How Hydrogen Bonds Affect Reactivity and Intervalence Charge Transfer in Ferrocenium-Phenolate Radicals

Article abstract of DOI:10.1002/ejic.201501471

The ferrocenyl-phenol 2,4-di-tert-butyl-6-(ferrocenylcarbamoyl)phenol (H-1) forms intramolecular hydrogen bonds which are absent in its constitutional isomer 2,6-di-tert-butyl-4-(ferrocenylcarbamoyl)phenol (H-2). Their corresponding bases 1^-and 2^-show intra- and intermolecular NH?O hydrogen bonds, respectively. The phenolate 1^- is reversibly oxidized to 1^·, whereas 2^- only undergoes a quasi-reversible oxidation to 2^·, which suggests a higher reactivity. The radical pools of 1^· and 2^· formed by the oxidation/deprotonation of H-1 and H-2 have been probed by (rapid-freeze) electron paramagnetic resonance (EPR) spectroscopy and by spin-trapping techniques to elucidate the types of radicals present. Ferrocenium phenolate [1a]^· featuring an NH?O intramolecular hydrogen bond is the most stable radical and undergoes thermal and photoinduced valence isomerization to the phenoxyl radical valence isomer [1b]^· with participation of the NH stretching mode (proton-coupled electron transfer). A ferrocenium iminolate radical [1c]^· is present as well and equilibrates with the carbon-centered ferrocenyl radicals [1^Cp]^· and [1^β]^·. The latter radicals are intercepted by nitrobenzene to give the corresponding stable nitroxyl radicals [6^Cp]^· and [6^β]^·. All the radicals 2^·, which lack intramolecular hydrogen bonds, are transient in nature due to rapid follow-up reactions. However, rapid-freeze EPR spectroscopy indicated the presence of ferrocenium iminolate [2c]^·, the phenoxyl radical [2b]^·, and/or carbon-centered radicals [2^Cp]^· and [2^β]^·. The carbon-centered radicals [2^Cp]^· and [2^β]^· are selectively trapped as the corresponding nitroxide radicals [7^Cp]^· and [7^β]^·. These diverse reactivity patterns are relevant for cytostatic ferrocenyl-phenols such as ferrocifen.

Full text of DOI:10.1002/ejic.201501471

DOI: 10.1002/ejic.201501471
Full Paper
Metallocenes
How Hydrogen Bonds Affect Reactivity and Intervalence Charge
Transfer in Ferrocenium-Phenolate Radicals
Andreas Neidlinger,[a] Christoph Förster,[a] and Katja Heinze*[a]
Abstract: The ferrocenyl-phenol 2,4-di-tert-butyl-6-(ferrocenyl- isomer [1b]· with participation of the NH stretching mode (pro-
carbamoyl)phenol (H-1) forms intramolecular hydrogen bonds ton-coupled electron transfer). A ferrocenium iminolate radical
which are absent in its constitutional isomer 2,6-di-tert-butyl-4- [1c]· is present as well and equilibrates with the carbon-cen-
(ferrocenylcarbamoyl)phenol (H-2). Their corresponding bases tered ferrocenyl radicals [1Cp]· and [1ꢀ]·. The latter radicals are
1– and 2– show intra- and intermolecular NH···O hydrogen intercepted by nitrobenzene to give the corresponding stable
bonds, respectively. The phenolate 1– is reversibly oxidized to nitroxyl radicals [6Cp]· and [6ꢀ]·. All the radicals 2·, which lack
1·, whereas 2– only undergoes a quasi-reversible oxidation to intramolecular hydrogen bonds, are transient in nature due to
2·, which suggests a higher reactivity. The radical pools of 1· rapid follow-up reactions. However, rapid-freeze EPR spectro-
and 2· formed by the oxidation/deprotonation of H-1 and H- scopy indicated the presence of ferrocenium iminolate [2c]·, the
2 have been probed by (rapid-freeze) electron paramagnetic phenoxyl radical [2b]·, and/or carbon-centered radicals [2Cp ·
resonance (EPR) spectroscopy and by spin-trapping techniques and [2ꢀ]·. The carbon-centered radicals [2Cp]· and [2ꢀ]· are se-
]
to elucidate the types of radicals present. Ferrocenium phenol- lectively trapped as the corresponding nitroxide radicals [7Cp ·
]
ate [1a]· featuring an NH···O intramolecular hydrogen bond is and [7ꢀ]·. These diverse reactivity patterns are relevant for cyto-
the most stable radical and undergoes thermal and photoin- static ferrocenyl-phenols such as ferrocifen.
duced valence isomerization to the phenoxyl radical valence
Introduction
ferrocene-free salicylamide H-4 (Scheme 1, b). The deprotona-
tion and oxidation of H-1 yields the ferrocenium phenolate
zwitterion [1a]·, the valence isomeric ferrocenyl phenoxyl radi-
cal [1b]·, and the ferrocenium iminolate tautomer [1c]·. In the
latter tautomer, the OH···O IHB is retained whereas an NH···O
IHB is established in the valence isomers [1a]· and [1b]·.
Electron-transfer (ET) reactions continue to be a focus of inter-
est due to their importance in a plethora of chemical and bio-
logical processes.[1–7] Combination with proton-transfer (PT)
reactions facilitates ET (proton-coupled electron transfer,
PCET).[8–18] As a prominent example, in photosystem II, the
tyrosyl residue (phenoxyl radical) is reduced to phenol(ate) by
the oxygen-evolving complex through PCET with the assistance
of a nearby histidine base as proton shuttle.[19–24] This PCET is
a key process in the stepwise oxidation of the oxygen-evolving
complex, which finally results in the oxidation of water to dioxy-
gen at the oxygen-evolving complex, which is itself based on
PCET processes.[19–24]
We have been interested in small model systems comprising
a phenoxyl radical[25] and ferrocene[26–30] (and in the corre-
sponding ferrocenium/phenolate valence isomers) to investi-
gate PCET intervalence charge-transfer (IVCT)[31] reactions in
redox-asymmetric systems.[16] In the parent ferrocenyl-phenol
conjugate H-1,[16] an intramolecular hydrogen bond (IHB) is
formed between the phenolic OH group and the carbonyl oxy-
gen of the amide bond (OH···O IHB, Scheme 1, a), similarly to
The IHB of the zwitterion [1a]· is highly stabilized with re-
spect to the IHB of the phenoxyl radical [1b]· for two reasons:
First, the phenolate is a better hydrogen-atom acceptor than a
phenoxyl radical, which is also observed for the IHB of the ferro-
cene-free reference phenolate/phenoxyl pair 4– and 4·
(Scheme 1, b).[25] Secondly, the ferrocenium-amide is a stronger
hydrogen-atom donor than the ferrocenyl-amide due to the
positive charge. This phenomenon has been previously ob-
served, for example, in oligoferrocene amides H-A and the cor-
responding mixed-valent systems [H-A]·+ (Scheme 2, a).[28,29]
Hence, the zwitterion [1a]· is considerably stabilized with re-
spect to [1b]· (by 30 kJ mol–1 according to DFT calculations).[16]
The proton in the IHB is involved in the thermal intramolecular
ET between [1a]· and [1b]·, with the proton being closer to the
oxygen atom in [1a]· and closer to the nitrogen atom in [1b]·.
Optical ET (IVCT absorption band at 1040 nm) from the ground
state of [1a]· to give the metastable phenoxide [1b]· can only
occur when the proton is already closer to the nitrogen atom,
as found in [1b]·. Hence, the ET is coupled with the NH vibra-
tional mode of the [1a]·/[1b]· valence isomers.
[a] Institute of Inorganic and Analytical Chemistry, Johannes Gutenberg
University,
Duesbergweg 10–14, 55128 Mainz, Germany
E-mail: katja.heinze@uni-mainz.de
http://www.ak-heinze.chemie.uni-mainz.de/
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/ejic.201501471.
Charges stabilize the IHB and vice versa the hydrogen bond
affects the redox potentials of the redox centers connected by
Eur. J. Inorg. Chem. 2016, 1274–1286
1274
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 2. a,b) IHBs in the ferrocene/ferrocenium amide redox couples H-A/
[H-A]·+ and H-B/[H-B]·– and c) ET and PT from the ferrocenyl-phenol ferroci-
fen (H-C).
Scheme 1. Hydrogen bonding in salicyl amides H-1, H-4, and H-4Me, their
corresponding phenolates 1–, 4–, and [4Me]–, and their corresponding
phenoxyl radicals 1·, 4·, and [4Me · [16,25]
] .
one dyad H-B (Scheme 2, b).[32] Upon reduction of the quinone
to the semi-quinonate [H-B]·–, the NH···O IHB is strengthened
the IHB. Large potential differences are observed between and the semiquinone radical anion is stabilized. Similarly, the
hydrogen-bonded and non-hydrogen-bonded phenolate/phen- irradiation of H-B at 388 nm yields the charge-separated state
oxyl couples such as 4–/4· (E1/2 = –160 mV vs. FcH/FcH+ in [H-B]* with ferrocenium-semiquinonate character effecting a
CH3CN/[nBu4N][BF4])[25] and [4Me]–/[4Me]· (E1/2 = –425 mV vs. strong NH···O IHB. Owing to the stabilizing effect of this IHB,
FcH/FcH+ in CH3CN/[nBu4N][BF4];[25] Scheme 1, b, c).[25] Simi- the lifetime of this zwitterionic state [H-B]* is quite high
larly, the hydrogen-bonded ferrocenyl-NH moiety in H-A (E1/2
=
(Scheme 2, b).[32]
–115 mV vs. FcH/FcH+ in CH3CN/[nBu4N][PF6])[27,28] is oxidized
Ferrocenyl-phenols such as ferrocifen (H-C; Scheme 2, c)
have attracted considerable interest as potent anticancer
drugs.[33] Indeed, oxidation and deprotonation yields the meso-
meric phenoxyl radical/C-centered radical [Cb]· and the con-
at a lower potential than N-acetylaminoferrocene H-3 (E1/2
=
–50 mV vs. FcH/FcH+ in CH2Cl2/[nBu4N][B(C6F5)4];[26,27,29,30]
Scheme 2, a).
