Essays

DOI: 10.1002/anie.200503492
Spin Configurations

How High the Spin? Allowed and Forbidden Spin States

in Transition-Metal Chemistry**
Santiago Alvarez* and Jordi Cirera
Keywords:
coordination compounds · ligand field theory ·
molecular orbital theory · spin states ·
transition metals

“When I use a word,” Humpty Dumpty methyl ligands were reported by Seppelt and chemical properties of such com-
said in rather a scornful tone, “it means and co-workers.[2] Homoleptic cyano pounds are not so well understood,
just what I choose it to mean—neither compounds have been known for a long despite the tremendous advances in
more nor less.” time,[3] but some new elegant examples our capabilities to analyze them both
“The question is,” said Alice, “whether were recently reported[4–6] in view of crystallographically and computational-
you can make words mean different their potential use as building blocks for ly. Especially for those electron config-
things.” polynuclear and extended structures urations that allow for more than one
Lewis Carroll, Through The Looking Glass with interesting magnetic properties.[7] spin state, it is not always easy to predict
(1871) Even such large families of complexes which will be the ground state for a
as the metal carbonyls[8] and alkoxides[9] given combination of metal ion and
are continuously expanding. To date, ligands. The concepts of high-spin and
Introduction hundreds of thousands of transition- low-spin states, and the related classifi-
metal complexes have been structurally cation of low- and high-field ligands,
The quest for new homoleptic com- characterized, and the mainstream lines derived from the crystal-field treatment
plexes (those with identical ligands) is a of research in that area are concerned of an octahedral compound, are usually
common target of several research with the self-assembly of large polynuc- extrapolated to other cases in a straight-
groups in the field of synthetic transi- lear systems that contain hundreds of forward way, leading in some instances
tion-metal chemistry. For instance, Men- atoms, tens or even hundreds of which to ambiguities.
j n, Forni!s, and co-workers have pre- can be transition-metal atoms.[10] The The case of the homoleptic cyano
pared a series of organometallic [MRx] question is thus asked: Why then is complexes illustrates the problems and
complexes (M = transition metal; R = there a renaissance of the chemistry of lack of a unified use of the low-spin and
polyhalophenyl group; x = 4–6)[1] that mononuclear complexes with only one high-spin labels. The [Mn(CN)4]2
anion
cover almost all the dn electron config- type of relatively simple ligand? reported recently by Miller and co-
urations and in most cases are para- Clearly, there is a synthetic chal- workers is a textbook example of the
magnetic. Even simpler in composition, lenge as the preparation of compounds general tendency of tetrahedral com-
a variety of homoleptic complexes with with only one type of ligand is not plexes to have their d electrons arranged
always an easy task. In our view, though, with parallel spins, resulting in the high-
there is also the perception that the est possible spin state (Figure 1 a).[4] (We
[*] Prof. S. Alvarez, J. Cirera stereochemistry, electronic structure, will see in the last section of the Essay
Departament de Qu7mica Inorg8nica
and
Centre Especial de Recerca en Qu7mica
Te<rica (CeRQT)
Universitat de Barcelona
Mart7 i Franqu@s 1–11
08028 Barcelona (Spain)
Fax: (+ 34) 93-490-7725
E-mail: santiago@qi.ub.es
[**] Financial support of this work was pro-
vided by the Ministerio de Ciencia y
Tecnolog7a (project no. CTQ2005-08123-
C02-02/BQU) and through an MECD
doctoral grant to J.C. The authors thank Figure 1. Splitting of d orbitals in a) a tetrahedral high-spin complex, [Mn(CN)4]2
, b) a square-
R. Poli and R. Hoffmann for illuminating pyramidal complex, [Cr(CN)5]3
, with high-spin (left) or low-spin configuration (right), and c) an
discussions. octahedral low-spin complex, [Cr(CN)6]4
.

3012  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2006, 45, 3012 – 3020
 Angewandte
Chemie

