same line of the periodic table. These notions will be refined later, but we will NO (bent bond with the metal) (see Chap. 7.5.4).
first examine the electron count given by the ligands to the transition metal. O
• •
We will use the convention that considers all ligands as neutral in transition-metal N
complexes. It is simple and close to reality for most transition metals. The ionic
M
convention is less used, but appropriate for the scandium, lanthanide and actinide
complexes. The characteristics of the complexes (see section 2) are the same
whatever convention is used. In this chapter, we will examine the mononuclear 2-electron X2 biradical-type ligands
complexes whereas the bi- and polymetallic complexes are described in Chap. 2. =CH2, =CR2 (carbenes or alkylidenes), =C=CR2 (vinylidenes) (see Chap. 9.1)
=NR (amido), =PR (phosphinidenes), =O (oxo or oxene) (see Chap. 9.3)
1. THE LIGANDS –CH2(CH2)nCH2– (cycloalkyls), –O–O– (peroxide), CO2, CS2
CH2 O O S
Given the convention according to which all the ligands are considered as neutral,
M (CH2)n M M M
there are two classes of ligands. First, there are ligands bringing one or several
O C C
electron pairs to the metals; they are even ligands and are designated as L or Ln, CH2
O S
n being the number of electron pairs given to the metal. Then, there are ligands
bringing one electron or an odd number of electrons to the metal, i.e. the radical-
type ligands; they are designated as X (one electron) or XLn (odd number of
2-electron L ligands
electrons). The L or Ln ligands do not accept valence electrons from the metal to They are:
make the metal-ligand bond, because the bond involved is of the donor-acceptor The donors of a non-bonding electron pair of an heteroatom
type. On the other hand, the X or XLn ligands require one valence electron from the
metal to form the metal-ligand bond. Thus, the M-X bond resembles the covalent H2O, H2S, ROH, RSH, R2O, THF, R2S
bond in organic chemistry as each partner brings one electron to form the bond NH3, NR3, PR3 (phosphine), P(OR)3 (phosphite), AsR3 (arsine), SbR3 (stilbine)
electron pair. The triplet carbenes or alkylidenes (CR2), oxo (oxene, O) and nitrido
1 – MONOMETALLIC TRANSITION-METAL COMPLEXES 25 26 PART I – STRUCTURES OF THE TRANSITION-METAL COMPLEXES
N2, O2 (can also be L ligands via each of the two heteroatoms of the molecule), 3-electron X3 ligands (trivalent)
CO (carbonyl), CS (thiocarbonyl), (CH3)2CO, CH3CN, RCN (nitrile), RNC CR (carbynes or alkylidynes), N (nitrido), P (phosphido) (see Chap. 9.2)
(isonitrile), CH2Cl2 (non-bonding electron pair of a chlorine atom)
+
3-electron radical LX ligands
CH2=CH–CH2 (allyl), CHCHCH (cyclopropenyl) (see Chap. 9.4)
Fe
O Me H
OC C CH2 C
CO HC HC
Me C
CH2
H
M M
Triplet carbenes such as Bertrand’s non-N-heterocyclic carbenes (non-NHCs) and
Arduengo’s N-heterocyclic carbenes (NHCs) CO2R (carboxylato), O2CNR2 (carbamato), S2CNR2 (dithiocarbamato),
R R
CH3COCHCOCH3 (acetylacetonato)
Me3Si R2P N N H
: : : : R
S C
R2N P R’2N N N H3C CH3
C N C C
NR2 R R S R
O O
Bertrand’s carbenes (non-NHCs) Arduengo’s carbenes (NHCs)
NO (linear bond with the metal), can also be a bent 1-electron X ligand
The donors of a electron pair (see Chap. 7.5.4)
Alkenes, alkynes (can also be 4-electron L2 ligands), double C=O bond or oxygen
M N O
non-bonding electron pair of an aldehyde or a ketone (also C=S), O2
X Y The ligands halogeno, alkoxy (or aryloxy), alkylthio (or arylthio), amino,