Another intriguing example of this kind of PCET with the ceivable ferrocenium phenolate zwitterionic valence isomer
crucial NH···O IHB is found for the amide-linked ferrocenyl-quin- [Ca]·, both lacking IHBs.
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1275
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
In this context, we recently succeeded in capturing transient corresponding ferrocenyl-phenols H-1 and H-2 in the presence
carbon-centered ferrocenyl radicals, as proposed for [Cb]·, by of base and oxidant was probed by the spin-trapping tech-
using nitrosobenzene (PhNO) as spin trap.[26] Relevant to the nique. The possible presence of different radical types is rele-
present study is the formation of trace amounts of C-centered vant for the reactivity of cytostatic drugs based on the ferro-
radicals [3x]· derived from N-acetylaminoferrocene (H-3) in the cenyl-phenol motif, such as ferrocifen (H-C).
presence of a base and an oxidant, in addition to the more
stable ferrocenium iminolate radical [3c]· (Scheme 3). The EPR
Results and Discussion
spectra of the nitroxyl radicals [8x]· resulting from the trapping
Synthesis and Characterization of Ferrocenyl Phenols H-1
of C-centered radicals [3x]· by PhNO revealed the preferred site
and H-2 and Reference Phenols H-4 and H-5
of attack, namely on the unsubstituted cyclopentadienyl ring
and the ꢀ position with respect to the acetylamino group
([8Cp]·, [8ꢀ]·; Scheme 3).[26] The ferrocenium iminolate [3c]· does
not react with PhNO, neither at the nitrogen atom nor at the
metal center.[26]
The phenols H-1 and H-4 were prepared according to literature
procedures[16,25] (Scheme 1). The constitutional isomer of H-1,
2,6-di-tert-butyl-4-(ferrocenylcarbamoyl)phenol (H-2; Scheme 4,
a), was obtained by condensation of aminoferrocene[27] with
the commercially available benzoic acid derivative 3,5-di-tert-
Scheme 3. Radical reactivity of [H-3]·+ in the presence of a base and PhNO.[26]
In this study we addressed the effect of the NH···O IHB in 1·
on the reactivity and distribution of valence isomers (ferroce-
nium phenolate zwitterion [1a]· vs. phenoxyl radical [1b]·) and
tautomers (ferrocenium iminolate [1c]·, C-centered radicals
[1x]·). To eliminate all the IHBs while essentially maintaining the
overall geometry and electronic structure, the phenolic OH
group is formally shifted from the 2-position in H-1 to the 4-
position in the constitutional isomer H-2 (Scheme 4). The prop-
erties of phenol H-2, its conjugate base 2–, its oxidation product
[H-2]·+, and the conceivable radical species [2a]·, [2b]·, [2c]·,
and [2x]· were studied by (rapid-freeze) EPR spectroscopy and
DFT calculations and contrasted with the corresponding species
derived from H-1. The potential presence of the C-centered
radicals [1x]· and [2x]· (in analogy to [3x]·) derived from the
Scheme 4. Generation of radicals from a) H-2 and b) H-5 (atomic numbering
for NMR assignment given).
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1276
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
butyl-4-hydroxybenzoic acid with HATU (1-[bis(dimethylamino)-
The presence of the OH···O IHB in H-1, as suggested by NMR
methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexa- and IR studies in solution,[16] was further corroborated by sin-
fluorophosphate) as coupling reagent. The ferrocene-free refer- gle-crystal XRD analysis. H-1 crystallizes from ethyl acetate at
ence phenol H-5 [2,6-di-tert-butyl-4-(2,4,6-trimethylphenyl- 4 °C in the monoclinic space group P21/c (Figure 1). The ferro-
carbamoyl)phenol, Scheme 4, b] was accessible by an analo- cenyl moiety shows the usual structural features. The OH···O
gous coupling of the same benzoic acid derivative with 2,4,6- IHB has an O1···O2 distance of 2.580(3) Å. Furthermore, inter-
trimethylaniline. Both phenols were fully characterized by NMR molecular hydrogen bonds between the amide proton and the
and IR spectroscopy as well as by mass spectrometry and ele- carbonyl oxygen O1′ of neighboring molecules of H-1 lead to
mental analysis.
the typical hydrogen-bonded chain of amides in the solid state
with an N1···O1′ distance of 3.220(3) Å.[27,28] Additional crystal-
lographic details are given in the Table S1 and Figure S2 in the
Supporting Information.
The molecular ion peak of H-2 is at m/z = 433 in the FD mass
1
spectrum. The H NMR spectrum of H-2 displays the expected
number of resonances with correct integral ratios. Owing to the
low solubility of H-2 in dichloromethane, more detailed NMR
studies (13C NMR, 2D NMR) were also conducted in [D8]THF. In
this coordinating solvent, the NH and OH protons resonate at
δNH = 8.50 ppm and δOH = 6.59 ppm, respectively. These two
resonances were unambiguously assigned by NOE spectroscopy
(see Figure S1a in the Supporting Information). Contacts be-
tween the NH proton and the α-ferrocenyl protons 2,5-H (δ =
4.71 ppm) as well as the aromatic protons 7-H (δ = 7.73 ppm)
are observed (Scheme 4). As expected, the OH proton shows a
contact with the tBu groups (δ = 1.47 ppm). These chemical
shifts suggest the absence of hydrogen bonds in H-2, in con-
trast to its constitutional isomer H-1, which features an OH···O
Figure 1. Molecular structure of H-1 in the solid state. Thermal ellipsoids are
drawn at the 50 % probability level. CH hydrogen atoms have been omitted
for clarity.
IHB (δNH
= 9.01 ppm, δOH = 13.37 ppm in [D8]THF;
Scheme 1).[16,35] Similarly, in CD2Cl2, the resonances for the NH
and OH protons are found at δNH = 7.09 and 7.36 ppm and
δ
OH = 5.63 and 12.69 ppm for H-2 and H-1, respectively. Hence,
Radical Formation and Radical Characterization
weak intermolecular hydrogen bonds from NH/OH to THF are
likely, but IHBs of any kind are clearly absent in H-2.
The PT and ET reactions of H-1/H-2 and their corresponding
The IR spectrum of H-2 in CH2Cl2 confirms the absence of reference phenols H-4/H-5 are depicted in Schemes 1 and 4.
IHBs involving NH or OH groups. The absorption for the NH The chemical oxidations were performed by using silver hexa-
stretching vibration is observed at νNH = 3440 cm–1, similarly to fluoroantimonate (AgSbF6) as oxidant [E1/2(CH2Cl2) = 650 mV vs.
˜
those of H-3 (νNH
=
3435 cm–1)[27] and H-1 (νNH
=
FcH/FcH+],[36] which is capable of oxidizing the ferrocenyl as
˜
˜
3450 cm–1),[16] characteristic of a free NH group. The free OH well as the phenolate unit. Deprotonation was achieved by us-
group of H-2 shows an absorption at νOH = 3625 cm–1. In con- ing the non-nucleophilic and non-coordinating phosphazene
˜
trast, the OH stretching vibration of H-1 is not observed in the base P1tBu [tert-butyliminotris(dimethylamino)phosphorane,
spectral region between 3800–3100 cm–1 due to its involve- pKa(MeCN) = 26.98][37] as proton acceptor. All analyses, oxid-
ment in the OH···O IHB (Scheme 1).[16]
ations, and deprotonations were performed under an inert at-
mosphere. After filtration of precipitated silver when applicable,
the resulting solutions were immediately subjected to spectro-
scopic analyses (IR, UV/Vis, and EPR spectroscopy).
The reference radical 4· obtained from H-4 by PT and ET
gives the expected sharp EPR resonance of a phenoxyl radical
The reference phenol H-5 was similarly characterized (see
the Exp. Sect. and Figure S1b in the Supporting Information).
