Ligand-centered redox activity in cobalt(II) and nickel(II) bis(phenolate)-dipyrrin complexes

Article abstract of DOI:10.1002/chem.201202882

One for all: A trianionic ligand containing the biologically relevant moieties phenolate and porphyrin was designed and synthesized. One-electron oxidation of the nickel and cobalt complexes of these ligands affords an unprecedented and highly stable hybrid porphyrinyl-phenoxyl radical bound to the metal center. Two-electron oxidation of these complexes leads to the M ²⁺-(close-shell two-electron oxidized ligand) species. Copyright

Full text of DOI:10.1002/chem.201202882

DOI: 10.1002/chem.201202882
Ligand-Centered Redox Activity in Cobalt(II) and Nickel(II)
Bis(phenolate)–Dipyrrin Complexes
Amꢀlie Kochem,[a] Linus Chiang,[b] Benoit Baptiste,[a] Christian Philouze,[a]
Nicolas Leconte,[a] Olivier Jarjayes,[a] Tim Storr,[b] and Fabrice Thomas*[a]
The chemistry of coordinated radicals has recently attract-
ed considerable interest.[1] This research has improved the
understanding of the interaction between transition metals
and non-innocent ligands, and spurred transition-metal cata-
lyst development incorporating redox-active ligands.[2]
Metal-coordinated ligand radicals are found naturally in the
active site of a number of metalloenzymes and the most
ubiquitous ones are porphyrinyls (as in cytochrome P450)[3]
and tyrosinyls (as in galactose oxidase).[4] These two classes
of radicals are formed by oxidation of either coordinated co-
factors or amino-acids and exhibit distinct properties, stabili-
ties, and generation pathways. In principle, one-electron oxi-
dation of a Mn+–L species involving a redox-active ligand
both ligand families, but their electronic structures remain
speculative because of their high reactivity.
Bloom and Garcia in the early 1970s introduced a fasci-
nating class of ligands that combine a “half-porphyrin” (di-
pyrrin) and two phenol moieties in a single and highly elec-
tron-rich trianionic ligand.[9] The use of this ligand system
has increased recently, especially for catalytic purposes.[10,11]
Although this ligand architecture may support an extraordi-
narily rich ligand-based redox activity, no data supporting
this fact has yet been reported in the literature, and oxida-
tion catalysis is proposed to occur through metal-centered
processes.[10,11]
We present in this paper the first evidence for ligand-cen-
tered, as opposed to metal-centered oxidative redox activity
leads to either M(n+1)+–L or the species Mn+–L , where the
C
radical is centered on the ligand. Modeling studies on repre-
sentative Mn+–phenolate and Mn+–porphyrin systems reveal
that several interdependent factors such as the solvent, tem-
perature, nature of the metal ion, and substituents may in-
fluence the site of oxidation. As an example, square-planar
Ni2+–bis(phenolate)[5] and Ni2+–porphyrin[6] complexes un-
dergo a reversible one-electron oxidative electron transfer,
which is assigned to the Ni3+/Ni2+ redox couple in coordinat-
in the bisACTHNGUTRENNU(G phenol)–dipyrrin ligand family. For that purpose
we investigated the oxidative chemistry of the bis(di-tert-
butyl-phenol)–dipyrrin ligand, H3L, coordinated to Ni2+ and
Co2+ metal ions (Scheme 1). The one-electron (a radical)
C
ing solvent, whereas it corresponds to the L /L redox couple
in CH2Cl2. Further, the Co2+ ion is oxidized more easily
than coordinated phenolates but less easily than porphyrin
macrocycles in CH2Cl2. Consequentely one-electron oxi-
dized Co2+–salen complexes exhibit a marked Co3+–bis-
ACHTUNGTRENNUNG
(phenolate) character,[7] whereas oxidized Co2+–porphyrin
complexes exhibit a Co2+–(porphyrinyl radical) character.[8]
Interestingly, two-electron oxidized species are accessible in
Scheme 1. Bis(di-tert-butyl-phenol)–dipyrrin ligand, H3L, and metal com-
plexes thereof.
and two-electron oxidized forms of these complexes were
structurally characterized by X-ray diffraction and their
electronic structure confirmed by spectroscopy and DFT cal-
culations.