X, Y = C, O, N, etc. phosphido and NO that are all 1-electron ligands can also be 3-electron
M LX ligands when the metal is electron-deficient; i.e. has less than 18 valence
electrons (vide supra)
The donors of a bond electron pair
H-H (dihydrogen ligand; see Chap. 8.6), H-SiR3 (Si-H silane bond) M X
X Y Analogously, the acyl and methoxy ligands are 3-electron ligands in electron-
X = H, C, Si, etc. deficient early transition-metal complexes such as [Ti(OCH3)4]
M Y= H
R OCH3
H-CR3 (intramolecular: C-H bond belonging to another ligand of the same metal C O
by another bond); the C-H-M bond is called “agostic” (see Chaps 6.3, 6.4 and 8.2) Ti
C H M H3CO OCH3
M M M M
C5H5 (cyclopentadienyl, noted Cp) C6H7 (cyclohexadienyl)
4-electron LX2 ligands
The oxo and imido ligands, that are X2 ligands, can become LX2 ligands when the Heteroatoms can be incorporated in rings, which gives rise to a large variety of
ligands. Mathey’s phospholyl C4H4P, comparable to the cyclopentadienyl, and
metal is electron-deficient, i.e. has less than 18 valence electrons.
Herberich’s borabenzene, in which 5p electrons are delocalized in 6 contiguous
The 2-electron imido ligand forms an angle with the metal, because the p orbital orbitals are among the best known 5-electron L2X ligands.
does not interact with the metal. On the other hand, in the 4-electron imido
ligand, the p orbital of the nitrogen atom interacts with the vacant d metal orbital, RB(C3H3N2)3, tris-1- (pyrazolyl)borato, noted Tp, with R = H, alkyl, aryl.
and the M-N-R chain is linear, as for instance in [Mo(NAr)3]. H
• •
Ar B
N N N
M N M N R
N N
Mo N
R N
2e 4e ArN NAr
Hawthorne has introduced the carborane ligands and prepared many metallo- Tp
carborane complexes, in particular the dianion nido-carborane C2B9H112– whose
frontier orbitals are analogous to those of the cyclopentadienyl (see Wade’s rules 5-electron LX3 ligands
for the stereoelectronic structure of the carboranes, Chap. 2.2.4). In its neutral
The X3 nitrido and phosphido ligands can aso become LX3 if the metal is electron
form, this ligand is counted as a 4-electron LX2 ligand.
deficient
2– M N
6-electron L3 ligands
9 10 11
C6H6 (benzene), other aromatics and polyaromatics, cyclic trienes, borazine
8 7
2 3 B N
= CH N B
6 4
5 = BH B N
1 Cr Cr
OC CO OC CO
CO CO
C5H5N (pyridine), C5H5P (phosphinine), C4H4S (thiophene)– these three ligands 6-electron L2X2 ligands (divalent)
are most often L ligands, however, via the lone pair of the heteroatom N4-ligands: porphyrins and phthalocyanins
H H
H H
H C H H C H C C H
C C C C C N
C C N N
C C C C H H N N
H N H H P H S
HC M CH N M N
Trisulfides, for instance (CH2CH2S)3; triamines, for instance Wiegardt’s
triazacyclononane (CH2CH2NR)3; triphosphines, for instance CH(CH2PPh2)3 N N N N
H
C N
R N N R S S H
C
Porphyrin Phthalocyanin
N S P P
P
The porphyrin ligand is involved, as an iron complex, in cytochromes (for instance
R
in cytochrome P450), hemoglobin and myoglobin (see Chap. 18).
[9]ane [9]ane S3
7-electron L3X radical ligand
Tris(pyrazolyl)methanes and the little-developed tris(pyrazolyl)silanes are the
C7H7 (cycloheptatrienyl)
neutral analogs of the widely known tris(pyrazolyl)borates
H R’ CH3
R C R C H3C Si CH3 CH3
M
N RN N N N N
N N N N N N N N N
L4, L5 and L6 inorganic macrocyclic and cage ligands
R’ N R’ R N R H3C N CH3
Providing respectively 8, 10 and 12 electrons with the heteroatoms O (ether), N
R’ R CH3
(amine), S (sulfide) and P (phosphines).