As expected, no IHBs were detected for H-5 in CDCl3 by 1H
NMR spectroscopy (δOH = 5.60 ppm and δNH = 7.18 ppm) or in
CH2Cl2 by IR spectroscopy (ν
= 3625 cm–1 and ν
=
˜
˜
OH
NH
3425 cm–1).
at g1,2,3 = 2.0117, 2.0071, 2.0037 (Table 2) at 77 K and at giso =
Table 1. EPR parameters obtained by simulation of the experimental X-band EPR spectra recorded at 298 K.[a]
Radical
giso
A(14N) [G] A(1Ho) [G] A(1Hm) [G] A(1Hp) [G]
Other A(1H) [G]
Gauss pp linewidth [MHz] Lorentz pp linewidth [MHz]
[1b]·
4·
2.0059
2.0058
2.0060
2.0071
2.0069
2.0072
0.4
0.1
0.02
0.08
0.15
0.06
0.4
1.8/0.95[b]
0.07
0.02
0.04
0.05
0.02
5·
1.90[c]
11.10
11.30
10.90
1.7 (2×)[d]
[6Cp ·
]
2.80 (2×)
2.90 (2×)
2.77 (2×)
0.90 (2×)
0.85 (2×)
0.90 (2×)
2.60
2.70
2.64
1.20, 1.00, 0.90 (2×)[e]
1.10 (2×), 0.85 (2×)[e]
1.20/1.10, 0.70 (2×)[e]
[7Cp ·
[8Cp [f]
]
]
[a] For atomic numbering of the nitroxyl radicals, see Scheme 7. [b] Hydrogen atoms of the 3,5-di-tert-butyl-2-hydroxybenzoic acid moiety. [c] Amide nitrogen
atom. [d] Hydrogen atoms of the 2,6-di-tert-butyl-4-hydroxybenzoic acid moiety. [e] Hydrogen atoms of the ferrocenyl moiety. [f] See ref.[26]
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1277
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
2.0058 at 298 K in CH2Cl2 (Figure 2, Table 1).[16] The most stable hyperfine coupling (hfc) constants to ferrocenyl protons (Fig-
valence isomer of 1·, obtained by oxidation and deprotonation ure 2). The yield of [1b]· with respect to the initial H-1, obtained
of H-1, is the zwitterionic [1a]·. The zwitterion is characterized by double integration of the EPR resonance referenced against
by its 77 K EPR resonance (g1,2,3 = 3.0400, 1.9490, 1.9030; the 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical, was approxi-
Table 2), which significantly differs from the EPR pattern of the mately 0.05 %.[26,39] This low yield explains the difficulty of its
parent ferrocenium ion [H-1]·+ (g1,2,3 = 3.4200, 1.8552, 1.7550; observation.
Table 2).[16] Attempts to observe the expected trace amounts
We also attempted to detect the conceivable IVCT band of
of the phenoxyl radical valence isomer [1b]· were recently suc-
cessful and it shows a very weak EPR resonance at giso = 2.0059
at 298 K (Figure 2, Table 1). At this temperature, the resonance
of the ferrocenium-phenolate [1a]· is not observed. This is typi-
cal for ferrocenium radicals as these are EPR-silent at room tem-
perature due to fast spin-lattice relaxation.[38] The signal of [1b]·
is broader than that of 4·, possibly due to additional unresolved
[1b]· in addition to the IVCT band of [1a]· (1040 nm in
CH2Cl2),[16] even though [1b]· is only present in trace amounts.
Indeed, by close inspection of the NIR spectral region we were
able to observe two further very weak NIR bands at 1970 and
1820 nm (ν
= 5076 and 5491 cm–1, Δν
= 548 and
˜
˜
max
1/2
815 cm–1, ε = 45 and 27 –1 cm–1) in CH2Cl2 (obtained by Gaus-
M
sian deconvolution; see Figure S3 in the Supporting Informa-
tion). However, the width, Δν1/2, of these bands is rather low
˜
and smaller than expected for IVCT transitions.[40] These two
low-energy bands are fairly similar to those observed for NHR-
substituted ferrocenium ions such as [H-3]·+[41a] and other sub-
stituted Fc+ ions[41b] and have been attributed to Laporte-for-
bidden d–d transitions of the ferrocenium moiety. To further
confirm this assignment, a time-dependent density functional
theory (TD-DFT) calculation [B3LYP/SV(P), COSMO CH2Cl2] was
performed for [1a]·, which provided low-energy transitions at
2765, 2055, 1317, 1091, 997, and 771 nm (3617, 4865, 7594,
9168, 10031, and 12973 cm–1, respectively; see Table S2 in the
Supporting Information). The calculated lowest-energy band
corresponds to an IVCT [phenolate → dxy(Fc+), f = 0.00623] giv-
ing the ferrocenyl phenoxyl radical ground state of [1b]·. This
band was not observed experimentally. The 1091 nm band
[phenolate → dxz(Fc+), f = 0.01999] corresponds to the experi-
mentally observed IVCT band of [1a]· (1040 nm).[16] The calcu-
lated transitions at 2055 and 1317 nm have been assigned to
the experimentally observed weak NIR bands at 1970 and
1820 nm and are clearly ferrocenium d–d transitions (see
Table S2 and Figure S4 in the Supporting Information). Two
further IVCT bands were calculated by TD-DFT at 997 and
771 nm (10031 and 12973 cm–1). These represent phenol-
ate → dyz(Fc+) and lone pair (phenolate) → dxy(Fc+) transitions.
As they are close to the 1040 nm IVCT band and very weak in
Figure 2. a) X-band EPR spectrum of [1b]· (20 m
M solution of H-1 in CH2Cl2)
at 298 K, including a simulation obtained by using the following parameters:
field: 3346.20 G, sweep: 499.77 G, sweep time: 90 s, modulation: 5000 mG,
MW attenuation: 10 dB. b) X-band EPR spectrum of 4· (5 m
M solution of H-4
in CH2Cl2) at 298 K, including a simulation obtained by using the following
parameters: field: 3346.20 G, sweep: 94.79 G, sweep time: 90 s, modulation:
1000 mG, MW attenuation: 10 dB.[16]
Table 2. EPR parameters obtained by simulation of experimental X-band spectra recorded at 77 K.
Radical (mixture)
g1,2,3
A(14N) [G] (nitroxyl N) Fraction [%] Gauss pp linewidth [MHz] Lorentz pp linewidth [MHz]
[H-1]·+[a]
3.4200, 1.8552, 1.7550
3.0400, 1.9490, 1.9030
3.4000, 1.8780, 1.7965
3.3500, 1.8750, 1.7870
2.0105, 2.0070, 2.0025
2.9600, 1.9530, 1.9330
≈2.011, 2.008, 2.005
N/A, ≈1.98, ≈1.96
2.0105, 2.0060, 2.0045
N/A, 1.9620, 1.9450
2.0095, 2.0065, 2.0030
N/A, 1.9650, 1.9400
1.5
0.5
1.0
1.5
0.5
0.5
0.3
1.7
0.3
0.3
0.3
0.1
0.6
0.3
0.5
0.5
1.5
0.5
1.0
1.5
0.5
0.5
0.3
1.7
0.4
0.2
0.2
0.1
0.5
0.3
0.5
0.6
[1a]·[a]
[H-2]·+
[H-3]·+[a]
[1x]· (x = α, ꢀ, Cp)
1.3
98.7
30
70
17
83
0.8
99.2
[1c]·
[2b]·/[2x]· (x = α, ꢀ, Cp)
[2c]·
[8Cp ·[b]
]
3.5, 3.5, 28.0
[3c]·[b]
[3x]· (×= α, ꢀ, Cp, Me)[b]
[3c]·[26]
4·
5·
2.0117, 2.0071, 2.0037
2.0096, 2.0065, 2.0036
2.0103, 2.0069, 2.0040
2.0094, 2.0067, 2.0048
[6Cp ·
]
3.0, 3.0, 28.0
4.0, 4.0, 26.0
[7Cp ·
[a] See ref.[16] [b] See ref.[26]
]
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1278
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
intensity (f < 0.0001) they are not distinctly observed. In es-
sence, the relevant bands of [1a]· are well understood, whereas
the IVCT bands of [1b]· are not detectable due to its very low
concentration. Clearly, [1a]· is significantly preferred over [1b]·
due to the strong NH···O IHB in the zwitterion.
The phenol amides H-1 and H-4 were easily deprotonated
by using 1 equivalent of P1tBu in CH2Cl2. The analogous reac-
tion with the constitutional isomers H-2 and H-5 proved to
be more challenging, based on the results of UV/Vis and IR
spectroscopic studies, namely an incomplete shift of the π–π*
phenol/phenolate band[42,43] and an incomplete decrease of
the ν = 3625 cm–1 IR absorption band (see Figures S6a–S9a
˜
OH
in the Supporting Information). Increasing the ionic strength of
the CH2Cl2 solution by the addition of [nBu4N][B(C6F5)4] (similar
to the electrochemical experiments;[44,45] see below) resulted
in the almost complete deprotonation of H-5 to 5– and the
quantitative deprotonation of H-2 to 2– by using 1 equivalent
of P1tBu, based on the π–π* phenolate bands and the absence
of OH vibrational stretching bands (see Figures S6b–S9b in the
Supporting Information). Clearly, proton transfer benefits from
a higher ionic strength of the solution.
Scheme 5. a) Formation of [22]2– and [52]2– from 2– and 5– through intermo-
lecular hydrogen bonds and b) intermolecular hydrogen bond between H-3
and phenolate in CH2Cl2 solution.