[a] A. Kochem, Dr. B. Baptiste, Dr. C. Philouze, Dr. N. Leconte,
Dr. O. Jarjayes, Prof. F. Thomas
Dꢀpartement de Chimie Molꢀculaire
Chimie Inorganique Redox Biomimꢀtique (CIRE)
UMR CNRS 5250, Universitꢀ Joseph Fourier
B. P. 53, 38041 Grenoble Cedex 9, France
Fax : (+33)47651-4836
The bis(di-tert-butyl-phenol)–dipyrrin ligand, H3L, was
synthesized according to a published procedure.[11] Mixing
H3L with either NiACHTUNTRGENN(UG OAc)2·4 H2O or CoCAHTUNTGERN(NUGN OAc)2·4 H2O in the
E-mail: Fabrice.Thomas@ujf-grenoble.fr
presence of air results in the formation of 1 and 2, respec-
tively. Both complexes were isolated and crystallized in the
neutral form, as demonstrated by the absence of a counter
ion in the crystal cell. The structure of 1 is depicted in
Figure 1 whereas the structurally similar complex, 2, is
shown in the Supporting Information, Figure S1.
[b] L. Chiang, Prof. T. Storr
Simon Fraser University
8888 University Drive, Burnaby
British Columbia V5A-1S4 (Canada)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202882.
14590
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 14590 – 14593

COMMUNICATION
Figure 2. Mean bond lengths in one half of the ligand: (a) titanium, man-
ganese, and aluminum complexes of a trianionic bis(phenolate)—dipyrrin
ligand that does not have t-butyl groups;[9,10] (b) 1 and 2; (c) 1+; (d) 2+.
Unless a length range is indicated the values have an error of ꢁ0.01 ꢂ.
(e) Simplified canonical forms of one phenol–pyrrole radical unit (full
delocalization occurs over the second unit).
Figure 1. ORTEP views of (a) 1 and (b) 2+; thermal ellipsoids are shown
at 30% probability and hydrogen atoms are omitted for clarity, except
those of the water molecule.
from oxidation of the ligand could be so clearly evidenced
in 1 and 2: Extensive delocalization in the radical complexes
[13]
C+[12]
C+
Ni—OEP
and Ni–(di-tert-butyl-salen) , which belong
to rare Ni2+—(ligand-centered radical) complexes of the
two families structurally characterized, attenuates the struc-
tural variations resulting from oxidation of the neutral pre-
cursors.
Owing to the trianionic nature of the fully deprotonated
ligand L3ꢀ the isolated complexes involve either a M2+-coor-
dinated ligand radical or a M3+–bis(phenolate) species. The
fact that the metal center resides in a square-planar geome-
try in 1 and 2 rules out a formal +3 oxidation state for the
metal ion. The ligand thus exhibits radical character, and
The 100 K EPR spectrum of 1 displays an anisotropic (S=
= ) signal at g1 =2.025, g2 =2.008, and g3 =1.975 (gav =2.003),
1
2
thus being consistent with a radical character of the ligand.
Interestingly, the g value is lower for 1 than for Ni—(radical
salen) complexes involving similar di-tert-butyl-phenol moi-
eties.[13,14] These data highlights a much smaller Ni contribu-
tion to the SOMO (calculated as ca. 6% versus 16% for
C2ꢀ
ꢀ
will be referred to as L . Detailed investigations on the C
ꢀ
ꢀ
C, C N, and C O bond lengths indicate that the two halves
of the ligand exhibit very similar structural parameters, con-
sistent with a radical that is delocalized. Notably, the bond
lengths of the ligand show little change upon metal-ion sub-
C2ꢀ
C+
Ni–(di-tert-butyl-salen) complexes,[13] and shows that the
stitution. Comparison of the L
bond lengths with the
radical form of bis(di-tert-butyl-phenol)–dipyrrin ligand is
more accessible than that of salen ligands. Complex 2 exhib-
its magnetic coupling between the radical and the Co2+
spins, thus resulting in EPR silence at 100 K. Interestingly,
DFT calculations on 2 predict the triplet state (S=1) to be
only 0.4 kcalmolꢀ1 lower in energy in comparison to the
broken symmetry (BS)[15] state, suggesting that the high spin
and antiferromagnetically-coupled spin states are similar in
energy.
The Vis/NIR spectra of 1 (Figure 3) and 2 (see the Sup-
porting Information, Figure S8) are dominated by an intense
transition at the approximate range 600–700 nm, as well as
lower intensity transitions in the region, 1000–1400 nm.
