When Do Strongly Coupled Diradicals Show Strongly Coupled Reactivity? Thermodynamics and Kinetics of Hydrogen Atom Transfer Reactions of Palladium and Platinum Bis(iminosemiquinone) Complexes

Article abstract of DOI:10.1021/acs.inorgchem.8b00068

The 2,2′-biphenylene-bridged bis(iminosemiquinone) complexes (^tBuClip)M [^tBuClipH₄ = 4,4′-di-tert-butyl-N,N′-bis(3,5-di-tert-butyl-2-hydroxyphenyl)-2,2′-diaminobiphenyl; M = Pd, Pt] can be reduced to the bis(aminophenoxide) complexes (^tBuClipH₂)M by reaction with hydrazobenzene (M = Pd) or by catalytic hydrogenation (M = Pt). The palladium complex with one aminophenoxide ligand and one iminosemiquinone ligand, (^tBuClipH)Pd, is generated by comproportionation of (^tBuClip)Pd with (^tBuClipH₂)Pd in a process that is both slow (0.06 M^-1 s^-1 in toluene at 23 °C) and only modestly favorable (K_com = 1.9 in CDCl₃), indicating that both N-H bonds have essentially the same bond strength. The mono(iminoquinone) complex (^tBuClipH)Pt has not been observed, indicating that the platinum analogue shows no tendency to comproportionate (K_com < 0.1). The average bond dissociation free energies (BDFE) of the complexes have been established by equilibration with suitably substituted hydrazobenzenes, and the palladium bis(iminosemiquinone) is markedly more oxidizing than the platinum compound, with hydrogen transfer from (^tBuClipH₂)Pt to (^tBuClip)Pd occurring with ?G° = ?8.9 kcal mol^-1. The palladium complex (^tBuClipH₂)Pd reacts with nitroxyl radicals in two observable steps, with the first hydrogen transfer taking place slightly faster than the second. In the platinum analogue, the first hydrogen transfer is much slower than the second, presumably because the N-H bond in the monoradical complex (^tBuClipH)Pt is unusually weak. Using driving force-rate correlations, it is estimated that this bond has a BDFE of 55.1 kcal mol^-1, which is 7.1 kcal mol^-1 weaker than that of the first N-H bond in (^tBuClipH₂)Pt. The two radical centers in the platinum, but not the palladium, complex thus act in concert with each other and display a strong thermodynamic bias toward two-electron reactivity. The greater thermodynamic and kinetic coupling in the platinum complex is attributed to the stronger metal-ligand ? interactions in this compound.

Full text of DOI:10.1021/acs.inorgchem.8b00068

Forum Article
pubs.acs.org/IC
Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX
When Do Strongly Coupled Diradicals Show Strongly Coupled
Reactivity? Thermodynamics and Kinetics of Hydrogen Atom
Transfer Reactions of Palladium and Platinum Bis(iminosemiquinone)
Complexes
Kyle M. Conner, AnnaMaria C. Arostegui, Daniel D. Swanson, and Seth N. Brown*
Department of Chemistry and Biochemistry, University of Notre Dame, 251 Nieuwland Science Hall, Notre Dame, Indiana
46556-5670, United States
S
* Supporting Information
ABSTRACT: The 2,2′-biphenylene-bridged bis(iminosemiquinone) complexes (tBuClip)M [tBuClipH4 = 4,4′-di-tert-butyl-
N,N′-bis(3,5-di-tert-butyl-2-hydroxyphenyl)-2,2′-diaminobiphenyl; M = Pd, Pt] can be reduced to the bis(aminophenoxide)
complexes (tBuClipH2)M by reaction with hydrazobenzene (M = Pd) or by catalytic hydrogenation (M = Pt). The palladium
complex with one aminophenoxide ligand and one iminosemiquinone ligand, (tBuClipH)Pd, is generated by comproportionation
of (tBuClip)Pd with (tBuClipH2)Pd in a process that is both slow (0.06 M−1 s−1 in toluene at 23 °C) and only modestly favorable
(Kcom = 1.9 in CDCl3), indicating that both N−H bonds have essentially the same bond strength. The mono(iminoquinone)
complex (tBuClipH)Pt has not been observed, indicating that the platinum analogue shows no tendency to comproportionate
(Kcom < 0.1). The average bond dissociation free energies (BDFE) of the complexes have been established by equilibration with
suitably substituted hydrazobenzenes, and the palladium bis(iminosemiquinone) is markedly more oxidizing than the platinum
compound, with hydrogen transfer from (tBuClipH2)Pt to (tBuClip)Pd occurring with ΔG° = −8.9 kcal mol−1. The palladium
complex (tBuClipH2)Pd reacts with nitroxyl radicals in two observable steps, with the first hydrogen transfer taking place slightly
faster than the second. In the platinum analogue, the first hydrogen transfer is much slower than the second, presumably because
the N−H bond in the monoradical complex (tBuClipH)Pt is unusually weak. Using driving force−rate correlations, it is
estimated that this bond has a BDFE of 55.1 kcal mol−1, which is 7.1 kcal mol−1 weaker than that of the first N−H bond in
(tBuClipH2)Pt. The two radical centers in the platinum, but not the palladium, complex thus act in concert with each other and
display a strong thermodynamic bias toward two-electron reactivity. The greater thermodynamic and kinetic coupling in the
platinum complex is attributed to the stronger metal−ligand π interactions in this compound.
with the opportunity for the molecule to engage in one-electron
chemistry of various sortsdimerization, atom abstraction,
addition to unsaturated molecules, etc.in order to increase its
bonding. Many radicals can easily form additional strong bonds,
and it is this high driving force that creates reactivity.1 When
steric or electronic effects combine to decrease the propensity
of radicals to form additional bonds, then persistent or isolable
radicals of modest reactivity result.2 Familiar examples from
organic chemistry include nitroxyl radicals. Of more relevance
to this Forum, transition metals are well-known to be able to
stabilize radicals. A classic example of this stabilization is found
with semiquinone radicals. These can exhibit some stability as
free organic radicals, but they form highly stable inorganic
INTRODUCTION
■
The term “radical” is freighted with a connotation of “reactive”;
it is ingrained in chemists that free radicals are highly reactive.
Yet a moment’s reflection by any inorganic chemist will indicate
the peculiarity of this reflex: the defining feature of a radical, the
presence of an unpaired electron, is not in and of itself
associated with any particular reactivity. Transition-metal
complexes with one or more unpaired electrons are common
and are not possessed of special reactivity. Indeed, Hund’s rule
indicates that species with greater numbers of unpaired
electrons should, all other things being equal, be more stable
than those with fewer unpaired electrons. This expectation is
borne out chemically, for example, in the notably higher
thermodynamic and kinetic reactivity of singlet dioxygen
compared to triplet dioxygen.
Special Issue: Applications of Metal Complexes with Ligand-
Centered Radicals
Of course, many free radicals are highly reactive, but this is
not principally due to unpaired electrons per se but rather the
common association of the presence of an unpaired electron
Received: January 8, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Forum Article
complexes where the metal center can dramatically tune their
electronic properties and give rise to species of controlled
reactivity.3 This moderation of the reactivity is of considerable
interest in the development of reactive species based on
coordinated radicals4 because efficient catalysis requires
reagents that operate with low kinetic barriers at only modest
thermodynamic driving forces.
(II).9−14 The electronic structures of these compounds are well
described as “singlet diradicals”: the compounds are diamag-
netic but have open-shell character with substantial config-
uration interaction.15 We recently undertook a study to
measure the singlet−triplet gaps in these compounds and
found gaps of ∼1800 cm−1 for the palladium compounds and
∼3000 cm−1 for the platinum compounds. In both cases, the
radicals are thus very strongly coupled, with the larger coupling
in the platinum compounds is due to stronger metal−ligand
(anti)bonding interactions.16
Given such strong interradical communication, these bis-
(iminosemiquinone) complexes would seem to be ideal
candidates to test whether diradical complexes can achieve
the kind of kinetic or thermodynamic coupling in chemical
reactions that could lead to efficient multielectron trans-
formations. To date, the only chemistry reported for these
complexes is outer-sphere redox chemistry, where one- and
two-electron-oxidized or -reduced compounds can be generated
and, in many cases, isolated.13 In contrast, monoligated
complexes appear to be somewhat more reactive; for example,
a bimetallic (cyclooctadiene)platinum complex of a bis-
(amidophenoxide) has been shown to deliver electrons from
the amidophenoxide and protons from the bound olefin to
reduce an o-quinone.17 Here we prepare the chelating
bis(aminophenoxide) complexes of palladium(II) and
platinum(II) based on the tetradentate 4,4′-di-tert-butyl-N,N′-
bis(3,5-di-tert-butyl-2-hydroxyphenyl)-2,2′-diaminobiphenyl
(tBuClipH4)18 ligand. The bis(aminophenoxide) complexes
represent the formal product of double hydrogen atom addition
to the known bis(iminosemiquinone) complexes; interconver-
sions among these species thus open a window into the
thermodynamics and kinetics of hydrogen atom transfer
processes in strongly coupled diradical complexes.
