Structural, 103Rh NMR and DFT Studies of a Bis(phosphane)Rh III-Porphyrin Derivative

Article abstract of DOI:10.1002/ejic.200800837

The synthesis and characterisation of a novel bis(phosphane):rhod.ium.(III) -porphyrm, [Rh(TPP)(PEtPh₂J₂]SbF₆ (compound. 1; TPP = dianion of 5,10,15,20-tetraphenylporphyrin) is described. The X-ray structure of the bis(dic

Full text of DOI:10.1002/ejic.200800837

FULL PAPER
DOI: 10.1002/ejic.200800837
Structural, 103Rh NMR and DFT Studies of a Bis(phosphane)RhIII–Porphyrin
Derivative
Orde Q. Munro,*[a] Greville L. Camp,[a] and Laurence Carlton[b]
Keywords: 103Rh NMR spectroscopy / Porphyrin / Density functional calculations / GIAO / Phosphanes
The synthesis and characterisation of a novel bis(phosphane)-
rhodium(III)–porphyrin, [Rh(TPP)(PEtPh2)2]SbF6 (compound
1; TPP = dianion of 5,10,15,20-tetraphenylporphyrin) is de-
scribed. The X-ray structure of the bis(dichloromethane)
solvate of the complex has a moderately ruffled porphyrin
mations are possible for 1 and were modelled by using DFT
simulations at the PBE1PBE/3-21G** level of theory with the
cation [Rh(TPP)(PEtPh2)2]+. Accurate structural parameters
(calculated bond lengths and out-of-plane porphyrin core
atom displacements within 2% and 0.02 Å of the experimen-
core conformation, which is primarily determined by the rela- tal values, respectively) and moderately accurate 103Rh iso-
calcd.
exp.
tive orientations of the axial phosphane ligands and their ste-
ric interaction with the porphyrin phenyl rings. The mean
tropic shielding tensors (δRh
within 13% of δRh
at 0 K
using the GIAO method) were calculated with this hybrid
functional and relatively small all-electron basis set. The DFT
simulations indicate unusually high fractional electron pop-
ulations for the formally antibonding 4dz2 and 4dx2–y2 orbitals
of both the ruffled and planar conformers of [Rh(TPP)-
(PEtPh2)2]+, consistent with the relatively low energies of
these metal-character orbitals that are evidently well-mixed
with ligand orbitals.
Rh–N and Rh–P bond lengths for
1 are 2.036(5) and
2.401(3) Å, respectively. The mean absolute perpendicular
displacements of the porphyrin α-, β- and meso-carbon atoms
from the 24-atom porphyrin mean plane measure 0.15(3),
0.11(8) and 0.27(2) Å, respectively. The 103Rh NMR chemical
shifts of 1, determined by means of indirect detection
through polarisation transfer from 31P, were δ = 2480, 2558
and 2590 ppm at 213, 300 and 333 K, respectively. The 31P
chemical shifts measured δ = 10.68, 10.79 and 10.84 ppm at
the same temperatures. Ruffled and planar porphyrin confor-
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
tions with P(OMe)2Ph and P(OEt)2Ph as axial ligands,[8]
whereas the bis(phosphane) derivative [Fe(TPP)(PMe2Ph)2]-
ClO4 exhibits a planar porphyrin core conformation.[9] Of
the 4d and 5d elements within groups 7–9, a 17-electron
ReII derivative has been structurally characterised and only
a small number of 18-electron OsII[10] and IrIII[11] complexes
such as [Os(TPP)(PPh3)2][10a] have been reported. The anal-
ogous 18-electron RhIII complexes have, similarly, not been
extensively characterised, and only five X-ray structure de-
terminations have been reported,[12] the most recent study
illustrating the elegant use of RhIII–porphyrins as building
blocks for the construction of supramolecular arrays.[13]
Eighteen-electron RhIII–porphyrins undergo ligand ex-
change reactions and anions such as iodide (but not meth-
ylide or chloride) may be substituted by phosphanes and
related P-donor ligands.[13a] (Hydrido)RhIII derivatives are
reactive toward RCHO insertions,[14] simple (chlorido)-
RhIII–porphyrins are good photocatalysts for alcohol dehy-
drogenation reactions,[15] whereas other RhIII–porphyrins
efficiently catalyse the decomposition of ethyl diazoacetate
and the transfer of (ethoxycarbonyl)carbene to generate cy-
clopropane derivatives.[16] The electrochemical behaviour of
[Rh(TPP)(PR3)2]PF6 salts, where R is an alkyl or aryl
group, has been studied and is characterised by the forma-
tion of reactive RhIII π anion radicals or the RhII dimer
Introduction
From a structural standpoint, group 7–9 metalloporphyr-
ins with P-donor axial ligands have not been as extensively
characterised as their N-donor axial ligand counterparts.
The majority of the reported structures are six-coordinate
18-electron RuII complexes such as [Ru(OEP)(PPh3)2],[1,2]
[Ru(OEP)Br(PPh3)],[3] [Ru(TF5PP)(PPh3)2][4] and other
porphyrins with elaborate axial phosphane ligands [e.g. di-
phenyl(2-phenylethynyl)phosphane].[5] Structures of the iso-
electronic low-spin FeII complexes typified by
[Fe(OEP)(PMe3)2] and [Fe(TPP)(PBu3)2][6] are also known,
not only with simple phosphanes but also with phosphites
such as P(OMe)3.[7] These FeII complexes all have planar
porphyrin core conformations. The analogous bis(phos-
phonite)FeIII derivatives have ruffled porphyrin conforma-
[a] School of Chemistry, University of KwaZulu-Natal,
Private Bag X01, Scottsville, 3209, Pietermaritzburg, South
Africa
Fax: +27-33-260-5009
E-mail: MunroO@ukzn.ac.za
[b] School of Chemistry, University of the Witwatersrand,
P. O. Wits 2050, Johannesburg, South Africa
Fax: +27-11-717-6749
E-mail: Laurence.Carlton@wits.ac.za
Supporting information for this article is available on the
WWW under http://www.eurjic.org/ or from the author.
2512
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2009, 2512–2523

Studies of a Bis(phosphane)RhIII–Porphyrin
[(TPP)Rh]2.[17] Interestingly, sterically hindered phosphanes steric interactions between the coordinated phosphane li-
such as PPh3 bind fairly strongly to RhIII–porphyrins gands and the porphyrin ring. Other RhIII complexes with
{logK293K ≈ 3.1 for substitution of the solvent ligand in a trans-RhN4P2 coordination group and sterically hindered
[Rh(TPP)(PPh3)(THF)]+}.[17a]
phosphane ligands exhibit similar axial bond lengths to 1,
[25]
Despite the increasing importance of RhIII–porphyrins in for example [Rh(NCCH3)3(NO)(PPh3)2]SbF6
[Rh–P
catalysis and coordination chemistry, none have been char- 2.404(3) Å] and [Rh(TPP)(PPh2{CCC6H5})2]I·(CHCl3)
acterised by 103Rh NMR spectroscopy,[18] a situation that [Rh–P 2.371(2) Å].[26]
reflects the experimental difficulties associated with direct
detection of this I = ½ nucleus.[19,20] Most studies on simple
Rh complexes have therefore employed indirect detection[21]
of the 103Rh nucleus in double- or triple-resonance HMQC
experiments by employing polarisation transfer from 31P,
1H or 11B.[22] Given that the RhIII–porphyrins investigated
hitherto[13,23] have not been probed by 103Rh NMR spec-
troscopy, we describe here the first combined 103Rh NMR
and DFT study of a phosphane-ligated RhIII–porphyrin,
namely [Rh(TPP)(PEtPh2)2]SbF6 (1, Scheme 1). Compound
1 has, moreover, been characterised by single-crystal X-ray
diffraction as its bis(dichloromethane) solvate. The geome-
tries, electronic structures and NMR shielding tensors of
the planar and ruffled conformational isomers of
[Rh(TPP)(PEtPh2)2]+ have been determined by DFT and
DFT-GIAO methods to (i) establish their relative energies
and (ii) to gain a fundamental description of the electronic
structure of the RhIII ion in a porphyrin-based N4P2 ligand
field.
Figure 1. Selectively labelled thermal ellipsoid diagram[64] of the X-
ray crystal structure of 1·2(CH2Cl2). H atoms, solvent molecules
and the counterion have been omitted for clarity (30% ellipsoid
probabilities are shown). Key bond lengths [Å] and angles [°]: Rh–
N1 2.035(3), Rh–N2 2.033(3), Rh–N3 2.043(3), Rh–N4 2.031(3),
Rh–P1 2.3986(13), Rh–P2 2.4031(13), N1–Rh–P1 90.5(1), N1–Rh–
P2 91.5(1); P1–Rh–P2 177.96(4), C51–P1–Rh 112.1(2), C71–P1–
Rh 112.3(2), C61–P1–Rh 116.3(2), C81–P2–Rh 112.2(1), C100–P2–
Rh 112.1(2), C91–P2–Rh, 116.9(2).
The long Rh–P bond for 1 differs substantially from the
short [2.306(3) Å] distance reported for [Rh(OEP)Cl(PPh3)]·
2(CHCl3).[27] The shorter Rh–P distance of the latter com-
plex reflects the fact that the RhIII ion is displaced out of
the porphyrin mean plane towards the single PPh3 ligand,
which permits a closer interaction with PPh3, even though
PPh3 (Tolman’s[28] cone angle = 145°) is sterically more hin-
Scheme 1.
Results and Discussion
Synthesis and Crystallography
Metalloporphyrin
1
was synthesised by treating dered than the PEtPh2 ligands of 1 (cone angle = 140°).
[Rh(TPP)Cl] with AgSbF6 in dry THF to generate The average Rh–N distance for 1 is more typical for a
[Rh(TPP)(THF)2]SbF6 in situ prior to removal of the sol- metalloporphyrin at 2.036(5) Å and is experimentally equiv-
vent in vacuo and treatment of this labile intermediate with alent (within 3σ) to that seen in [Rh(OEP)Cl(PPh3)]
an excess of ethyldiphenylphosphane (PEtPh2) in dry [2.024(35) Å].[28] Although the porphyrin meso-phenyl
dichloromethane. The X-ray crystal structure of 1·2CH2Cl2 groups are canted relative to the 24-atom porphyrin mean
is shown in Figure 1. The porphyrin ligand exhibits a pre- plane (dihedral angles = 65.6, 64.1, 82.8 and 74.2° for the
dominantly ruffled core conformation with mean absolute Ph groups appended to meso-carbon atoms C1m through
perpendicular displacements of the Rh, nitrogen and por- C4m, respectively), they do not significantly saddle[25] the
phyrin α-, β- and meso-carbon atoms of 2(0), 2(2), 0.15(3), already ruffled conformation of the porphyrin ring. This
0.11(8) and 0.28(1) Å, respectively. The porphyrin is thus suggests that the conformation of 1 is probably largely gov-
moderately distorted from planarity (Figure 2).[24] The erned by axial ligand···porphyrin core and axial
modest ruffling (porphyrin meso-carbon atoms alternately ligand···porphyrin phenyl group nonbonding interactions,
displaced above and below the 24-atom mean plane) pre- an effect confirmed by the DFT-calculated structure of 1 in
sumably reflects the rather long mean axial Rh–P coordina- the presence and absence of the porphyrin phenyl groups
tion distance of 2.401(3) Å and thus somewhat diminished (vide infra).
Eur. J. Inorg. Chem. 2009, 2512–2523
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2513