that low-spin tetrahedral complexes do into account the typical d-orbital split- number and geometry and establish
exist under certain circumstances.) Such ting pattern,[17] two alternative spin accordingly the possible spin states for
a tendency is explained by the small states can be expected in the presence different coordination geometries and
splitting of the d orbitals in a tetrahedral of weak-field (S = 2) or strong-field dn configurations. Last, we address the
environment, even in the presence of ligands (S = 1; Figure 2). A ground spin possibility of less-common orbital-split-
strong-field ligands such as cyanide. The ting situations and provide some ration-
octahedral [Cr(CN)6]4
anion, on the ale for known exceptions to the rules
other hand, presents a low-spin config- proposed.
uration with S = 1 (Figure 1 c),[4] which
can be explained by the large splitting of
the d orbitals due to an octahedral Splitting of the d Orbitals
environment of strong-field ligands. In
both of these cases, two different factors The use of high-spin and low-spin
determine the magnitude of the d-orbi- labels for electron configurations is
tal splitting and, hence, the choice of the associated with the well-known splitting
spin state: the molecular (coordination) Figure 2. Splitting of d orbitals in a square- pattern of the metal d orbitals into two
geometry and the nature of the ligands. planar d6-ML4 complex and its high-spin (left) degenerate sets in octahedral and tetra-
(Another factor, which we will comment or low-spin (right) configurations. hedral complexes (Figure 1), when crys-
on later, is the interelectronic repul- tal-field or molecular-orbital theories
sion.) are applied. In both cases, there are
If we consider yet a third type of state with all electrons paired (S = 0) only two distinct choices of electron
structure, such as the square-pyramidal could be conceived, but, to the best of configuration, characterized by the
or trigonal-bipyramidal geometries re- our knowledge, there is no experimental highest and lowest possible spin for
ported recently for the [Cr(CN)5]3
evidence for such a spin state in a n electrons in the five d orbitals. How-
anion,[5] should we expect all electrons square-planar d6-metal complex. Should ever, the situation is less clear for other
to have parallel spins or not? Should the the configuration as shown on the right coordination geometries, for which
relevant energy gap between the a1g and in Figure 2 be termed the low-spin state more than two sets of symmetry-equiv-
eg orbitals in the presence of strong or the intermediate-spin state? Is the alent orbitals can be found, as in square-
ligands such as cyanide be large (Fig- diamagnetic state possible? planar (Figure 2) and square-pyramidal
ure 1 c) or small (Figure 1 a)? It has The correct characterization and complexes (Figure 1 b). Therefore, let us
been shown experimentally that the description of the spin state of a tran- try to establish some rules that allow us
high-spin electron configuration is pre- sition-metal complex is very important to treat these less symmetric cases in a
ferred (Figure 1 b).[5] However, could we given the different magnetic, structural, simple but reliable way. We do so by
have predicted the spin state before- and chemical properties associated with considering the molecular orbitals of an
hand by using qualitative molecular- different spin states.[18] However, in the idealized complex with predominant
orbital arguments? Moreover, if the most-recent revision of the IUPACEs metal d character.
[Mn(CN)5]3
anion with an extra elec- Nomenclature of Inorganic Chemistry[19] When d orbitals interact with the
tron could be prepared, should we the naming of spin configurations or symmetry-adapted linear combinations
expect the fifth electron to lie parallel states is completely overlooked. There- (SALC)[20] of ligand orbitals, they form
(occupying a b1g orbital) or antiparallel fore, we believe that it is timely to reflect bonding and antibonding molecular or-
(occupying a b2g orbital)? Similarly, we on when and why are high-, intermedi- bitals for each symmetry representation.
could turn to lower coordination num- ate-, or low-spin configurations possible. Atomic d orbitals that do not match the
bers and ask if the high-spin configura- It would be convenient that the scien- symmetry of the SALCs are strictly
tions found for three-coordinated d7-FeI tific community establishes some nam- nonbonding, as is the case for the
and d6-FeII cations with the diketiminate ing convention to avoid ambiguities and t2g metal set in an octahedral environ-
bidentate ligand[11] is something we misunderstandings. For this reason, we ment (assuming only s-donor ligands) or
could have predicted without calcula- present herein a qualitative revision of for the b2g and eg orbitals in a square-
tions or experiment. the orbital principles underlying the planar four-coordinate complex (Fig-
Another possible source of ambigu- choice of spin states in transition-metal ure 3 a). For the purpose of the present
ity comes from labeling some com- compounds. As the interplay of d-orbital discussion, we focus on the antibonding
pounds as having “intermediate” spin splitting and electron repulsion deter- combinations, whose largest contribu-
states. This label has been applied to mines the spin configuration, we start by tion comes in most cases from the metal
Jahn–Teller-distorted six-coordinated looking at some general rules that d orbitals, such as the dz2 and
metal ions, such as CoIII in LaCoO3,[12] govern the relative magnitude of these dx2
y2 orbitals in a square-planar com-
to FeII and FeIII porphyrin complexes in two effects. We present a general mo- plex. For simplicity, we refer to these
which the Fe ion is four-, five-, or six- lecular-orbital scheme that provides a molecular orbitals as d-MOs to differ-
coordinated,[13–15] and to a square-planar sound qualitative description of the entiate them from the unperturbed
FeII dithiolato complex.[16] In the case of relative energies of the metal d orbitals atomic d orbitals. Thus, now we can
a square-planar FeII complex, taking in complexes with any coordination classify the d-MOs into two sets: those

Angew. Chem. Int. Ed. 2006, 45, 3012 – 3020  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 3013
Essays
split into two distinct sets, with the lower
energy set containing the nonbonding
(both strictly and formally) orbitals and
the higher energy set comprising the
metal–ligand antibonding orbitals. The
sets of orbitals are separated by a
significant d-orbital energy gap.
LetEs pretend that the splitting be-
tween nonbonding and antibonding d-
MOs is much larger than the energy
Figure 3. d Molecular orbitals in square-planar ML4 complexes: a) without hybridization; b) with differences within each of these two
s+dz2 hybridization. orbital sets; it is therefore this energy
gap that should be taken into account
when considering alternative spin con-
with antibonding character and those with metal p orbitals that have the same figurations.[21] There are good reasons
with strictly nonbonding character (Fig- symmetry (Figure 4 c). for this assumption, as illustrated by the
ure 3 a). In those cases the d-MOs are hybri- splitting of the d orbitals in different
However, there are two different dized such that their electron density is homoleptic cyano complexes (Figure 6)
situations regarding the antibonding d- concentrated away from the ligands, according to density functional theory
MOs, depending on whether or not appearing similar to a balloon when (DFT) calculations.[22] In each case, the
there are valence s or p metal orbitals subject to pressure. The effect of press- symmetry of the coordination sphere
of the same symmetry. In the presence ing a balloon in axial and equatorial determines which orbitals belong to
of symmetry-matched valence s or p or- directions (Figure 5 a, b) can be com- each of these two sets (see Table 1). It
bitals, energy minimization of the sys- pared with the hybridization found for can be seen that the number of anti-
tem imposes extensive mixing of orbitals the dz2 orbital in ML2 and ML4 com- bonding d-MOs corresponds to the
of the same symmetry, in such a way as plexes (Figure 5 c, d). In summary, we number of d orbitals that participate in
to make the lower one as little metal– should expect the d-MO manifold to be the valence-bond hybridization
ligand antibonding as possible, while the
upper, unoccupied one (formerly s or p),
accumulates most of the antibonding
character. Such a hybridization converts
the corresponding d-MOs into formally
nonbonding orbitals, as happens with dz2
in the case of a square-planar complex;
it is hybridized with the s orbital (both
being A1g-symmetric; Figure 4 a). Con-
versely, in a linear complex hybridiza-
tion occurs with the opposite sign (Fig-
ure 4 b). A similar situation appears, for
example, for the t2 set of d-MOs in a
tetrahedral complex which hybridize

Figure 4. Minimization of antibonding charac-

ter through hybridization of d-MOs:
s+dz2 hybridization in a) square-planar ML4 Figure 5. Effect of applying pressure to a balloon along the z axis (a) and circularly in the
and b) linear ML2 complexes, and c) p+d hy- xy plane (b), compared to the shape of the dz2 orbital (see Figure 4 a, b) in linear [Ag(CN)2]
(c)
bridization in tetrahedral complexes. and in square-planar [Ni(CN)4]2
(d), as obtained from DFT calculations.