Pincer N-heterocyclic carbenes recently introduced by Crabtree and Eisenstein Cyclooctatetraene C8H8 (COT) can be an 8-electron L4 ligand. With the alkaline
metals, lanthanides and actinides however, COT is found as the aromatic COT2–
dianion with 10 electrons (Hückel’s rule: 4n + 2 electrons, n = 2).
There are even 16-electron L8 ligands (for example O4N4) for the large actinide
N N N H 2C N H2 C ions.
N .. .. N N N
Z ligands
R R
..
..
N N These Z ligands are Lewis acids that do not bring electrons to the metal, but on the
Various R, in particular R = n-Bu for solubility
contrary, bind to the metal in order to get their missing electron pair. Example: BH3,
and other pyridine-linked pincer ligands (phosphines, amines, thiols) AlMe3.
1 – MONOMETALLIC TRANSITION-METAL COMPLEXES 31 32 PART I – STRUCTURES OF THE TRANSITION-METAL COMPLEXES
In conclusion, the modes described here are the most frequent ones, but others can 2. THE CHARACTERISTICS OF THE TRANSITION-METAL
be found. The ligands can have variable coordination modes. If one hesitates, one IN THE COMPLEXES
can follow the 18-electron rule (vide infra). Such ambiguities exist for the ligands
NO, polyenes and alkynes. NO is frequently found as either a bent 1-electron or a
The four main characteristics of the transition-metal in a complex that allow to
linear 3-electron ligand depending on the metal electron count. Alkynes often are
define this complex well are the Number of Valence Electrons (NVE), the Number
4-electron ligands but the 2-electron count is found when the metal has 18 valence
of Non-Bonding Electrons (NNBE), the Oxidation State (OS) and the Coordination
electrons. Then, the second alkyne orbital does not find a vacant metal orbital of
number (C).
adequate symmetry. Other coordination modes will be found in polynuclear
compounds for the same ligands. Organometallic chemistry is very versatile and the
coordination modes can be numerous for the same ligand. 2.1. THE NUMBER OF VALENCE ELECTRONS (NVE)
Summary of ligands: most common modes It is the total number of electrons in what is defined as the transition-metal valence
shell, i.e. the group of n s, (n – 1) d, n p sublayers. It contains the electrons that are
Ligands Type Number of electrons given to the metal initially in the metal valence shell and the electrons brought by the ligands.
H, CH3, Ph X 1 NEV = nM + 2 nL + nX – q
C2H4, CO, PPh3, O2, N2 L 2 with: nL = number of L ligands, nX = number of X ligands, q = charge of the
-allyl, -enyl, dtc LX 3 complex, nM = number of electrons initially in the metal valence shell, i.e.
diolefin, diphosphine, diamine L2 4 2 electrons from the ns sublayer and the electrons of the sublayer (n – 1) d which
depends on the nature of the metal sublayer.
Cp, dienyl L2X 5
porphyrinato L2X2 6
nM = 2 + y for a transition element having the electronic structure ns2 (n – 1) dy np0.
alkyne RC≡CR L 2 according to metal need This is the number of electrons remaining in the valence electron shell after
L2 4 formation of the bonds with the ligands. Inorganic chemists often use the symbol dn
1 – MONOMETALLIC TRANSITION-METAL COMPLEXES 33 34 PART I – STRUCTURES OF THE TRANSITION-METAL COMPLEXES
for a metal atom or ion (or the complex itself) that has n non-bonding electrons on Some carbenes (triplet ones) are clearly X2 ligands whereas others (singlet ones) are
this metal in the complex (n = NENL). rather L ligands, but in complexes, there is a whole area in between. The decision is
NNBE = nM – nX – q = nM – OS clear-cut, but the reality is often not so obvious. CO is an L ligand, thus [Ni(CO) 4]
has OS = 0, but if it was considered an X2 ligand, the oxidation state Ni would be 8,
For instance, [FeCp(CO)2]–, for which NNBE = 8 – 1 – (–) = 8 is said to be a which would make an enormous difference. Is the metal so electron-rich that it
d8 complex. deserves a 0 oxidation state or does the reality correspond to a somewhat larger OS?