The NH stretching vibrational bands of 2– and 5– (νNH
=
˜
the metal-centered oxidation (Table 2). The UV/Vis spectrum
of [H-2]·+, prepared under the same conditions, features the
expected ferrocenium absorption band at λmax = 770 nm (see
Figure S14 in the Supporting Information), similarly to [H-
1]·+.[16,28,29] The IR spectrum of [H-2]·+, prepared in the same
fashion, shows an essentially unperturbed OH vibration
3435 cm–1) are essentially unperturbed in comparison with
those of H-2 and H-5 (see Figures S7b and S9b in the Support-
ing Information). However, an additional NH stretching band is
found at ν = 3365 and 3380 cm–1 for 2– and 5–, respectively.
˜
NH
These bands have been assigned to an NH group intermolecu-
larly hydrogen bonded to a phenolate ion to give at least di-
meric aggregates [22]2– and [52]2– in CH2Cl2 solution
(Scheme 5, a; intermolecular NH···O hydrogen bond). Higher
aggregates of these ionic systems 2– and 5– might also be
present, but these are not shown for clarity. The presence of
intermolecular hydrogen bonds is supported by the analogous
association of H-3 and phenolate (Scheme 5, b and Figure S10
(νOH = 3620 cm–1) and a slightly shifted NH vibration (νNH
=
˜
˜
3365 cm–1). This suggests a weakening of the NH bond in the
cation [H-2]·+ and hence an acidification of the NH group due
to the positive charge on the ferrocenium moiety close to the
NH group, which results in coordination of the counter ion to
the NH group [NH···F(SbF6)].[16]
in the Supporting Information; νNH = 3380 cm–1). As expected
In the presence of approximately equimolar amounts of
P1tBu, H-2, that is, phenolate 2–, shows several distinct redox
processes in CH2Cl2/[nBu4N][B(C6F5)4] solution. An irreversible
oxidation at Ep = –500 mV versus FcH/FcH+ is observed followed
˜
from the different pKa values of phenol and amide[35,46] and
corroborated by DFT calculations, with the phenolate favored
over the iminolate by 65 and 71 kJ mol–1 for 2– and 5–, respec-
tively (see Figure S11 in the Supporting Information), no evi-
dence for the presence of conceivable iminolate tautomers is
found (Scheme 5). That the NH group of amide 2– should be
sterically accessible for hydrogen bonding is also evident from
the solid-state structure of H-1 and other substituted N-ferro-
cenyl amides (see Figure S2 in the Supporting Information). The
intermolecular NH···O hydrogen bond observed between the
amide and phenolate is electronically congruent with the intra-
by a reversible, one-electron oxidation process at E1/2
=
–110 mV versus FcH/FcH+ and another irreversible follow-up
wave at Ep = 70 mV (Figure 3, bottom). The potential of the
reversible wave (E1/2 = –110 mV) is close to that of the H-2/[H-
2]·+ pair and the irreversible wave resembles that of the 5–/5·
molecular NH···O hydrogen bond in 1– (Scheme 1; νNH
=
˜
3380 cm–1).[16]
H-2 was reversibly oxidized to [H-2]·+ at E1/2 = –105 mV ver-
sus FcH/FcH+ at 298 K in CH2Cl2/[nBu4N][B(C6F5)4] solution (see
Figure S12 in the Supporting Information). Under the same con-
ditions, the H-1/[H-1]·+ redox couple was observed at E1/2
=
–150 mV.[16] The EPR spectrum of [H-2]·+ in frozen CH2Cl2/
[nBu4N][B(C6F5)4] at 77 K, obtained by the oxidation of H-2 with
AgSbF6, shows a characteristic, nearly axial ferrocenium reso-
nance (see Figure S13 in the Supporting Information), similar to
that of the ferrocenium ion [H-1]·+,[16] thereby fully confirming
Figure 3. Square-wave (top) and cyclic voltammograms (bottom) of 2– in
CH2Cl2 containing [nBu4N][B(C6F5)4] as supporting electrolyte at 298 K. Rela-
tive peak areas A are given.
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1279
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
reference couple (E1/2 = –490 mV vs. FcH/FcH+; see Figure S15b broad resonance at g ≈ 1.96 (see Figure S17 in the Supporting
in the Supporting Information). The reversible wave has indeed Information). We tentatively assigned the resonance at g ≈ 2.00
been assigned to the H-2/[H-2]·+ pair. As both 5– and 2– are to the phenoxyl radical valence isomer [2b]· and/or the carbon-
present as (at least dimeric) aggregates (Scheme 5), we as- centered radicals [2x]· by comparison with the resonance of the
signed the first wave to the irreversible oxidation of the phenoxyl radical 5· and radicals [3x]· (see Table 2, Scheme 3,
hydrogen-bonded aggregate [22]2– to the hydrogen-bonded and Figure S18 in the Supporting Information). The high-field
radical [22]·–. At higher scan rates (up to 800 mV s–1), the irre- resonance is assigned to the ferrocenium part of the iminolate
versible oxidation of [22]2– becomes partially reversible (see Fig- zwitterion [2c]· by comparison with the resonances of the ferro-
ure S16 in the Supporting Information). The fact that the peak cenium iminolates [1c]· and [3c]· (Schemes 1, 3, 4, and
area of the two high potential waves is approximately twice Table 2).[16,26] Annealing this sample to room temperature for
that of the area of the [22]2–/[22]·– oxidation wave (Figure 3, one minute and refreezing to 77 K resulted in an EPR-silent
top) suggests that the oxidation of [22]2– releases the phenol mixture. Clearly, the radicals 2· are much more reactive than
H-2 after proton transfer and another species (Scheme 6). Fur- the radicals 1· and they could not be observed by UV/Vis or IR
thermore, the wave of the initially present aggregated phenol- spectroscopy.[16]
ate [22]2– decreases upon further scanning, which confirms a
DFT geometry optimizations of the ferrocenium phenolate
follow-up reaction of [22]·– after electron transfer. In contrast to
the behavior of [22]2–, monomeric 1– displays a (quasi)reversi-
ble oxidation at E1/2 = –535 mV and an irreversible oxidation at
Ep = +100 mV.[16] This difference in reactivity will be discussed
below.
zwitterion [2a]·, the ferrocenyl phenoxyl radical [2b]·, and the
ferrocenium iminolate zwitterion [2c]· showed that these va-
lence isomers and tautomers are very close in energy (Figure 4).
The quite high stability of the ferrocenium iminolate [2c]· can
be accounted for by the increased acidity of the positively
charged ferrocenium amides, similar to the stability of [3c]·
(Scheme 3).[26,28,29] In contrast, zwitterion [1a]· is significantly
stabilized relative to [1b]· and [1c]· due to its strong NH···O
IHB.[16] The charge-transfer and d–d transitions of [1a]· and the
elusive [2a]· derived by TD-DFT calculations can be compared
in Table S2 in the Supporting Information.
Figure 4. DFT-calculated geometries. energies, and spin densities (red, isosur-
face value 0.01 a.u.) of the valence isomers and tautomers of 2· (CH2Cl2; CH
hydrogen atoms have been omitted).
The high reactivity of 2· in comparison with 1· is not immedi-
ately obvious from the DFT calculations. The radicals [2a]·, [2b]·,
and [2c]· should be quite persistent compared with typical phe-
noxyl and ferrocenium radicals such as 4·, 5·, FcH·+, [H-2]·+, and
[3c]· (see Figures S18 and S13 in the Supporting Information
and refs.[16,26,47–52]). The main difference between 1· and 2· is
the type of NH···O hydrogen bond between the amide and phe-
nolate in their precursor phenolates 1– and 2–, namely an intra-
Scheme 6. Oxidation of the aggregate [22]2– to [22]·– followed by proton
transfer and dissociation to H-2 and [2a-H]·–/[2b-H]·– radical anions.
Attempts to spectroscopically observe the radicals 2· met molecular hydrogen bond in 1– (Scheme 1) and an intermolecu-
with limited success due to the transient nature of these radical lar hydrogen bond in 2– ([22]2–, Scheme 5, a). Given that 2–
species. Independent of the preparative pathway (oxidation/de- forms (at least) dimeric aggregates [22]2– in CH2Cl2/
protonation or deprotonation/oxidation), only EPR-silent reac- [nBu4N][B(C6F5)4] solution (Scheme 5, a), the oxidation of this
tion mixtures were obtained. However, the addition of approxi- aggregate to [22]·– should occur at the ferrocenyl moiety that
mately 1 equivalent of P1tBu to a pre-cooled solution (ca. 193 K) forms the NH···O bond (Scheme 6) as this hydrogen bond in-
of [H-2]·+ and rapidly freezing the sample to 77 K delivered an creases the phenolate/phenoxyl potential but lowers the ferro-
EPR spectrum displaying a sharp resonance at g ≈ 2.00 and a cene/ferrocenium potential (see above).