O. Q. Munro, G. L. Camp, L. Carlton
FULL PAPER
K[Rh(TPP)(CN)2] as a result of complete substitution of
the PEtPh2 ligands. The electronic spectrum of
K[Rh(TPP)(CN)2] shows the same number of bands and
overall spectroscopic intensity pattern as 1, particularly in
the visible region, confirming that the spectroscopic fea-
tures for 1 are consistent with a six-coordinate RhIII–por-
phyrin. The main absorption maxima of K[Rh(TPP)(CN)2]
are, however, blueshifted relative to those of 1 (by up to
9 nm). One possible interpretation of the marked redshift
for the Q and B bands of 1 relative to those of the dicyanido
complex is that such a spectroscopic perturbation might
well correlate with the order of CN– and PEtPh2 in the
nephelauxetic series. Evidently, the softer axial ligand
(PEtPh2) favours a narrower HOMO–LUMO gap and con-
sequently lower Q and B state energies.
Figure 2. Formal diagram displaying the perpendicular displace-
ments in units of 0.01 Å (esd = 0.003 Å) of the porphyrin ring
atoms from the 24-atom mean plane of 1·2(CH2Cl2). Chemically
unique mean bond lengths [Å] and angles [°] as well as the relative
orientations of the axial phosphane ligands are shown (heavy and
dashed lines reflect the above- and below-plane axial ligands,
respectively). Angles shown without esd’s (for clarity) have an esd
of 0.2°.
The aryl groups of the coordinated phosphane ligands
adopt dihedral angles of 26.3(1), 35.4(1), 24.7(2) and
42.3(1)° relative to the porphyrin mean plane for the rings
containing carbon atoms C51, C61, C81 and C91, respec-
tively. For the phosphane phenyl rings C51–C56 and C81–
C86, these dihedral angles are rather close to being parallel
with the porphyrin ring. Indeed, there are several short
phenyl ring···pyrrole ring intramolecular nonbonding con-
tacts that point to the existence of stabilising intramolecular
π–π interactions in 1.[29] Interestingly, the axial phosphane
ligands exhibit an approximate anti configuration, in which
the ethyl substituents nearly eclipse the in-plane Rh–N
bonds [N4–Rh–P1–C71 4.9(2)° and N2–Rh–P2–C100
10.3(2)°]. We believe that this relative orientation for the
axial ligands in 1, coupled with their significant steric bulk,
might be responsible for the observed ruffled porphyrin
conformation, particularly since we previously found
[Co(TPP)(1-MepipZ)2]SbF6 (1-MepipZ = 1-methylpipera-
zine[1]) to be ruffled for similar reasons.[30]
Figure 3. Normalised electronic spectra of selected RhIII–porphy-
rins recorded in dry dichloromethane at 298 K. The inset shows an
expansion of the visible spectra of the three complexes.
The foregoing argument is strongly supported by the
spectroelectrochemical data of Kadish and co-workers[17a]
which provide an experimental probe of the HOMO–
LUMO gap from the one-electron redox reactions of related
RhIII–porphyrins. Specifically, [Rh(TPP)(PF3)(OH)] has a
Soret band maximum at 420 nm and ∆EHOMO–LUMO
2.29 V, whereas [Rh(TPP)(PMePh2)2]PF6 has λmax
=
=
=
446 nm and ∆EHOMO–LUMO = 2.22 V (∆EHOMO–LUMO
ox
E
– E red). Note that the one-electron first oxidation
½
½
and reduction reactions of these complexes involve the por-
phyrin ring and form π cation and π anion radicals, respec-
tively. Clearly, an axial ligand combination high up in the
nephelauxetic series correlates with a redshifted Soret band
and narrower HOMO–LUMO gap.
One problem encountered during spectroscopic analysis
of 1 was partial aquation of the complex if dry solvents
Electronic Spectroscopy
The electronic absorption spectrum of 1 recorded in dry were not employed. Indeed, addition of water-saturated
dichloromethane (Figure 3) over the full spectroscopic dichloromethane to a solution of 1 in dry dichloromethane
range is very similar to that reported by Kadish et al.[17a] afforded a small saturating fraction of [Rh(TPP)(OH2)-
for [Rh(TPP)(PPh3)2]PF6 (λmax = 448, 557, 597 nm) from (PEtPh2)]SbF6 with an increase in the water content of the
380–700 nm in the same solvent. The Soret (448 nm) and solution, as evidenced by the appearance of the B and Q
two sharply resolved Q bands (557, 597 nm) are markedly bands at ca. 420 and 527 nm, respectively, bands that are
redshifted relative to those of the five-coordinate precursor characteristic of this species (Figure S1, Supporting Infor-
[Rh(TPP)Cl]. Addition of solid KCN to the solution of 1 mation). The 1H NMR spectra of 1 recorded in CDCl3
followed by equilibration of the heterogeneous mixture at (Figures 5 and S3) also confirmed the presence of this spe-
room temperature for 12 h afforded the spectrum of cies. Formation of a stable mixed-ligand RhIII–porphyrin
2514
www.eurjic.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2009, 2512–2523