The value of NNBE is important in understanding the reactivity of a complex. For Another example is the family of the hydrides: ReH92– is a ReVII complex, but is the
instance, the metals in complexes whose NNBE is 0 cannot be oxidized, and Re metal so electron-poor in this complex that it really deserves such a high
reactions on the metal in such complexes will be limited to those bringing electrons. oxidation state?
The NNBE will also be useful to define the geometry of the complex, because the Note, however, that a strength of this formalism is that the OS cannot be higher than
orbitals of the non-bonding electrons will have to be taken into account in the the number of the metal group. For instance, for Sc and the lanthanides, the OS
organization of space around the metal center. cannot be larger than +3; for the Ti column, it cannot be larger than +4; for the
Note - do not confuse dy, given by the electronic structure ns2 (n – 1) dy np6 of the V column, it cannot be larger than +5. For instance, this corroborates the fact that
atom and dn indicating the NNBE (equal to n) of the metal in a complex. This CO is conventionally an L ligand, not X2: in [Cr(CO)6], Cr (OS = 0) could not have
double d nomenclature (atomic and molecular) is unfortunate. Thus, it is important OS = +12, because the Cr atom can only provide 6 electrons.
to clearly distinguish these two meanings; n = y only if the oxidation state is 2. For
instance, ferrocene, [FeCp2] is a d6 complex, n = y = 2. On the other hand, 2.4. THE COORDINATION NUMBER (C)
[Fe0(CO)5] is a d8 complex (n = NNBE = 8) in which the iron atom has 6d electrons
(y = 6). It is the number of occupied coordination sites on the transition-metal center. Thus,
a L or X ligand occupies one site, a L2 ligand (and most often a LX ligand) occupies
2.3. THE OXIDATION STATE (OS) two sites, and the L3 or L2X ligands occupies 3 sites. When the complex is of the
form MLnXpq, the coordination number is:
The oxidation state (OS) is obvious when the complex contains only L or Ln C = n L + pX
ligands, because it is then equal to the charge q of the complex. For instance, in
–
[Fe(H2O)6]2+, it clearly appears, just by looking at the formula, that the OS is +2. Example: [FeCp(CO)2]
Also, the OS is –2 in Collman’s reagent [Fe(CO)4]2–. The OS is not so apparent, [FeCp(CO)2]– is a FeL4X– complex: C = 4 + 1 = 5.
however, when the complex also contains X or Xn ligands, such as in [FeCp(CO)2]–,
This simple formula does not work, however, when the complex contains one or
for example. It is then necessary to add the number nX of X ligands to the charge q
several X2 or X3 ligands or a single-site LX ligand such as linear NO. Applying the
of the complex:
OS = nX + q formula gives a result in excess. In these cases, one must count the number of sites
carefully.
Thus, for the above example of the monoanionic iron complex, OS = 1 + (–1) = 0.
Example: [RuCp(CO)2(=CH2)]+
The transition metals all have numerous oxidation states, which accounts for the
[RuCp(CO)2(=CH2)]+ has a coordination number equal to 6 (3 for Cp, 2 for CO and
richness of their chemistry. For purely organometallic compounds, the oxidation
1 for the carbene).
states are low, zero or negative. On the other hand, purely inorganic complexes
always have positive, even high oxidation states. For instance, in [Fe(S2CNMe2)3] +, Summary of the characteristics
each dithiocarbamato ligand is LX, and the complex is of the type FeL3X3+;
OS = 3 + 1 = +4. NVE = nM + 2 nL + nX – Q
NNBE = nM – nX – q = nM – DO
Let us recall that the OS is a formalism that we have decided to adopt for the whole
OS = nX + q (mononuclear complexes)
area of inorganic and organometallic chemistry as a very practical form of language
to communicate in the chemical community, and any controversy about oxidation C = nL + nX (except if X2, X3, single-site LX)
states is meaningless. One can take the example of carbenes to illustrate this point.
1 – MONOMETALLIC TRANSITION-METAL COMPLEXES 35 36 PART I – STRUCTURES OF THE TRANSITION-METAL COMPLEXES
6 pseudo-octahedral M M H
Here (above) are some classic examples among many others for which it is
indispensable to write the hapticity of the ligands.
capped octahedron M [ReH(PR3)3(MeCN)3]+
Note 1 - The ligand hapticity should not be confused with the number of electrons
7 given by the ligand to the metal. For hydrocarbons, these two numbers usually are
identical, especially for hapticities larger than 1. The C2H4 ligand is bound to the
pentagonal biprism M [IrH5(PPh3)2] metal by its 2 carbon atoms and brings 2 electrons, the Cp is bound to the metal by
its 5 carbon atoms and brings 5 electrons, etc. The monohapto ligands can bring a
variable number of electrons: 1 electron for X ligands such as CH3, 2 electrons for
1 – MONOMETALLIC TRANSITION-METAL COMPLEXES 37 38 PART I – STRUCTURES OF THE TRANSITION-METAL COMPLEXES
L ligands such as CO and for X2 ligands such as the carbenes, 3 electrons for Mo = 6
carbynes and linear NO, 4 electrons for the O ligand (LX2) in d0 complexes, and Mo MoL6 6L = 12
5 electrons for the N ligand (LX3) in d0 complexes. 18
Note 2 - With the ionic convention that we are not using here for the transition-
metal complexes, the X ligands are counted with a negative charge, i.e. X–. Then
they bring two electrons from X– (just like L ligands) to the metal cation. The Cr = 6
characteristics NVE, NNBE, OS and C remain the same whatever the convention. Cr CrL6 6L = 12
For instance, with the ionic convention in ferrocene, iron is considered as Fe2+ and 18
has thus 6 valence electrons. Each Cp– ligand brings 6 electrons. NVE = 6 + (2 6)
= 18; NBBE = 6; OS = 2; C = 6 . Our convention is justified with transition-metal Re = 7
complexes, because they are essentially covalent (for instance, ferrocene is pentane
5L = 10
soluble and the metal does not bear a charge), but this is no longer the case for the Re ReL5X
X = 1
Sc, Ln and Ac complexes (see Chap. 12.4) that are essentially ionic and for which OC CO
CO 18
the ionic convention is preferred. For instance, in uranocene, [U(COT)2], the
dianionic COT2– ligands are genuine 10-electron ligands for the ion U4+ (see
Chap. 12.4). Fe = 8
Fe FeL5 5L = 10
OC CO 18
4. THE “18-ELECTRON RULE”: CO
TENDENCIES AND EXCEPTIONS
The NVE is often equal to 18 for transition-metal organometallics and for many Ta = 5
inorganic complexes. This “18-electron rule” should be better viewed as a strong CH3
4L = 8
tendency than a rule, but it is followed by a majority of complexes (despite many Ta TaL4X5
CH2 5X = 5
exceptions, vide infra). The 18-electron electronic structure often brings a good
18
stability for complexes. For instance, this is the case for the metal carbonyl
complexes, possibly the largest family. Transition-metal sandwich complexes are
more stable in the 18-electron count than in others. The 18-electron electronic + Co = 9
structure is also found most of the time in complexes containing a mixture of 4L = 8
carbonyls, hydrocarbons, carbenes, hydrides, etc.1.1-1.3 Co CoL4X2+ 2X = 2
The 18 electrons correspond to the filling of the 9 molecular orbitals (one pair for – (+ 1) = –1
each of them) coming from the 9 atomic orbitals of the transition-metal 18
{5(n – 1) d orbitals, 1n s orbital and 3n p orbitals}. Some of these 9 molecular
4–
orbitals are bonding, and some others are non-bonding or antibonding. The CN Fe = 8
interaction of atomic orbitals with the ligand orbitals of the same symmetry usually NC CN 6X = 6
gives rise to occupied molecular orbitals and to unoccupied antibonding orbitals.1.4 Fe FeX64–
NC CN – (– 4) = +4
The NVE is sometimes not 18, however, for various reasons: CN 18
The more at the right of the periodic table the metal is located, the more for steric reasons, the valence shell be far away from 18 electrons. The NEV can
d electrons it has, and it will be easier to complete its coordination sphere with come down to 8, for instance in [Ti(CH2Ph)4] or 10 in [TaMe5], 12 in
18 valence electrons. On the other hand, the early transition metals may have an [Cr(CH2SiMe3)6]. In all these complexes, however, the metal tries by all means to
average or weak tendency to fulfill the 18-electron shell. It can even happen that,
1 – MONOMETALLIC TRANSITION-METAL COMPLEXES 39 40 PART I – STRUCTURES OF THE TRANSITION-METAL COMPLEXES
accept more electrons by binding in ways that are not included in the formal
electron count (agostic interactions, distal interactions with aromatic rings).
e g*
“Noble” transition metals (2nd and 3rd periods of transition metals: Ru and Os, Rh
and Ir, Pd and Pt) whose complexes are excellent catalysts, often have a square- z
planar structure and NVE = 16. This is due to the fact that the pz orbital cannot be y
∆ dz2 dx2-y2
occupied in the square-planar geometry, because it has a high energy. Thus, the x
complexes are stable with the 16-electron count.
t2g
z pz
L
Mn+ ML6n+ dxy dyz dxz
4 more electrons, in addition to 18. For example, for n = 1, theoretical = 1.73 B; for n = 2, theoretical = 2.83 B; etc.
1 – MONOMETALLIC TRANSITION-METAL COMPLEXES 41 42 PART I – STRUCTURES OF THE TRANSITION-METAL COMPLEXES
In conclusion, departures from the 18-electron rule are much more frequent than the electron pair on the carbon of empty π*
olefin orbital occupied d
exceptions to the octet rule (NVE = 8) in organic and main-group inorganic the carbonyl σ donor
metal orbital
chemistry.
C
5. BONDING BETWEEN THE TRANSITION METAL AND
M C O M
THE -ACCEPTOR (CO, C2H4) AND -DONOR LIGANDS
(HALOGENO, ALKOXY, AMINO) C
CO and C2H4 are two among the most important ligands in organometallic
chemistry: they are representative of the type of bond encountered. They are occupied d occupied π olefin orbital
metal orbital empty π*
L ligands, and as such, they coaxially give a pair of electrons to the metal: this is the CO orbital used as a σ donor
bond of a p orbital of CO or of a orbital of C2H4 towards a vacant d metal
orbital. It is this bond that is formally taken into account in the characteristics of the The consequences of this backbonding are experimentally observable. X-ray crystal
complexes. There is another bond that is not considered in the electron count, but structures show that the C-O and C-C bonds of CO and C2H4 respectively are longer
which is also very important: it is the lateral backbonding of type from a filled when these molecules are bonded to a transition metal than when they are free,
d metal orbital to a vacant antibonding orbital of CO or C2H4. This bond is in the which means a decrease of the bond order upon coordination. This bond lengthe-
opposite direction to the bond, as indicated by its name. It partly compensates the ning is due to the population of the antibonding ligand orbital in the complex. Also,
bond and allows the metal to give back part of its excess of electron density and the infrared frequency of the carbonyl is lower for the CO ligand in the metal-
therefore to exist in a low or even negative oxidation state. carbonyl complex than for free CO, as a consequence of the decrease of the C-O
bond order upon coordination.