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1280
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
In the radical aggregate [22]·–, a phenolate 2– is coordinated (Ep = 70 mV). The [22]2–/[22]·– redox process is followed by
to the positively charged [Fc-NH]·+ site of radical [2a]·, which an intermolecular proton transfer leading to dissociation (EC
could easily result in a proton transfer from the acidic [Fc-NH]·+ mechanism). Oxidation of 1–, however, leads to stable [1a]· with
group to the phenolate. Dissociation furnishes H-2 and the the hydrogen-bonded proton simply moving between the ni-
doubly deprotonated ferrocenium phenolate iminolate [2a-H]·– trogen and oxygen atoms.[16] To conclude, the fundamental dif-
or the phenoxyl iminolate species [2b-H]·– (Scheme 6). The lat- ference in reactivity between 2· and 1· is based on the nature
ter radical anions might be susceptible to decomposition. This of the hydrogen bond.
mechanistic scenario fits perfectly with the electrochemical ex-
periments (Figure 3): One-electron oxidation of the dimer [22]2–
Spin-Trapping of Carbon-Centered Radicals
(Ep = –500 mV) leads to [22]·– followed by the formation of H- The ferrocenium iminolate [3c]· equilibrates with a pool of car-
2 (E1/2 = –110 mV) and the follow-up products [2a-H]·–/[2b-H]·– bon-centered radicals [3x]· (x = α, ꢀ, Cp, Me) and some of these
Scheme 7. Suggested radical reactivity of [H-1]·+, [H-2]·+, and [H-3]·+ in the presence of a base and PhNO. The EPR data (298 K, 77 K) of identified radical
intermediates are given.
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1281
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
radicals can be intercepted by PhNO to give the nitroxides [8Cp ·
]
and [8ꢀ]· (Scheme 3). Carbon-centered metallocenyl radicals
might also be involved in the decarboxylation of cobaltocenium
carboxylate.[53] Although PhNO has been reported to trap phe-
noxyl radicals,[54–57] the phenoxyl reference radicals 4· and 5·
are inert towards PhNO under our conditions. Furthermore, sim-
ple ferrocenium ions such as FcH·+ or [H-3]·+ are also resistant
towards PhNO.[26] Hence, the formation of nitroxide radicals se-
lectively indicates the presence of ferrocenium iminolates,
which equilibrate with C-centered radicals.
Indeed, [H-1]·+ and [H-2]·+ are unreactive towards PhNO and
only the ferrocenium resonances are observed by EPR spectro-
scopy at 77 K (Table 2, Scheme 7). In the presence of base, the
valence isomeric radicals [1a]·/[2a]· and [1b]·/[2b]· are formed
(O deprotonation) along with the ferrocenium iminolates [1c]·/
[2c]· (N deprotonation; Schemes 1, 4, a, and 7). The latter might
equilibrate with C-centered radicals [1x]·/[2x]· (x = α, ꢀ, Cp; C
deprotonation), which should be susceptible to the spin-trap-
ping reaction giving nitroxide radicals [6x]·/[7x]· (Scheme 7). In-
deed, EPR triplet resonances characteristic of nitroxide radicals
are observed at 298 K in both cases (Figure 5) with a spectro-
scopic yield of less than 1 % (by double integration of the EPR
resonance referenced against the DPPH radical).[26,39] The EPR
parameters obtained for [6x]·/[7x]· closely resemble those of
[8Cp]· (Scheme 3, Table 1), which suggests a preferred substitu-
tion at the Cp ring ([6Cp]·/[7Cp]·) with possibly some substitu-
tion at the ꢀ position ([6ꢀ]·/[7ꢀ]·; Scheme 7). Similarly to the
case of [3x]·/[8x]· and in agreement with DFT calculations, sub-
stitution at the nitrogen atom ([6N]·/[7N]·) is thermodynamically
uphill whereas substitution reactions at the Cp ring and the α/ꢀ
positions to give [6Cp,α,ꢀ]· and [7Cp,α,ꢀ]· are thermodynamically
Figure 5. X-band EPR spectra (top) and simulated spectra (bottom) of a) [6Cp ·
(25 m ) in CH2Cl2 . All the spectra were re-
) in CH2Cl2 and b) [7Cp]· (3 m
corded by using the following parameters: T = 298 K, field: 3346.20 G, sweep:
94.79 G, sweep time: 90 s, modulation: 250 mG, MW attenuation: 5 dB ([6Cp]·)
and modulation: 5000 mG, MW attenuation: 10 dB ([7Cp]·)].
]
M
M
feasible (Figure 6). The driving forces are similar for the [6Cp,α,ꢀ ·
]
and [7Cp,α,ꢀ]· nitroxides derived from the radicals [1x]·/[2x]·
(Scheme 7). According to DFT calculations, the conceivable ni-
troxide radicals [6b]·/[7b]· derived from the phenoxyl radicals
[1b]·/[2b]· are thermodynamically unfavorable with respect to
the starting radical and PhNO (see Figures S19 and S20 in the
Supporting Information), in agreement with the experimental
observations that the phenoxyl radicals 4· and 5· are inert to-
wards nitroxide radical formation with PhNO (see above).
DFT calculations were also performed on the radicals [1x]·/
[2x]· (x = α, ꢀ, Cp; see Figures S21 and S22 in the Supporting
Information). As expected, all these radicals are high-energy
species, independent of the presence or absence of NH···O or
OH···O IHBs. The spin densities in radicals [1α]· and [2α]· are
essentially localized at the iron center (Mulliken spin density at
iron: 1.25/1.25), which suggests a dominant contribution of a
ferrocenium carbanion resonance structure, similar to [3α]·.[26]
Hence, a reaction of [1α]·/[2α]· with PhNO is not expected. In
contrast, [1Cp]·/[2Cp]· and [1ꢀ]·/[2ꢀ]· feature large spin densities
at the respective carbon atoms with Mulliken spin densities
above 0.65 and hence should be susceptible to attack by PhNO.
This is essentially independent of the type of IHB (NH···O,
OH···O) in the radicals [1x]· (see Figures S21 and S22 in the
Supporting Information).
Figure 6. DFT-optimized geometries with spin densities for [7x]· (x = α, ꢀ, Cp,
N) (0.01 a.u. isosurface value) and energies in CH2Cl2 continuum solvent for
[6x]· and [7x]· (x = N, α, ꢀ, Cp) as well as their Lewis structures.
The presence of [1x]· and [2x]· radicals has been probed by
rapid-freeze EPR techniques. Indeed, adding P1tBu to a pre-
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1282
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
cooled solution containing [H-1]·+ and PhNO (ca. 193 K) and
rapid freezing to 77 K allowed the detection of a broad, aniso-
tropic large ferrocenium resonance at around g = 1.96 (98.7 %)
as well as a sharp resonance signal at around g = 2.00 without
discernible hfc constants (1.3 %; Table 2; Figure 7, a). The former
closely resembles the resonance of the ferrocenium iminolate
[3c]·, whereas the latter is similar to that of the carbon-centered
radicals [3Cp,ꢀ]· (Table 2, Schemes 3 and 7).[26] Hence, we as-
signed the former to the ferrocenium iminolate zwitterion [1c]·
and the latter to traces of the radicals [1Cp,ꢀ]·. Clearly, [H-1]·+ is
deprotonated mainly at the amide nitrogen atom to give the
ferrocenium iminolate [1c]· and to some extent at the Cp rings
to give [1Cp,ꢀ]· as the phenolic OH is engaged in an OH···O IHB
(Scheme 1). On this time scale, the kinetic product [1c]· has not
yet been quantitatively isomerized to radicals [1a]· and [1b].
Figure 7. X-band EPR spectra (top) and simulated spectra (bottom) of a) a
rapidly freezed mixture of [1Cp,ꢀ]·/[1c]· (25 m
M
H-1) in CH2Cl2, b) [6Cp]· (25 m
M
H-1) in CH2Cl2 after annealing at room temperature, and c) [7Cp]· (3.0 m
M
H-
2) in CH2Cl2/[nBu4N][B(C6F5)4]. All the spectra were recorded by using the
following parameters: temperature = 77 K, field = 3346.20 G, sweep =
499.77 G, sweep time = 90 s, modulation = 5000 mG, MW attenuation =
20 dB ([1Cp,ꢀ]·/[1c]· and [6Cp]·) and 10 dB ([7Cp]·).
Scheme 8. a) Possible deprotonation of the ferrocenium ion [H-2]·+ to give
the dimeric aggregate [(2c)2]··, followed by proton transfer and dissociation
into [H-2]·+ and [2a-H]·–/[2b-H]·– with subsequent electron transfer to give
H-2 and diradical [2-H]··. b) Oxidation of phenolate 5– with the ferrocenium
ion [FcH]·+.
No nitroxide radicals were discernible in this sample. Clearly,
the following trapping reaction of [1Cp,ꢀ]· with PhNO is slow
enough to allow detection of radicals [1Cp,ꢀ]·. After annealing
the sample to room temperature for five minutes and refreezing
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1283
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
to 77 K, the [1c]· and [1Cp,ꢀ]· resonances vanished in favor of
the nitroxide EPR resonance of [6Cp,ꢀ]· with a large nitrogen
to use. Absolute DMF was used as received from Acros. 1-[Bis(di-
methylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide
hexafluorophosphate (HATU) was commercially available from No-
vabiochem. P1tBu, sodium hydroxide, 2,6-di-tert-butyl-4-carboxy-
phenol, N,N-diisopropylethylamine (DIPEA), 2,2-diphenyl-1-picryl-
hydrazyl (DPPH), nitrosobenzene (PhNO), ferrocenium hexafluoro-
phosphate, decamethylcobaltocene, and 2,4,6-trimethylaniline were
commercially available from Sigma–Aldrich. DIPEA was dried with
NaOH overnight under argon and applied to the coupling reactions
through a syringe. Aminoferrocene,[27,34] 2,4-di-tert-butyl-6-(ferro-
cenylcarbamoyl)phenol (H-1),[16] 2,4-di-tert-butyl-6-(methylcarb-
amoyl)phenol (H-4),[25] N-acetylaminoferrocene (H-3),[27] and
[nBu4N][B(C6F5)4][44] were prepared according to literature proce-
dures.
hfc for the g3 resonance, similar to nitroxide radicals [8Cp,ꢀ ·[26]
]
(Table 2, Figure 7, b).