Time-dependent density-functional theory (TD-DFT) calcu-
lations were used to gain insight into the nature of the low-
energy transitions of 1 and 2.[16] For 1, two low-energy bands
were predicted at 1090 nm and 1260 nm (Figure 3), thus
matching the envelope of low-energy transitions in the ex-
perimental spectrum. Both transitions involve orbitals delo-
calized over the ligand radical framework with very little Ni
character (see the Supporting Information, Figure S12), and
values published for the Ti4+, Mn3+, and Al3+ complexes in-
volving nonoxidized trianionic bis(phenolate)–dipyrrin li-
gands is very instructive.[10,11] From the mean bond lengths
listed in Figure 2 (individual bond lengths for 1 and 2 are
given in the Supporting Information, Table S1) several es-
ꢀ
sential features are notable: The C1 O1 (and the corre-
ꢀ
sponding bond on the other side of the metal, C20 O2),
ꢀ
ꢀ
ꢀ
ꢀ
C2 C15 (and C21 C34), and C16 C17 (and C35 C36)
bond lengths are shorter in the ligand radical. In contrast,
the C1 C6 (and C20 C25) bond is longer in L 2ꢀ. These
structural parameters, which are nicely predicted by DFT
calculations (see the Supporting Information, Figure S7), are
clear evidence that the radical is delocalized among the
ꢀ
ꢀ
C
C2ꢀ
phenol and pyrrole rings (Figure 2). L thus exhibits unpre-
cedented mixed porphyrinyl–phenoxyl radical character.
The predicted spin densities for both 1 and 2 (see the Sup-
porting Information, Figure S10 and S11) are symmetrically
distributed with significant spin density on the dipyrrin
(30%, 1; 20%, 2) and phenolate (65%, 1; 60%, 2) moieties.
It is remarkable that the changes in bond distances resulting
Chem. Eur. J. 2012, 18, 14590 – 14593
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14591

F. Thomas et al.
Scheme 2. Possible formulations for the chelated and two-electron oxi-
ꢀ
dized bis(di-tert-butyl-phenol)–dipyrrin ligand: (a) closed shell, M–L
and (b) diradical, M–LDꢀ.
Figure 3. Vis/NIR spectra of 1 (solid lines) and 1+ (dotted lines) in
CH2Cl2 solution. T=298 K. Insert: b-HOMO!b-LUMO transition
which contributes to the NIR band of 1 (lclcd =1270 nm, f=0.072). Pre-
dicted transitions are shown by vertical bars (1, black bars; 1+, grey
bars).
Notably, the monoanionic ligand in 1+ may exist as either a
diradical, LDꢀ, or as a closed shell, Loxꢀ (Scheme 2). Although
the crystal structure supports both forms, the DFT-based en-
ergetic analysis of 1+ predicts that the closed-shell singlet
state (S=0) is approximately 12 kcalmolꢀ1 lower in energy
than the triplet state (S=1). Broken-symmetry optimiza-
tions for 1+ collapsed to a singlet state in all cases, suggest-
ing that the closed-shell form, Loxꢀ, form is lower in energy
than the antiferromagnetically coupled diradical state, LDꢀ.
The structure of 2+ (Figure 1) contrasts sharply with the
previously described compounds because the metal ion
exists in a square pyramidal geometry containing an axially
are thus assigned as ligand-to-ligand charge-transfer (LLCT)
transitions. A TD-DFT calculation on 2 (triplet state) pre-
dicts a low-energy transition at 940 nm, which is in good
agreement with the broad experimental feature at approxi-
mately 1000 nm. The low-energy transition for 2 involves or-
bitals delocalized over the ligand-radical framework with a
small Co contribution (see the Supporting Information, Fig-
ure S14).
ꢀ
The cyclic-voltammetry curves of 1 and 2 in CH2Cl2 (with
coordinated water molecule with a Co O3 bond distance of
1
0.1m TBAP) display a reversible reduction wave at E1/2
=
2.252 ꢂ. The discrimination between a (+2) and (+3) oxi-
dation level of the Co ion is therefore not possible solely on
the basis of the geometry at the metal center. Further ex-
amination of the bond lengths within the ligand framework
ꢀ0.48 and ꢀ0.61 versus Fc/Fc+, respectively; this wave is as-
signed to the L 2ꢀ/L3ꢀ redox couple. The potentials required
C
for the formation of the ligand radical in 1 and 2 are thus
much lower in comparison to M2+–(di-tert-butyl-salen), M2+
–TPP, and M2+–OEP complexes.[4,5,13,14] This result could be
rationalized by considering the additional negative charge of
the deprotonated bis(di-tert-butyl-phenol)–dipyrrin ligand
(ꢀ3 instead of ꢀ2 in its initial nonradical form), which coun-
teracts the electron deficiency common to oxidized radical
complexes. When scanning towards positive potentials a re-
versible one-electron redox process is observed at E1/22 = +
0.03 V versus Fc/Fc+ for 1, whereas two reversible waves
are observed for 2 at E1/22 =ꢀ0.01 V and E1/23 = +0.58 V.
The former is attributed to a second ligand-centered redox
process, as demonstrated below.[17]
shows an overall distribution very similar to the one ob-
+
ꢀ
ꢀ
served for 1 , although the C1 O1 and C20 O2 bond
lengths (1.285 and 1.297 ꢂ, respectively) are intermediate
between those of 2 and 1+. The EPR spectrum of 2+ at
100 K displays a major anisotropic (S=1= ) signal at gk =
2
2.005 and g? =2.258 with cobalt hyperfine structure
(Ak(Co) =10.4 mT and A?(Co) =2.0 mT), which is comparable
to other Co2+ complexes involving a closed-shell ligand.[18,19]
In agreement with the experimental data, DFT calculations
ꢀ
of 2+ predict that the singlet state (S=1/2) of Co2+–Lox is
lower in energy by approximately 11 kcalmolꢀ1.
The Vis/NIR spectra are dominated by intense transitions
at 432 (18480), 760 (16050), and 904 nm (10540 Mꢀ1cmꢀ1)
for 1+ (Figure 3), and 508 (13 750) and 864 nm
(17410 Mꢀ1cmꢀ1) for 2+. In both cases the UV/Vis bands
shift to higher energy upon oxidation to the cations 1+ and
2+. TD-DFT calculations on the closed-shell singlet state
(S=0) for 1+ predicts three bands in the 600–950 nm range,
thus matching the experiment data, in which there was a
shift to higher energy upon oxidation (Figure 3). The pre-
dicted transitions are assigned as delocalized LLCT bands
(see the Supporting Information, Figure S13). TD-DFT cal-
culations on the oxidized 5-coordinate Co complex, Co2+
The cations 1+ and 2+ were generated chemically by oxi-
dation with AgSbF6 (E8=0.65 V) in CH2Cl2 and crystallized
by slow diffusion of pentane into the solution. The metal ion
exists in a square-planar geometry in 1+, similar to 1, sug-
gesting that the Ni atom remains in the +2 oxidation state,
and indicating that the ligand is two-electron oxidized.
ꢀ
ꢀ
ꢀ
When comparing the C C, C N, and C O bond lengths of
1+ with the corresponding values measured for 1, a trend
similar to that obtained on going from the nonoxidized L3ꢀ
derivatives[10,11] to 1 is observed. The C1 O1, C20 O2, C2
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
C5, C21 C34, C17 C17, and C36 C37 bonds have thus a
more marked double-bond character. In addition, the bond
lengths within the rings of each half of the ligand are very
similar, thus showing that the ligand is highly conjugated.
ꢀ
–Lox (2+), also predicts three bands in the 600–950 nm
range. The predicted bands for 2+ are similar to those pre-
dicted for 1+, thus providing further evidence that the
14592
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 14590 – 14593

Ligand-Centered Redox Activity
COMMUNICATION
ligand exists in the closed-shell form, Loxꢀ, in both 1+ and
2+.
[1] P. Chaudhuri, K. Wieghardt, Prog. Inorg. Chem. 2001, 50, 151; F.
Thomas in Stable Radicals: Fundamentals and Applied Aspects of
Odd-Electron Compounds (Ed.: R. G. Hicks), John Wiley and Sons,
Chichester, 2010, pp. 281–316.
[2] P. J. Chirik, K. Wieghardt, Science 2010, 327, 794.
[3] J. Rittle, M. T. Green, Science 2010, 330, 933.
In conclusion, we have demonstrated by combined X-ray
diffraction, spectroscopy, and theoretical calculations that
the oxidative redox processes occur at the ligand and not at
the metal center in Ni2+ and Co2+ complexes containing the
bis(di-tert-butyl-phenolate)–dipyrrin ligand, L3ꢀ, to give M2+
[4] J. W. Whittaker, Chem. Rev. 2003, 103, 2347.
[5] M. Orio, O. Jarjayes, H. Kanso, C. Philouze, F. Neese, F. Thomas,
Angew. Chem. 2010, 122, 5109; Angew. Chem. Int. Ed. 2010, 49,
4989.
[6] K. M. Kadish, E. van Caemelbecke, G. Royal in The Porphyrin
Handbook Vol. 8, Electron Transfers (Eds.: K. M. Kadish, K. M.
Smith, R. Guilard), Academic Press, London, 2000, pp. 1–97.
[7] L. P. C. Nielsen, C. P. Stevenson, D. G. Blackmond, E. N. Jacobsen,
J. Am. Chem. Soc. 2004, 126, 1360.
C2ꢀ
–L and M2+–Loxꢀ. Full delocalization of the SOMO gives
the ligand an unprecedented hybrid porphyrinyl–phenoxyl
C2ꢀ
radical character and a highly stable L form. In addition,
both the extensive conjugation and negative charge of L3ꢀ
greatly facilitates oxidation of the complexes, an oxidation
that is more facile than that of salen and porphyrin com-
pounds. The fact that the ligand can be easily accept and re-
lease electrons, combined with its ability to spontaneously
oxidize into a radical in air suggests fascinating applications
in sensing, spintronics, and radical catalysis.[1,2]
[8] A. Salehi, W. A. Oertling, G. T. Babcock, C. K. Chang, J. Am.
Chem. Soc. 1986, 108, 5630.
[9] S. M. Bloom, P. P. Garcia, US Pat. 3691161, 1972.
[10] C. Ikeda, S. Ueda, T. Nabeshima, Chem. Commun. 2009, 2544; S. El
Ghachtouli, K. Wojcik, L. Copey, F. Szydlo, E. Framery, C. Goux-
Henry, L. Billon, M.-F. Charlot, R. Guillot, B. Andrioletti, A. Au-
kauloo, Dalton Trans. 2011, 40, 9090. S. Rausaria, A. Kamadulski,
N. P. Rath, L. Bryant, Z. Chen, D. Salvemini, W. L. Neumann, J.
Am. Chem. Soc. 2011, 133, 4200.
Experimental Section
[11] K. Nakano, K. Kobayashi, K. Nozaki, J. Am. Chem. Soc. 2011, 133,
10720.
[12] H. Song, R. D. Orosz, C. A. Reed, W. R. Scheidt, Inorg. Chem. 1990,
29, 4274.
[13] T. Storr, E. C. Wasinger, R. C. Pratt, T. D. P. Stack, Angew. Chem.
2007, 119, 5290; Angew. Chem. Int. Ed. 2007, 46, 5198.
[14] Y. Shimazaki, F. Tani, K. Fului, Y. Naruta, O. Yamauchi, J. Am.
Chem. Soc. 2003, 125, 10512.
[15] L. Noodleman, J. Chem. Phys. 1981, 74, 5737. L. Noodleman, E. R.
Davidson, Chem. Phys. 1986, 109, 131. L. Noodleman, D. A. Case,
Adv. Inorg. Chem. 1992, 38, 423.
The neutral complexes, 1 and 2, were obtained by mixing the ligand with
the appropriate metal acetate in air and then allowing the complexes to
crystallize by slow evaporation of a either a concentrated ethanolic or
ꢀ
ꢀ
methanolic solution. Compounds 1+·SbF6 and 2+·SbF6 were obtained
by addition of one equivalent of AgSbF6 to a CH2Cl2 solution of the neu-
tral precursor followed by crystallization by diffusion of pentane into the
solution. Full characterization data for the compounds, experimental de-
tails, instrumentation, crystallographic details, and computational details
are given in the Supporting Information.
[16] M. E. Casida in Recent Advances in Density Functional Methods
(Ed.: D. P. Chong), World Scientific, Singapore, 1995, pp. 155; R. E.
Stratmann, G. E. Scuseria, M. J. Frisch, J. Chem. Phys. 1998, 109,
8218.
3
[17] The oxidation process at E1/2 for 2 is assigned mainly to the Co3+
/Co2+ redox couple. It results in the formation of an EPR-silent
Co3+ complex that is coordinated to a two-electron oxidized closed-
shell ligand.
Acknowledgements
This work is supported by an NSERC Discovery Grant to T. S. from
Canada, and a doctoral fellowship to A. K. from the DCM. L. C. thanks
the Governments of Canada and France for an exchange scholarship.
[18] A minor signal (less than 5%) could be detected at giso =2.005 (see
the Supporting Information, Figure S5–6), which is assigned to a
contaminant monoradical species.
[19] A. Wolberg, J. Manassen, J. Am. Chem. Soc. 1970, 92, 2982.
Keywords: cobalt · dipyrrin · nickel · porphyrins · redox
Received: August 10, 2012
chemistry
Published online: October 5, 2012
Chem. Eur. J. 2012, 18, 14590 – 14593
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14593