Studies of a Bis(phosphane)RhIII–Porphyrin
derivative (P,O axial donor atom combination) is not with- wavelength absorption band for compound 1 (615 nm) is
out precedent. Specifically, Kadish and co-workers studied ca. 1/7th of the intensity of the Q(0,0) band and, although
solvolysis of [Rh(TPP)(PPh3)2]PF6 in THF which leads to of unknown orbital parentage, we note that a similar fea-
the formation of the mixed-ligand complex [Rh(TPP)- ture is found on the long-wavelength side of the Q(0,0)
(THF)(PPh3)]PF6.[17a] Consistent with the aquation reac- band for K[Rh(TPP)(CN)2]. This feature is absent in
tion here, formation of [Rh(TPP)(THF)(PPh3)]PF6 was [Rh(TPP)Cl].
characterised by the loss of the Soret band intensity at
448 nm, with concomitant appearance of a new Soret band
for the mixed-ligand solvolysis product at 425 nm.[17a] We
also found that [Rh(TPP)(OH2)(PEtPh2)]SbF6 was the sole
species in a dry dichloromethane solution if aged and pow-
dered crystalline material was used for the preparation of
NMR Spectroscopy
1H NMR Spectroscopic Data
An interesting question that arises for 1 is whether the
the solution (Figure S2). The evidence suggests that pow- relatively bulky axial ligands still allow dynamic exchange
dered 1 is susceptible to aquation [specifically formation of phenomena at ambient temperature. Inspection of the
1
the mono(aqua) complex], particularly since this problem downfield region of the H NMR spectrum of 1 (Figure 5)
could be averted if large single crystals of 1 were used for suggests that the axial phosphane ligands undergo free rota-
solution preparation in dry dichloromethane.
tion about the Rh–P bonds leading to a single mean pyrrole
The strongly redshifted Soret band of 1 occurs in the β-proton resonance at δ = 8.810 ppm (a static configuration
visible spectroscopic region and this, in conjunction with for the axial phosphanes would render the eight pyrrole β-
narrow bandwidths (Table S1, Supporting Information), protons inequivalent and hence split the signal). Further-
permits facile resolution of the Q, B, N and L electronic more, the pyrrole β-proton resonance changed negligibly
states[31] for the complex. As shown in Figure 4, the com- from –20 °C (δ = 8.809 ppm) to 50 °C (δ = 8.814 ppm).
plete spectroscopic envelope could be deconvoluted into a Coupled with the fact that no NMR signals from free
number of constituent Voigt functions. The quantised vi- PEtPh2 appeared at elevated temperatures, we conclude that
brational levels of the Q state exhibit an energy separation dissociation of 1 to form the five-coordinate species
of ca. 1203–1427 cm–1 based on the energies of the Q(0,0), [Rh(TPP)(PEtPh2)]SbF6 does not occur to any significant
Q(1,0) and Q(2,0) bands (16750, 17953 and 19380 cm–1, extent for T Յ 50 °C. The large magnetic anisotropy of the
respectively). The vibrational energy levels of the B state for porphyrin ring current is also apparent in the 1H NMR
1 are more narrowly spaced (ca. 406–475 cm–1). A distinc- spectrum of 1. Specifically, the axial ligand methylene pro-
tive spectroscopic feature for compound 1 is the markedly tons exhibit a marked upfield shift from δ = 2.116 ppm in
good resolution of the N and L bands at 367 and 309 nm, the free ligand to δ = –2.829 ppm in the RhIII-bound ligand.
respectively. These bands are also present for Similarly, the ethyl group ϪCH3 resonance shifts from δ =
K[Rh(TPP)(CN)2] but occur at slightly higher energy and 1.147 to –1.388 ppm in the free and coordinated ligands,
are not as sharp or well-resolved as those of 1. The longest- respectively. Finally, we note the marked 3.738 ppm ring
current-induced shielding for the ortho protons of the axial
ligand aryl groups as a result of their spatial projection
towards the central core of the RhIII porphyrin (the o-H
signal shifts from δ = 7.484 to 3.746 ppm upon coordina-
tion of PEtPh2 to [Rh(TPP)]+ in the present solvent sys-
tem). Similar porphyrin ring current induced shieldings are
known for other low-spin d6 metalloporphyrins.[31,32] From
1
the H NMR spectra of 1 above 253 K (Figure S4), the ax-
ial ligands and meso-phenyl groups are in the fast exchange
limit on the 500 MHz NMR timescale.
1
The H NMR spectrum of 1 (Figure 5) indicates a sam-
ple of acceptable purity (Ͼ 97%), although we did find evi-
dence for two contaminant species in solution, namely ca.
0.7% of the mono(aqua) adduct, [Rh(TPP)(OH2)(PEtPh2)]-
SbF6, and ca. 2–2.5% of a second bis(phosphane)RhIII
–
porphyrin species with a distinct pyrrole proton signal at δ
= 9.032 ppm (293 K). The identity of the latter minor com-
ponent was difficult to establish. However, the variable-tem-
1
perature H NMR spectra for the system showed no evi-
Figure 4. Deconvolution of the electronic spectrum of 1 into its
constituent bands (Voigt functions, CH2Cl2, 298 K). The inset
shows an expansion of the visible spectrum of the complex. The
thick grey line is the experimental absorption envelope; the dashed
line is the fit by the linear sum of the constituent bands (solid lines).
Band maxima and assignments are given.
dence of an exchange equilibrium involving this species and
1, suggesting that the minor component is not a conforma-
tional isomer of 1 but a separate six-coordinate RhIII–por-
phyrin. As discussed below, we have assigned the minor spe-
cies in the system to the complex [Rh(TPP)(PHEtPh)-
Eur. J. Inorg. Chem. 2009, 2512–2523
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2515

O. Q. Munro, G. L. Camp, L. Carlton
FULL PAPER
free water in the solvent. We also used the isotropic shield-
ing tensor for the Rh–OH2 protons calculated at the
PBE1PBE/3-21G** level of theory by GIAO methods (δ =
–5.63 ppm at 0 K) to confirm the experimentally observed
strong shielding of the aqua ligand protons by the porphy-
rin ring current and thus the assignment of this signal/spe-
cies in solution.
31P and 103Rh NMR Spectra
Proton-decoupled 103Rh-31P NMR spectra (2D) for 1
were obtained at three temperatures in CDCl3 by using the
classic pulse sequence described by Bax et al.[33] The 1D
31P{1H} and 2D 31P-103Rh{1H} spectra for 1 are shown in
Figure 6a and b, respectively. The 31P NMR spectrum exhi-
bits a distinct doublet consistent with spin coupling of the
magnetically equivalent 31P nuclei of the two PEtPh2 li-
gands to the I = ½ 103Rh nucleus. The 31P-103Rh coupling
constant is most easily measured from the 31P spectrum and
has a value of 82.1 Hz at 300 K and an experimentally neg-
ligible temperature dependence from 213 to 333 K. As
1
noted above, the H NMR spectrum for 1 also showed the
presence of [Rh(TPP)(PHEtPh)(PEtPh2)]SbF6 (ca. 2–
2.5%), a compound that is clearly structurally very similar
to 1. This species gave a 31P doublet at δ = 13.3 ppm
(300 K) and a 31P-103Rh coupling constant of 82.4 Hz that
is clearly consistent with a six-coordinate complex.[17a] It
was also evident in the 103Rh NMR spectra of 1 (see
Table S6 for chemical shifts as a function of T). Notwith-
standing the inconsequentially small presence of both the
Figure 5. 1H NMR spectrum and signal assignments for 1 (CDCl3,
293 K). The minor species in solution (Ͻ 3%) was assigned with
reasonable certainty to the mixed-ligand complex [Rh(TPP)-
(PHEtPh)(PEtPh2)]SbF6 based on its observed and DFT-calcu-
lated 31P and 103Rh NMR spectra and the fact that PHEtPh is
present (1.2%) in the commercial reagent (PEtPh2) used to synthe-
sise 1.
(PEtPh2)]SbF6 based on its 103Rh and 31P chemical shifts monoaqua species and [Rh(TPP)(PHEtPh)(PEtPh2)]SbF6,
and the chemical shift of the P–H proton for the Rh-bound the 2D 31P-103Rh{1H} spectrum for 1 at 300 K clearly
contaminant ligand (δ = –3.32 ppm). This diagnostic pro- shows two well-resolved cross peaks, consistent with spin-
ton signal is shifted upfield relative to its chemical shift in coupling of the two nuclei. Although the 31P-103Rh cou-
the free ligand (δH = 5.66 ppm) upon coordination to RhIII
,
pling is cleanly resolved in the 31P{1H} spectrum, the ex-
consistent with marked shielding by the porphyrin ring cur- pected triplet resonance pattern is not resolved on the 103Rh
rent (see Supporting Information for spectroscopic details). axis, consistent with observations for many simple RhIII
The source of PHEtPh in the system, which is evidently complexes.[34]
present at a high enough concentration during the synthesis
of 1 to give some of the mono(PHEtPh) adduct, was traced 10.84 ppm at 213, 300 and 333 K, respectively. The linear
to the commercial reagent which is only 97.5% PEtPh2. increase in δP with increasing temperature had a slope, in-
The 31P chemical shifts were δ = 10.68, 10.79 and
Returning to the aquation adduct [Rh(TPP)(OH2)- tercept and correlation coefficient of 1.32(5)ϫ10–3
(PEtPh2)]SbF6, we found that the extent of equilibration to ppmK–1, 10.40(2) ppm and 0.999, respectively. The 103Rh
the monoaqua complex could be reduced, but not com- chemical shifts were 2480, 2558 and 2590 ppm at 213, 300
pletely eliminated, by first passing the CDCl3 used for solu- and 333 K, respectively. The 103Rh chemical shift therefore
tion preparation through a short column of activated alu- increases linearly with increasing temperature with a slope,
mina. This problem unfortunately persisted even when intercept and correlation coefficient of 0.91(2) ppmK–1,
using clean crystalline material with good elemental analy- 2285(5) ppm and 0.9999, respectively. Interestingly, the tem-
sis data for solution preparation. Unambiguous spectro- perature dependence of δRh in 1 is roughly three times the
scopic identification of [Rh(TPP)(OH2)(PEtPh2)]SbF6 is value typically observed for simple RhIII coordination com-
1
possible from the distinct, exchange-broadened, upfield H pounds (slope ≈ 0.3–0.4 ppmK–1)[19] and is experimentally
resonance (δ = –4.8 ppm region, Figure S3) that shifts equivalent to that reported for mer-[RhCl3(SMe2)3] [slope
downfield with increasing temperature in an analogous = 1.0(1) ppmK–1].[19,35] Furthermore, the 103Rh chemical
fashion to signals of other protons in the molecule. The shifts for 1 are upfield relative to those of simple RhIII
temperature dependence of the signal for the RhIII–OH2 coordination compounds such as [RhCl6]3– (δRh
=
protons [δH = –5.05(2) ppm + 8.0(7)ϫ104 ppmK–1 (T), cor- 8000 ppm)[19,36] but fall in the same range as those of mer-
relation coefficient R = 0.993] is thus opposite to that for [RhCl3(PBu3)3] (δRh = 2770 ppm) and mer-[RhCl3(PMe3)3]
2516
www.eurjic.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2009, 2512–2523

Studies of a Bis(phosphane)RhIII–Porphyrin
inverse cube of the non-s orbital radius (͗1/r3͘) and the in-
verse mean excitation energy for electronic transitions from
the ground state (∆E).[39]
γX
νX
=
B0(1 – σX)
(1)
2π
ΣQXY
σX = σd –
͗1/r3͘
(2)
∆E
In the case of 31P, ∆E is the mean singletǞtriplet state
excitation energy, which decreases slightly with increasing
temperature due to the population of excited vibrational
levels of the ground singlet state. This causes a slight in-
crease in the Larmor frequency and thus δP with increasing
temperature. A similar argument prevails for δRh with the
exception that ∆E in Equation (2) corresponds to the mean
excitation energy for transitions involving the singlet state
ligand field terms for the RhIII ion. Because both nuclides
exhibit temperature-dependent changes in chemical shifts of
the same sign, a plot of δP vs. δRh shows a linear trend with
increasing temperature (Figure 6c) and succinctly summa-
rises the changes in nuclear shielding for the two nuclides.
Although Equation (2) is often invoked (with emphasis on
the ∆E–1 parameter) to account for the temperature depen-
dence of chemical shifts, one cannot exclude other explana-
tions. The numerator term ΣQXY in Equation (2), in par-
ticular, is not insignificant for coordination compounds, be-
cause dative covalent bonds are more pliable than covalent
bonds and thus deform more readily with changes in tem-
perature and even the dielectric constant of the medium
(which is also temperature-dependent). An interesting ex-
ample of an alternative explanation of the large thermal
shielding derivatives and NMR isotope shifts of many tran-
sition metal nuclei was put forward by Jameson et al.[39]
using a model in which the derivatives of the nuclear shield-
ing are dependent on the equilibrium M–L bond lengths
according to a classical Morse-type function. This model
was superior to Equation (2) because it not only accounted
1
Figure 6. (a) H-decoupled 31P NMR spectrum of 1 in CDCl3, (b)
2D 31P-103Rh{1H} NMR spectrum of 1 and (c) plot of the 31P
and 103Rh chemical shifts vs. temperature for 1. The minor doublet for chemical shift changes with temperature but also those
component observed at δP = 13.3 ppm in the 31P NMR spectrum
that attend isotopic substitution in ML6 complexes. Thus,
coupled with changes in ground state vibrational energy
1
of 1 has a similar JRh,P coupling constant (82.4 Hz) to the major
component and is therefore assigned to the less abundant (Ͻ 3%)
six-coordinate
level populations, the marked thermal shielding derivatives
mixed-ligand
complex
[Rh(TPP)(PHEtPh)-
for the 31P and 103Rh nuclei of 1 probably also reflect rather
complex changes in the mean Rh–N and Rh–P distances
and internuclear charge densities with temperature.
(PEtPh2)]SbF6. The solution was 8 m in concentration.
(δRh = 2207 ppm).[19,37] Evidently, coordination of phos-
phane ligands by RhIII increases the covalent nature of the
complex and thus the shielding of the 103Rh nucleus.
The variation of δP and δRh with temperature may be
explained with Equation (1), where νX, γX, B0 and σX are
DFT Calculations
In-vacuo DFT simulations at the PBE1PBE[40]/3-
the Larmor frequency, magnetogyric ratio, applied mag- 21G**[41] and PBE1PBE/DGDZVP[42] levels of theory were
netic field and nuclear shielding constant for nuclide X, used to establish (a) the electronic structure, (b) the relative
respectively.[19,38] The term σX in Equation (1) is given by stabilities of the conformations of 1 with ruffled (ruf) and
Equation (2), where σd is the diamagnetic shielding con- planar (flat) porphyrin macrocycles and (c) the 103Rh and
stant (governed only by the core electron density), and the 31P nuclear shielding tensors (by using the GIAO theory[43]
)
second term is the “paramagnetic” shielding constant, of the system. Simulations with the 3-21G** basis set were
which depends on the charge density and bond order be- the most accurate for both the geometry and shielding ten-
tween nuclei X and Y (ΣQXY), the expectation value of the sor determinations (Figure 7 and Tables 1 and 4). For ex-
Eur. J. Inorg. Chem. 2009, 2512–2523
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2517

O. Q. Munro, G. L. Camp, L. Carlton
FULL PAPER
ample, bond lengths for the coordination group were within
0.6% of those determined crystallographically with the gra-
dient-corrected PBE1PBE hybrid functional. Interestingly,
the in vacuo geometry of ruf-[Rh(TPP)(PEtPh2)2]+ was
slightly less distorted than that observed crystallographi-
cally (cf. the conformational parameter |Cm| in Table 1),
presumably because of the omission of neighbouring mole-
cules (i.e. crystal-packing effects) during the DFT simula-
tions. Indeed, it is noteworthy that crystal packing induced
distortions of axial and porphyrin ligand substituents are
well-known determinants of the molecular conforma-
tion.[44] Simulations carried out on the ruf conformer of
cationic 1 without the meso-phenyl groups of the porphyrin,
i.e. by using a simple porphine model (data not shown),
afforded good agreement between crystallographic and cal-
culated bond lengths but poorer agreement between the ob-
served and calculated porphyrin ring conformations. This
suggests that intramolecular steric interactions (especially
porphyrin-aryl···phosphane-aryl van der Waals interac-
tions) play a role alongside axial-ligand···porphyrin-core
nonbonded interactions in shaping the experimentally ob-
served ruffled conformation of 1.
Population Analysis
NBO analysis[45] of the converged wave function for the
optimised geometry of ruf-[Rh(TPP)(PEtPh2)2]+ reflects the
formal low-spin 4d6 electron configuration for the system
(Figure 7). It is, however, noteworthy (Table 2) that the val-
ence-electron population of the RhIII ion substantially ex-
ceeds 6 electrons presumably due to σ-donation from the
porphyrin and axial phosphane ligands and substantial
mixing of the metal and ligand orbitals (significant co-
valency), as shown in Figure 8. Indeed, the electron occu-
pancies of the σ-antibonding MOs with significant 4dz2 and
4dx2–y2 character far exceed zero, contrary to expectations
from simple ligand-field arguments. Well-mixed metal and
ligand orbitals for ruf-[Rh(TPP)(PEtPh2)2]+ apparently al-
low excess electron density donated to the metal ion to be
housed by the σ* MOs as these metal-character orbitals are
Figure 7. (a) Left: root mean square fit[65] of the X-ray cation of 1
(black) to the DFT-calculated C1 symmetry geometry of ruf-
[Rh(TPP)(PEtPh2)2]+ at the PBE1PBE/3-21G** level of theory
(grey). Right: 4d orbital energy levels and electron populations
(NBO method) for the calculated cation structure on the left. (b)
Left: DFT-calculated C2 symmetry geometry of flat-
[Rh(TPP)(PEtPh2)2]+ (PBE1PBE/3-21G**). Right: 4d orbital en-
ergy levels and electron populations (NBO method) for the calcu-
lated cation structure on the left. H atoms have been omitted for
clarity in all structures.
Table 1. Comparison of selected geometrical parameters for the experimental and DFT-calculated ruf and flat conformers of cationic
1.[a]
Exp.
3-21G**, ruf
Diff.[d]
3-21G**, flat
DGDZVP, ruf
Diff.[d]
[b]
Rh–Np
Rh–P
Ca–Np
Cb–Ca
Cb–Cb
Ca–Cm
|Rh|[c]
|N|
2.036(5)
2.401(3)
1.374(4)
1.439(4)
1.342(4)
1.397(6)
0.01(0)
0.02(2)
0.15(3)
0.11(8)
0.27(2)
2.043(6)
2.392(4)
1.381(2)
1.449(1)
1.365(1)
1.396(1)
0.00(0)
0.02(1)
0.14(2)
0.10(6)
0.25(1)
176.0
–0.007
–0.009
–0.007
–0.010
–0.023
0.001
0.01
0
0.01
0.01
0.02
2.046(7)
2.399(0)
1.381(1)
1.448(1)
1.365(2)
1.395(2)
0.00(0)
0.03(2)
0.04(3)
0.10(4)
0.02(4)
177.2
2.051(4)
2.452(2)
1.369(1)
1.441(1)
1.361(1)
1.404(1)
0.00(0)
0.02(1)
0.12(3)
0.09(8)
0.22(1)
176.4
–0.019
–0.056
0.008
–0.003
–0.020
0.003
0.01
0.02
0.07
0.05
0.13
|Ca|
|Cb|
|Cm|
P–Rh–P
177.96(4)
0
–1.46
[a] DFT method = PBE1PBE; distances in Å, angles in °. [b] Np, Ca, Cb, and Cm = porphyrin nitrogen, α-, β-, and meso-carbon,
respectively. [c] |X|: mean absolute perpendicular displacement of atom type X from the 24-atom porphyrin mean plane. [d] Difference
relative to the experimental value.
2518
www.eurjic.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2009, 2512–2523

Studies of a Bis(phosphane)RhIII–Porphyrin
particularly low-lying in energy. This phenomenon is evi- NBO population analysis for the two conformers of
dent for both conformations of cationic 1 but perhaps best [Rh(TPP)(PEtPh2)2]+ is the lack of any real difference in
illustrated in the case of ruf-[Rh(TPP)(PEtPh2)2]+ where the the partial atomic charges and valence-electron populations
4dz2 orbital exhibits a population of ca. 1.4 e due to its of the RhN4P6 coordination sphere (Table 2). However, the
rather low energy (Ͻ –9.3 eV; Figure 7) for an otherwise slightly lower partial atomic charge and slightly higher val-
somewhat atypical “rhombic symmetry” metalloporphyrin ence-electron population for Rh in ruf-[Rh(TPP)(PEt-
d-orbital energy-level sequence (the typical d-orbital en- Ph2)2]+ probably reflects the marginally shorter Rh–N and
ergy-level sequence for low-spin ferric porphyrins axially Rh–P distances (Table 1). The Rydberg populations of the
ligated by a pair of strong-field ligands, for example, is: dxy Rh-bound P atoms (0.064 e in both conformers) are signifi-
Ͻ dxz Ͻ dyz Ͻ dx2–y2 Ͻ dz2).[46]
cantly higher than that of free PEtPh2 (0.042 e, calculated
at the same level of theory in vacuo) and likely reflect at
least some quantifiable backbonding from Rh to P. The ex-
act nature of the backbonding interaction is difficult to es-
tablish with certainty from the present data but could in-
volve either one or more vacant 3d orbitals on P or indeed
a σ*-type MO on the phosphane ligand. From the full NBO
atomic orbital population analysis data given in Table S2,
we find that the 4p (Rydberg) orbitals of phosphorus in
PEtPh2 gain the most electron density upon coordination
of the ligand to RhIII, suggesting that the 3d orbitals of P
are not the dominant MǞL electron-transfer destination
for either conformer of 1. Lastly, the partial positive charge
on P exceeds that on RhIII and that of the metal-free ligand
in the gas phase (0.917 e), a situation that strikingly high-
lights the significant covalency of the Rh–P bonds in con-
trast to the Rh–N bonds in this complex, which are seem-
ingly slightly more ionic in character.
Table 2. Key NBO charge and electron populations for the confor-
mations of [Rh(TPP)(PEtPh2)2]+ with ruffled and planar porphyrin
cores calculated at the PBE1PBE/3-21G** level of theory.
Atom
Charge
Core
Valence
Rydberg
Total
Ruffled (C1 symmetry)
Rh
N
N
N
N
P
0.783
35.985
1.999
1.999
1.999
1.999
9.997
9.997
8.170
5.487
5.495
5.496
5.496
3.802
3.795
0.063
0.007
0.007
0.007
0.007
0.064
0.064
44.217
7.493
7.502
7.502
7.502
13.862
13.855
–0.493
–0.502
–0.502
–0.502
1.138
P
1.145
Planar (C2 symmetry)[a]
Rh
N
N
P
0.792
35.985
1.999
1.999
9.997
8.162
5.490
5.498
3.802
0.061
0.007
0.007
0.064
44.208
7.496
7.504
–0.496
–0.504
1.138
13.862
Conformational Energetics
[a] Only the symmetry-unique charges are shown.
The ruffled and planar conformers of cationic 1 are not
markedly different in energy, as shown by the thermochemi-
cal data in Table 3. At 298.15 K, the free energy difference
between the conformers is only 2.23 kJmol–1, with the
ruffled conformer being the more stable of the two. Interest-
ingly, the energy order of the conformers is reversed if the
meso-phenyl groups are replaced with H atoms in the calcu-
lations, i.e. ruf-[Rh(porphine)(PEtPh2)2]+ is less stable than
flat-[Rh(porphine)(PEtPh2)2]+ by ∆H ≈ 0.77 kJmol–1 at
298.15 K (data not shown). The latter is consistent with an
earlier report based on molecular mechanics (MM) calcula-
tions that ruffled metal–porphine conformers are always
more strained than their planar counterparts and that the
magnitude of the energy difference increases as a function
of the extent of ruffling of the macrocycle.[47] From the data
in Table 3, we surmise that nonbonded interactions between
the axial phosphane ligands and the meso-phenyl groups in
Figure 8. Mixing of the formally σ antibonding 4d metal and li-
gand orbitals for ruf-[Rh(TPP)(PEtPh2)2]+. (a) Antibonding MO
2
with ca. 18% 4dz character. (b) Antibonding MO with ca. 45%
2
2
4dx –y character.
The markedly different 4d-orbital energy-level sequence the TPP derivative dominate the conformational energetics,
for flat-[Rh(TPP)(PEtPh2)2]+ relative to that of ruf- at least for the minimum energy conformations shown in
[Rh(TPP)(PEtPh2)2]+ appears odd at first glance (Figure 7), Figure 7. A more complete understanding of the potential
particularly the observation that the highest-energy metal- energy surface (∆H vs. the two dihedral angles controlling
character MO in ruf-[Rh(TPP)(PEtPh2)2]+ (4dx2–y2) differs the orientations of the axial phosphane ligands relative to
from that in flat-[Rh(TPP)(PEtPh2)2]+ (4dyz). This change the porphyrin core) could be gained from a surface scan
in the energy order of the 4d orbitals merely reflects the effected by standard dihedral angle driving methods. How-
fact that the Cartesian axes have different orientations rela- ever, even at the PBE1PBE/3-21G** level of theory, an ade-
tive to the RhN4P2 coordination sphere framework in the quately sampled grid of 372 conformers to define the poten-
two conformers (as shown in Figure S8), a situation that is tial energy surface (PES) is computationally highly expens-
probably chemically inconsequential but nonetheless of ive and well beyond the scope of the present article (though
some theoretical interest. One worthwhile comment on the clearly of interest to supercomputer enthusiasts).
Eur. J. Inorg. Chem. 2009, 2512–2523
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2519

O. Q. Munro, G. L. Camp, L. Carlton
FULL PAPER
Table 3. Thermochemical analysis for the DFT-calculated ruf and Table 4. In-vacuo DFT-calculated (PBE1PBE-GIAO method)
flat conformers of cationic 1 (298.15 K, 1 atm).[a]
ruf-[Rh(TPP)(L)2]+ flat-[Rh(TPP)(L)2]+
shielding tensors, diagonal matrix elements and anisotropy param-
eters (all in ppm) for key conformations of [Rh(TPP)(PEtPh2)2]+.[a]
[d]
[d]
Atom δiso
δxx
δyy
δzz
δanisot.
∆
iso
∆Јiso
Eelec
ZPE
–8323.282868
1.127888
–8323.281658
1.127678
1.193017
1.193961
1.024346
–8322.15398
–8322.088641
–8322.087697
–8322.257312
2.626
Ruffled (C1 symmetry)
Thermal corr. to Eelec 1.193132
Rh[b] 2683
Rh[c] 1272
2655 2659 2734 89.9
1140 1193 1484 422
398
125
Thermal corr. to H
Thermal corr. to G
E0 (0 K)
E (stp)
H (stp)
1.194076
1.024706
–8322.15498
–8322.089736
–8322.088792
–8322.258162
0
0
0
0
–1013 –1286
20.1
7.16
P[b]
P[c]
30.49
17.6
–42.1 –10.5 144
–4.25 –47.4 104
32.0
93.0
19.7
6.77
Planar (C2 symmetry)
Rh[b] 2605
2679 2517 2618 119
P[b]
26.2 –54.4 –33.8 167 22.4
G (stp)
320
15.8
47
15.4
Erel (0 K)[b]
Erel (stp)[b]
Hrel (stp)[b]
Grel (stp)[b]
2.875
2.875
2.232
[a] NMR reference compounds: 103Rh, [RhCl6]3–; 31P, H3PO4. [b] 3-
21G** basis set. [c] DGDZVP basis set. [d] Experimental isotropic
shielding: δRh = 2285(5) ppm at 0 K (δ = 2558 ppm at 300 K); δP
= 10.40(2) ppm at 0 K (δ = 10.79 ppm at 300 K). ∆iso and ∆Јiso are
the differences between the calculated and experimental isotropic
shieldings at 0 and 300 K, respectively.
[a] DFT method = in vacuo PBE1PBE/3-21G**; L = PEtPh2; all
energies are in Hartree units unless otherwise noted (1 Hartree =
2625.50 kJmol–1). The energy parameters Eelec, H, G, and ZPE are
the electronic energy, enthalpy, Gibbs free energy and zero-point
energy, respectively. E0 = Eelec + ZPE; E = E0 + Evib + Erot + Etransl
,
H = E + RT; G = H – TS; stp = standard temperature (298.15 K)
and pressure (1 atm). [b] Relative energy, enthalpy or Gibbs free
energy in kJmol–1.
A noteworthy point concerning the data in Table 4 is that
because the DFT simulations were performed in vacuo, the
DFT-calculated chemical shifts are compared with the ex-
perimental chemical shifts extrapolated to 0 K in the first
instance for two reasons. First, the temperature of the sys-
tem in a vacuum tends towards zero; second, current
GIAO-DFT methods on static conformations do not take
into account thermal shielding derivatives despite the fact
that most nuclei have temperature-dependent isotropic and
anisotropic shielding tensors. At 0 K then, the calculated
in-vacuo 103Rh chemical shift of ruf-[Rh(TPP)(PEtPh2)2]+
is within 20% of the experimental chemical shift for 1 for
the calculation with the 3-21G** basis set. At 300 K, the
agreement between theory and experiment is even better
(5%) but fundamentally unsatisfactory because of the in-
trinsic temperature difference between the two methods.
Other workers have reported DFT-based GIAO-calcu-
lated 103Rh chemical shifts for several RhI and RhIII coordi-
nation compounds that are typically within 300 ppm, or
roughly 10–15%, of the experimental chemical shift.[50,51]
Our results for 1 are evidently typical for the DFT method
employed and show that a relatively small basis set may be
What we can articulate from the data in Table 3 is that at
298 K, both the ruffled and planar conformers of cationic 1
should have almost equal populations. Our variable-tem-
perature NMR-spectroscopic studies in fact confirm this (a
single pyrrole 1H resonance is observed from 253 to 323 K)
and indicate a fairly low energy barrier connecting the two
stable conformers on the PES. Conformational surfaces of
the type discussed above have been computed previously at
the MM level of theory with appropriately derived force
fields for iron(III)–[48] and cobalt(III)–porphyrin[31] com-
plexes. In these studies, with both sterically compact and
bulky porphyrin ligands, in-vacuo potential-energy barriers
delineating the pathways that connect ruffled and planar
conformers tend to be Ͻ 12 kJmol–1, thus permitting free
rotation of the axial ligands at ambient temperature.
Shielding Tensor Determinations
The isotropic shielding tensor (δiso) for the 103Rh nucleus used with some measure of confidence for future simula-
of 1 was adequately calculated at the PBE1PBE/3-21G** tions on this comparatively large-scale problem. Although
level of theory (Table 4). The shielding tensor determination the GIAO-calculated chemical shifts for 103Rh in 1 were
employing the larger basis set (DGDZVP) afforded a sur- relatively accurate for simulations with the 3-21G** basis
prisingly poor prediction of δRh. This result is somewhat set, the 31P chemical shift determinations for the planar and
unexpected given the generally good performance of the ruffled conformers of the [Rh(TPP)(PEtPh2)2]+ cation were
DGDZVP basis set for geometry optimisations and fre- significantly less accurate. As indicated in Table 4, δP devi-
quency calculations for most complexes of transition metal ates from the experimental chemical shift by up to 20 ppm.
ions.[49] Furthermore, the 31P shielding for 1 was consider- This more than likely highlights shortfalls in the 3-21G**
ably more accurately determined relative to that calculated basis set for shielding tensor calculations on light elements,
with the 3-21G** basis set. Such a problem can only result because neither the use of a truncated porphyrin model
from an inherent deficiency in the core electron param- such as [Rh(porphine)(PEtPh2)2]+ (Table S3) nor a chloro-
eterisation for heavier nuclei such as Rh because geometry form solvent model (see below) yielded markedly improved
optimisations and frequency calculations, which are primar- agreement between the calculated and observed chemical
ily dependent on the accuracy of interactions involving the shifts. (It is normal to use the much larger 6-311G** basis
valence electrons, are not in any way problematic for this set for accurate GIAO simulations on H through Kr, com-
basis set.
putational resources permitting.) Whereas the accuracy of
2520
www.eurjic.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2009, 2512–2523

Studies of a Bis(phosphane)RhIII–Porphyrin
niques. THF and hexane were distilled from sodium/benzophe-
none. Dichloromethane and pyrrole (Aldrich) were distilled from
CaH2. Benzaldehyde, silver hexafluoroantimonate(V) and ethyldi-
phenylphosphane (PEtPh2) were used as received (Aldrich). H2TPP
was synthesised according to published procedures.[56] [Rh(TPP)Cl]
was prepared by metallation of H2TPP with rhodium(III) chloride
in DMF heated to reflux.[57]
GIAO calculations with large basis sets tends to be rather
good for organic compounds, it is quite clear that in the
presence of transition metal ions, δP predictions with large
basis sets are typically only accurate to within 10–25 ppm
of the experimental chemical shift[52–54] and therefore no
better than those employing the 3-21G** basis set reported
here.
Instrumental Methods: Electronic spectra were recorded with a Shi-
madzu UV-1800 double-beam scanning spectrophotometer by
using dry dichloromethane solutions prepared from single crystals
of 1 without excess phosphane in a 1.0 cm path-length quartz cu-
vette. FTIR spectra of one or more single crystals of 1 were re-
corded with a Bruker Alpha diamond ATR spectrometer (48 scans,
spectroscopic resolution = 1.0 cm–1). C, H and N elemental analy-
sis (combustion) data were obtained from Galbraith Laboratories
(USA) on a polycrystalline sample of 1.
The NMR shielding tensor determinations for the ruffled
and planar conformers of 1 were also calculated in a chloro-
form solvent continuum (PCM method,[55] Table S4) in an
effort to gauge the solvent contribution to the chemical
shifts. Because the 103Rh and 31P nuclei of 1 are deeply bur-
ied within the framework of a large molecule, the presence
of the solvent had a negligible effect on the calculated iso-
tropic shielding values for these two nuclei. Thus, δRh and
δP computed in a chloroform polarisation continuum were
1% worse and 1.5% better compared with the experimental
chemical shifts of these nuclei, respectively. Interestingly,
and as noted earlier, use of a truncated porphyrin model
such as [Rh(porphine)(PEtPh2)2]+ (Table S3) had practi-
cally no effect on the accuracy of the calculated 31P chemi-
Synthesis of 1: To [Rh(TPP)Cl] (150 mg, 0.20 mmol) and AgSbF6
(82 mg, 0.24 mmol) in a 250 mL Schlenk tube under nitrogen was
added freshly distilled THF (50 mL). The solution was stirred at
room temperature for ca. 12 h. The THF was then removed in
vacuo and the green-brown solid redissolved in dichloromethane
(50 mL). The solution was then filtered, to remove precipitated sil-
cal shifts. The calculated 103Rh chemical shifts were, on the ver chloride, into a 250 mL Schlenk tube into which PEtPh2
(0.82 mL, 4.0 mmol) had been added. The solution was stirred at
room temperature for ca. 10 min. The purple-red solution was then
transferred into 12 Schlenk tubes in ca. 4 mL aliquots and layered
with hexane. X-ray-quality crystals were observed after 4 d. Iso-
lated yield: 0.2532 g (92%). C74H62Cl4F6N4P2RhSb (1549.76):
other hand, closer to the experimental values for 1 (this
improvement for the truncated structural model is of course
counter-intuitive). The meso-phenyl groups of the porphy-
rin ligand are evidently not inconsequential to the electronic
structure of the compound, and it is clear that an RhIII
complex as large as 1 presents a formidable challenge for
current DFT and GIAO theory that we have not yet satis-
factorily met with the present PBE1PBE/3-21G** simula-
tions. The problem is that very few all-electron basis sets
1
calcd. C 57.35, H 4.03, N 3.62; found C 56.80, H 4.48, N 4.19. H
NMR (500 MHz, CDCl3, 298 K): δ = 8.78 (s, 8 H, pyrrole β-H),
7.76 (m, 20 H, TPP o-, m-, p-H), 7.01 (t, 3J = 7.3 Hz, 4 H, L p-H),
3
3
6.57 (t, J = 7.5 Hz, 8 H, L m-H), 3.73 (p, J = 3.8 Hz, 8 H, L o-
3
3
H), –1.40 (p, J = 7.7 Hz, 6 H, CH3), –2.84 (q, J = 7.7 Hz, 4 H,
for heavy nuclei suitable for conducting nuclear shielding CH2) ppm. 13C{1H} NMR (125 MHz, CDCl3, 298 K): δ = 142.494
(TPP Ca), 142.074 (TPP Ca), 140.913 (TPP Ca), 140.842 (TPP Ca),
140.691 (Cm-Cphenyl), 137.913 (TPP Cb), 137.844 (TPP Cb),
134.360, 134.238, 134.112, 133.253, 133.092, 132.831, 132.639,
132.499, 132.257, 131.589, 131.570, 130.763, 130.690, 128.567,
128.353, 128.300, 127.048 (P-Cphenyl), 126.943, 126.742 (P-Cphenyl),
122.26 (TPP Cm), 121.994 (TPP Cm), 20.543 (CH2), 20.483 (CH2),
calculations have been developed.
Conclusions
We have synthesised and characterised a novel ruffled
RhIII–porphyrin bearing moderately bulky axial PEtPh2 li-
gands as its hexafluoroantimonate(V) salt, [Rh(TPP)-
(PEtPh2)2]SbF6. Indirect detection of the 103Rh NMR sig-
nal by means of polarisation transfer from 31P has enabled
us to gauge the nuclear shielding of the rhodium ion as a
function of temperature – a first for RhIII–porphyrins. DFT
simulations at the PBE1PBE/3-21G** level of theory were
surprisingly accurate for the coordination geometry of the
metal ion and porphyrin ligand and moderately accurate at
prediction of the 103Rh isotropic shielding tensor. We are
currently extending this study to include several bis(phos-
phane)-, bis(phosphinite)- and bis(phosphonite)RhIII–por-
phyrins in an effort to establish empirical correlations be-
tween δRh, J(P,Rh) and parameters derived from X-ray crys-
tal structures of these stable, yet conformationally flexible
compounds.
9.933 (CH ) ppm. IR (crystalline solid): ν = 518, [m, ν(Rh–P)], 652
˜
3
[s, ν(P–Et)], 1012 [s, ν(C=C) porphyrin], 1439 [w, ν(P–Ph)] cm–1 (s,
m, w = strong, medium, weak). UV/Vis CH2Cl2 at 298 K (ε): λ
= 597 [2.78(7)ϫ104], 557 [2.45(4)ϫ104], 515 [1.16(9)ϫ104], 448
[3.35(14)ϫ105], 367 [8.40(24)ϫ104], 308 [7.47(20)ϫ104 –1
cm–1] nm.
Crystallography: X-ray data were collected on a red crystal of the
compound with dimensions 0.38ϫ0.40ϫ0.50 mm by using an En-
raf–Nonius CAD4 diffractometer operating at 1.65 kW X-ray
power (Mo-Kα radiation, λ = 0.71073 Å) with a Bruker LT3 low-
temperature attachment. The data were reduced by using
XCAD4[58] and the structure solved with SHELXS-97[59] (running
under WinGX[60]) by using direct methods. The structure was re-
fined with SHELXL-97.[61] All non-hydrogen atoms were refined
anisotropically, and H atom coordinates were calculated by using
the standard riding model of SHELXL. Towards the end of the
structure refinement for 1, it became clear that the phenyl group
appended to meso-C4 (C4m) was disordered about two orienta-
tions. A model, in which two orientations for this phenyl group
were present, gave a satisfactory and stable refinement upon fixing
the geometry of each component ring to a regular hexagon (AFIX
66 command in SHELXL). Crystallographic data for 1·2CH2Cl2:
Experimental Section
General: All manipulations were carried out under nitrogen by
using a double manifold vacuum line, Schlenk and cannula tech-
Eur. J. Inorg. Chem. 2009, 2512–2523
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2521

O. Q. Munro, G. L. Camp, L. Carlton
FULL PAPER
C74H62Cl4F6N4P2RhSb (1549.68 gmol–1), monoclinic, space group
[Rh(TPP)(OH2)(PEtPh2)]SbF6 and 1 as a function of temperature,
P21/n, a = 17.678(5), b = 17.747(7), c = 22.173(6) Å, β = 106.52(2)°, respectively; Figures S4–S6, additional X-ray structure diagrams;
V = 6669(4) Å3, Z = 4, Dc = 1.543 gmL–1, µ = 0.925 mm–1, T = Figure S7, frontier MOs for cationic 1; Figure S8, Cartesian axes
259(2) K, 14060 total reflections, 11689 independent reflections
(Rint = 0.0168), 9815 reflections with I Ͼ 2σ(I), R1 = 0.0535, wR2
= 0.1476 [for I Ͼ 2σ(I)]. CCDC-223496 contains the supplementary
crystallographic details for 1 in CIF format. These data can be
obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
for conformers of cationic 1; Figure S9, calculated geometry of
[Rh(TPP)(OH2)(PEtPh2)]+; Figure S10, 1H NMR spectroscopic
data for of [Rh(TPP)(PHEtPh)(PEtPh2)]+; Tables S1–S6, ad-
ditional experimental and theoretical numerical data.
Acknowledgments
NMR Spectroscopy: 1H, 13C and 31P NMR spectra of 1 were re-
corded by using ca. 6–10 mg of polycrystalline material dissolved
in 0.60 mL of dry CDCl3 under nitrogen. Data were acquired with
a 500 MHz Varian Unity Inova spectrometer equipped with an Ox-
The authors thank the National Research Foundation of South
Africa (NRF, Pretoria), the University of the Witwatersrand and
the University of KwaZulu-Natal for financial support. We are also
especially indebted to Mr. Craig Grimmer for recording some of
the variable-temperature 1H and 31P NMR spectroscopic data in
this work and Wen Wen Suo for her contribution to sample prepa-
rations.
1
ford magnet (11.744 T) at 298 K. Standard H, 13C and 31P pulse
sequences were used for 1D and 2D spectra. 103Rh NMR spectra
were recorded at 213, 300 and 333 K with a Bruker DRX 400 spec-
trometer equipped with a 5 mm triple-resonance inverse probe with
a dedicated 31P channel and extended decoupler range, operating
at 161.98 MHz (31P) and 12.65 MHz (103Rh). Samples were pre-
pared by dissolving 6 mg of polycrystalline material in 0.55 mL of
CDCl3 under nitrogen. 2D 103Rh-31P spectra were obtained by
using the pulse sequence[61] π/2(31P)–1/[2J(103Rh-31P)]–π/2(103Rh)–
τ–π(31P)–τ–π/2(103Rh)–Acq(31P). A spectroscopic width in f2 (31P)
of 8 ppm and an acquisition time of 0.396 s gave a digital resolution
of 1.26 Hz per point; in f1 (103Rh), a spectroscopic width of 40 ppm
and a time domain of 256 (reduced, for spectra of the minor com-
ponent to 60) gave, after zero filling, a digital resolution of 0.49 Hz
per point. With a relaxation delay of 2 s and 4 scans per increment,
the data collection required 27 min. To eliminate the possibility of
a folded signal in f1, the spectra were first recorded with a spectro-
scopic width of 2000 ppm. Chemical shifts were referenced to the
[1] Abbreviations: DFT = density functional theory, esd = esti-
mated standard deviation, GIAO = gauge-independent atomic
orbitals; TF5PP = dianion of 5,10,15,20-tetrakis(pentafluoro-
phenyl)porphyrin, TPP = dianion of 5,10,15,20-tetraphen-
ylporphyrin, OEP = 2,3,7,8,12,13,17,18-octaethylporphyrin di-
anion, 1-MepipZ
= 1-methylpiperazine, THF = tetra-
hydrofuran; nd = not determined.
[2] S. Ariel, D. Dolphin, G. Domazetis, B. R. James, T. W. Leung,
S. J. Rettig, J. Trotter, G. M. Williams, Can. J. Chem. 1984, 62,
755–762.
[3] B. R. James, D. Dolphin, T. W. Leung, F. W. B. Einstein, A. C.
Willis, Can. J. Chem. 1984, 62, 1238–1245.
[4] C.-M. Che, J.-L. Zhang, R. Zhang, J.-S. Huang, T.-S. Lai, W.-
M. Tsui, X.-G. Zhou, Z.-Y. Zhou, N. Zhu, C. K. Chang, Chem.
Eur. J. 2005, 11, 7040–7053.
[5] E. Stulz, M. Maue, N. Feeder, S. J. Teat, Y.-F. Ng, A. D. Bond,
S. L. Darling, J. K. M. Sanders, Inorg. Chem. 2002, 41, 5255–
5268.
generally accepted standards of H3PO4 and Ξ(103Rh)
=
3.16 MHz[19] with positive values indicating deshielding. The chem-
ical shift of H3PO4 (85%, 300 K) with a CDCl3 external lock corre-
sponds to a frequency of 161.975491 MHz in a field (4.395 T) in
which the protons of TMS (in CD2Cl2 at 300 K) resonate at
400.130020 MHz.
[6] R. M. Belani, B. R. James, D. Dolphin, S. J. Rettig, Can. J.
Chem. 1988, 66, 2072–2078.
[7] L. Toupet, N. Legrand, A. Bondon, G. Simonneaux, Acta
Crystallogr., Sect. C 1994, 50, 1014–1016.
[8] M.-A. Pilard, M. Guillemot, L. Toupet, J. Jordanov, G. Simon-
neaux, Inorg. Chem. 1997, 36, 6307–6314.
[9] L. Toupet, P. Sodano, G. Simonneaux, Acta Crystallogr., Sect.
C 1990, 46, 1631–1633.
[10] a) C.-M. Che, T.-F. Lai, W.-C. Chung, W. P. Schaefer, H. B.
Gray, Inorg. Chem. 1987, 26, 3907–3911; b) J. Xie, J.-S. Huang,
N. Zhu, Z.-Y. Zhou, C.-M. Che, Chem. Eur. J. 2005, 11, 2405–
2416.
[11] a) K. M. Kadish, J.-L. Cornillon, P. Mitaine, Y. J. Deng, J. D.
Korp, Inorg. Chem. 1989, 28, 2534–2542; b) K. M. Kadish, Y. J.
Deng, J. D. Korp, Inorg. Chem. 1990, 29, 1036–1042.
[12] Cambridge Structural Database (CSD), version 5.29 (July
2008); F. H. Allen, Acta Crystallogr., Sect. B 2002, 58, 380–
388.
[13] a) E. Stulz, S. M. Scott, Y.-F. Ng, A. D. Bond, S. J. Teat, S. L.
Darling, N. Feeder, J. K. M. Sanders, Inorg. Chem. 2003, 42,
6564–6574; b) K. Fukushima, K. Funatsu, A. Ichimura, Y. Sa-
saki, M. Suzuki, T. Fujihara, K. Tsuge, T. Imamura, Inorg.
Chem. 2003, 42, 3187–3193; c) E. Stulz, S. M. Scott, A. D.
Bond, S. J. Teat, J. K. M. Sanders, Chem. Eur. J. 2003, 9, 6039–
6048.
DFT Simulations: All calculations were performed with
Gaussian 03W.[62] Input atomic coordinates were either Cartesian
coordinates derived from the fractional atomic coordinates of the
X-ray structure of 1 or structures prepared with GausView 3.09.[63]
The DFT method employed (PBE1PBE) was the gradient-cor-
rected hybrid method of Perdew et al.[42] with either the
DGDZVP[44] or 3-21G**[43] basis sets. Frequency calculations were
performed on geometry-optimised structures to establish the nature
of the stationary points located on the potential energy surface for
each structure/conformer. All structures reported in this paper were
stable minima (no frequency eigenvalues Ͻ 0). Shielding tensors
were determined by the GIAO method[45] on the optimised 3-
21G** geometries. Chemical shifts were determined by using calcu-
lated shielding tensors for H3PO4 and [RhCl6]3– as the 31P and
103Rh reference compounds, respectively. In order to compare our
experimental 103Rh chemical shifts (which are referenced to met-
allic rhodium) with those obtained by the GIAO theory, a correc-
tion factor of –4819.1 ppm was used. This factor corresponds to
the difference between δRh calculated for [RhCl6]3– at the
PBE1PBE/3-21G** level of theory (–12819.1 ppm) and the chemi-
cal shift value (–8000 ppm) needed to zero the experimental chemi-
cal shift of [RhCl6]3– (+8000 ppm relative to metallic rhodium).
[14] B. B. Wayland, B. A. Woods, V. M. Minda, J. Chem. Soc.,
Chem. Commun. 1982, 634–635.
[15] X. Li, R. Irie, S. Shinoda, Y. Saito, Nippon Kagaku Kaishi 1984,
Supporting Information (see footnote on the first page of this arti-
cle): Figures S1 and S2, additional electronic spectra; Figures S3a
and 3b, 1H NMR spectra for the water protons of
271–276.
[16] a) H. J. Callot, C. Piechocki, Tetrahedron Lett. 1980, 21, 3489–
3492; b) H. J. Callot, F. Metz, C. Piechocki, Tetrahedron 1982,
2522
www.eurjic.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2009, 2512–2523

Studies of a Bis(phosphane)RhIII–Porphyrin
38, 2365–2369; c) H. J. Callot, F. Metz, Nouv. J. Chim. 1985,
9, 167–171.
a) K. M. Kadish, Y. Hu, T. Boschi, P. Tagliatesta, Inorg. Chem.
1993, 32, 2996–3002; b) R. Guilard, K. M. Kadish, Chem. Rev.
1988, 88, 1121–1146.
103Rh NMR studies of RhI and RhIII complexes in general
include the following works: a) M. Soleilhavoup, L. Viau, G.
Commenges, C. Lepetit, R. Chauvin, Eur. J. Inorg. Chem. 2003,
2, 207–212; b) M. Herberhold, H. Yan, W. Milius, B. Wrackme-
yer, Chem. Eur. J. 2002, 8, 388–395; c) S. Haep, M. Nieger,
D. Gudat, M. Betke-Hornfeck, D. Schramm, Organometallics
2001, 20, 2679–2685.
R. G. Kidd, R. J. Goodfellow, in NMR and the Periodic Table
(Eds.: R. K. Harris, B. E. Mann), Academic Press, New York,
1978, pp. 195–278.
[43] K. Ruud, T. Helgaker, K. L. Bak, P. Jørgensen, H. J. A. Jensen,
J. Chem. Phys. 1993, 99, 3847–3859.
[44] a) O. Q. Munro, H. M. Marques, P. G. Debrunner, K. Mohan-
rao, W. R. Scheidt, J. Am. Chem. Soc. 1995, 117, 935–954; b)
O. Q. Munro, P. S. Madlala, R. A. F. Warby, T. S. Seda, G.
Hearne, Inorg. Chem. 1999, 38, 4724–4736.
[45] J. E. Carpenter, F. Weinhold, THEOCHEM 1988, 169, 41–62.
[46] F. A. Walker, in The Porphyrin Handbook (Eds.: K. M. Kadish,
K. M. Smith, R. Guilard), Academic Press, San Diego, 2000,
vol. 5, pp. 81–183.
[47] O. Q. Munro, J. C. Bradley, R. D. Hancock, H. M. Marques,
F. Marsicano, P. W. Wade, J. Am. Chem. Soc. 1992, 114, 7218–
7230.
[48] O. Q. Munro, J. A. Serth-Guzzo, I. Turowska-Tyrk, K. Mohan-
rao, T. K. Shokhireva, F. A. Walker, P. G. Debrunner, W. R.
Scheidt, J. Am. Chem. Soc. 1999, 121, 11144–11155.
[49] C. Sosa, J. Andzelm, B. C. Elkin, E. Wimmer, K. D. Dobbs,
D. A. Dixon, J. Phys. Chem. 1992, 96, 6630–6636.
[50] M. Bühl, M. Håkansson, A. H. Mahmoudkhani, L. Öhrstrom,
Organometallics 2000, 19, 5589–5596.
[17]
[18]
[19]
[20]
[21]
[22]
J. R. Houston, M. M. Olmstead, W. H. Casey, Inorg. Chem.
2006, 45, 7799–7805.
D. J. Law, G. Bigam, R. G. Cavell, Can. J. Chem. 1995, 73, 635–
642.
a) B. R. Bender, M. Koller, D. Nanz, W. von Philipsborn, J.
Am. Chem. Soc. 1993, 115, 5889–5890; b) L. Carlton, Magn.
Reson. Chem. 1997, 35,153–158; c) W. von Philipsborn, Chem.
Soc. Rev. 1999, 28, 95–105; d) M. Herberhold, H. Yan, W. Mi-
lius, B. Wrackmeyer, Chem. Eur. J. 2002, 8, 388–395.
[51] M. Bühl, Chem. Phys. Lett. 1997, 267, 251–257.
[52] X. Xu, L. Fang, Z.-X. Chen, G.-C. Yang, S.-L. Sun, Z.-M. Su,
J. Organomet. Chem. 2006, 691, 1927–1933.
[53] Y. Ruiz-Morales, T. Ziegler, J. Phys. Chem. A 1998, 102, 3970–
3976.
[23]
[24]
[25]
[26]
[27]
[54] K. Eichele, R. E. Wasylishen, J. F. Corrigan, N. J. Taylor, A. J.
Carty, K. W. Feindel, G. M. Bernard, J. Am. Chem. Soc. 2002,
124, 1541–1552.
[55] M. Cossi, G. Scalmani, N. Rega, V. Barone, J. Chem. Phys.
2002, 117, 43–54.
[56] G. H. Barnett, M. F. Hudson, K. M. Smith, J. Chem. Soc. Per-
kin Trans. 1 1975, 1401–1403.
[57] A. D. Adler, F. R. Longo, F. Kampas, J. Kim, J. Inorg. Nucl.
Chem. 1970, 32, 2443–2445.
[58] K. Harms, S. Wocadlo, XCAD4 – Program for Processing
CAD4 Diffractometer Data, University of Marburg, Germany,
1995.
[59] G. M. Sheldrick, SHELXS-97 and SHELXL-97 – Programs for
Crystal Structure Solution and Refinement, release 97-2, Uni-
versity of Göttingen, Germany, 1997.
[60] L. J. Farrugia, J. Appl. Crystallogr. 1999, 32, 837–838.
[61] A. Bax, R. H. Griffey, B. L. Hawkins, J. Magn. Reson. 1983,
55, 301–315.
[62] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Mont-
gomery, R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Mil-
lam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J.
Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pom-
elli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y.
Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe,
P. M. W. Gill, B. G. Johnson, W. Chen, M. W. Wong, J. L.
Andres, M. Head-Gordon, E. S. Replogle, J. A. Pople,
Gaussian 03, revision D.01, Gaussian, Inc., Pittsburgh PA,
2004.
[63] GausView 3.09, Graphical Interface to Gaussian 03, Gaussian
Inc., Carnegie Office Park, Building 6, Pittsburgh PA 15106,
USA, http://www.gaussian.com.
[64] L. J. Farrugia, J. Appl. Crystallogr. 1997, 30, 565–565.
[65] The root mean square fit employed the Rh atom, axial P atoms
and 24 porphyrin core atoms from each structure.
Received: August 21, 2008
T. Boschi, S. Licoccia, P. Tagliatesta, Inorg. Chim. Acta 1987,
126, 157–160.
W. R. Scheidt, Y. J. Lee, Struct. Bonding (Berlin) 1987, 64, 1–
70.
B. A. Kelly, A. J. Welch, P. Woodward, J. Chem. Soc., Dalton
Trans. 1977, 2237–2242.
E. Stulz, S. M. Scott, A. D. Bond, S. Otto, J. K. M. Sanders,
Inorg. Chem. 2003, 42, 3086–3096.
D. C. Thackray, S. Ariel, T. W. Leung, K. Menon, B. R. James,
J. Trotter, Can. J. Chem. 1986, 64, 2440–2446.
C. A. Tolman, Chem. Rev. 1977, 77, 313–348.
The relevant nonbonded contacts include C4a···C52 (3.197 Å),
N2···C52 (3.275 Å), C1a···C66 (3.334 Å), and N1···C61
(3.298 Å) for the uppermost axial ligand in Figure 1.
[28]
[29]
[30] O. Q. Munro, S. C. Shabalala, N. J. Brown, Inorg. Chem. 2001,
40, 3303–3317.
[31] A. J. Parusel, S. Grimme, J. Porphyrins Phthalocyanines 2001,
5, 1–8.
[32] I.-K. Choi, Y. Liu, Z. Wei, M. D. Ryan, Inorg. Chem. 1997, 36,
3113–3118.
[33] A. Bax, R. H. Griffey, B. L. Hawkins, J. Am. Chem. Soc. 1983,
105, 7188–7190.
[34] W. von Philipsborn, Chem. Soc. Rev. 1999, 28, 95–105.
[35] H. C. E. McFarlane, W. McFarlane, R. J. Wood, Bull. Soc.
Chim. Belg. 1976, 85, 864–871.
[36] L. Öhrström, Comments Inorg. Chem. 1996, 18, 305–323.
[37] T. H. Brown, P. J. Green, J. Am. Chem. Soc. 1970, 92, 2359–
2362.
[38] P. S. Pregosin, R. W. Kunz, “31P and 13C NMR of Transition
Metal Phosphine Complexes” in NMR 16, Basic Principles and
Progress (Eds.: P. Diehl, E. Fluck, R. Kosfeld), Springer-Ver-
lag, Berlin, 1979, pp. 47–49.
[39] C. J. Jameson, D. Rehder, M. Hoch, J. Am. Chem. Soc. 1987,
109, 2589–2594.
[40] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996,
77, 3865–3868.
[41] R. Ditchfield, W. J. Hehre, J. A. Pople, J. Chem. Phys. 1971, 54,
724–728.
[42] N. Godbout, D. R. Salahub, J. Andzelm, E. Wimmer, Can. J.
Chem. 1992, 70, 560–571.
Published Online: Aoril 29, 2009
Eur. J. Inorg. Chem. 2009, 2512–2523
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
2523