One intrinsic problem, however, in realizing this efficient
reactivity is the disconnect between the one-electron reactivity
of radicals and the thermodynamic predilection of organic
substrates to undergo two-electron reactions. For example, gas-
phase dehydrogenation of ethane to form ethylene (and H2) is
endothermic by a modest 32.6 kcal mol−1, but formation of the
ethyl radical (and 0.5 H2) requires 48.5 kcal mol−1.5 If one is
restricted to one-electron reactions such as hydrogen atom
abstraction, one builds in an intrinsic thermodynamic barrier
(an overpotential, in electrochemical terms) that necessarily
implies a waste of energy. Any oxidant capable of oxidizing
ethane to an ethyl radical is grossly overpowered (by 64 kcal
mol−1!) for oxidizing an ethyl radical to ethylene. Insofar as
organic radicals with β-hydrogen atoms are unstable with
respect to disproportionation, this analysis is general for other
examples of dehydrogenation (such as alcohols to aldehydes or
ketones). Other two-electron oxidations, such as hydroxylation,
also suffer this thermodynamic bias; for example, oxidation of
ethane to an ethyl radical by a hydroxyl radical is 75 kcal mol−1
more endothermic than the subsequent oxidation of an ethyl
radical to ethanol.5 Even in enzymes that operate by a rebound
mechanism, where the oxidant that reacts with the initially
formed radical need not be as powerful as the one that abstracts
the initial hydrogen atom, this overpotential appears to be
common. For example, calculations on the cytochrome P450
oxidation of methane estimate that the rebound step is about
65 kcal mol−1 more exothermic than the initial hydrogen atom
abstraction.6 This exothermicity cannot appreciably assist the
kinetics of the reaction, which are dominated by the barrier to
the initial hydrogen atom abstraction. Of course, many
reactions of hydrocarbons do take place by radical mechanisms,
but they must have a significant driving force if they are to
occur at reasonable rates. Reactions with low driving forces,
such as acceptorless dehydrogenation of alkanes, do not involve
one-electron intermediates if they occur at reasonable rates at
moderate temperatures.7
If one-electron oxidants such as metal-bound radicals are to
react efficiently with organic substrates, they must have some
way of coordinating their action so that two one-electron
centers can act simultaneously on a substrate. One way to favor
such concerted two-electron reactivity is for the reactivity of the
two one-electron centers to be strongly coupled; in other
words, engagement of the first redox center in one-electron
reactions (such as hydrogen atom transfer) should strongly
increase the propensity of the second redox center to do so as
well. This kind of linkage between two one-electron events
finds a parallel in hydrogen atom transfer chemistry, where the
strong interdependence between proton- and electron-transfer
events in a chemical system creates a strong thermodynamic
bias for concerted, proton-coupled electron transfer over
stepwise pathways.8
RESULTS
■
Preparation of Palladium and Platinum Bis-
(aminophenoxide) Complexes (tBuClipH2)M. Attempted
t
metalation of palladium acetate with BuClipH4 gives only the
ligand-oxidized iminosemiquinone complex (tBuClip)Pd, even
under anaerobic conditions, with Pd(OAc)2 being reduced to
palladium black in the process.16 Once formed and purified, the
dark green bis(iminosemiquinone) complex (tBuClip)Pd reacts
with hydrazobenzene to form trans-azobenzene and the yellow
bis(aminophenoxide) complex (tBuClipH2)Pd (eq 1). In situ
monitoring of the reaction by NMR spectroscopy indicates that
the reaction is quantitative, but there are significant losses
because of the solubility of the product upon isolation. The
compound with an unsubstituted 2,2′-biphenylene bridge,
(Clip)Pd,10 reacts analogously.
The products show the properties expected for square-planar
palladium(II) complexes. They are diamagnetic and air-stable,
with normal 1H and 13C NMR spectra, and show essentially no
optical absorption at λ > 450 nm. The presence of the added
hydrogen atoms on the amine nitrogen atoms is suggested by
NH stretches in IR (3215 cm−1) and NH resonances in the 1H
In fact, when multiple organic radicals are coordinated to a
single transition-metal center, very strong communication
between the radicals is the rule rather than the exception. A
canonical example of this phenomenon is given by bis-
(iminosemiquinone) complexes of palladium(II) and platinum-
B
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Forum Article
NMR at δ 5.80. The relatively upfield position of this resonance
strongly suggests that the molecule does not form hydrogen-
bonded dimers in solution, in contrast to the structurally
characterized palladium bis(aminophenoxide) complexes
may be the source of the slight asymmetry of the Pd−N
distances. In contrast to the bis(iminosemiquinone) complex
(Clip)Pd, the 2,2′-biphenylene linker does not induce a
significant deviation of the coordination sphere from planarity,
with the angle between ligand planes in (ClipH2)Pd of only
2.7° compared to 30.1° in (Clip)Pd.10 Reduction of the ligand
framework does not affect the Pd−O distances significantly but
results in a 0.09 Å elongation of the average Pd−N distances
[from 1.954(5) Å in (Clip)Pd to 2.040(9) Å in (ClipH2)Pd].
The platinum complex (tBuClip)Pt also reacts with
hydrazobenzene to form trans-azobenzene and (tBuClipH2)Pt,
but the reaction is not preparatively useful because it does not
go to completion [Keq = 0.043(7) in CDCl3; eq 2].
16
t
{([3,5-tBu2C6H3NH]C6H2 Bu2O)2Pd}2 (with NH at δ 8.57)
t
or {(CH3)2C(CH2NHC6H2 Bu2O)2Pd}2 (with NH at δ
7.54).12
The 2-fold symmetry shown by the NMR spectra suggests
that the compound adopts a C2-symmetric geometry, which is
confirmed by the solid-state structure of (ClipH2)Pd (Table 1
Table 1. X-ray Crystallography of (tBuClipH2)Pd·C4D8O
empirical formula
temperature (K)
λ (Å)
C44H50D8N2O3Pd
200(2)
1.54178 (Cu Kα)
cryst syst
monoclinic
space group
total data collected
no. of indep reflns
Rint
P21/n
66059
7984
0.0461
Synthetically, catalytic hydrogenation over palladium on carbon
gives complete reduction and allows the isolation of
(tBuClipH2)Pt in 70% yield (eq 3). No reaction takes place
obsd reflns [I > 2σ(I)]
a (Å)
7316
9.6206(2)
b (Å)
14.6796(2)
c (Å)
29.2837(4)
α (deg)
90
β (deg)
91.7843(9)
γ (deg)
V (Å3)
90
4133.63(12)
Z
4
μ (mm−1
)
3.914
cryst size (mm)
no. of refined param
R indices [I > 2σ(I)]
R indices (all data)
GOF
0.17 × 0.12 × 0.05
606
with H2 in the absence of the heterogeneous catalyst. In
contrast to the palladium analogue, (tBuClipH2)Pt is quite air-
sensitive in solution, rapidly oxidizing to the bis-
(iminosemiquinone).
Generation of the Mono(iminosemiquinone) Com-
plex (tBuClipH)Pd. Treatment of a solution of (tBuClip)Pd
with a large excess of (tBuClipH2)Pd leads to slow bleaching of
the intense band at 967 nm and formation of a new
R1 = 0.0418, wR2 = 0.1026
R1 = 0.0462, wR2 = 0.1049
1.185
and Figure 1). The molecule crystallizes with a molecule of
tetrahydrofuran (THF) hydrogen-bonded to N1−H, which
Figure 1. Thermal ellipsoid plot of (ClipH2)Pd. Hydrogen atoms (except for the NH hydrogen atoms), lattice THF, and minor components of the
disordered tert-butyl groups have been omitted for clarity. Selected bond distances (Å): Pd−O1, 1.993(2); Pd−O2, 1.994(2); Pd−N1, 2.033(2);
Pd−N2, 2.046(3).
C
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Forum Article
chromophore in the near-IR region (Figure 2) as the
bis(iminosemiquinone) and bis(aminophenoxide) compropor-
solution, ambient-temperature electron paramagnetic resonance
(EPR) spectrum of the reaction mixture (Figure 3) is likewise
Figure 3. X-band EPR spectrum of (tBuClipH)Pd, generated in situ by
comproportionation of (tBuClipH2)Pd and (tBuClip)Pd (toluene, 296
Figure 2. Successive optical spectra of (tBuClip)Pd (2.6 × 10−5 M)
treated with (tBuClipH2)Pd (2.0 × 10−3 M). The initial spectrum is
solid red, and subsequent spectra are at 2400 s intervals, with the final
spectrum (solid purple) at t = 36000 s. Inset: A960 as a function of
time. The solid line is the best fit to first-order decay.
K). Simulated hyperfine coupling constants: AN = 21.8 MHz, AH‑5
−10.6 MHz, AH‑3 = −3.0 MHz, and APd = 13.2 MHz.
=
characteristic of an organic iminosemiquinone radical (g =
2.0037), with hyperfine couplings observable to nitrogen,
hydrogen, and palladium nuclei. It is very similar to spectra
previously observed for other neutral palladium monoimino-
semiquinones20−22 and is nearly superimposable on that shown
by (tBu2C6H2ONCMe2C5H4N)Pd(N3).19
tionate to form the monoaminophenoxide−monoiminosemi-
quinone complex (tBuClipH)Pd (eq 4). Formation of the
When mixtures of (tBuClip)Pd and (tBuClipH2)Pd are
examined by NMR, the signals of the two compounds remain
distinct and unshifted. Their intensities decrease (relative to an
internal standard) over time but do not go to zero, indicating
that comproportionation (eq 4) is not strongly favored
thermodynamically. The signals for the aromatic hydrogen
atoms of the odd-electron complex (tBuClipH)Pd are
apparently too broad to be observed, although a broadened
signal, presumably due to a tert-butyl group, is observed at 2.0
ppm. By measuring the decrease in the signal intensity of the
diamagnetic compounds compared to the internal standard at
equilibrium, one can measure a value of Kcom = 1.9 0.3 in
CDCl3.
monoradical complex is clean, as indicated by the presence of
isosbestic points at 488, 498, 530, 588, and 715 nm. The
reaction adheres to pseudo-first-order kinetics under these
conditions of excess (tBuClipH2)Pd (Figure 2, inset). The
observed rate constants generally increase with increasing
concentration of (tBuClipH2)Pd, and although reproducibility
is mediocre, one can roughly estimate an overall second-order
In contrast to the palladium analogue, there is no evidence
for appreciable comproportionation of (tBuClip)Pt with
(tBuClipH2)Pt. The optical spectrum of (tBuClip)Pt is
unchanged upon treatment with (tBuClipH2)Pt; significant
changes would have been expected because platinum
monoiminosemiquinones have spectral features similar to
those of their palladium analogues.23 The NMR spectral
intensities and chemical shifts of the two compounds are also
unchanged upon mixing. Comproportionation is thus con-
cluded to be thermodynamically unfavorable in the case of the
platinum complexes, with Kcom < 0.1.
rate constant kcom = 0.06
toluene.
0.02 L mol−1 s−1 at 23 °C in
The product (tBuClipH)Pd has not been isolated, but its
spectroscopic features unequivocally support its assignment as a
monoiminosemiquinone complex. In particular, the optical
spectrum of the equilibrium reaction mixture in the presence of
excess (tBuClipH2)Pd, with its structured, long-wavelength
(λmax = 869 nm) transition with a vibronic progression with a
spacing of ∼1250 cm−1, is extremely similar to those of other
well-characterized neutral palladium mono(iminosemiquinone)
complexes prepared by van der Vlugt and co-workers.19−21
Cationic [(ISQ)Pd(bpy)]+ has a similar band.11 The fluid
Thermodynamics of Hydrogenation of (tBuClip)M.
The platinum bis(iminosemiquinone) complex (tBuClip)Pt
reacts reversibly with hydrazobenzene to give azobenzene (eq
2), giving ΔG°2 = +1.85 kcal mol−1. The average bond
dissociation free energy (BDFE) of hydrazobenzene, upon
dehydrogenation to give trans-azobenzene as its product, is 59.6
D
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Forum Article
kcal mol−1, based upon its solution enthalpy of reaction with
excess 2,4,6-tri-tert-butylphenoxyl radical (BDFE = 76.7 kcal
mol−1)8 of −34.27 kcal mol−1 in C6H624 and neglecting entropy
changes in the reaction of the aroxyl radical with
hydrazobenzene, which appears to be reasonable in organic
hydrogen atom transfer reactions.25 This implies that the
average BDFE for the two N−H bonds in (tBuClipH2)Pt is
58.6 kcal mol−1 (Scheme 1). The fact that comproportionation
come to equilibrium in THF-d8 within 2 days in the presence of
solid palladium on carbon as a catalyst. By interposing several
different intermediary levels of oxidizing power, one can
indirectly determine the difference in the average BDFEs
between hydrazobenzene and 4,4′-dinitrohydrazobenzene as
3.56 kcal mol−1 (Table 2). Combined with the value of K5 and
Table 2. Relative Reducing Ability of Substituted
Hydrazobenzenes
Scheme 1. Determination of the Average N−H BDFEs of
average N−H BDFE in
Hydrazobenzene and (tBuClipH2)Pt
YC6H4NHNHC6H4Y′, relative to Y
X, X′
H, H
Y, Y′
H, H
K6
= Y′ = H (kcal mol−1
)
1.00
6.3(5)
64(7)
0.00
0.54
1.77
1.95
H, H
Br, Br
Br, Br
H, NO2
H, NO2
CO2CH3,
CO2CH3
1.90(5)
CO2CH3,
CO2CH3
CN, CN
13.2(7)
17.6(4)
2.71
3.56
CN, CN
NO2, NO2
of (tBuClipH2)Pt and (tBuClip)Pt is thermodynamically
unfavorable, with Kcom < 0.1, implies that the BDFE of an
N−H bond in (tBuClipH2)Pt must be greater than that of the
N−H bond in putative (tBuClipH)Pt by at least 1.3 kcal mol−1.
The reaction of hydrazobenzene with the palladium complex
(tBuClip)Pd (eq 1) goes to completion. Qualitatively, this
indicates that the N−H bonds in (tBuClipH2)Pd are
significantly stronger than those in its platinum analogue. In
order to make a quantitative estimate of the strength of the N−
H bonds in the palladium complex, a weaker reductant than
PhNHNHPh is required in order to allow the reaction to go to
a measurable equilibrium. This was achieved using 4,4′-
dinitrohydrazobenzene (eq 5), where NMR monitoring of
the average BDFE in hydrazobenzene, this gives an average
BDFE for the N−H bonds in (tBuClipH2)Pd of 63.1 kcal mol−1
(Figure 4). Using the value of ΔG°com determined in CDCl3,
−0.38(9) kcal mol−1, one may conclude that breaking an N−H
bond in (tBuClipH2)Pd requires slightly less free energy (62.9
kcal mol−1) than breaking the N−H bond in (tBuClipH)Pd
(63.3 kcal mol−1).
Kinetics of Hydrogen Atom Transfer Reactions of
(tBuClipH2)M with Oxygen Radicals. In order to assess the
kinetic reactivity of the N−H bonds in the bis-
(aminophenoxide) complexes, we assayed their reactions with
various nitroxyl and aroxyl radicals (Scheme 2). These radicals
are all stable monomers in solution, and the corresponding O−
H compounds span a range of about 11 kcal mol−1 in BDFE
(Table 3). In all cases, reactions of an excess (>2 equiv) of
reagent with (tBuClipH2)M resulted in the quantitative
formation of (tBuClip)M by NMR. These reagents are
known to have a strong predilection for single hydrogen
atom transfer reactions; as one-electron oxidants, they are not
capable of removing both hydrogen atoms simultaneously from
(tBuClipH2)M. The reactions thus presumably involve an initial
hydrogen atom transfer to form (tBuClipH)M, followed by a
second hydrogen atom transfer to form the final product
(tBuClip)M.
In the case of M = Pd, the observed reaction traces with most
oxidants display discernibly biphasic kinetics (Figure 5), with an
initially accelerating phase in the growth of A960, where
(tBuClip)Pd absorbs strongly (Figure 2; ε = 37500 L mol−1
cm−1 in toluene) and (tBuClipH)Pd absorbs only modestly (ε
= 3600 L mol−1 cm−1). The data can be fit to successive
irreversible pseudo-first-order reactions26 using known ex-
tinction coefficients to determine k1(obs) and k2(obs). In all
cases, k1 is observed to be somewhat faster than k2. On the basis
of the low concentrations of palladium and the relatively slow
kinetics measured earlier, comproportionation/disproportiona-
tion reactions cannot contribute significantly to the time
evolution of the palladium species; comproportionation would
take place on the time scale of hundreds of hours under these
conditions, and even the slowest of the reactions is complete in
under 5 h. Therefore, k1 is assigned to hydrogen atom transfer
from (tBuClipH2)Pd to the oxyl radical, while k2 is assigned to
the reaction between (tBuClipH2)Pd and O2NC6H4N
NC6H4NO2 (in THF-d8, due to the very low solubility of
nitrohydrazobenzene in aromatic solvents) gives K5 = 0.89(3).
In order to compare the BDFE of the palladium complex to
that of the platinum complex, one needs to be able to compare
the thermodynamics of the azobenzene/hydrazobenzene
couple to that of the 4,4′-dinitroazobenzene/4,4′-dinitro-
hydrazobenzene couple. This was done by constructing a
thermodynamic ladder, whereby successively more strongly
oxidizing azoarenes were allowed to equilibrate with more
reducing hydrazoarenes. These reactions (eq 6) were found to
E
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Forum Article
Figure 4. Thermodynamic ladder connecting (tBuClipH2)Pd to (tBuClipH2)Pt via intermediary substituted hydrazobenzenes. The values in bold are
the average N−H BDFE (kcal mol−1 at 296 K) for the indicated compound. Italicized values in red are measured differences in free energy based on
measurement of the equilibrium of eq 2, eq 5, or eq 6, with ΔBDFE = −0.5 × RT ln(Keq). The factor of 0.5 arises because two N−H bonds are
broken and formed in each equilibrium.
Scheme 2. Reactions of Oxygen Radicals with (tBuClipH2)M (M = Pd, Pt)
hydrogen atom transfer from the intermediate (tBuClipH)Pd to
the oxyl radical.
In the reactions of (tBuClipH2)Pt, only monophasic kinetics,
to form fully oxidized (tBuClip)Pt, are observed (Figure 6).
F
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Forum Article
Table 3. Kinetics of Reactions of (tBuClipH2)M (M = Pd, Pt) with Oxygen Radicals
a
radical RO•
galvinoxyl
BDFE of RO−H (kcal mol−1
)
ref for BDFE
27
k1{Pd} (L mol−1 s−1
)
k2{Pd} (L mol−1 s−1
)
k1{Pt} (L mol−1 s−1
)
74.1
67.5
65.6
65.1
63.3
228(5)
940(30)
38.5(9)
41.7(20)
2.77(6)
n.d.
745(10)
2390(80)
78(4)
b
PTIO
28
207(6)
14.6(2)
12.70(16)
1.64(2)
c
4-oxo-TEMPO
TEMPO
29
30
90.2(18)
b
3-carbamoyl-PROXYL
31
8.46(5)
a
b
Determined in benzene or toluene, except as noted. Anchored to the value of 76.7 kcal mol−1 for 2,4,6-tri-tert-butylphenoxyl in C6H6.8 Based on
c
aqueous electrochemistry, corrected to C6H6 based on the difference between the aqueous and C6H6 BDFEs of TEMPO-H. Based on equilibration
between 4-oxo-TEMPO and TEMPO-H in heptane.
Given the ligand-localized structure of the monoiminosemiqui-
none complex (in the case of palladium), one would expect
(tBuClipH)Pt to show similar spectroscopic features, and no
such spectral signatures (such as long-wavelength absorption)
are observed at any point in the reaction. The conclusion is
thus that the monoradical complex (tBuClipH)Pt does not
accumulate to a significant extent. Presumably, the oxygen
radical reacts substantially more rapidly (by at least a factor of
10) with (tBuClipH)Pt than it does with (tBuClipH2)Pt, with
kobs corresponding to the initial hydrogen atom transfer from
(tBuClipH2)Pt. The reaction of the palladium complex with
galvinoxyl also shows monophasic kinetics.
In all cases, varying the concentration of the oxidant results
in proportionate increases in all of the observed pseudo-first-
order rate constants, consistent with identification of the
reactions as overall second order (first order in both the oxidant
and aminophenoxide complex). The second-order rate
constants are compiled in Table 3.
Figure 5. Absorbance (at 960 nm) versus time for the reaction of
DISCUSSION
(tBuClipH2)Pd (2.3 × 10−5 M) with 3-carbamoyl-PROXYL (2.0 ×
■
10−3 M) (toluene, 23 °C). The curve is fit to successive irreversible
pseudo-first-order reactions. Inset: kobs versus [3-carbamoyl-
PROXYL].
Thermodynamics of Hydrogen Atom Transfer of
Palladium and Platinum Bis(iminosemiquinone) Com-
plexes. The bis(iminosemiquinone) complexes of group 10
metals are generally considered to have a central metal(II) ion
bonded to ligand-localized radicals; accessible redox reactions
are invariably assigned to ligand-centered reductions or
oxidations.13 Binding to the metal has a significant impact on
the properties of the aminophenoxide/iminosemiquinone
redox couple. The free organic aminophenol 4,6-di-tert-butyl-
2-(tert-butylamino)phenol has a BDFE of 76.7 kcal mol−1.32
Metal coordination greatly stabilizes the iminosemiquinone
relative to the aminophenol, weakening the N−H BDFEs of the
palladium- and platinum-bound aminophenoxides by 15−20
kcal mol−1.
The nature of the metal also has a significant thermodynamic
impact: the palladium(II) complex (tBuClip)Pd is some 8.9 kcal
mol−1 better able to abstract the elements of H2 than is the
platinum(II) complex (tBuClip)Pt (eq 7), as measured
Figure 6. Absorbance (at 863 nm) versus time for the reaction of
(tBuClipH2)Pt (2 × 10−5 M) with 3-carbamoyl-PROXYL (4.4 × 10−4
M) (toluene, 23 °C). The curve is fit to a simple pseudo-first-order
reaction. Inset: kobs versus [3-carbamoyl-PROXYL].
G
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Forum Article
experimentally via the thermodynamic ladder connecting the
two compounds (Figure 4). The difference is qualitatively
evident in their reactions with hydrazobenzene, as (tBuClip)Pd
oxidizes hydrazobenzene irreversibly to trans-azobenzene, while
(tBuClip)Pt only oxidizes hydrazobenzene to a small extent at
equilibrium.
independently of one another. Comproportionation between
(tBuClipH2)Pd and (tBuClip)Pd takes place readily, with the
equilibrium constant for comproportionation Kcom not too
different from the statistical value of 4 expected if the two
bonds were completely independent of one another. Kineti-
cally, a comparison of the rate constants for the reaction with
nitroxyl radicals of (tBuClipH2)Pd to those of (tBuClipH)Pd
shows that the fully reduced compound enjoys a kinetic
advantage not too different from the factor of 2 expected on
statistical grounds.
The situation is quite different for the platinum complexes.
Comproportionation of (tBuClipH2)Pt with (tBuClip)Pt is not
observed, and only single-exponential decay is observed in the
reactions of (tBuClipH2)Pt with oxygen radicals, consistent
with the presumed intermediate (tBuClipH)Pt having much
greater kinetic reactivity than the fully reduced compound. In
other words, both thermodynamically and kinetically, removal
of the first hydrogen atom from (tBuClipH2)Pt greatly weakens
the remaining N−H bond; the two ligands do not behave
independently.
Upon reflection, it makes sense that the metal should exert a
significant influence on the thermodynamics even of a
nominally ligand-localized reaction. It is well appreciated that
ancillary ligands can have dramatic effects on the metal-
centered reactivity, so it should not be too surprising that
ancillary metals can have dramatic effects on ligand-centered
reactivity. Spectroscopic studies indicate that the bonding
interactions in these bis(iminosemiquinone) complexes are
markedly stronger for the platinum complexes than for the
palladium complexes, with the differences in the energies of the
metal−ligand nonbonding and antibonding orbitals increasing
from 4800 cm−1 in (tBuClip)Pd to 6300 cm−1 in (tBuClip)Pt (a
difference of 4.2 kcal mol−1).16 Reduction of the iminosemi-
quinones results in increased occupation of the metal−ligand π
antibonding orbitals, which is more costly for the platinum
complexes, thus decreasing the relative stability of the
hydrogenated bis(aminophenoxide) complexes. This trend is
evident in the redox potentials of the complexes, where the
(tBuClip)M0/− redox potential is 80 mV more positive for
palladium and the (tBuClip)M−/2− redox potential is 220 mV
more positive. This 150 mV (average) difference translates to a
3.5 kcal mol−1 difference in the average BDFEs, accounting for
most of the 4.5 kcal mol−1 total difference. The balance of the
difference must be due to the palladium anions being more
basic than the platinum anions, which may be due to stronger
metal−nitrogen σ bonding in the platinum compounds.
Computationally, eq 7 has been analyzed using density
functional theory (DFT), using the B3LYP and SDD basis sets
for palladium and platinum and a 6-31G* basis set for other
atoms, on gas-phase molecules with all of the tert-butyl groups
replaced by hydrogen atoms. The computed value for ΔG°7,
−8.7 kcal mol−1, agrees exceptionally well with the
experimental value of −8.9 kcal mol−1. The good agreement
is surprising because the strongly multireference character of
the bis(iminosemiquinone) complexes15 does not inspire
confidence that their energetics could be accurately modeled
using a B3LYP functional. It is possible that some fortuitous
cancellation of errors in the descriptions of the two
bis(iminosemiquinone) complexes in eq 7 contributes to the
apparent accuracy of the DFT methods. This cannot, however,
be the full explanation because computations on the reactions
of bis(iminosemiquinones) with hydrazoarenes (eqs 2 and 5),
where only a single reagent has significant multireference
character, also give good results. The agreement of the
calculated values (ΔG°2 = 0.9 kcal mol−1; ΔG°5 = −2.1 kcal
mol−1) with the experimental values (1.8 and 0.1 kcal mol−1,
respectively) is not as sparkling as the agreement on eq 7 but is
still quite reasonable. We therefore conclude that the DFT/
B3LYP method is remarkably tenable for computing the
energetics of reactions involving these bis(iminosemiquinone)
complexes.