The rapid-freeze EPR spectrum (77 K) of a mixture of [H-2]·+,
PhNO, and P1tBu, prepared as described for [H-1]·+ yielded only
the resonance of the nitroxide radicals [7Cp,ꢀ]· without indica-
tion of ferrocenium phenolate [2a]·, phenoxyl radical [2b]·, fer-
rocenium iminolate [2c]·, or C-centered radicals [2Cp,ꢀ]· (Fig-
ure 7, c). Hence, under these conditions, the follow-up reactions
are too fast and only some [2Cp,ꢀ]· radicals have been inter-
cepted by PhNO to yield [7Cp,ꢀ]· (Scheme 7). Again, a conceiva-
ble follow-up reaction of [2c]· and its conceivable hydrogen-
bonded dimer [(2c)2]·· (Scheme 8) might be intermolecular pro-
ton transfer from the phenol to the iminolate to give ferroce-
nium cation [H-2]·+ and the doubly deprotonated ferrocenium
phenolate iminolate [2a-H]·– or phenoxyl iminolate species [2b-
H]·– (similar to the reactivity of [22]·–, Scheme 6). As the ferroce-
nium ion [FcH]·+ oxidizes phenolate 5– to the phenoxyl radical
5· (Scheme 8, b), a similar reaction might occur between [H-2]·+
and the proposed doubly deprotonated ferrocenium phenolate
iminolate [2a-H]·– to give diradical [2-H]·· (Scheme 8, a). Such a
diradical might be prone to decomposition yielding only EPR-
silent products.
Filtrations from precipitated silver after oxidation were performed
with syringe filters (Rotilabo-Spritzenfilter, Ø = 15 mm, pore size:
0.20 μm; Carl Roth GmbH + Co. KG, Germany). NMR spectra were
recorded with a Bruker Avance DRX 400 spectrometer at 400.13 (1H)
and 100.03 MHz (13C{1H}) at 25 °C. All resonances are reported in
ppm versus the solvent signal as internal standard: CD2Cl2 (1H: δ =
5.32 ppm; 13C: δ = 54.0 ppm), CDCl3 (1H: δ = 7.26 ppm; 13C: δ =
77.2 ppm), and [D8]THF (1H: δ = 3.58 ppm; 13C: δ = 67.6 ppm).
IR spectra were recorded with a Varian Excalibur Series 3100 FT-IR
spectrometer using KBr cells in CH2Cl2 and as KBr disks. Electro-
chemical experiments were carried out with a BioLogic SP-50 vol-
tammetric analyzer using a platinum working electrode, a platinum
wire as counter electrode, and a 0.01
M Ag/AgNO3 electrode as
Conclusions
reference electrode. The measurements were carried out at a scan
rate of 100 mV s–1 for cyclic and square-wave voltammetry experi-
Amide-bridged ferrocenyl phenols H-1 and H-2 can be oxidized
to the respective ferrocenium phenol radical cations [H-1]·+ and
[H-2]·+, respectively. These are deprotonated at the phenol oxy-
gen atom, the amide nitrogen atom, or cyclopentadienyl car-
bon atoms to give the corresponding neutral radicals [1a]·/[2a]·,
[1b]·/[2b]· (@O), [1c]·/[2c]· (@N), and [1x]·/[2x]· (@C; Scheme 7).
The ferrocenium iminolate [1c]· and the carbon-centered radi-
cals [1Cp,ꢀ]· are the kinetic products of the [H-1]·+ deproton-
ation. The latter species are intercepted by PhNO as nitroxide
radicals [6Cp,ꢀ]·. The NH···O IHB in [1a]· strongly stabilizes the
ferrocenium phenolate zwitterion, yet traces of the valence iso-
meric phenoxyl radical [1b]· are also observed at 298 K. The
IVCT between [1a]· and [1b]· is coupled to the NH···O vibration.
For [H-2]·+ lacking IHBs, all the radicals derived by deprotona-
tion are transient. Only some of them can be observed by rapid-
freeze EPR techniques ([2b]·, [2c]·) or indirectly by spin-trapping
techniques ([2Cp,ꢀ]· as [7Cp,ꢀ]·). Suggested follow-up reactions
of these transient radicals 2· are intermolecular proton (and
electron) transfer reactions in hydrogen-bonded assemblies
leading finally to EPR-silent products. This contrasts with the
reversible intramolecular PCET reaction (IVCT) between [1a]·
and [1b]·. The absence of the stabilizing IHB and the resulting
high reactivity precludes the observation of IVCT between the
transient valence isomers [2a]· and [2b]·. In any case, ferrocenyl
phenols generate a pool of radicals under oxidative and basic
conditions. This observation might also be relevant for biologi-
cally active ferrocenyl phenols such as ferrocifen (H-C).
ments unless noted otherwise using 0.1
M [nBu4N][B(C6F5)4] as sup-
porting electrolyte and 0.001 of the sample in CH2Cl2. Potentials
M
are given relative to the ferrocene/ferrocenium couple. Referencing
was achieved by addition of decamethylcobaltocene (E1/2 = –2.04 V
vs. FcH/FcH+, CH2Cl2, [nBu4N][B(C6F5)4])) to the sample.[36] UV/Vis/
NIR spectra were recorded with a Varian Cary 5000 spectrophoto-
meter using 1.0 cm cells unless mentioned otherwise (Hellma, Su-
prasil). FD mass spectra were recorded with a Thermo Fisher DFS
mass spectrometer with a LIFDI upgrade. CW EPR spectra (X-band;
ca. 9.4 GHz) were recorded with a Miniscope MS 300 spectrometer
at 298 and 77 K with cooling in liquid nitrogen in a finger Dewar
(Magnettech GmbH, Berlin, Germany). Settings are given as indi-
cated in the spectra. The g factors are referenced to external Mn2+
in ZnS (g = 2.118, 2.066, 2.027, 1.986, 1.946, 1.906). The EPR spectra
were simulated by using EasySpin (v 5.0.0)[58] for MatLab (R2015a).
Melting points were determined by using a Gallenkamp capillary
melting point apparatus (MFB 595 010M). Elemental analyses were
performed by the microanalytical laboratory of the chemical insti-
tutes of the University of Mainz.
Crystal Structure Determination: Intensity data were collected
with a Bruker AXS Smart1000 CCD diffractometer equipped with an
APEX II detector and an Oxford cooling system using Mo-Kα radia-
tion (λ = 0.71073 Å) at 173(2) K and were corrected for absorption
and other effects. The diffraction frames were integrated by using
the SAINT software package, and most were corrected for absorp-
tion with MULABS.[59,60] The structures were solved by direct meth-
ods and refined by the full-matrix method based on F2 using the
SHELXTL software package.[61,62] All non-hydrogen atoms were re-
fined anisotropically, and the positions of all hydrogen atoms were
generated with appropriate geometric constraints and allowed to
ride on their respective parent atoms with fixed isotropic thermal
parameters.
Experimental Section
General: All reactions were performed under argon unless noted
otherwise. Dichloromethane was dried with CaH2 and distilled prior
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1284
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
CCDC 1426154 (for H-1) contains the supplementary crystallo- (CDCl3): δ = 7.75 (s, 2 H, 7-H), 7.18 (s, 1 H, NH), 6.93 (s, 2 H, 3-H),
graphic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre.
5.60 (s, 1 H, OH), 2.29 (s, 3 H, 5b-H), 2.25 (s, 6 H, 5a-H), 1.49 (s, 18
H, 10a-H) ppm. 13C NMR (CDCl3): δ = 166.6 (C=O), 157.2 (C-9), 137.0
(C-4), 136.2 (C-8), 135.4 (C-2), 131.7 (C-1), 129.1 (C-3), 126.0 (C-6),
124.6 (C-7), 34.6 (C-10), 30.3 (C-10a), 21.1 (C-5b), 18.7 (C-5a) ppm.
Density Functional Calculations: DFT calculations were carried
out with the ORCA 3.0.2/DFT series[63] of programs. For geometry
optimizations and energy calculations, the B3LYP formulation of
density functional theory was used employing the SV(P)[64,65] basis
set, the RIJCOSX approximation, the approximate second order SCF
(SOSCF),[66,67] the zero-order regular approximation (ZORA),[68–70]
and the KDIIS algorithm at GRIDX4. No symmetry constraints were
imposed on the molecules. Solvent modeling was carried out by
employing the conductor-like screening model (COSMO, CH2Cl2).[71]
The approximate free energies at 298 K were obtained by thermo-
chemical analysis of the frequency calculations using the thermal
correction to Gibbs free energy as reported by ORCA 3.0.2. TD-
DFT calculations were performed by using the same basis set and
functional as for the geometry optimizations, calculating 50 states
with MaxDim = 250. The 4-methyl group of H-5 and all its ET and PT
products were omitted from the DFT calculations for convergence
reasons.