3014 www.angewandte.org  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2006, 45, 3012 – 3020
 Angewandte
Chemie

Figure 6. Relative energies of the d molecular orbitals in several homoleptic cyano complexes, from left to right: linear [Ag(CN)2]
, trigonal-planar
[Cu(CN)3]2
, square-planar [Ni(CN)4]2
, tetrahedral [Mn(CN)4]2
, square-pyramidal [Cr(CN)5]2
, nearly trigonal-bipyramidal [Cr(CN)5]2
, octahedral
[Cr(CN)6]4
, pentagonal-bipyramidal [V(CN)7]4
, and dodecahedral and square-antiprismatic [Mo(CN)8]4
. To facilitate comparison of the energy
gaps, the orbital energies of each compound are referred to that of the highest formally nonbonding d-MOs (0.1 a.u. = 2.19 K 104 cm
1).

Table 1: Expected hybridization of the metal atom in MLn complexes and formal bonding nature of state of the metal and also upon de-
the metal d orbitals when considering metal–ligand s interactions only. scending a periodic group.)
Symbol[a] Geometry Hybridization Nonbonding[b] Antibonding The choice of the spin state in a
particular case, however, depends on
L-ML2 linear sp all none
TP-ML3 trigonal plane sp2 all none both the d-orbital energy gap and the
T-ML4 tetrahedron sp3 all none electron repulsion associated with the
SP-ML4 square plane dsp2 dz2, dxy, dxz, dyz dx2
y2 occupation of the same orbital by two
TBPY-ML5 trigonal bipyramid dsp3 dxz, dyz, dx2
y2, dxy dz2 electrons with opposed spins (“pairing
SPY-ML5 square pyramid dsp3 dxz, dyz, dxy, dz2 dx2
y2 energy”). Thus, a large gap and a small
OC-ML6 octahedron d2sp3 dxy, dxz, dyz dx2
y2, dz2 pairing energy favor a low-spin state,
TPR-ML6 trigonal prism d2sp3 dxy, dx2
y2, dz2 dxz, dyz
whereas a small gap combined with
PBPY-ML7 pentagonal bipyramid d3sp3 dxz, dyz dz2, dx2
y2, dxy
OCF-ML7 capped octahedron d3sp3 dx2
y2, dxy dz2, dxz, dyz large pairing energy favors a high-spin
TPRS-ML7 capped trigonal prism d3sp3 dx2
y2, dxy dz2, dxz, dyz state. Intermediate situations are harder
CU-ML8 cube d3sp3 dx2
y2, dxy dz2, dxz, dyz to predict and are best exemplified with
DD-ML8 dodecahedron d4sp3 dx2
y2 dz2, dxy, dxz, dyz spin-crossover complexes, which can
SAPR-ML8 square antiprism d4sp3 dz2 dxy, dxz, dyz, dx2
y2 change their spin state through changes
[a] IUPAC polyhedral symbol. [b] The metal atomic orbitals identified as “nonbonding” may be in temperature or pressure or by irradi-
strictly so, as imposed by symmetry (e.g., the t2g set in octahedral complexes), or only formally ation.[25] To get a feel for how much the
nonbonding as discussed in the text. pairing energy can change from one
metal to another and with oxidation
state, one can take a look at the values of
scheme.[23] As an example, the interaction between the metal center the Racah electron-repulsion parame-
dsp2 hybridization expected for a and the donor atoms, and is controlled ters for free ions.[26] The following trends
square-planar complex is translated into by the overlap and difference in electro- are observed: 1) for metal ions with the
a molecular-orbital scheme with only negativity of both s- and p-type metal– same oxidation state, the pairing energy
one antibonding d orbital, whereas the ligand interactions. Its magnitude ob- increases with atomic number along a
d4sp3 hybridization expected for a do- tained from visible spectra, for example, transition series; 2) for metal ions be-
decahedral, eight-coordinated metal is empirically expressed through the longing to the same group, the pairing
corresponds to a splitting into one non- spectrochemical series of the ligands energy decreases down the group; and
bonding and four antibonding d-MOs. and of the metal ions.[24] On the basis 3) for the same metal atom, the pairing
of such information, cyanide, phos- energy increases with increasing oxida-
phines, or CO are considered strong tion state. In Figure 7, the factors that
High-Spin and Low-Spin ligands, while O-donor ligands (includ- influence the magnitudes of these two
Configurations ing water) and halides are referred to as parameters are summarized according
weak ligands. (The nature of the metal to the magnitudes of the pairing energy
The size of the d-energy gap is also has an effect on the magnitude of and the d-orbital energy gap. A qualita-
determined by the strength of the orbital the gap; it increases with the oxidation tive map is also shown that indicates the

Angew. Chem. Int. Ed. 2006, 45, 3012 – 3020  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 3015
Essays
these cases is exemplified by [Fe-
(H2O)6]3+, [Fe(CN)6]3
, and
2

[Mn(CN)4] . For the two octahedral
complexes, the S = 5/2 ([Fe(H2O)6]3+)
and S = 1/2 ([Fe(CN)6]3
) states ob-
served correspond to the two allowed
situations at the extremes of weak-field
(H2O) and strong-field (CN
) ligands.
The S = 5/2 spin state observed for
Figure 7. Factors that influence the magnitudes of the pairing energy and the d-orbital energy
gap. [Mn(CN)4]2
, even with a strong ligand,
corroborates the negligible effect of the
splitting within nonbonding orbitals on
regions of high- and low-spin states, the highest spin (S = 5/2) is attained if the electron configuration. Thus, for the
separated by a region in which spin the d-orbital splitting (energy gap) is case of a d5 metal ion, it is sensible to
crossover can occur. With so many small compared to the pairing energy, refer to the S = 5/2 and S = 1/2 states of
factors in action, it is not easy to predict and the lowest spin (S = 1/2) is obtained octahedral complexes as the high-spin
the most-stable spin configuration for a in the opposite case. In contrast, the S = and low-spin configurations, respective-
given complex. However, we can try to 3/2 state is forbidden because the pairing ly, while for isoelectronic tetrahedral
predict at least which spin states can be of the fourth and fifth electrons in compounds there is no need to refer to a
reasonably expected in the limiting t2g orbitals is either favorable for both “high-spin” situation as there is only one
cases of strong and weak fields. (S = 1/2 preferred) or unfavorable for observed spin state. The intermediate
Taking into account the classifica- both (S = 5/2 preferred), but cannot be spin for that configuration is not ex-
tion of d orbitals in nonbonding and preferred for one electron and unfavor- pected to exist as a ground state, at least
antibonding sets, we can establish the able for the other one (as required for not for homoleptic complexes.
possible spin configurations for a given S = 3/2 to be the ground state). For the With a “map” in hand of the possible
coordination number and geometry (Ta- tetrahedron (T-ML4), both the S = 3/2 spin states for a variety of coordination
ble 2). As an example, let us consider and S = 1/2 spin states are forbidden geometries and different d-electron con-
the case of a d5 ion in tetrahedral and because the energy difference between figurations, and considering only the
octahedral geometries, for which three the e and t2 orbital sets is small regard- splitting of the d orbitals in two broad
spin states can be envisaged: S = 5/2, 3/2, less of the nature of the metal and sets of nonbonding and antibonding
or 1/2. For the octahedron (OC-ML6), ligands. The experimental realization of orbitals (Table 2), we can establish some
generalities as well as some specificities.
p a) Three spin states are possible only
Table 2: Allowed spin configurations ( ) for different coordination geometries and numbers of
d electrons, considering only an energy gap between formally nonbonding and s-antibonding for the d4, d5, and d6 electron config-
orbitals.[a] urations. Nevertheless, for a given
coordination geometry only two or
Configuration Spin T-ML4 SP-ML4 SPY-ML5 OC-ML6 PBPY-ML7 DD-ML8
p p p p p p even just one of those spin states are
2
d 1 allowed, if we disregard spin-pairing
p
0 – – – – –
within the nonbonding set, before it
p p p p p p is half-filled.
d3 3/2
p p b) For d2, d3, d7, and d8 configurations,
1/2 – – – –
only two spin states are possible, but
p p p p p p
d4 2 in many instances the spin state with
p p
1 – – – – the highest multiplicity only is al-
p
0 – – – – – lowed.
p p p p p p c) From Table 2, a general rule for
d5 5/2
p p p tetrahedral complexes (T-ML4) is
3/2 – – –
p p that only the highest spin state
1/2 – – – –
should be expected in all cases (see
p p p p p p
d6 2 the discussion in the next section).
p p p
1 – – – d) The highest spin state should be
p
0 – – – – –
expected also for linear two-coordi-
p p p p p p nate (L-ML2) and trigonal-planar
d7 3/2
p p p three-coordinate (TP-ML3) com-
1/2 – – –
plexes. In these cases, all the d mo-
p p p p p p
d8 1 lecular orbitals are formally non-
p p
0 – – – – bonding (Figure 6) and a negligible
[a] The forbidden cases correspond to spin configurations that imply electron pairing within the set d-orbital energy gap should be ex-
of (formally and strictly) nonbonding d orbitals. Polyhedral symbols as reported in Table 1. pected. Recently reported examples

3016 www.angewandte.org  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2006, 45, 3012 – 3020
 Angewandte
Chemie

of two-coordinate complexes with d4

to d8 electron configurations all pre-
sented high-spin states.[27] Similarly,
three-coordinate complexes seem in
most cases to appear in their high-
spin configurations.[28, 29] (Exceptions
are discussed in the next section.)
e) For eight-coordinate complexes with
four or more d electrons (DD-ML8
or SAPR-ML8), the state with the
lowest spin multiplicity is always Figure 8. Tridentate ligands that appear, supplemented with a monodentate p-donor ligand, in
forbidden because it would imply pseudotetrahedral FeII and CoII complexes with low-spin configurations.
unnecessary electron-pairing within
the antibonding orbitals (Figure 6).
f) Given the different d-orbital split- ure 8). These compounds adopt a low- An explanation for the stability of
ting patterns, the allowed spin states spin configuration when the fourth li- the low-spin state in these systems was
change with the coordination num- gand, X, is a p donor. A somewhat proposed on the basis of DFT calcula-
ber. Thus, for the d6 configuration different case is that of diamagnetic tions by Peters and co-workers.[32] The
the state with all spins paired is tetrahedral NiIV compound, [NiBr(nor- distortion of the coordination tetrahe-
forbidden in square-planar (SP- bornyl)3],[35] in which all ligands are dron results in the splitting of the t2 set
ML4), square-pyramidal (SPY- monodentate. In all these compounds of d orbitals (Figure 9), and the a1
ML5), and pentagonal-bipyramidal there are varying degrees of umbrella- (dz2) orbital thereby loses its antibond-
(PBPY-ML7) complexes, while it is type distortions of the tetrahedron. This ing character as the E donor atoms
allowed for octahedral geometries. distortion is in contrast to the more approach its nodal cone (at S = 2708). In
On the contrary, the S = 1 state is commonly found spread distortion addition, the p-donor nature of X adds
allowed in the former cases but path[36] that ultimately leads to a p-antibonding character to the 2e (dxz,
forbidden in the latter. A similar square-planar complex. Such distortions dyz) degenerate pair. Such a p interac-
situation arises for the can be calibrated by the sum of the E-M- tion is enhanced as a result of the
d5 configuration. E bond angles (S), which is expected to balloon effect discussed before (Fig-
g) A spin-paired configuration is for- be 328.48 for the ideal tetrahedron but is ure 4 c). Note also that such a hybrid-
bidden for octahedral complexes of found to be 300.28 for the NiIV complex ization increases with the umbrella dis-
d2 or d3 metals. However, the high– and between 2648 and 2988 in the tortion, as the overlap between the
low spin state dichotomy exists for compounds with tridentate ligands. ligand lone pairs and metal dxz or
seven- and eight-coordinated metal
ions with these electron configura-
tions, and d2-configured octacyano-
metallates are indeed low-spin com-
plexes.

Electron-Pairing within the

Nonbonding Orbital Set

So far, we have relied on the as-

sumption that differences in energy
within the set of nonbonding d orbitals
are not large enough as to induce spin
pairing. However, we have neglected
metal–ligand p interactions in our argu-
ments. Let us now consider some exam-
ples in which p bonding may invalidate
such a hypothesis. The exceptions cor-
respond mostly to pseudotetrahedral
FeII and CoII complexes of the type
[M(BE3)X], where BE3 represents tri-
dentate ligands such as trispyrazolyl-
borates (E = N),[30] triphosphineborates Figure 9. Effect of an umbrella distortion and of a p-donor apical ligand X on the d-orbital
(E = P),[31, 32] trithioborates (E = S),[33] or splitting. The loss of s overlap in the a1 orbital and the enhanced p overlap in the 2e orbital
tris(carbene)borates[34] (see 1–3, Fig- explain the deviation of the orbital splitting from the e + t2 pattern of the tetrahedron.

Angew. Chem. Int. Ed. 2006, 45, 3012 – 3020  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 3017
Essays
dyz orbitals increases. In summary, the effectively antibonding. Another inter-
umbrella distortion stabilizes the a1 esting recent case is that of a four-
(dz2) orbital, while the presence of a p- coordinate d4-RuIV nitrido complex with
donor ligand X combined with the a tridentate ligand that adopts an ap-
umbrella distortion strongly destabilizes proximately square-planar geometry
the 2e set. The resultant wide energy (SP-ML4) but a low-spin state (S =
gap stabilizes the low-spin configura- 0).[38] Three factors that favor the un-
tion. usual low-spin state can be identified in
Consistent with the orbital picture that compound: a second-series transi-
just discussed, all the d6 and tion metal, a high oxidation state, and
Figure 10. Destabilization of the e’ and e’’ sets
d7 complexes of ligands 1–3 with a s- asymmetric p bonding.
by strong p-donor ligands in d2 tricoordinate
donor or p-acceptor ligand X appear in Three-coordinate complexes also imido complexes.
their high-spin configurations, as the present exceptions that require closer
2e orbitals lack the p-antibonding char- inspection. According to the discussion
acter that leads to a significant d-orbital above, all the d orbitals should be non- Some Final Considerations
energy gap. On the other hand, those bonding in such a case (TP-ML3 in
complexes in which X is a p-donor Table 1 and Figure 6). In fact, most The general tendency to label spin
ligand present a good correlation be- three-coordinate complexes for which states as either high or low spin reflects
tween the degree of umbrella distortion magnetic moments have been reported the duality of the choice of spin state
and the spin of the ground state: Com- present their maximum spin multiplici- that is rooted in the general rules that
pounds with strong distortions have a ty.[28] However, there are a few with a govern the energy patterns for the d-
low-spin ground state, whereas those spin-paired configuration, such as [Rh- MOs in transition-metal complexes.
with weaker distortions appear in their (PPh3)3]+,[39] [Ni(mesityl)3]
,[40] [M- These orbitals can, in general, be divid-
high-spin configuration. The borderline (mesityl)3] (M = Rh, Ir),[41] or d2 imido ed into two subsets with nonbonding
appears at S  2778. complexes.[42] The coordination geome- and antibonding character relative to
If we turn now to the pseudotetra- try in the [Rh(PPh3)3]+ and [Ni- the metal–ligand bonds. These two sub-
hedral complex [NiBr(norbornyl)3],[35] (mesityl)3]
complexes should be de- sets are separated by a significant energy
we note that it has two of the requisites scribed as T-shaped rather than trigonal gap, the magnitude of which, together
to present a low-spin state: an umbrella planar, and therefore the orbital scheme with the electron-pairing energy, ulti-
distortion and a p-donor ligand. How- that applies[28] is similar to that found for mately determines the relative energies
ever, as the umbrella distortion is not square-planar complexes (SP-ML4 in of the allowed spin states. Although
strong (S = 3008) it does not comply Figure 6). A singlet state should be some of the formally nonbonding MOs
with the third requisite. Current studies expected for d8 ions. In the case of the have antibonding contributions, these
in our research group indicate that even Rh and Ir mesityl complexes the coor- are minimized by hybridization with
with four s-donor ligands (e.g., alkyl dination sphere is far from planar, and metal s and p orbitals of the same
groups), a tetrahedrally coordinated agostic interactions with methyl groups symmetry, resulting in only a small
metal in its + iv oxidation state presents of the ligands have been identified. increase in energy relative to the strictly
a large enough gap between the t2 and These account for an effective octahe- nonbonding orbitals. As a consequence,
e orbitals as to favor the low-spin state. dral ligand field for which an S = 0 electron pairing within the nonbonding
This behavior has been experimentally situation is conceivable (OC-ML6 in subset is forbidden before all its orbitals
found for the [Co(norbornyl)4]y com- Table 2). Such a situation has been are half-occupied. This simplified orbi-
plexes (y =
1, 0, + 1), which appear in shown through computational studies tal scheme, which is strongly dependent
their low-spin states.[37] for a CrII alkoxy complex, in which the on the coordination number and geom-
We have already noted above that dx2
y2 orbital is strongly destabilized by etry and disregards p interactions in a
the d-orbital splitting depends not only p interactions, even if in this case the first approximation, allows us to qual-
on the donor ability of the ligands but d4 configuration precludes a low-spin itatively predict two or three allowed
also increases with the metal oxidation state.[43] As for the d2 imido complexes, spin states for a variety of electron
state. Hence, the different behavior of again the strong p-donor character of configurations and coordination geome-
the tetrahedral [Mn(CN)4]2
and the ligands in an anisotropic arrange- tries (Table 2). Exceptions to these rules
[NiBr(norbornyl)3] complexes, which ment, combined with high oxidation arise from the promotion of one or more
present high- and low-spin configura- states in those WIV, ReV, and OsVI of the formally nonbonding orbitals to
tions, respectively, are well accounted complexes, should result in a strong the antibonding set and are caused by
for by the general rules of d-orbital destabilization of both the e’ and e’’ sets one or more of three possible causes:
splitting if we assume that for metal of orbitals (Figure 10). As discussed 1) Severe distortions from the ideal
oxidation states of + iv or higher the above for the case of pseudotetrahedral geometries, 2) strongly anisotropic in-
metal–ligand interaction is so strong complexes with one p-donor ligand, the teraction with p-donor ligands, and 3) a
that the p+d hybridization is insufficient hybridization required to alleviate the s- high oxidation state of the metal atom in
to make the t2 orbitals formally non- antibonding nature of some d orbitals combination with strong s-donor li-
bonding. In other words, they become enhances the p-antibonding effect. gands.

3018 www.angewandte.org  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2006, 45, 3012 – 3020
 Angewandte
Chemie

Eventual references to intermedi- only two possibilities (d2, d3, d7, and d8), GarcOa-Monforte, A. MartOn, B. Menj n,
ate-spin states are not in contradiction and to high-, intermediate-, and low- G. RodrOguez, Chem. Commun. 1997,
with the dual nature of the spin state but spin states for the d4, d5, and d6 cases, 1721; P. J. Alonso, J. Forni!s, M. A.
GarcOa-Monforte, A. MartOn, B. Menj n,
lies with the fact that three spin states regardless of whether such states in
Chem. Commun. 2001, 2138; P. J. Alon-
are conceivable for some dn configura- different geometries are allowed. There so, J. Forni!s, M. A. GarcOa-Monforte,
tions. Then, it is the specific geometrical are some precedents for referring to A. MartOn, B. Menj n, C. Rillo, Chem.
and bonding situation of each com- square-pyramidal[15] or square-planar[16] Eur. J. 2002, 8, 4056; P. J. Alonso, J.
pound that determines which of these complexes as having intermediate spin, Forni!s, M. A. GarcOa-Monforte, A.
are the two allowed spin states. Refer- but it would certainly change the way we MartOn, B. Menj n, Chem. Commun.
ences to intermediate-spin states are usually name spin states in, for example, 2002, 728; P. J. Alonso, L. R. Falvello, J.
Forni!s, M. A. GarcOa-Monforte, B.
found mostly in metal porphyrin com- octahedral complexes. For instance, all
Menj n, Angew. Chem. 2004, 116,
plexes. If two additional ligands are known octahedral “low-spin” d4-metal 5337; Angew. Chem. Int. Ed. 2004, 43,
coordinated to the metal atom besides complexes should be renamed as “inter- 5225; I. Ara, J. Forni!s, M. A. GarcOa-
the porphyrin ligand, only the highest mediate spin”. In return, such nomen- Monforte, A. MartOn, B. Menj n, Chem.
and lowest spin states should be expect- clature should allow for a clearer corre- Eur. J. 2004, 10, 4186; P. J. Alonso, J.
ed for a strictly octahedral coordination lation between complexes that are re- Forni!s, M. A. GarcOa-Monforte, A.
sphere (OC-ML6 in Table 2). However, lated by association or dissociation of a MartOn, B. Menj n, Organometallics
2005, 24, 1269; P. J. Alonso, J. Forni!s,
when the two axial ligands are much ligand, and the name of the spin state
M. A. GarcOa-Monforte, A. MartOn, B.
weaker donors compared to the porphy- would be unequivocally associated to Menj n, Chem. Eur. J. 2005, 11, 4713; J.
rin ring (e.g., tetrahydrofuran, ethanol, the S value. Forni!s, A. MartOn, L. F. MartOn, B.
or water),[14] the d-orbital splitting pat- Experience shows that changes in Menj n, Organometallics 2005, 24, 3266.
tern is closer to that of the square-planar nomenclature are rarely smoothly and [2] S. Kleinhenz, V. Pfennig, K. Seppelt,
complexes, for which the allowed spin universally adopted, and it may well be Chem. Eur. J. 1998, 4, 1687; S. Klein-
states are different for d5 and d6 electron that the introduction of a new naming henz, K. Seppelt, Chem. Eur. J. 1999, 5,
configurations (SP-ML4 in Table 2). A convention for the spin states finally 3573; B. Toessler, S. Kleinhenz, K.
Seppelt, Chem. Commun. 2000, 1039.
similar situation is found in LaCoO3, for brings about more confusion than it
[3] K. Dunbar, R. A. Henitz, Prog. Inorg.
which the intermediate-spin state is avoids. In any case, we hope that the Chem. 1997, 45, 283.
stabilized by a tetragonal, Jahn–Teller- reflections presented in this essay may [4] W. E. Buschmann, A. M. Arif, J. S. Mill-
driven distortion of the coordination help clarify the concepts that describe er, Angew. Chem. 1998, 110, 813; Angew.
octahedron.[12] spin states and that more data and new Chem. Int. Ed. 1998, 37, 781.
It is clear that the traditional way of ideas can be proposed in the near future [5] K. J. Nelson, I. D. Giles, W. W. Shum,
naming spin states for octahedral and by other researchers along these lines. A. M. Arif, J. S. Miller, Angew. Chem.
2005, 117, 3189; Angew. Chem. Int. Ed.
tetragonal complexes, referring to the Whatever convention is adopted in each
2005, 44, 3129.
two allowed spin states as high and low, particular case, to avoid ambiguities and [6] W. R. Entley, C. R. Treadway, S. R. Wil-
leads to inconsistencies when dealing misinterpretations the most important son, G. S. Girolami, J. Am. Chem. Soc.
with these systems, because one does point is to be aware of the strong 1997, 119, 6251.
not know beforehand whether a partic- interdependence of spin states and co- [7] M. Verdaguer, G. S. Girolami in Mag-
ular compound should be considered as ordination geometry and to be sure to netism: Molecules to Materials, Vol. V
octahedral or tetragonal from the orbi- have at hand a clear picture of the (Eds.: J. S. Miller, M. Drillon), Wiley-
tal point of view. Other inconsistencies splitting pattern of the d molecular or- VCH, Weinheim, 2005, p. 283; L. M. C.
Beltran, J. R. Long, Acc. Chem. Res.
result from the traditional labeling of bitals and of the factors that affect the
2005, 38, 325.
spin states. Consider, for example, the choice of spin state. No matter how high [8] E. Bernhardt, C. Bach, B. Bley, R.
lowest allowed spin states of d5-ML5 and or low one considers the spin of a given Wartchow, U. Westphal, I. H. T. Sham,
d5-MXL5 complexes, which are related compound, probably the wisest choice is B. von Ahsen, C. Wang, H. Willner,
by ligand association/dissociation. Even to specify in all cases the total spin or the R. C. Thompson, F. Aubke, Inorg.
if we refer to a low-spin configuration in spin multiplicity. Chem. 2005, 44, 4189; M. Finke, E.
both cases, we are actually talking about Bernhardt, H. Willner, C. W. Lehmann,
Published online: April 10, 2006 F. Aubke, Inorg. Chem. 2005, 44, 4206;
different spin states, S = 3/2 and S = 1/2,
H. Willner, F. Aubke, Organometallics
respectively (see SPY-ML5 and OC-ML6 2003, 22, 3612; J. E. Ellis, Organometal-
in Table 2). lics 2003, 22, 3322.
To avoid the above inconsistencies [1] R. Us n, J. Forni!s, M. TomLs, B. [9] M. V. Metz, Y. Sun, C. L. Stern, T. L.
in the naming of spin states, maybe some Menj n, K. SNnkel, R. Bau, J. Chem. Marks, Organometallics 2002, 21, 3691;
convention could be established that Soc. Chem. Commun. 1984, 751; J. For- T. C. Rosen, K. Kirschbaum, D. M. Gio-
considers all the situations discussed ni!s, B. Menj n, R. M. Sanz-Carrillo, M. lando, Dalton Trans. 2003, 120; M. C.
TomLs, N. G. Connelly, J. G. Crossley, Buzzeo, A. H. Iqbal, C. M. Long, D.
here as well as potential new develop-
A. G. Orpen, J. Am. Chem. Soc. 1995, Millar, A. Patel, M. A. Pellow, S. A.
ments. Because in all cases there are, at Saddoughi, A. L. Smenton, J. F. C. Turn-
117, 4295; P. J. Alonso, L. R. Falvello, J.
most, three possible spin states, one Forni!s, A. MartOn, B. Menj n, G. Ro- er, J. D. Wadhawan, R. G. Compton,
could refer to high- and low-spin states drOguez, Chem. Commun. 1997, 503; P. J. J. A. Golen, A. L. Rheingold, L. H.
for those electron configurations with Alonso, L. R. Falvello, J. Forni!s, M. A. Doerrer, Inorg. Chem. 2004, 43, 7709.

Angew. Chem. Int. Ed. 2006, 45, 3012 – 3020  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 3019
Essays
[10] N. T. Tran, D. R. Powell, L. F. Dahl, Cheeseman, J. A. Montgomery, Jr., T. Hikichi, Y. Moro-oka, Chem. Commun.
Angew. Chem. Int. Ed. 2000, 39, 4121; Vreven, K. N. Kudin, J. C. Burant, J. M. 1999, 417; N. Shirasawa, T. T. Nguyet, S.
A. MNller, P. KQgerler, C. Kuhlmann, Millam, S. S. Iyengar, J. Tomasi, V. Hikichi, Y. Moro-oka, M. Akita, Orga-
Chem. Commun. 1999, 1347; S. Dehnen, Barone, B. Mennucci, M. Cossi, G. nometallics 2001, 20, 3582.
A. Eichhofer, D. Fenske, Eur. J. Inorg. Scalmani, N. Rega, G. A. Petersson, H. [31] T. A. Betley, J. C. Peters, J. Am. Chem.
Chem. 2002, 279. Nakatsuji, M. Hada, M. Ehara, K. Toyo- Soc. 2003, 125, 10 782; T. A. Betley, J. C.
[11] S. A. Stoian, Y. Yu, J. M. Smith, P. L. ta, R. Fukuda, J. Hasegawa, M. Ishida, T. Peters, J. Am. Chem. Soc. 2004, 126,
Holland, E. L. Bominaar, E. MNnck, Nakajima, Y. Honda, O. Kitao, H. 6252; S. D. Brown, T. A. Betley, J. C.
Inorg. Chem. 2005, 44, 4915. Nakai, M. Klene, X. Li, J. E. Knox, Peters, J. Am. Chem. Soc. 2003, 125, 322;
[12] K. KnOzek, P. NovLk, Z. JirLk, Phys. Rev. H. P. Hratchian, J. B. Cross, C. Adamo, J. S. D. Brown, J. C. Peters, J. Am. Chem.
B 2005, 71, 054 420; J. Wang, Z. D. Wang, Jaramillo, R. Gomperts, R. E. Strat- Soc. 2004, 126, 4538.
W. Zhang, D. Y. Xing, Phys. Rev. B 2002, mann, O. Yazyev, A. J. Austin, R. Cam- [32] S. D. Brown, J. C. Peters, J. Am. Chem.
66, 064 406. mi, C. Pomelli, J. W. Ochterski, P. Y. Soc. 2005, 127, 1913; D. M. Jenkins, J. C.
[13] S. Obara, H. Kashiwagi, J. Chem. Phys. Ayala, K. Morokuma, G. A. Voth, P. Peters, J. Am. Chem. Soc. 2005, 127,
1982, 77, 3155; T. Ikeue, T. Saitoh, T. Salvador, J. J. Dannenberg, V. G. Zakr- 7148.
Yamaguchi, Y. Ohgo, M. Nakamura, M. zewski, S. Dapprich, A. D. Daniels, [33] P. J. Schebler, B. S. Mandimutsira, C. G.
Takahashi, M. Takeda, Chem. Commun. M. C. Strain, O. Farkas, D. K. Malick, Riordan, L. M. Liable-Sands, C. D. In-
2000, 1989; H. Goff, G. N. La Mar, C. A. A. D. Rabuck, K. Raghavachari, J. B. carvito, A. L. Rheingold, J. Am. Chem.
Reed, J. Am. Chem. Soc. 1977, 99, 3641. Foresman, J. V. Ortiz, Q. Cui, A. G. Soc. 2001, 123, 331; P. J. Schebler, C. G.
[14] J.-P. Simonato, J. P!caut, L. Le Pape, J.- Baboul, S. Clifford, J. Cioslowski, B. B. Riordan, I. A. Guzei, A. L. Rheingold,
L. Oddou, C. Jeandey, M. Shang, W. R. Stefanov, G. Liu, A. Liashenko, P. Pis- Inorg. Chem. 1998, 37, 4754.
Scheidt, J. Wojaczynski, S. Wolowiec, L. korz, I. Komaromi, R. L. Martin, D. J. [34] I. Nieto, F. Cervantes-Lee, J. M. Smith,
Latos-Grazynski, J.-C. Marchon, Inorg. Fox, T. Keith, M. A. Al-Laham, C. Y. Chem. Commun. 2005, 3811.
Chem. 2000, 39, 3978; T. Sakai, Y. Ohgo, Peng, A. Nanayakkara, M. Challa- [35] V. Dimitrov, A. Linden, Angew. Chem.
T. Ikeue, M. Takahashi, M. Takeda, M. combe, P. M. W. Gill, B. Johnson, W. 2003, 115, 2735; Angew. Chem. Int. Ed.
Nakamura, J. Am. Chem. Soc. 2003, 125, Chen, M. W. Wong, C. Gonzalez, J. A.
2003, 42, 2631.
13 028. Pople, Gaussian, Inc., Pittsburgh, PA,
[36] J. Cirera, P. Alemany, S. Alvarez, Chem.
[15] K. M. Barkigia, M. W. Renner, J. Fajer, 2003). A test calculation for the high-
Eur. J. 2004, 10, 190.
J. Porphyrins Phthalocyanines 2001, 5, spin state of [Cr(CN)6]4
indicates that
[37] E. K. Byrne, K. H. Theopold, J. Am.
415. the d-orbital energy gap is only slightly
Chem. Soc. 1989, 111, 3887.
[16] K. Ray, A. Begum, T. WeyhermNller, S. reduced and the semiquantitative pic-
[38] A. Walstrom, M. Pink, X. Yang, J.
Piligkos, J. van Slageren, F. Neese, K. ture of Figure 6 is still valid.
Tomaszewski, M.-H. Baik, K. G. Caul-
Wieghardt, J. Am. Chem. Soc. 2005, 127, [23] J. E. Huheey, E. A. Keiter, R. L. Keiter,
ton, J. Am. Chem. Soc. 2005, 127, 5330.
4403. Inorganic Chemistry: Principles of
[39] Y. W. Yared, S. L. Miles, R. Bau, C. A.
[17] T. A. Albright, J. K. Burdett, W.-H. Structure and Reactivity, 4th ed., Harper
Reed, J. Am. Chem. Soc. 1977, 99, 7076;
Whangbo, Orbital Interactions in Collins, New York, 1993.
C. B. Knobler, T. B. Marder, E. A. Miz-
Chemistry, Wiley, New York, 1985. [24] M. Gerloch, E. C. Constable, Transition
[18] R. Poli, Chem. Rev. 1996, 96, 2135; R. Metal Chemistry, VCH, Weinheim, 1994. usawa, R. G. Teller, J. A. Long, P. E.
Poli, J. N. Harvey, Chem. Soc. Rev. 2003, [25] J. A. Real, A. B. Gaspar, M. C. MuToz, Beehnken, M. F. Hawthorne, J. Am.
32, 1; J. N. Harvey, R. Poli, K. M. Smith, Dalton Trans. 2005, 2062. Chem. Soc. 1984, 106, 2990.
Coord. Chem. Rev. 2003, 238–239, 347; [26] J. S. Griffith, Theory of Transition Metal [40] R. Hay-Motherwell, G. Wilkinson, T. K.
R. Poli, J. Organomet. Chem. 2004, 689, Ions, Cambridge University Press, Cam- Sweet, M. B. Hursthouse, Polyhedron
4291. bridge, 1961. 1996, 15, 3163.
[19] Nomenclature of Inorganic Chemistry. [27] T. Nguyen, A. Panda, M. M. Olmstead, [41] D. M. Roddick, T. D. Tilley, A. L.
IUPAC Recommendations 2005 (Eds.: A. F. Richards, M. Stender, M. Brynda, Rheingold, S. J. Geib, M. B. Hursthouse,
N. G. Connelly, T. DAmhus, R. M. P. P. Power, J. Am. Chem. Soc. 2005, 127, G. Wilkinson, J. Chem. Soc. Chem.
Hartshorn, A. T. Hutton), Royal Society 8545; W. M. Reiff, A. LaPointe, E. H. Commun. 1990, 1242; R. S. Hay-Moth-
of Chemistry, 2005. Witten, J. Am. Chem. Soc. 2004, 126, erwell, G. Wilkinson, B. Hussain-Bates,
[20] F. A. Cotton, Chemical Applications of 10 206. M. B. Hursthouse, J. Chem. Soc. Dalton
Group Theory, 3rd ed., Wiley-Inter- [28] S. Alvarez, Coord. Chem. Rev. 1999, Trans. 1992, 3477.
science, New York, 1990. 193–195, 13. [42] D. F. Eppley, P. T. Wolczanski, G. D.
[21] “And here I wish I could tell you half the [29] J. Vela, J. M. Smith, Y. Yu, N. A. Ket- Van Duyne, Angew. Chem. 1991, 103,
things Alice used to say, beginning with terer, C. J. Flaschenriem, R. J. Lachi- 616; Angew. Chem. Int. Ed. Engl. 1991,
her favourite phrase ;Let;s pretend;.” cotte, P. L. Holland, J. Am. Chem. Soc. 30, 584; J. T. Anhaus, T. P. Kee, M. H.
Lewis Carroll, Through The Looking 2005, 127, 7857. Schofield, R. R. Schrock, J. Am. Chem.
Glass (1871). [30] J. L. Detrich, R. Konecny, W. M. Vetter, Soc. 1990, 112, 1642; D. S. Williams,
[22] The orbital patterns shown in Figure 6 D. Doren, A. L. Rheingold, K. H. Theo- R. R. Schrock, Organometallics 1993,
were obtained from DFT-B3LYP calcu- pold, J. Am. Chem. Soc. 1996, 118, 1703; 12, 1148; D. S. Williams, J. T. Anhaus,
lations on the low-spin states of the J. D. Jewson, L. M. Liable-Sands, M. H. Schofield, R. R. Schrock, W. M.
compounds indicated at their experi- G. P. A. Yap, A. L. Rheingold, K. H. Davis, J. Am. Chem. Soc. 1991, 113,
mental structures by using the Gaussian Theopold, Organometallics 1999, 18, 5480.
package (Gaussian 03 (Revision B.4), 300; O. M. Reinaud, A. L. Rheingold, [43] O. L. Sydora, P. T. Wolczanski, E. B.
M. J. Frisch, G. W. Trucks, H. B. Schle- K. H. Theopold, J. Am. Chem. Soc. 1994, Lobkovsky, C. Buda, T. R. Cundari,
gel, G. E. Scuseria, M. A. Robb, J. R. 116, 1703; N. Shirasawa, M. Akita, S. Inorg. Chem. 2005, 44, 2606.

3020 www.angewandte.org  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2006, 45, 3012 – 3020