One can make a quantitative estimate of the thermodynamic
separation in the N−H bond strength of the two N−H bonds
using the observed kinetics. The rates of reaction of the
aminophenoxide complexes with oxygen radicals are clearly
correlated with the BDFEs of the O−H bonds being formed
(Figure 7): as the driving force for atom transfer increases, the
Figure 7. Correlation of the free energies of activation for the
abstraction of N−H bonds with the BDFEs of the formed O−H bonds
of the abstracting oxygen radicals: reaction of (tBuClipH2)Pd, red
circles; reaction of (tBuClipH)Pd, blue diamonds; reaction of
(tBuClipH2)Pt, black squares. Open symbols represent the reactions
with galvinoxyl and were not used in the correlation.
activation energy for atom transfer decreases. This kind of
Evans−Polanyi relationship is widely observed in hydrogen
atom transfer reactions.1,33 The correlation is good for all of the
nitroxyl radicals, but the reactions with galvinoxyl are much
slower than expected based on this correlation, possibly
because of the greater bulk of the galvinoxyl radical. The
lines relating the reactions of (tBuClipH2)M with the nitroxyl
Thermodynamic and Kinetic Coupling between the
N−H Bonds in (tBuClipH2)M. Qualitatively, there is a striking
contrast between the two metals in the degree of
communication between the N−H bonds in the bis-
(aminophenoxide) complexes (tBuClipH2)M. In the palladium
complex, the two N−H bonds appear to act essentially
H
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Forum Article
Figure 8. Frontier Kohn−Sham orbitals of (Clip)Pt species. (a) Highest occupied molecular orbital of (ClipH2)Pt. (b) Singly occupied molecular
orbital of (ClipH)Pt. (c) Highest occupied molecular orbital of (Clip)Pt.
radicals are essentially parallel [slope for M = Pd, −0.80(8);
slope for M = Pt, −0.78(10)], while the reaction of (tBuClipH)
Pd with the oxygen radicals is perhaps slightly less sensitive to
the driving force of the reaction [slope = −0.67(5)].
ligand redox-active orbitals has essentially no overlap with any
of the metal d orbitals (Figure 8c);11 population of this orbital
by the two ligand-centered electrons avoids antibonding
interactions.
These concerns will be more pressing for platinum than for
palladium. Spectroscopic studies establish that the difference in
energy between the nonbonding and antibonding combinations
in (tBuClip)M are larger for platinum (6300 cm−1) than
palladium (4800 cm−1),16 and presumably this difference carries
over to the antibonding interaction in (tBuClipH)M as well.
Because of the modest difference in energy between the
nonbonding and metal−ligand π* combinations of the redox-
active orbitals, both are significantly occupied in the open-shell
singlet ground states of the bis(iminosemiquinone) complexes.
However, the greater energy difference for platinum means that
there is less occupancy of the antibonding orbital, further
enhancing the energetic advantage of the bis-
(iminosemiquinone) oxidation state for platinum.
It is thus not surprising that the platinum compound should
show greater coupling than the palladium compound in its
hydrogen atom transfer reactions, but it is surprising that the
palladium compound shows essentially no coupling at all. The
palladium bis(iminosemiquinone) (tBuClip)Pd, with its
singlet−triplet gap of 1940 cm−1,16 shows very strong coupling
between the two ligand radicals. Yet this does not seem to
transfer over to its behavior in hydrogen atom transfer
reactions; only small differences from the behavior expected
on the basis of pure statistics are observed. Where (tBuClip)Pd
acts like a pair of ligand radicals, the iminosemiquinones in
(tBuClip)Pt act as a single entity, with two-electron chemistry
strongly favored over one-electron chemistry. Why this
cooperation is turned on for platinum but essentially absent
for palladium is unclear.
The observation of good correlations between the reaction
rates of the nitroxyl radicals and the reactions’ driving force
indicates that differences in the rate of reaction are largely
determined by differences in the bond strengths of the bonds
being broken and formed. This explanation may plausibly be
extended to explain the differences in reactivity between the
palladium and platinum complexes; that is, it is likely that the
faster rate of reaction of the oxygen radicals with (tBuClipH2)Pt
compared to its palladium congener is due to the weaker N−H
bond in the platinum complex. In this case, one can use the
observed kinetic correlations to estimate the difference in the
bond strengths for the scissile N−H bonds in the two
compounds. For the five reactions studied, ΔΔG⧧ for reactions
of (tBuClipH2)Pd compared to those of (tBuClipH2)Pt
[corresponding to −RT ln(k1{Pd}/k1{Pt})] averages 0.55
0.12 kcal mol−1. Given the slope of the Evans−Polanyi plot in
Figure 6 of about −0.8, this corresponds to a difference in the
driving force ΔΔG° = 0.70 0.15 kcal mol−1, with the reaction
of the platinum compound more exergonic than that of the
palladium compound. Given a BDFE of 62.9 kcal mol−1 for an
N−H bond in (tBuClipH2)Pd, this implies a BDFE of 62.2 kcal
mol−1 for an N−H bond in (tBuClipH2)Pt. Given the average
BDFE in the platinum complex of 58.6 kcal mol−1, this means
that the BDFE of the N−H bond in (tBuClipH)Pt is only 55.1
kcal mol−1, 7.1 kcal mol−1 weaker than the first N−H bond.
This corresponds to a Kcom of 6 × 10−6 for the
comproportionation of (tBuClipH2)Pt with (tBuClip)Pt,
consistent with the unobservability of the monoradical complex
(tBuClipH)Pt. On the basis of these thermodynamics, the
correlation in Figure 7 predicts that (tBuClipH)Pt would react
with nitroxyl radicals ∼104 times faster than (tBuClipH2)Pt
would, consistent with the observation of simple pseudo-first-
order kinetics in these reactions.
One possible reason for the platinum complex to shun the
monoradical is that it presents a uniquely unfavorable π-
bonding arrangement. The singly occupied π orbital in the
iminosemiquinone (the so-called redox-active orbital) is
relatively high in energy because of its C−O and C−N
antibonding character and therefore interacts strongly with the
metal dπ orbitals. In contrast, the filled ligand-centered π orbital
in the fully reduced aminophenoxide ligand is mostly oxygen in
character and is C−O nonbonding, hence lower in energy, and
so experiences a weaker antibonding interaction with the filled
dπ orbitals (Figure 8a). In the monoiminosemiquinone, the
ligand redox-active orbital is singly occupied and must be
metal−ligand antibonding (Figure 8b). In contrast, in the
bis(iminosemiquinone) complexes, one combination of the
CONCLUSIONS
■
A diradical is not just two radicals; interactions between the two
radicals, mediated, for example, by a transition metal, can make
a strongly coupled diradical an entity with unique properties.
The group 10 bis(iminosemiquinonate) complexes (tBuClip)M
(M = Pd, Pt) are excellent candidates to examine whether the
very strong coupling evident spectroscopically in both
compounds transfers to their reactivity. In hydrogen atom
transfer reactions, the two compounds behave markedly
differently. The palladium compound readily compropor-
tionates with the reduced (tBuClipH2)Pd to form (tBuClipH)-
Pd, and in hydrogen atom transfer reactions, the first hydrogen
atom is removed from (tBuClipH2)Pd only slightly faster than
the second hydrogen atom. These patterns are consistent with
the two ligand radicals acting essentially independently of one
another, both thermodynamically and kinetically. In contrast,
no comproportionation is ever observed with the platinum
I
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Forum Article
0.023 mmol Pd, 10 mol %), and a stirbar. Dry benzene (10 mL) was
vacuum transferred into the bomb and the mixture allowed to warm to
room temperature under vacuum. The bomb was then backfilled with
an atmosphere of hydrogen gas and the solution stirred for 4 h to give
a colorless solution. The bomb was taken into the drybox and the
reaction mixture filtered through a fritted funnel to remove Pd/C. The
eluate was collected in a 25 mL round-bottom flask and the solvent
removed in vacuo on the vacuum line. The solid was suspended in 3
mL of acetonitrile, collected via vacuum filtration on a fine-porosity
fritted funnel, and washed with 2 × 3 mL of acetonitrile. The product
complexes, and removal of the first hydrogen atom from
(tBuClipH2)Pt is much slower than removal of the second
hydrogen atom. Using linear free-energy correlations of the
reaction rates with driving forces, one can estimate that the
second N−H bond is 7.1 kcal mol−1 weaker than the first,
giving the platinum compound a strong thermodynamic
propensity toward two-electron chemistry. Both the attenuated
oxidizing power of the platinum compound (ΔΔG° for
hydrogenation = 8.9 kcal mol−1) and its greater tendency to
disproportionate from the monoradical state are due to
stronger metal−ligand π interactions in the platinum
compound.
1
was air-dried for 1 h, yielding 0.1444 g of (tBuClipH2)Pt (70%). H
t
NMR (CDCl3): δ 1.20, 1.28, 1.47 (s, 18H each, Bu), 6.58 (d, J = 2.1
Hz, 2H, H-3), 6.60 (s, 2H, NH), 7.04 (d, J = 2.2 Hz, 2H, H-5), 7.20
(d, J = 1.4 Hz, 2H, H-3′), 7.32 (d, J = 7.9 Hz, 2H, H-6′), 7.35 (dd, J =
7.9 and 1.7 Hz, 2H, H-5′). 13C{1H} NMR (CDCl3): δ 29.71, 31.30,
31.96 (C(CH3)3), 34.13, 35.33, 35.60 (C(CH3)3), 115.72, 119.01,
123.48, 123.54, 124.78, 127.61, 131.00, 137.02, 139.55, 146.93, 154.31,
168.40 (CO). IR (cm−1): 3200 (w, νNH), 2953 (m), 2928 (w), 2900
(w), 2863 (w), 1616 (w), 1602 (w), 1473 (m), 1442 (m), 1409 (m),
1364 (m), 1358 (w), 1352 (w), 1298 (m), 1270 (m), 1259 (m), 1252
(s), 1234 (w), 1202 (w), 1116 (w), 1034 (m), 968 (w), 921 (w), 901
(m), 873 (m), 829 (s), 814 (w), 769 (m), 743 (m), 718 (w), 691 (m).
Anal. Calcd for C48H66N2O2Pt: C, 64.19; H, 7.41; N, 3.12. Found: C,
64.45; H, 7.65; N, 3.00.
EXPERIMENTAL SECTION
■
Unless otherwise noted, procedures were carried out under a nitrogen
atmosphere using glovebox or vacuum-line techniques. NMR spectra
were measured on a Bruker Avance DPX 400 or 500 spectrometer.
Chemical shifts for 1H and 13C{1H} NMR spectra are reported in ppm
downfield of tetramethylsilane, with spectra referenced using the
known chemical shifts of the solvent residuals. IR spectra were
measured on a Jasco 6300 Fourier transform infrared spectrometer as
powders on attenuated-total-reflectance plates. UV−visible−NIR
spectra were recorded in 1 cm quartz cells on a ThermoFisher
Evolution Array spectrophotometer. Elemental analyses were
performed by M-H-W Laboratories (Phoenix, AZ). The bis-
(iminosemiquinone) complexes (tBuClip)M (M = Pd, Pt)16 and
(Clip)Pd10 were prepared by literature procedures.
EPR Spectroscopy. In the drybox, 9.9 mg of (tBuClipH2)Pd and
10.4 mg of (tBuClip)Pd were mixed in 0.6 mL of toluene-d8, and the
solution was monitored by NMR to assess the degree of equilibration
of the comproportionation reaction. After the reaction had reached
equilibrium, a 100 μL aliquot of this solution was diluted to 10 mL in
toluene, and 1.0 mL of this solution was further diluted to 5.0 mL in
toluene. This most dilute solution (∼80 μM in total palladium) was
analyzed by EPR spectroscopy in a screw-cap-sealed 3 mm quartz tube
at ambient temperature in a Bruker EMX X-band (9.624 GHz) EPR
spectrometer. Spectra were acquired with a power of 20.02 mW and a
modulation amplitude of 1.00 G. Spectra were simulated using the
MATLAB toolbox program EasySpin 5.2.11 (easyspin.org).
Azoarene/Hydrazoarene Equilibration Experiments.
Hydrazobenzene, azobenzene, and 4,4′-azobenzenedicarboxylate
dimethyl ester were commercially available (TCI) and were used as
received. 4-Nitroazobenzene was prepared by the condensation of
nitrosobenzene with 4-nitroaniline.34 Other 4,4′-disubstituted azoben-
zenes (X = Br, NO2, or CN) were prepared by the oxidative coupling
of the corresponding anilines using tert-butyl hypoiodite.35 In order to
measure the hydrogen atom transfer equilibrium between two azo
compounds, the following procedure was used. To a J. Young NMR
tube in the drybox were added approximately 0.007 g of the less
oxidizing azo compound, 0.002 g of 10% Pd/C, and 0.6 mL of THF-
d8. The tube was removed from the drybox and the solution frozen in
liquid nitrogen. The Teflon seal was unscrewed, and 50 μL of a 5% (v/
v) solution of hydrazine hydrate in THF-d8 was immediately syringed
into the NMR tube, after which the tube was resealed. While the
solution was still frozen, the NMR tube was attached to a vacuum line
and degassed for 5 min. The reaction with hydrazine was monitored
via 1H NMR, and if the reaction was incomplete in 1 day, an additional
30 μL of 5% hydrazine hydrate in THF-d8 was added, which invariably
caused the reaction to go to completion after 1 day of further reaction.
(When azobenzene was the less oxidizing azo compound,
hydrazobenzene was used directly instead of being generated in situ
by hydrazine reduction.) Chemical shifts in THF-d8 of the azo- and
hydrazobenzenes studied are given in Table S1.
(tBuClipH2)Pd. In the drybox, 0.1235 g of (tBuClip)Pd (0.153
mmol) and 0.0616 g of hydrazobenzene (0.344 mmol) were added to
a 25 mL round-bottomed flask and dissolved in 6 mL of
dichloromethane to give a dark-green solution. The solution was
stirred for 1.5 h on a stir plate to give a yellow solution. The solution
was opened to the air and the solvent removed using a rotary
evaporator. The yellow residue was slurried in 3 mL of acetonitrile,
collected via vacuum filtration on a fine-porosity fritted funnel, and
washed with 3 mL of acetonitrile. The product was then air-dried for
15 min to yield 0.0796 g of (tBuClipH2)Pd as orange crystals (64%).
t
1H NMR (CDCl3): δ 1.17, 1.30, 1.43 (s, 18H each, Bu), 5.80 (s, 2H,
NH), 6.48 (d, J = 2.1 Hz, 2H, H-3), 7.04 (d, J = 2.2 Hz, 2H, H-5), 7.26
(d, J = 7.9 Hz, 2H, H-6′), 7.28 (d, J = 1.5 Hz, 2H, H-3′), 7.39 (dd, J =
7.9 and 1.8 Hz, 2H, H-5′). 13C{1H} NMR (CDCl3): δ 29.72, 31.38,
32.00 (C(CH3)3), 34.19, 35.40, 35.60 (C(CH3)3), 116.52, 119.59,
123.55, 123.65, 124.68, 126.39, 130.85, 136.58, 140.20, 145.30, 154.48,
166.89 (CO). IR (cm−1): 3215 (w, νNH), 2952 (m), 2864 (w), 1617
(w), 1600 (w), 1562 (w), 1469 (m), 1441 (m), 1410 (m), 1362 (m),
1300 (m), 1261 (s), 1203 (m), 1163 (w), 1117 (w), 1023 (w), 1000
(m), 974 (w), 897 (m), 874 (m), 820 (s), 728 (s), 687 (m). Anal.
Calcd for C48H66N2O2Pd: C, 71.22; H, 8.22; N, 3.46. Found: C, 71.02;
H, 8.11; N, 3.53.
(ClipH2)Pd. Using the procedure for preparing (tBuClipH2)Pd,
0.1872 g of (Clip)Pd (0.269 mmol) and 0.1024 g of hydrazobenzene
(0.556 mmol) were reacted in 10 mL of dichloromethane for 2 h to
give, after slurrying in 3 mL of acetonitrile, vacuum filtration, and
washing with 4 × 3 mL of acetonitrile, 0.1805 g of (ClipH2)Pd as
1
orange crystals (96%). H NMR (CDCl3): δ 1.18, 1.41 (s, 18H each,
tBu), 5.73 (s, 2H, NH), 6.48 (d, J = 2.1 Hz, 2H, H-3), 7.05 (d, J = 2.2
Hz, 2H, H-5), 7.30 (d, J = 8.1 Hz, 2H, H-6′), 7.36 (dd, J = 7.3 and 1.8
Hz, 2H, H-3′), 7.41 (td, J = 7.3 and 1.2 Hz, 2H, H-4′), 7.46 (td, J = 7.8
and 1.75 Hz, 2H, H-5′). 13C{1H} NMR (CDCl3): δ 29.55, 31.93
(C(CH3)3), 34.19, 35.56 (C(CH3)3), 118.99, 119.49, 123.69, 126.11,
126.84, 127.35, 130.96, 130.99, 136.84, 140.29, 145.58, 166.96 (CO).
IR (cm−1): 3215 (w, νNH), 2951 (m), 2905 (w), 2866 (w), 1607 (w),
1565 (w), 1472 (m), 1440 (m), 1409 (m), 1360 (m), 1300 (m), 1256
(s), 1202 (w), 1161 (w), 1122 (w), 1001 (w), 962 (w), 917 (w), 871
(m), 830 (s), 766 (s), 691 (m). Anal. Calcd for C40H48N2O2Pd: C,
69.10; H, 6.96; N, 4.03. Found: C, 68.93; H, 6.82; N, 4.40.
The NMR tube was then brought into the drybox, where
approximately 7 mg of the more oxidizing azo compound was
added, and the solution thoroughly mixed. The equilibrium was then
1
measured over the course of several days via integration of the H
NMR spectra; in all cases, equilibration was achieved within 2 days, as
judged by the subsequent constancy of the NMR spectra.
Kinetics of Reactions of Oxygen Radicals with (tBuClipH2)M.
In the drybox, a 1 mM (tBuClipH2)M solution was prepared by
dissolving 0.01 g of the complex in 10 mL of toluene in a 20 mL glass
vial. In reactions with lower concentrations of oxygen radicals, a 10-
(tBuClipH2)Pt. Into a glass bomb were placed 0.2065 g of (tBuClip)
Pt (0.230 mmol), 24.6 mg of 10 wt % palladium on carbon (Aldrich;
J
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Forum Article
fold dilution of the above solution was carried out to maintain pseudo-
first-order conditions during the kinetic runs. Stock solutions of the
organic radical of the desired concentration were prepared in toluene.
In the drybox, 2 mL of the appropriate organic radical stock solution
was dispensed into the cuvette compartment of an angled-reservoir
Teflon-sealed (ARTS) cuvette (see the Supporting Information for
photographs), followed by the addition of 50 μL of the (tBuClipH2)M
solution to the reservoir of the cuvette. The cuvette was then sealed
with a Teflon plug and brought to the UV−vis instrument. The cell
was equilibrated at 23 °C in a cell block whose temperature was
maintained using a recirculating bath, and the spectrometer was
blanked using a solution of the radical. The reaction was initiated by
inverting the ARTS cell, and spectra were obtained while illuminating
the sample using only the visible lamp. Absorbances at λmax of the
(tBuClip)M complex as a function of time were fit either to first-order
kinetics or, in the case of reactions of the palladium complex with
nitroxyl radicals, to successive first-order reactions with fixed ε values
for (tBuClip)Pd and (tBuClipH)Pd. Pseudo-first-order rate constants
varied linearly with the concentration of the oxygen radicals (Figures
S1), and second-order rate constants were determined from the slopes
of the plots of kobs versus [RO•].
X-ray Crystallography. Crystals of (ClipH2)Pd·C4D8O were
deposited after a reaction mixture of (Clip)Pd with PhNHNHPh in
THF-d8 was mixed with hexane and allowed to evaporate slowly in air.
Crystals were placed in Paratone oil before being transferred to the
cold dinitrogen stream of the diffractometer. Initial attempts to run the
experiments at 120 K resulted in crystal fracturing, so data were
collected at 200 K. Data were reduced and corrected for absorption
using the program SADABS. The structure was solved using direct
methods, and non-hydrogen atoms not apparent from the initial
solutions were found on difference Fourier maps.
Both the tert-butyl group centered on C18 and the one centered on
C28 were disordered over two orientations. In each case, the disorder
was modeled by constraining opposite methyl groups in the two
components to have equal thermal parameters and allowing the
occupancies to refine; the refined occupancies of the major
components were 68.6(6)% and 55.7(6)%, respectively. The THF-d8
molecule displayed two different orientations for the middle CH2CH2
moiety (C52/C521 and C53/C531), which were modeled analo-
gously. The major orientation refined to 65.5(15)% occupancy.
Deuterium atoms on the solvent and hydrogen atoms on the
disordered tert-butyl groups were placed in calculated positions with
thermal parameters 1.5 times (for CH3) and 1.2 times (for CD2) that
of the carbon atoms to which they were attached. All other hydrogen
atoms were found on difference maps and refined isotropically.
Calculations used SHELXTL (Bruker AXS),36 with scattering factors
and anomalous dispersion terms taken from the literature.37 Further
details about the structure are given in Table 1.
Computational Methods. Computationally, the tert-butyl groups
in (tBuClip)M and (tBuClipH2)M were replaced with hydrogen atoms.
Geometry optimizations were performed on all compounds as singlet
gas-phase molecules using a B3LYP functional, with an SDD basis set
for the palladium or platinum atom and a 6-31G* basis set for all other
atoms, using the program Gaussian09.38 Geometries were optimized
by minimizing the energies of the compounds, and optimized
geometries were confirmed to be minima by the lack of imaginary
frequencies in vibrational analysis.
Accession Codes
CCDC 1815092 contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
via www.ccdc.cam.ac.uk/data_request/cif, or by emailing data_
request@ccdc.cam.ac.uk, or by contacting The Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge CB2
1EZ, UK; fax: +44 1223 336033.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: Seth.N.Brown.114@nd.edu.
ORCID
Seth N. Brown: 0000-0001-8414-2396
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the U.S. National Science
Foundation (Grant CHE-1465104). We gratefully acknowledge
the assistance of Dr. Allen G. Oliver with X-ray crystallography.
REFERENCES
■
(1) Mayer, J. M. Hydrogen Atom Abstraction by MetalOxo
Complexes: Understanding the Analogy with Organic Radical
Reactions. Acc. Chem. Res. 1998, 31, 441−450.
(2) Forrester, A. R.; Hay, J. M.; Thomson, R. H. Organic Chemistry of
Stable Free Radicals; Academic Press: New York, 1968.
(3) (a) Pierpont, C. G. Studies on charge distribution and valence
tautomerism in transition metal complexes of catecholate and
semiquinonate ligands. Coord. Chem. Rev. 2001, 216−217, 99−125.
(b) Pierpont, C. G. Unique properties of transition metal quinone
complexes of the MQ3 series. Coord. Chem. Rev. 2001, 219−221, 415−
433. (c) Pierpont, C. G. Ligand Redox Activity and Mixed Valency in
First-Row Transition-Metal Complexes Containing Tetrachlorocate-
cholate and Radical Tetrachlorosemiquinonate Ligands. Inorg. Chem.
2011, 50, 9766−9772.
(4) (a) Lyaskovskyy, V.; de Bruin, B. Redox Non-Innocent Ligands:
Versatile New Tools to Control Catalytic Reactions. ACS Catal. 2012,
́
2, 270−279. (b) Munha, R. F.; Zarkesh, R. A.; Heyduk, A. F. Group
transfer reactions of d0 transition metal complexes: redox-active
ligands provide a mechanism for expanded reactivity. Dalton Trans.
2013, 42, 3751−3766. (c) Broere, D. L. J.; Plessius, R.; van der Vlugt,
J. I. New avenues for ligand-mediated processes  expanding metal
reactivity by the use of redox-active catechol, o-aminophenol and o-
phenylenediamine ligands. Chem. Soc. Rev. 2015, 44, 6886−6915.
(d) Sarkar, B.; Schweinfurth, D.; Deibel, N.; Weisser, F. Coord. Chem.
Rev. 2015, 293−294, 250−262.
(5) Tsang, W. Heats of Formation of Organic Free Radicals by
Kinetic Methods. In Energetics of Organic Free Radicals; Martinho
Simoes, J. A., Greenberg, A., Liebman, J. F., Eds.; Blackie Academic
̃
and Professional: London, 1996; pp 22−58.
(6) Ogliaro, F.; Harris, N.; Cohen, S.; Filatov, M.; de Visser, S. P.;
Shaik, S. A Model “Rebound” Mechanism of Hydroxylation by
Cytochrome P450: Stepwise and Effectively Concerted Pathways, and
Their Reactivity Patterns. J. Am. Chem. Soc. 2000, 122, 8977−8989.
(7) Kumar, A.; Bhatti, T. M.; Goldman, A. S. Dehydrogenation of
Alkanes and Aliphatic Groups by Pincer-Ligated Metal Complexes.
Chem. Rev. 2017, 117, 12357−12384.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.8b00068.
(8) Warren, J. J.; Tronic, T. A.; Mayer, J. M. Thermochemistry of
Proton-Coupled Electron Transfer Reagents and its Implications.
Chem. Rev. 2010, 110, 6961−7001.
1
Plots of kobs versus [RO•], descriptions of the H NMR
spectra of azo- and hydrazobenzenes in THF-d8,
photographs of the ARTS cuvettes, and energies and
Cartesian coordinates of DFT-optimized structures
(PDF)
(9) Chaudhuri, P.; Verani, C. N.; Bill, E.; Bothe, E.; Weyhermuller,
̈
T.; Wieghardt, K. Electronic Structure of Bis(o-
iminobenzosemiquinonato)metal Complexes (Cu, Ni, Pd). The Art
of Establishing Physical Oxidation States in Transition-Metal
K
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Forum Article
Complexes Containing Radical Ligands. J. Am. Chem. Soc. 2001, 123,
2213−2223.
Quinone or o-Iminoquinone; pap = Phenylazopyridine). Inorg.
Chem. 2014, 53, 1021−1031.
(24) Mahoney, L. R.; Ferris, F. C.; DaRooge, M. A. Calorimetric
Study of the 2,4,6-Tri-t-butylphenoxy Radical in Solution. J. Am. Chem.
Soc. 1969, 91, 3883−3889.
(10) Mukherjee, C.; Weyhermuller, T.; Bothe, E.; Chaudhuri, P.
̈
Targeted Oxidase Reactivity with a New Redox-Active Ligand
Incorporating N2O2 Donor Atoms. Complexes of Cu(II), Ni(II),
Pd(II), Fe(III), and V(V). Inorg. Chem. 2008, 47, 11620−11632.
(25) Mader, E. A.; Manner, V. W.; Markle, T. F.; Wu, A.; Franz, J. A.;
Mayer, J. M. Trends in Ground-State Entropies for Transition Metal
Based Hydrogen Atom Transfer Reactions. J. Am. Chem. Soc. 2009,
131, 4335−4345.
(26) Jackson, W. G.; Harrowfield, J. M.; Vowles, P. D. Consecutive,
Irreversible First-Order Reactions. Ambiguities and Practical Aspects
of Kinetic Analysis. Int. J. Chem. Kinet. 1977, 9, 535−548.
(27) Lucarini, M.; Pedulli, G. F.; Cipollone, M. Bond Dissociation
Enthalpy of α-Tocopherol and Other Phenolic Antioxidants. J. Org.
Chem. 1994, 59, 5063−5070.
(28) Goldstein, S.; Russo, A.; Samuni, A. Reactions of PTIO and
Carboxy-PTIO with •NO, •NO2, and O2•−. J. Biol. Chem. 2003, 278,
50949−50955.
(29) Billone, P. S.; Johnson, P. A.; Lin, S.; Scaiano, J. C.; DiLabio, G.
A.; Ingold, K. U. Accurate O−H Bond Dissociation Energy Differences
of Hydroxylamines Determined by EPR Spectroscopy: Computational
Insight into Stereoelectronic Effects on BDEs and EPR Spectral
Parameters. J. Org. Chem. 2011, 76, 631−636.
(30) Mahoney, L. R.; Mendenhall, G. D.; Ingold, K. U. Calorimetric
and Equilibrium Studies on Some Stable Nitroxide and Iminoxy
Radicals. Approximate O−H Bond Dissociation Energies in Hydroxyl-
amines and Oximes. J. Am. Chem. Soc. 1973, 95, 8610−8614.
(31) Sen, V. D.; Tikhonov, I. V.; Borodin, L. I.; Pliss, E. M.; Golubev,
V. A.; Syroeshkin, M. A.; Rusakov, A. I. Kinetics and thermodynamics
of reversible disproportionation−comproportionation in redox triad
oxoammonium cations − nitroxyl radicals − hydroxylamines. J. Phys.
Org. Chem. 2015, 28, 17−24.
(32) Carter, S. M.; Sia, A.; Shaw, M. J.; Heyduk, A. F. Isolation and
Characterization of a Neutral Imino-semiquinone Radical. J. Am.
Chem. Soc. 2008, 130, 5838−5839.
(33) Evans, M. G.; Polanyi, M. Inertia and Driving Force of Chemical
Reactions. Trans. Faraday Soc. 1938, 34, 11−24.
(34) Slavov, C.; Yang, C.; Schweighauser, L.; Boumrifak, C.; Dreuw,
A.; Wegner, H. A.; Wachtveitl, J. Connectivity matters  ultrafast
isomerization dynamics of bisazobenzene photoswitches. Phys. Chem.
Chem. Phys. 2016, 18, 14795−14804.
(35) Okumura, S.; Lin, C.-H.; Takeda, Y.; Minakata, S. Oxidative
Dimerization of (Hetero)aromatic Amines Utilizing t-BuOI Leading to
(Hetero)aromatic Azo Compounds: Scope and Mechanistic Studies. J.
Org. Chem. 2013, 78, 12090−12105.
(36) Sheldrick, G. M. A short history of SHELX. Acta Crystallogr.,
Sect. A: Found. Crystallogr. 2008, 64, 112−122.
(37) International Tables for Crystallography; Kluwer Academic
Publishers: Dordrecht, The Netherlands, 1992; Vol C.
(11) Sun, X.; Chun, H.; Hildenbrand, K.; Bothe, E.; Weyhermuller,
̈
T.; Neese, F.; Wieghardt, K. o-Iminobenzosemiquinonato(1−) and o-
Amidophenolato(2−) Complexes of Palladium(II) and Platinum(II):
A Combined Experimental and Density Functional Theoretical Study.
Inorg. Chem. 2002, 41, 4295−4303.
(12) Min, K. S.; Weyhermuller, T.; Bothe, E.; Wieghardt, K.
̈
Tetradentate Bis(o-iminobenzosemiquinonato(1−)) π Radical Ligands
and Their o-Aminophenolate(1−) Derivatives in Complexes of
Nickel(II), Palladium(II), and Copper(II). Inorg. Chem. 2004, 43,
2922−2931.
(13) Kokatam, S.; Weyhermuller, T.; Bothe, E.; Chaudhuri, P.;
̈
Wieghardt, K. Structural Characterization of Four Members of the
Electron-Transfer Series [PdII(L)2]n (L = o-Iminophenolate Deriva-
tive; n = 2−, 1−, 0, 1+, 2+). Ligand Mixed Valency in the Monocation
and Monoanion with S = 1/2 Ground States. Inorg. Chem. 2005, 44,
3709−3717.
(14) Swanson, D. D.; Conner, K. M.; Brown, S. N. A chelating
bis(aminophenol) ligand bridged by a 1,1′-ferrocene-bis(para-phenyl-
ene) linker. Dalton Trans. 2017, 46, 9049−9057.
(15) Herebian, D.; Wieghardt, K. E.; Neese, F. Analysis and
Interpretation of Metal-Radical Coupling in a Series of Square Planar
Nickel Complexes: Correlated Ab Initio and Density Functional
Investigation of [Ni(LISQ)2] (LI SQ =3,5-di-tert-butyl-o-
diiminobenzosemiquinonate(1-)). J. Am. Chem. Soc. 2003, 125,
10997−11005.
(16) Conner, K. M.; Perugini, A. L.; Malabute, M.; Brown, S. N.
Group 10 Bis(iminosemiquinone) Complexes: Measurement of
Singlet-Triplet Gaps and Analysis of the Effects of Metal and
Geometry on Electronic Structure. Inorg. Chem. 2018, DOI:
10.1021/acs.inorgchem.8b00062.
(17) Deibel, N.; Hohloch, S.; Schweinfurth, D.; Weisser, F.; Grupp,
A.; Sarkar, B. Three-Way Cooperativity in d8 Metal Complexes with
Ligands Displaying Chemical and Redox Non-Innocence. Chem. - Eur.
J. 2014, 20, 15178−15187.
(18) Kopec, J. A.; Shekar, S.; Brown, S. N. Molybdenum(VI)
Complexes of a 2,2′-Biphenyl-bridged Bis(amidophenoxide): Com-
petition between Metal−Ligand and Metal−Amidophenoxide π
Bonding. Inorg. Chem. 2012, 51, 1239−1250.
(19) Broere, D. L. J.; Demeshko, S.; de Bruin, B.; Pidko, E. A.; Reek,
J. N. H.; Siegler, M. A.; Lutz, M.; van der Vlugt, J. I. Dinuclear
Palladium Complexes with Two Ligand-Centered Radicals and a
Single Bridging Ligand: Subtle Tuning of Magnetic Properties. Chem. -
Eur. J. 2015, 21, 5879−5886.
(20) Broere, D. L. J.; Metz, L. L.; de Bruin, B.; Reek, J. N. H.; Siegler,
M. A.; van der Vlugt, J. I. Redox-Active Ligand-Induced Homolytic
Bond Activation. Angew. Chem., Int. Ed. 2015, 54, 1516−1520.
(21) Broere, D. L. J.; van Leest, N. P.; de Bruin, B.; Siegler, M. A.; van
der Vlugt, J. I. Reversible Redox Chemistry and Catalytic C(sp3)−H
Amination Reactivity of a Paramagnetic Pd Complex Bearing a Redox-
Active o-Aminophenol-Derived NNO Pincer Ligand. Inorg. Chem.
2016, 55, 8603−8611.
(22) Broere, D. L. J.; de Bruin, B.; Reek, J. N. H.; Lutz, M.; Dechert,
S.; van der Vlugt, J. I. Intramolecular Redox-Active Ligand-to-Substrate
Single-Electron Transfer: Radical Reactivity with a Palladium(II)
Complex. J. Am. Chem. Soc. 2014, 136, 11574−11577.
(38) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian09,
revision A.02; Gaussian, Inc.: Wallingford, CT, 2009.
(23) (a) Deibel, N.; Schweinfurth, D.; Hohloch, S.; Fiedler, J.; Sarkar,
B. Donor−acceptor systems of Pt(II) and redox-induced reactivity
towards small molecules. Chem. Commun. 2012, 48, 2388−2390.
(b) Deibel, N.; Schweinfurth, D.; Hohloch, S.; Delor, M.; Sazanovich,
I. V.; Towrie, M.; Weinstein, J. A.; Sarkar, B. Electrochemistry,
Chemical Reactivity, and Time-Resolved Infrared Spectroscopy of
Donor−Acceptor Systems [(Qx)Pt(papy)] (Q = Substituted o-
L
DOI: 10.1021/acs.inorgchem.8b00068
Inorg. Chem. XXXX, XXX, XXX−XXX