UV/Vis/NIR (CH2Cl2): λ (ε) = 275 nm (15800
M
–1 cm–1). IR (KBr): ν =
˜
˜
3625 (m, OH), 3190 (m, NH), 1635 (m, CO) cm–1. IR (CH2Cl2): ν =
3625 (m, OH), 3425 (m, NH), 1665 (m, CO) cm–1. MS (FD): m/z (%) =
367.2 (100) [M]+. C24H33NO2 (367.25): calcd. C 78.43, H 9.05, N 3.81;
found C 78.16, H 9.31, N 3.73.
Electron Transfer Between Ferrocenium Hexafluorophosphate
([FcH][PF6]) and Phenolate 5–: A stoichiometric amount of a solu-
tion of [FcH][PF6] was added to a solution of 5– in CH2Cl2/
[nBu4N][B(C6F5)4]. An immediate color change from blue to orange
(FcH) was observed. The EPR spectra of a sample of this solution
recorded at 298 and 77 K show resonances arising from 5· (see
Figure S18 in the Supporting Information).
Acknowledgments
2,6-Di-tert-butyl-4-(ferrocenylcarbamoyl)phenol (H-2): Amino-
ferrocene (250 mg, 1.24 mmol, 1.0 equiv.) and 2,6-di-tert-butyl-4-
carboxyphenol (311 mg, 1.24 mmol, 1.0 equiv.) were dissolved in
absolute CH2Cl2 (10 mL). DIPEA (241 mg, 320 μL, 1.87 mmol,
1.5 equiv.) was added to the mixture followed by HATU (472 mg,
1.243 mmol, 1.0 equiv.). The solution was stirred at room tempera-
ture for 2 h. Water (10 mL) was added and the product was ex-
tracted with CH2Cl2 (3 × 30 mL). The organic phase was washed
with water (10 mL) and brine (10 mL) and dried with MgSO4. After
filtration the solvent was removed under reduced pressure. The raw
product was purified by column chromatography [SiO2, petroleum
ether (40–60 °C)/CH2Cl2, 1:0 → 1:4] to give a yellow solid, yield 79 %
(423 mg, 0.98 mmol). Rf (CH2Cl2) = 0.17, m.p. > 350 °C (decomp.).
The authors are grateful to Regine Jung-Pothmann for the X-
ray data collection, Dipl.-Chem. Torben Kienz for helpful discus-
sions, as well as Dipl.-Chem. Mathias Enders and Dipl.-Chem.
Maximilian Lauck for preparative assistance. Parts of this re-
search were conducted using the supercomputer Mogon and
the advisory services offered by Johannes Gutenberg-University
Mainz (www.hpc.uni-mainz.de), which is a member of the AHRP
and the Gauss Alliance e.V.
Keywords: Hydrogen bonds · Charge transfer · Radicals ·
EPR spectroscopy · Sandwich complexes · Valence isomers
1
E1/2 = –105 mV. H NMR (CD2Cl2): δ = 7.60 (s, 2 H, 7-H), 7.09 (s, 1 H,
NH), 5.63 (s, 1 H, OH), 4.71 (s, 2 H, 2,5-H), 4.18 (s, 5 H, Cp), 4.05 (s,
2 H, 3,4-H), 1.48 (s, 18 H, 10a-H) ppm. 1H NMR ([D8]THF): δ = 8.50
(s, 1 H, OH), 7.73 (s, 2 H, 7-H), 6.59 (s, 1 H, NH), 4.74 (pt, 2 H, 2,5-H),
4.09 (s, 5 H, Cp), 3.92 (pseudo-t, 2 H, 3,4-H), 1.47 (s, 18 H, 10a-
H) ppm. 13C NMR (CD2Cl2): δ = 157.4 (C-9), 136.7 (C-8), 124.6 (C-7),
[1] R. A. Marcus, N. Sutin, Biochim. Biophys. Acta Rev. Bioenerg. 1985, 811,
265–322.
[2] G. L. Closs, J. R. Miller, Science 1988, 240, 440–447.
[3] J. R. Winkler, H. B. Gray, Annu. Rev. Biochem. 1996, 65, 537–561.
[4] P. F. Barbara, T. J. Meyer, M. A. Ratner, J. Phys. Chem. 1996, 100, 13148–
13168.
[5] T. Kienz, C. Förster, K. Heinze, Organometallics 2014, 33, 4803–4812.
[6] M. A. Fox, Chem. Rev. 1992, 92, 365–368.
[7] C. Creutz, H. Taube, J. Am. Chem. Soc. 1969, 91, 3988–3989.
[8] T. F. Markle, J. M. Mayer, Angew. Chem. Int. Ed. 2008, 47, 738–740; Angew.
Chem. 2008, 120, 750–752.
[9] J. J. Warren, T. A. Tronic, J. M. Mayer, Chem. Rev. 2010, 110, 6961–7001.
[10] a) J.-M. Savéant, Energy Environ. Sci. 2012, 5, 7718–7731; b) C. Costentin,
M. Robert, J.-M. Savéant, Phys. Chem. Chem. Phys. 2010, 12, 11179–11190.
[11] a) S. Hammes-Schiffer, Chem. Rev. 2010, 110, 6937–6938; b) S. Hammes-
Schiffer, A. Stuchebrukhov, Chem. Rev. 2010, 110, 6939–6960; c) S.
Hammes-Schiffer, J. Am. Chem. Soc. 2015, 137, 8860–8871.
[12] I. J. Rhile, T. F. Markle, H. Nagao, A. G. DiPasquale, O. P. Lam, M. A. Lock-
wood, K. Rotter, J. M. Mayer, J. Am. Chem. Soc. 2006, 128, 6075–6088.
[13] a) O. S. Wenger, Chem. Eur. J. 2011, 17, 11692–11702; b) W. Herzog, C.
Bronner, S. Löffler, B. He, D. Kratzert, D. Stalke, A. Hauser, O. S. Wenger,
ChemPhysChem 2013, 14, 1168–1176.
[14] D. R. Weinberg, C. J. Gagliardi, J. F. Hull, C. F. Murphy, C. A. Kent, B. C.
Westlake, A. Paul, D. H. Ess, D. G. McCafferty, T. J. Meyer, Chem. Rev. 2012,
112, 4016–4093.
[15] C. Costentin, M. Robert, J.-M. Savéant, J. Am. Chem. Soc. 2006, 128, 4552–
4553.
69.8 (Cp), 65.1 (C-3,4), 62.1 (C-2,5), 34.9 (C-10), 30.5 (C-10a) ppm. 13
C
NMR ([D8]THF): δ = 166.3 (C=O), 157.8 (C-9), 137.6 (C-8), 127.8 (C-6),
125.4 (C-7), 98.4 (C-1), 68.8 (Cp), 64.7 (C-3,4), 61.7 (C-2,5), 35.5 (C-
10), 30.7 (C-10a) ppm. UV/Vis/NIR (CH2Cl2): λ (ε) = 230 (13825), 257
(16785), 325 (1815), 440 nm (325
M
–1 cm–1). IR (KBr): ν = 3625 (m,
˜
OH), 3285 (m, NH), 1635 (m, CO) cm–1. IR (CH2Cl2): ν = 3625 (m,
˜
OH), 3440 (m, NH), 1665 (m, CO) cm–1. MS (FD): m/z (%) = 433.2
(100), 434.2 (28.7) [M]+, 866.2 (11.8) [2M]+. C25H31FeNO2 (433.17):
calcd. C 69.29, H 7.21, N 3.23; found C 69.64, H 7.48, N 3.41.
2,6-Di-tert-butyl-4-(2,4,6-trimethylphenylcarbamoyl)phenol (H-
5): 2,4,6-Trimethylaniline (168 mg, 174 μL, 1.24 mmol, 1.0 equiv.)
and 2,6-di-tert-butyl-4-carboxyphenol (311 mg, 1.24 mmol,
1.0 equiv.) were dissolved in CH2Cl2 (15 mL). DIPEA (241 mg, 320 μL,
1.87 mmol, 1.5 equiv.) was added to the mixture followed by HATU
(472 mg, 1.243 mmol, 1.0 equiv.). The solution was stirred at room
temperature for 2 h. Water (15 mL) was added and the solution was
extracted with CH2Cl2 (2 × 30 mL). The organic phase was washed
with a saturated aqueous sodium hydrogen carbonate solution
(15 mL) and brine (15 mL) and dried with MgSO4. After filtration,
the solvent was removed under reduced pressure. The raw product
was purified by column chromatography (SiO2, CH2Cl2/MeOH,
1:0 → 95:5) to give a colorless solid, yield 72 % (328 mg, 0.89 mmol).
[16] A. Neidlinger, V. Ksenofontov, K. Heinze, Organometallics 2013, 32, 5955–
5965.
1
Rf (CH2Cl2/MeOH, 98:2) = 0.47, m.p. 301–303 °C (decomp.). H NMR
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1285
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[17]
[18]
[19]
[41] a) D. Siebler, C. Förster, K. Heinze, Eur. J. Inorg. Chem. 2010, 3986–3992;
b) T. Tagg, H. G. Kjaergaard, J. R. Lane, C. J. McAdam, B. H. Robinson, J.