It is this backbonding that makes the big difference between inorganic and
organometallic chemistry. For inorganic complexes, the backbonding cannot exist, If the backbonding becomes even more important, the lengthening of the C-C
because of the too high energy level of the antibonding N and O orbitals (the P and bond will be such that this bond will reach the length of a single bond. In this
S atoms are weak acceptors). The fact that there is no backbonding with O and extreme case, it is appropriate to change its representation. The system should then
N ligands explains, for instance, why complexes such as M(H2O)6 or M(NH3)6 no longer be viewed as a metal-olefin complex, but as a metallacyclopropane. This
(M = Cr, Mo or W) do not exist even with NVE = 18. These complexes would have reasoning can be extended to metal-alkyne complexes and to the bond between the
OS = 0, and the metal would be too electron-rich, since backbonding would be carbonyl of ketones and very oxophilic early transition-metals such as Sc and Ti.
absent. On the other hand, the only solution to compensate the enormous electronic
gain provided by the inorganic ligands is that the metal be oxidized at least to the O
oxidation state +2. The complexes [Fe(H2O)6]2+ and [Fe(NH3)6]2+ are stable whereas M M M
the 18-electron complex [Fe(CO)6]2+ is extremely fragile. Indeed, the metal is very
R R O
electron-poor in the latter complex because of the loss of electron density through
oxidation to FeII, and therefore the backbonding to all the CO ligands decreases. Mo
The Dewar-Chatt-Duncanson model that takes into account this bonding mode,
CH2
including the backbonding with the CO and C2H4 ligands, is represented below O
for classic metal-carbonyl and metal-ethylene bonds. Of course, the other unsatu- M M M
rated hydrocarbons bind the transition-metals according to the same backbonding R
model. R M = Sc, Ti or Zr
zero. In d0 complexes, this backbonding cannot take place, which considerably 6. MOLECULAR ORBITAL DIAGRAMS
destabilizes metal-ligand bonds that can then only be very weak and, in fact, occur
very rarely. This is the case, for instance, for Sc, lanthanide and actinide complexes Molecular orbital calculations allow to derive a molecular orbital diagram resulting
that are most often in the oxidation state +3. from the interaction of metal orbitals with the ligand orbitals of corresponding
The opposite effect is also well known. Some ligands provide electron density symmetry.1.4 The molecular orbital diagram of the complex contains bonding and
from the p electron pair of an heteroatom as ligand: these ligands are donors. It is antibonding molecular orbitals resulting from these interactions and non-bonding
necessary to have a geometrical overlap between the p ligand orbital and the empty metal orbitals that do not interact with ligand orbitals. Let us take, for instance, the
d metal orbital involved in the bonding. Therefore, the available p ligand orbital simplest case of an octahedral ML6 complex. Six metal orbitals, i.e. the s, p and two
must be perpendicular to that engaged as a donor with the metal. This interaction d orbitals interact with the six orbitals of the six ligands in the six directions of the
proceeds well with halogeno, alkoxy, amino, oxo and amido ligands and d0 metals octahedron, whereas the three other metal d orbital remain non-bonding. If L is a
such as ScIII, TiIV or TaV. One may notice that the early transition-metal chemistry acceptor such as CO in a metal hexacarbonyl complex [M(CO)6], M = Cr, Mo or
contains a plethora of such bonds in complexes that would have a NVE much lower W, the three non-bonding metal d orbital are those involved in the backbonding
than 18 valence electrons if this bonding was not taken into account. with the CO antibonding * orbitals. As a result, these three d orbitals will then
become bonding. The energy difference between the two groups of d orbitals, called
Examples: [ScCp*2Cl], [TiCp2Cl2], [Ta(mesityl)Cl4], [Ta(OAr)5] or 10 dQ, will consequently be increased (diagram on the right below).
Ar = 2,5-diisopropylphenyloxy. t1u* t1u*
ns ns
eg* eg*
On the other hand, if the metal has non-bonding d electrons, their orbitals and the
p orbital of the ligand exert mutual repulsion against each other, which leads to ∆°
some destabilization characterized by an increase of the energy level of these (n – 1)d σ (n – 1)d ∆° σ
t2g
orbitals. This does not prevent the compounds to exist with halides, alkoxy or amido t2g
ligands in late transition-metal complexes, but interesting effects are sometimes
M L M L
encountered. For instance, the complex [RuCp*(PMe3)(Cl)] is stabilized, due to the
donor chloride that gives a non-bonding p electron pair into the vacant d orbital.
Thus, the compound reaches NVE = 18 by counting the chloro as a 3-electron
LX ligand: a1g, eg, t1u a1g, eg, t1u
[PtCl3(C2H4)]–, square planar, [PtLX3]–, 16, 8, 2, 4; [V(CO)7]+: the addition of two disadvantages, a coordination number larger than 6
and the cationic nature of a binary metal carbonyl leads to the non-inexistence of
[CoCp2], sandwich structure with parallel rings, [CoL4X2], 19, 7, 2, 6; this complex in spite of its 18-electron structure.
[Fe( 6-C6Me6)2], sandwich structure with parallel rings, [FeL6], 20, 8, 0, 6; [Cr(H2O)6]: non-existent as all the binary water complexes in which the metal has a
[AuCl(PPh3)], linear, [AuLX], 14, 10, 1, 2; low oxidation state such as (0).
[Fe( 4-C8H8)(CO)3], trigonal bipyramid (piano stool), [FeL5], 18, 8, 0, 5; [Ni(H2O)6]2+: relatively stable 20-electrons complex as [CoCp2]–, vide supra, and in
addition because of the ideal coordination number (6) and oxidation state (II).
[Ru(NH3)5( 1-C5H5N)]2+, octahedral, [RuL6]2+, 18, 6, 2, 6;
[ReH9]2–: stable 18-electron complex in spite of its very high coordination number,
[Re(CO)4( 2-phen)]+, octahedral, [(ReL6]+, 18, 6, 1, 6; because hydrides, the only ligands present in the coordination sphere, are very
[FeCp*(CO)(PPh3)(CH2)]+, pseudo-octahedral (piano stool), [(FeL4X3]+, 18, 6, 4, 6; small.
[Ru(bpy)3]2+, pseudo-octahedral, [RuL6]2+, 18, 6, 2, 6. [Zr(CH2Ph)4]: in principle an 8-electron complex that is thermodynamically stable
but kinetically very reactive with many substrates. The stability, in spite of the very
1.2. In the form [FeCp*( 2-dtc)2], both dithiocarbamate ligands are chelated to iron,
low NVE, is due partly to the absence of H, and partly to the interactions
[FeL4X3], 19, 5, 3, 7. In the form [FeCp*( 2-dtc)( 1-dtc)], one of the two dithio-
between the phenyl rings and the metal decreasing the Zr-C-Ph angle below 109°
carbamate ligands is chelated to iron whereas the other is monodentate, [FeL3X3],
(the value of the NVE = 8, only is conventional and does not take this interactions
17, 5, 3, 6. The two forms having respectively 19 and 17 electrons are both to some
into account).
extent destabilized, being both one electron away (in excess and shortage
respectively) with respect to the robust 18-electron structure. The main reaction of [Zr(CH2Ph)4(CO)2]: does not exist as almost all the mono- and bimetallic metal-car-
19-electron complexes in the absence of other substrate is to partially decoordinate bonyl structures in which the metals do not have non-bonding electrons (NNBE = 0)
giving a 17-electron complex, and this reaction is very fast. The main reaction of (the M-CO bond needs backbonding from some filled metal d orbitals). There are
17-electron complexes in the presence of a ligand L (in the absence of reducing only exceptions in lanthanide chemistry in which the metal-CO bond is very weak
agent) is also very fast, addition giving the 19-electron complex. Thus, the (R. Anderson’s work).
ANSWERS TO EXERCISES 537