Simpson, Organometallics 2015, 34, 2662–2666.
[42] J. Disinger, S. E. Manahan, Anal. Lett. 1982, 15, 1017–1029.
[43] P. M. Tolstoy, B. Koeppe, G. S. Denisov, H.-H. Limbach, Angew. Chem. Int.
Ed. 2009, 48, 5745–5747; Angew. Chem. 2009, 121, 5855–5858.
[44] R. J. LeSuer, C. Buttolph, W. E. Geiger, J. Organomet. Chem. 2004, 76,
6395–6041.
M. H. V. Huynh, T. J. Meyer, Chem. Rev. 2007, 107, 5004–5064.
J. L. Dempsey, J. R. Winkler, H. B. Gray, Chem. Rev. 2010, 110, 7024–7039.
F. Lachaud, A. Quaranta, Y. Pellegrin, P. Dorlet, M.-F. Charlot, S. Un, W.
Leibl, A. Aukauloo, Angew. Chem. Int. Ed. 2005, 44, 1536–1540; Angew.
Chem. 2005, 117, 1560–1564.
J. Bonin, M. Robert, Photochem. Photobiol. 2011, 87, 1190–1203.
L. Hammarström, S. Styring, Energy Environ. Sci. 2011, 4, 2379–2388.
[20]
[21]
[22]
C. J. Gagliardi, A. K. Vannucci, J. J. Concepcion, Z. Chen, T. J. Meyer, Energy
Environ. Sci. 2012, 5, 7704–7717.
[45] N. Camire, U. T. Mueller-Westerhoff, W. E. Geiger, J. Organomet. Chem.
2001, 639, 823–826.
[23]
[24]
[25]
A. Migliore, N. F. Polizzi, M. J. Therien, D. N. Beratan, Chem. Rev. 2014,
114, 3381–3465.
[46] W. L. Mock, D. C. Y. Chua, J. Chem. Soc. Perkin Trans. 2 1995, 2069–2074.
[47] S. Panagiota, M. Louloudi, Y. Deligiannakis, Chem. Phys. Lett. 2009, 472,
85–89.
[48] P. Neta, R. W. Fessenden, J. Phys. Chem. 1974, 78, 523–529.
[49] L. W. Kiruri, L. Khachatryan, B. Dellinger, S. Lomnicki, Environ. Sci. Technol.
2014, 48, 2212–2217.
[50] J. W. Whittaker, Chem. Rev. 2003, 103, 2347–2363.
[51] F. Thomas, O. Jarjayes, H. Jamet, S. Hamman, E. Saint-Aman, C. Duboc,
J.-L. Pierre, Angew. Chem. Int. Ed. 2004, 43, 594–597; Angew. Chem. 2004,
116, 604–607.
[52] T. Maki, Y. Araki, Y. Ishida, J. Am. Chem. Soc. 2001, 123, 3371–3372.
[53] S. Vanicek, H. Kopacka, K. Wurst, S. Vergeiner, L. Oehninger, I. Ott, B.
Bildstein, Z. Anorg. Allg. Chem. 2015, 641, 1282–1292.
[54] M. Breza, J. Mol. Struct. 2004, 683, 167–169.
M. Sugiura, S. Ogami, M. Kusumi, S. Un, F. Rappaport, A. Boussac, J. Biol.
Chem. 2012, 287, 13336–13347.
R. Wanke, L. Benisvy, M. L. Kuznetsov, M. F. C. Guedes Da Silva, A. J. L.
Pombeiro, Chem. Eur. J. 2011, 17, 11882–11892.
A. Neidlinger, T. Kienz, K. Heinze, Organometallics 2015, 34, 5310–5320.
K. Heinze, M. Schlenker, Eur. J. Inorg. Chem. 2004, 2974–2988.
a) K. Heinze, D. Siebler, Z. Anorg. Allg. Chem. 2007, 633, 2223–2233; b) D.
Siebler, M. Linseis, T. Gasi, L. M. Carrella, R. F. Winter, C. Förster, K. Heinze,
Chem. Eur. J. 2011, 17, 4540–4551.
[26]
[27]
[28]
[29] D. Siebler, C. Förster, K. Heinze, Dalton Trans. 2011, 40, 3558–3575.
[30] H. Huesmann, C. Förster, D. Siebler, T. Gasi, K. Heinze, Organometallics
2012, 31, 413–427.
[31] M. B. Robin, P. Day, Adv. Inorg. Chem. Radiochem. 1968, 10, 247–422.
[32] a) S. Fukuzumi, K. Okamoto, Y. Yoshida, H. Imahori, Y. Araki, O. Ito, J. Am.
Chem. Soc. 2003, 125, 1007–1013; b) S. Fukuzumi, Y. Yoshida, K. Oka-
moto, H. Imahori, Y. Araki, O. Ito, J. Am. Chem. Soc. 2002, 124, 6794–
6795.
[33] a) G. Jaouen, S. Top, A. Vessieres, G. Leclercq, M. McGlinchey, Curr. Med.
Chem. 2004, 11, 2505–2517; b) E. Hillard, A. Vessières, L. Thouin, G. Jao-
uen, C. Amatore, Angew. Chem. Int. Ed. 2006, 45, 285–290; Angew. Chem.
2006, 118, 291–296; c) A. Nguyen, S. Top, P. Pigeon, A. Vessières, E. A.
Hillard, M. A. Plamont, M. Huché, C. Rigamonti, G. Jaouen, Chem. Eur. J.
2009, 15, 684–696; d) D. Hamels, P. M. Dansette, E. A. Hillard, S. Top, A.
Vessières, P. Herson, G. Jaouen, D. Mansuy, Angew. Chem. Int. Ed. 2009,
48, 9124–9126; Angew. Chem. 2009, 121, 9288–9290; e) D. Plażuk, A.
Vessières, E. A. Hillard, O. Buriez, E. Labbé, P. Pigeon, M.-A. Plamont, C.
Amatore, J. Zakrzewski, G. Jaouen, J. Med. Chem. 2009, 52, 4964–4967.
[34] B. Bildstein, M. Malaun, H. Kopacka, K. Wurst, K.-H. Ongania, G. Oprom-
olla, P. Zanello, Organometallics 1999, 18, 4325–4336.
[55] P. Pelikán, L. Omelka, K. Brudíková, M. Breza, J. Mol. Struct. 2003, 624,
251–255.
[56] L. Omelka, J. Kováčová, Magn. Reson. Chem. 1994, 32, 525–531.
[57] S. Terabe, R. Konaka, J. Chem. Soc. Perkin Trans. 2 1972, 2163–2172.
[58] S. Stoll, A. Schweiger, J. Magn. Reson. 2006, 178, 42–55.
[59] SMART Data Collection Software for the SMART System, various versions,
Bruker Analytical X-ray Instruments, Inc., Madison, WI, 2000; SAINT-Plus
Data Processing Software for the SMART System, various versions, Bruker
Analytical X-ray Instruments, Inc., Madison, WI, 2000.
[60] R. H. Blessing, Acta Crystallogr., Sect. A 1995, 51, 33–38.
[61] G. M. Sheldrick, SHELXTL, version 5.1, Bruker AXS, Madison, WI, 1998.
[62] G. M. Sheldrick, SHELXL-97, University of Göttingen, Germany, 1997.
[63] F. Neese, WIREs Comput. Mol. Sci. 2012, 2, 73–78.
[64] A. Schäfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97, 2571–2577.
[65] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297–3305.
[66] F. Neese, Chem. Phys. Lett. 2000, 325, 93–98.
[67] T. H. Fischer, J. Almlof, J. Phys. Chem. 1992, 96, 9768–9774.
[68] E. van Lenthe, E. J. Baerends, J. G. Snijders, J. Chem. Phys. 1993, 99, 4597–
4610.
[35] D. Kanamori, A. Furukawa, T. Okamura, H. Yamamoto, N. Ueyama, Org.
Biomol. Chem. 2005, 3, 1453–1459.
[69] C. van Wüllen, J. Chem. Phys. 1998, 109, 392.
[70] D. A. Pantazis, X. Y. Chen, C. R. Landis, F. Neese, J. Chem. Theory Comput.
[36] N. G. Connelly, W. E. Geiger, Chem. Rev. 1996, 96, 877–910.
[37] Y. Kondo, in: Supberbases for Organic Synthesis (Ed.: T. Ishikawa), Wiley,
Chichester, UK, 2009, p. 145–185.
2008, 4, 908–919.
[71] S. Sinnecker, A. Rajendran, A. Klamt, M. Diedenhofen, F. Neese, J. Phys.
Chem. A 2006, 110, 2235–2245.
[38] R. Prins, Mol. Phys. 1970, 19, 603–620.
[39] G. R. Eaton, S. S. Eaton, D. P. Barr, R. T. Weber, Quantitative EPR, Springer,
Vienna/New York, 2010.
Received: December 22, 2015
[40] B. S. Brunschwig, C. Creutz, N. Sutin, Chem. Soc. Rev. 2002, 31, 168–184.
Published Online: February 12, 2016
Eur. J. Inorg. Chem. 2016, 1274–1286
www.eurjic.org
1286
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim