Mechanism of Iodine(III)-Promoted Oxidative Dearomatizing Hydroxylation of Phenols: Evidence for a Radical-Chain Pathway

Article abstract of DOI:10.1002/chem.202002026

The oxidative dearomatization of phenols with the addition of nucleophiles to the aromatic ring induced by hypervalent iodine(III) reagents and catalysts has emerged as a highly useful synthetic approach. However, experimental mechanistic studies of this important process have been extremely scarce. In this report, we describe systematic investigations of the dearomatizing hydroxylation of phenols using an array of experimental techniques. Kinetics, EPR spectroscopy, and reactions with radical probes demonstrate that the transformation proceeds by a radical-chain mechanism, with a phenoxyl radical being the key chain-carrying intermediate. Moreover, UV and NMR spectroscopy, high-resolution mass spectrometry, and cyclic voltammetry show that before reacting with the phenoxyl radical, the water molecule becomes activated by the interaction with the iodine(III) center, causing the Umpolung of this formally nucleophilic substrate. The radical-chain mechanism allows the rationalization of all existing observations regarding the iodine(III)-promoted oxidative dearomatization of phenols.

Full text of DOI:10.1002/chem.202002026

Accepted Article
Accepted Article
Title: Mechanism of Iodine(III)-Promoted Oxidative Dearomatizing
Hydroxylation of Phenols: Evidence for Radical-Chain Pathway
Authors: Karol Kraszewski, Ireneusz Tomczyk, Aneta Drabinska,
Krzysztof Bienkowski, Renata Solarska, and Marcin Kalek
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.202002026
Link to VoR: https://doi.org/10.1002/chem.202002026
01/2020

10.1002/chem.202002026
Chemistry - A European Journal
FULL PAPER
Mechanism of Iodine(III)-Promoted Oxidative Dearomatizing Hy-
droxylation of Phenols: Evidence for Radical-Chain Pathway
Karol Kraszewski,[a,b] Ireneusz Tomczyk,[a,b] Aneta Drabinska,[c] Krzysztof Bienkowski,[a] Renata
Solarska,[a] and Marcin Kalek*[a]
[a]
K. Kraszewski, I. Tomczyk, Dr. K. Bienkowski, Dr. R. Solarska, Dr. M. Kalek
Centre of New Technologies
University of Warsaw
S. Banacha 2C, 02-097 Warsaw, Poland
E-mail: m.kalek@cent.uw.edu.pl
K. Kraszewski, I. Tomczyk
Faculty of Chemistry
University of Warsaw
[b]
[c]
L. Pasteura 1, 02-093 Warsaw, Poland
Dr. A. Drabinska
Faculty of Physics
University of Warsaw
L. Pasteura 5, 02-093 Warsaw, Poland
Supporting information for this article is given via a link at the end of the document.
Abstract: The oxidative dearomatization of phenols with the addition
of nucleophiles to the aromatic ring induced by hypervalent iodine(III)
reagents and catalysts has emerged as a highly useful synthetic
approach. However, experimental mechanistic studies of this
important process have been extremely scarce. In this report, we
describe systematic investigations of the dearomatizing hydroxylation
of phenols using an array of experimental techniques. Kinetics, EPR
spectroscopy, and reactions with radical probes demonstrate that the
transformation proceeds via a radical-chain mechanism, with a
phenoxyl radical being the key chain-carrying intermediate. Moreover,
UV and NMR spectroscopy, high-resolution mass spectrometry, and
cyclic voltammetry show that before reacting with the phenoxyl radical,
water molecule becomes activated by the interaction with the
iodine(III) center, causing the Umpolung of this formally nucleophilic
substrate. The radical-chain mechanism allows to rationalize all
existing observations regarding the iodine(III)-promoted oxidative
dearomatization of phenols.
indeed, several such enantioselective catalytic phenol
dearomatizations employing chiral iodine-containing compounds
have been recently reported.[1c-d,5]
Scheme 1. Oxidative dearomatizing addition of nucleophiles to phenols
promoted by iodine(III) reagents.
There are three distinct general mechanisms usually
proposed for these phenol dearomatizations with iodine(III)
reagents (Scheme 2). The most commonly invoked one assumes
the incipient coordination of the phenolic oxygen to the iodine(III)
center, forming intermediate A. This, due to the strongly electron-
withdrawing nature of the hypervalent iodine, primes the attack of
the nucleophile on the ring (pathway 1; TS1), resulting in
simultaneous dearomatization and reduction of iodine. Such a
mechanism has been supported by the recent computational
study by Houk and Xue on the asymmetric spirolactonization,
wherein it has been shown to proceed via a feasible energy
barrier (~21 kcal·mol-1) and to reproduce the experimentally
observed enantioselectivity.[6] In the second mechanistic pathway,
often referred to as dissociative, intermediate A undergoes a
unimolecular fragmentation, reducing iodine and yielding
phenoxenium ion B (pathway 2; TS2). The latter subsequently
reacts with the nucleophile to afford the dearomatized product.
Introduction
Hypervalent iodine(III)-promoted oxidative dearomatization of
phenols, coupled with the addition of nucleophiles to the ortho or
para position of the ring, constitutes a powerful synthetic strategy,
enabling access to important molecular scaffolds (Scheme 1).[1]
The reaction has been accomplished in both inter- and
intramolecular fashion with a variety of nucleophilic species, such
as water, alcohols, carboxylates, amides, C–C double bonds,
electron-rich aromatic rings, and other.[2] The utility of this process
is further reinforced by the fact that its products are versatile
synthetic intermediates, which can be directly engaged in a
subsequent build-up of the molecular complexity.[3] Particularly
noteworthy, the reaction can be made catalytic in the iodine-
containing compound by its in situ reoxidation, for instance with
m-chloroperbenzoic acid.[4] This provides an opportunity for the
development of practical asymmetric variants of the reaction and,
The dissociative pathway has been advocated for in
a
computational work by Harned on the hydroxylation of p-cresol
using phenyliodine(III) diacetate (PIDA), however the calculated
energy barrier was rather high (>28 kcal·mol-1).[7] This was later
followed up by a Hammett study, the only existing experimental
investigation on the hypervalent iodine-promoted dearomatization
1
This article is protected by copyright. All rights reserved.

10.1002/chem.202002026
Chemistry - A European Journal
FULL PAPER
of phenols, which has been also interpreted in favor of the
dissociative mechanism.[8] Namely, the reaction has been found
to proceed preferentially with phenols containing electron-
donating substituents, able to stabilize the positive charge, rather
than with these containing electron-withdrawing groups. The final
mechanistic alternative involves the initial electrophilic addition of
the iodine(III) reagent to the phenolic ring (pathway 3). Resulting
intermediate C is then attacked by the nucleophile, substituting
the iodine-based leaving group in an SN2’ fashion (TS3).
According to the calculations by Ariafard, this pathway is the most
energetically favorable option in the PIDA-promoted addition of
methanol to p-cresol, requiring to cross a viable ~24 kcal·mol-1
barrier.[9] The incorporation of an iodine atom into the phenol ring,
observed in some reactions with PIDA, also indirectly lends
support to this mechanistic possibility.[10]
species is not directly involved in the stereodetermining bond
forming step between phenoxenium ion B and the nucleophile.
The present mechanistic ambiguity and the contradictory
conclusions stemming from the computational studies hinder a
rational design of more efficient and selective catalysts for this
important reaction. In this context, we decided to undertake
systematic investigations of the mechanism of the hypervalent
iodine(III)-promoted oxidative dearomatization of phenols. The
obtained results provide fresh and significant insights. Foremost,
they strongly suggest that the reaction in fact follows a radical-
chain pathway, involving a nucleophilic phenoxyl radical as the
key intermediate.[11] Oppositely, the formally nucleophilic
substrate (NuH in Scheme 1) actually undergoes an Umpolung by
a coordination to the iodine(III) center and reacts as an
electrophile in an single-electron transfer (SET) process. The
putative radical-chain mechanism emerging from our
investigations allows to encompass all the experimental
observations, including the regioselectivity, the Hammett
relationship, and the ability to accomplish asymmetric
transformations.
Results and Discussion
Kinetics
We started our investigations by examining the kinetics of a model
oxidative dearomatizing hydroxylation of 2,4-di-tert-butylphenol 1
promoted by PIDA in MeCN/H2O mixture (Scheme 3). The
reaction uniformly yields the 4-hydroxylation product 2 in high
selectivity (>98%), with a trace formation of the 2-hydroxylated
side-product.
Scheme 3. A model oxidative dearomatizing hydroxylation of 2,4-di-tert-
butylphenol 1 promoted by PIDA in acetonitrile.
Scheme 2. General mechanistic pathways proposed for the oxidative
dearomatizing addition of nucleophiles to phenols, exemplified by a PIDA-
promoted hydroxylation.
First, the progress of the reaction was monitored over time at
two different H2O contents in the reaction mixture, namely 5% and
10% (v/v). The formation of the product is clearly faster in the
latter case (Figure 1a). The analysis of the decay of the starting
material 1 shows that the reaction displays an overall pseudo-first
order in the concentrations of the substrates, as the linear
dependences of LN([1]/[1]0) vs. time were observed (Figure 1b).
Interestingly, when the reaction was carried out under identical
conditions, but using phenyliodine(III) bis(trifluoroacetate) (PIFA),
instead of PIDA, the formation of product was virtually
instantaneous (full conversion in the first NMR spectra; see below
for a plausible explanation of this phenomenon).
Interestingly, none of the mechanisms presented in Scheme
2 can fully account for all the observed characteristics of the
reaction. Namely, the two associative pathways 1 and 3 do not
provide explanation why the addition of a nucleophile displays a
clear preference to occur at the position of the phenolic ring
containing a substituent, while the other ortho or para sites are
unsubstituted, thus more sterically available. They offer also
wrong predictions from the electronic viewpoint, as the
nucleophile addition via both TS1 and TS3 should be favored next
to an electron-withdrawing group, while the Hammett plot shows
a precisely opposite trend.[8] On the other hand, the major flaw of
the dissociative pathway 2 is that it cannot rationalize the
enantioselectivity of the reactions carried out in the presence of
chiral iodine(III) reagents or catalysts. This is because the iodine
2
This article is protected by copyright. All rights reserved.

10.1002/chem.202002026
Chemistry - A European Journal
FULL PAPER
Figure 1. (a) Time-course of the reaction shown in Scheme 3 carried out in d3-
MeCN/H2O 95:5 and 90:10 (v:v) mixtures; [1]0=10 mM, [PIDA]0=10 mM, 0 °C;
concentrations were determined by 1H NMR spectroscopy relative to internal
standard. (b) Plots of LN([1]/[1]0) vs. time.
To ascribe the orders to the concentrations of specific
reagents, we performed an initial rate study for the reaction shown
in Scheme 3. The analysis of the obtained data results in an
experimental rate law (1), wherein the reaction is zeroth order in
the concentration of phenol 1 (Figure 2a), first order in the
concentration of PIDA (Figure 2b), and second order in the
concentration of H2O (Figure 2c). This is in line with the overall
pseudo-first order, determined for the reactions monitored over a
longer period (Figure 1; H2O is present in a great excess,
rendering its concentration constant with conversion, hence its
influence on the rate does not appear in those experiments).
Figure 2. Initial rate measurements for the reaction shown in Scheme 3 with
varied concentration of (a) 1, (b) PIDA, and (c) H2O. The insets show
corresponding LN-LN plots; if not otherwise stated: [1]0=10 mM, [PIDA]0=10 mM,
[H2O]0=2.78 M, 0 °C, d3-MeCN; product concentration was determined by 1H
NMR spectroscopy relative to internal standard.
rate = k[PIDA][H2O]2
(1)
Assuming the above form of the rate law, the thermodynamic
parameters of the process were determined by the means of
Eyring plot (Figure 3). Notably, the reaction displays a strongly
negative entropy of activation (–42.4 cal·mol-1·K-1), implying an
associative character of the rate-determining events. This is in a
good agreement with rate law (1) that is overall third-order,
including two waters, whose involvement should be indeed
associated with a large loss of entropy. The free-energy barrier is
calculated to 21.9±0.7 and 23.0±0.9 kcal·mol-1 at 0 and 25 °C,
respectively.
The results of above kinetic experiments are not easily
matched with the mechanisms depicted in Scheme 2. Foremost,
in all three of them phenol substrate participates prior to the
respective rate-determining steps (TS1-TS3), in a direct contrast
with the observed absence of phenol concentration in the rate law.
Additionally, the observed intriguing second order dependence of
the rate on the concentration of H2O cannot be satisfactorily
explained on the basis of those mechanisms. In particular, the
substantial rate enhancement with the increase of water content
speaks strongly against the dissociative mechanism (Scheme 2,
pathway 2), in which H2O is engaged only after the rate-
determining dissociation of intermediate A. This is further
Finally, we established that the reactions carried out using
H2O and D2O show no measurable difference in the rates (Figure
S1).
3
This article is protected by copyright. All rights reserved.

10.1002/chem.202002026
Chemistry - A European Journal
FULL PAPER
reinforced by the large negative entropy of activation, unlikely for
a dissociative process. The measured experimental free energy
barrier (22-23 kcal·mol-1) is also considerably lower than that
determined computationally for pathway 2 (>28 kcal·mol-1),[7] but
in agreement to those computed for pathways 1 and 3 (21-24
kcal·mol-1).[6,9] However, the computational studies on the
associative mechanisms have shown that in both the transition
states of type TS1 and TS3 (Scheme 2, pathways 1 and 3) for the
barriers to be feasible, the attack of nucleophile must proceed with
a simultaneous deprotonation by an acetate anion (not shown in
the scheme).[6,9] This is, however, inconsistent with the lack of a
solvent kinetic isotope effect upon the exchange of H2O for D2O.
Overall, the data allows with high confidence to discard the
phenols to undergo a facile single-electron oxidation coupled with
a proton transfer, is well-known and stems from the favorable
formation of resonance-stabilized phenoxyl radicals.[12] It
constitutes the basis for the antioxidant activity of phenols
employed both by Nature and in technological applications.[13]
Phenoxyl radicals are also intermediates in numerous reactions
of biological and synthetic importance.[14] On the other side, the
free radical chemistry of iodine(III) species is also rich, in
particular, they can serve both as the initial source of radicals and
sustain the propagation of a free radical chain.[15] This is due to
the relative ease of formation of an iodanyl(II) radical by the
cleavage of the weak hypervalent bond, either via homolysis or a
SET from a reducing agent. Finally, radical pathways have been
pathways shown in Scheme 2 as the correct reaction mechanisms. explicitly proven for closely related I(III)-promoted processes
engaging aryl ethers.[16]
Figure 3. Temperature-dependence of the initial rate for the reaction shown in
Scheme 3. The inset shows corresponding Eyring plot, assuming rate law (1);
[1]0=10 mM, [PIDA]0=10 mM, [H2O]0=2.78 M, d3-MeCN; product concentration
was determined by 1H NMR spectroscopy relative to internal standard.
The only mechanistic possibility that we could conceive, able
to encompass the observed kinetic characteristics of the reaction
is a radical-chain process. Specifically, the engagement of phenol
in a kinetically irrelevant step of the chain would allow to attain the
observed zeroth kinetic order. This is perfectly feasible, assuming
that phenoxyl radical, derived from the phenol, has the highest
relative stability of all the chain-carrying radicals, so that the step
which consumes it determines the overall propagation rate.
Figure 4. (a) EPR spectrum of phenol 1 and PIDA in 1,4-dioxane/H2O 9:1 (v/v);
[1]0=100 mM, [PIDA]0=100 mM, 25 °C. (b) DFT-calculated spectrum of phenoxyl
Conversely, fast steps involving the other chain-carrying radicals,
including the one engaging the phenol and regenerating the
phenoxyl radical, would not contribute to the overall rate. To
achieve the second order in the concentration of H2O and at the
same time conform to the lack of solvent kinetic isotope effect,
two H2O molecules would need to be involved in fast pre-equilibria
within the rate-determining sequence, whose final irreversible
step, however, should not include the breaking of an O–H bond.
This is also viable within the radical-chain mechanism, which is
proposed and discussed below.
radical 3. (c) A simulated spectrum of a mixture of 3 and an additional non-
coupled radical based on fitting to the experimental data (see the SI for details).
To probe the presence of free radicals, we measured an
electron paramagnetic resonance (EPR) spectrum of the reaction
mixture for the model dearomatization depicted in Scheme 3. Due
to a strong dielectric absorption of microwave radiation by
MeCN/H2O mixtures, acetonitrile was replaced with less polar 1,4-
dioxane (the reaction also proceeds well in the latter solvent,
albeit at a lower rate; see Table S9 for details). The registered
EPR spectrum confirms that free radicals are indeed generated
under these conditions (Figure 4a).[17] The analysis of the
obtained signal discloses that it arises from a mixture of two
radical species present in ~1:1 ratio (Figure 4c). One of them is
the phenoxyl radical 3 (g factor 2.005), derived from phenol 1, as
identified by the density functional theory modelling (Figure 4b).
The other is an unknown radical (g factor 2.004), which does not
In the following sections we present investigations aiming at
substantiating the involvement of radicals in the investigated
reaction and elucidating further details of the mechanism.
EPR spectroscopy
The possible involvement of free radicals in the mechanism of
dearomatization of phenols with hypervalent iodine oxidants
should not in fact be surprising. On one side, the propensity of
4
This article is protected by copyright. All rights reserved.

10.1002/chem.202002026
Chemistry - A European Journal
FULL PAPER
exhibit any strong hyperfine coupling to hydrogens. Putatively, it
may be some iodanyl(II) species, similar to that generated in the
recently reported electrochemical oxidation of iodoarenes (see
The iodine(III)-promoted oxidative dearomatizing additions of
nucleophiles to phenols are generally quite robust and they do not
usually require the exclusion of air. In particular, the reaction
depicted in Scheme 3 works equally well under an inert
atmosphere and when exposed to oxygen (Table S9). This should
not be considered as opposing a radical mechanism, however,
because although some processes involving radicals are
sensitive to oxygen, other are not. The latter group often proceeds
via long radical chains, which once initiated can continue
unperturbed.[24] We evaluated whether the reaction can be
interrupted by a more potent radical trap than O2, namely 2,2,5,5-
tetramethylpiperidin-1-oxyl (TEMPO). It was found that
substoichiometric amounts of TEMPO do not impact the rate of
the reaction, and as much as 4 equiv. are required to slow down
the dearomatization by ~2-fold (Figure 5). While the inhibitory
effect of TEMPO is only moderate, probably due to the large steric
hindrance of phenoxyl radical 3,[25] it strongly points to the radical
mechanism of the transformation.
below for
a discussion on the possible identity of this
intermediate).[18,19]
The formation of phenoxyl radical 3 under the reaction
conditions does not yet prove that it is involved in the mechanism
of the reaction and the creation of product. However, its
intermediacy would provide good rationalization for the observed
selectivity of the addition at the substituted ortho- or para-position
of the phenolic ring, as the delocalized unpaired electron is
stabilized more efficiently next to
hyperconjugation or resonance.
a
substituent via
Reactions with radical probes
To test for the actual intermediacy of phenoxyl radical in the
investigated process, we have synthesized two phenols
containing in-built radical clocks.[20] Upon subjecting to the
standard dearomatization conditions, phenol 4 having a pendant
olefin, afforded exclusively a direct hydroxylation product 5,
without any signs of cyclization (Scheme 4A). Such outcome may
be due to the high stability of the phenoxyl radical, rendering the
cyclization slow or altogether energetically disfavored.[21]
Conversely, the PIDA-promoted hydroxylation of phenol 6,
bearing a much faster 2,2-dimethylcyclopropane radical clock,[22]
lead to the formation of compound 7 as the dominant product, with
no direct hydroxylation of the phenolic ring observed (Scheme 4B).
Product 7 likely origins from the sequence starting by the ring
opening of the cyclopropylcarbinyl radical and a subsequent
reaction of the resulting ternary radical with MeCN. After an
oxidative incorporation of oxygen, intermediate amide[23]
undergoes a facile intramolecular Michael addition, restoring the
aromaticity and forming the azetidine ring. Irrespective of the
unusual course of the reaction with phenol 6, the characteristic
opening of the cyclopropyl ring lends strong support to the
involvement of phenoxyl radical in this process.
Figure 5. Time-course of the reaction shown in Scheme 3 carried out at in the
absence and presence of TEMPO; [1]0=10 mM, [PIDA]0=10 mM, d3-MeCN/H2O
9:1 (v/v), 0 °C; product concentration was determined by 1H NMR spectroscopy
relative to internal standard; the fluctuations of the red data points are due to
interference from TEMPO.
UV spectroscopy, NMR, and HRMS studies on the interaction
of PIDA and H2O in MeCN
Having established that phenoxyl radical is implicated in the
mechanism of the investigated reaction, we pursued to gain
additional insight into how the bond between the ring and the
nucleophile is formed. Unlike phenoxenium ion (B in Scheme 2),
phenoxyl radical will not directly react with H2O providing the final
product, as such step is deemed energetically disfavored due to
the generation of a hydrogen atom (H·). Therefore, we conjecture
that H2O needs to undergo a prior activation, presumably by PIDA,
as implied by rate law (1). The necessity for the nucleophile to
interact with the hypervalent iodine species to become capable of
reacting with phenoxyl radical would explain the asymmetric
induction rendered by chiral hypervalent iodine reagents.
The interaction of iodine(III) compounds with nucleophilic
species, such as H2O and MeOH, has been a subject of some
investigations.[26] It has been for instance determined that in an
aqueous solution, PhI(OTs)OH (Koser’s reagent) is completely
transformed into an equilibrium mixture of fully dissociated
[PhI(OH2)OH]+ and [Ph(HO)I–O–I(OH2)Ph]+ ions.[27] A high-
resolution mass spectrometry (HRMS) study of the solution of
Scheme 4. PIDA-promoted dearomatizing hydroxylations of phenols containing
radical clocks. The yields were determined in crude reactions mixtures by 1H
NMR spectroscopy relative to internal standard.
5
This article is protected by copyright. All rights reserved.

10.1002/chem.202002026
Chemistry - A European Journal
FULL PAPER
PIDA in MeCN/H2O mixture has also qualitatively demonstrated
the existence of analogous species.[28] We set out to more
thoroughly investigate the behavior of PIDA when exposed to H2O
under the conditions relevant to the oxidative dearomatizing
hydroxylation of phenols.
First, we looked for an indication of any interaction between
PIDA and H2O using UV spectroscopy. The UV spectrum of the
solution of PIDA in MeCN indeed undergoes an evident alteration
in the presence of H2O (Figure 6).
d3-MeCN/H2O 99:1 (v/v) mixture (Figure 7b). This correlates with
the very high rate of phenol dearomatization promoted by PIFA
(see above).
In the HRMS spectrum of the solution of PIDA in MeCN/H2O
9:1 (v:v) mixture, the major species are [PhIOH]+ (220.9456) and
[Ph(HO)I–O–IPh]+ (440.8846), in line with the previous report.[28]
The only peak from acetate-containing species detected is a small
[Ph(AcO)I–O–IPh]+ (482.8945) signal. In the corresponding
solution of PIFA, just the former two ions are detected.
Collectively, above results show that under typical conditions
employed in the oxidative dearomatizing hydroxylation of phenols,
a considerable fraction of the hypervalent iodine reagent is
converted into species containing HO–IIII and IIII–O–IIII moieties.
These constitute the activated forms of H2O, possessing an
oxidative ability and exhibiting electrophilic character. As such,
they should be prone for a single electron transfer coupling
reaction with the nucleophilic phenoxyl radical, forming the
alcohol product and generating an iodanyl(II) radical.
Electrochemical properties of I(III)-H2O adducts
As the final part of our investigations, using cyclic voltammetry,
we studied the oxidizing ability of the I(III)-H2O adducts, whose
formation was described in the previous section.
First, as a reference, the reduction of PIDA was recorded
voltammetrically at a glassy carbon electrode in MeCN, in the
presence of ferrocene (Fc) as an internal redox standard (Figure
8a). The registered voltammogram displays two irreversible
reduction waves. The first one occurs at the peak potential of –
1.34 V relative to Fc/Fc+, in a full agreement with the previously
reported value.[16b]
Figure 6. UV spectrum of PIDA in pure MeCN and MeCN/H2O mixtures;
[PIDA]=0.2 mM.
We then performed a similar measurement for the solution of
PIDA in MeCN/H2O 9:1 (v:v) mixture, wherein the species of
interest are being present. In this case, the direct application of
Fc/Fc+ internal redox standard turned out to be impossible,
because in this solvent mixture the iodine(III) oxidants reacted
with ferrocene, as noticeable by the change of the solution color.
This is a very similar behavior to that reported previously by
Compton and co-workers, in the case of PIDA solution in
1,1,1,3,3,3-hexafluoroisopropanol (HFIP).[16b] Following the
procedure applied in the above work, the voltammetry was thus
recorded against a silver/silver hexafluorophosphate reference
electrode, without ferrocene. In a separate measurement, the
redox potential of ferrocene was determined to be 0.05 vs. Ag/Ag+
(Figure S4), allowing for referencing the data obtained in both
solvents. The recorded voltammogram contains three irreversible
reduction waves (Figure 8b). The first one is broad and it occurs
at the peak potential of –0.65 V relative to Fc/Fc+. The other two
waves most likely originate from the sequential reduction of PIDA,
as they display peak potentials analogous to those in Figure 8a.[29]
The electrochemical characteristics of the solution of PIDA in
MeCN/H2O mixture demonstrates that it contains species that are
considerably stronger oxidants than PIDA itself (by more than 0.6
V). The superior oxidizing ability of the I(III)-H2O adducts
compared to PIDA provides rationalization for the selective
transfer of the OH group to the phenoxyl radical, furnishing the
hydroxylation product, while its acetoxylated analog is not formed
under the conditions shown in Scheme 3.[30]
Figure 7. 500 MHz 1H NMR spectra of (a) PIDA and (b) PIFA in d3-MeCN/H2O
99:1 (v/v) mixture; [PIDA/PIFA]=50 mM. New signals appearing in the presence
of H2O are indicated with X.
To establish what the origins of the observed changes in the
UV absorption pattern of PIDA are, we turned to nuclear magnetic
resonance (NMR) spectroscopy. The 1H NMR spectra of PIDA in
d3-MeCN contains a set of 3 aromatic protons peaks at 8.18, 7.71,
and 7.59 ppm. Upon the addition of H2O, a new set of somewhat
broad signals of aromatic protons emerges (Figure 7a, indicated
with X). The novel NMR peaks are more shielded compared to
those of PIDA, but they are still considerably downfield, indicating
the presence of an I(III) substituent in the ring. In the case of PIFA,
a similar phenomenon was also observed, however, the new
NMR signals are much broader and the quantity of the
corresponding species is considerably higher, as much as 90% in
6
This article is protected by copyright. All rights reserved.

10.1002/chem.202002026
Chemistry - A European Journal
FULL PAPER
Here it is important to remark that Scheme 5 presents only the
propagation steps of the radical chain. With the current data, we
are not able to establish the specific course of initiation and
termination steps, i.e. how the radicals are created and how they
fade. These aspects of the mechanism are yet to be elucidated.
Scheme 5. Putative radical-chain mechanism for the dearomatizing
hydroxylation of 2,4-di-tert-butylphenol 1 promoted by PIDA.
Figure 8. Voltammograms of (a) PIDA and ferrocene in MeCN and (b) PIDA in
MeCN/H2O 9:1 (v:v) mixture; [PIDA]=1 mM, [ferrocene]=1 mM, [n-Bu4NPF6]=0.1
M, scan rate=0.1 V·s-1, glassy carbon electrode (radius = 1.5 mm). In (a)
ferrocene was used as an internal redox reference potential. In (b) a Ag/AgPF6
reference electrode (Eferrocene = 0.05 V vs. Ag/Ag+, measured separately in
MeCN/H2O 9:1 (v:v)) was used due to reaction between ferrocene and PIDA.
Voltammetric starting potential and initial scan direction are indicated with black
arrows.
Nonetheless, the mechanistic picture depicted in Scheme 5 is
able to accommodate all the reported characteristics of the
iodine(III)-promoted dearomatization of phenols. First, it explains
the regioselectivity of the reaction – the addition will take place
preferentially next to
a substituent providing the highest
stabilization of the delocalized unpaired electron in the phenoxyl
radical, overrunning the steric factors. Secondly, the Umpolung of
the nucleophile, by the association with the I(III) center, renders
electron-rich phenols more reactive, rationalizing the negative
slope of the Hammett plot.[8] Finally, the requirement for the
nucleophile to be activated by the I(III) reagent during the reaction
with the phenoxyl radical allows for the enantioselective formation
of the product, in the case when a chiral hypervalent iodine
reagent or catalyst is applied.
We would like to stress that above findings regarding the
radical-chain mechanism are only directly applicable to the
specific case of the hypervalent iodine-mediated dearomatizing
hydroxylation of phenols. To assess their relevance to the other
related dearomatizing phenol functionalizations, additional
investigations are required.
Overall mechanism
The results of the experiments presented in the previous
sections suggest that the iodine(III)-promoted oxidative
dearomatizing hydroxylation of phenols follows a radical-chain
mechanism, summarized in Scheme 5.
Phenoxyl radical 3 is the key chain-carrying intermediate,
which due to its high stability will be the dominant radical species
present in the reaction mixture. A single-electron oxidation of 3 by
an I(III)-H2O adduct (the strongest oxidant present in the reaction
mixture), coupled with the OH-transfer, yields product 2 and
generates a iodanyl(II) radical. Owing to the stability of 3, this step
is slow and, thus, it determines the overall rate of the reaction.
Based on rate law (1), we propose that the actual oxidant is
phenyliodine(III) dihydroxide 8, however, it may possibly be
another related IIII–O species. Importantly, in putative TS4, the O–
H bond is not being broken, hence there will be no KIE upon
exchanging H to D, in line with the experimental findings.[31]
Iodanyl(II) radical (9 or a related species) is moderately stable
and it is most likely the second radical detected by the EPR
(Figure 4). Its fragmentation leads to PhI and a hydroxyl radical,
which is quenched by the phenol, regenerating 3. The overall
transformation of 9 to 3 is fast, thus, the concentration of phenol
1 does not enter the rate law.[32]
Conclusion
In summary, the mechanism of iodine(III)-promoted oxidative
dearomatizing hydroxylation of phenols was investigated by an
array of experimental techniques. Obtained results disprove the
previously proposed pathways and they strongly point to the
radical-chain mechanism. Specifically, it was determined that the
7
This article is protected by copyright. All rights reserved.

10.1002/chem.202002026
Chemistry - A European Journal
FULL PAPER
Soc. 2006, 128, 16440-16441; d) N. T. Vo, R. D. M. Pace, F. O'Har, M.
J. Gaunt, J. Am. Chem. Soc. 2008, 130, 404-405; e) J. L. Frie, C. S.
Jeffrey, E. J. Sorensen, Org. Lett. 2009, 11, 5394-5397; f) H. Liang, M.
A. Ciufolini, Org. Lett. 2010, 12, 1760-1763; g) M. Tissot, R. J. Phipps, C.
Lucas, R. M. Leon, R. D. M. Pace, T. Ngouansavanh, M. J. Gaunt, Angew.
Chem. Int. Ed. 2014, 53, 13498-13501; h) R. R. Reddy, S. S. Gudup, P.
Ghorai, Angew. Chem. Int. Ed. 2016, 55, 15115-15119; i) A. E.
Williamson, T. Ngouansavanh, R. D. M. Pace, A. E. Allen, J. D.
Cuthbertson, M. J. Gaunt, Synlett 2016, 27, 116-120; j) R. Coffinier, M.
E. Assal, P. A. Peixoto, C. Bosset, K. Miqueu, J.-M. Sotiropoulos, L.
Pouységu, S. Quideau, Org. Lett. 2016, 18, 1120-1123; k) D. M. Rubush,
M. A. Morges, B. J. Rose, D. H. Thamm, T. Rovis, J. Am. Chem. Soc.
2012, 134, 13554-13557.
phenol substrate is converted into the corresponding phenoxyl
radical, which constitutes the key chain-carrying intermediate. In
turn, the nucleophilic coupling partner needs to become activated
by the coordination to the I(III) center in order to react with the
phenoxyl radical, to generate the product together with a
iodanyl(II) radical that further propagates the chain. The radical-
chain mechanism, as the only one, is able to rationalize all the
features of the reaction. This conclusion is also in a good
agreement with the known propensity of both phenols and
hypervalent iodine(III) regents to undergo single-electron redox
processes. Computational studies aimed at elucidating further
details of the radical pathway are currently underway.
We are convinced that the established mechanism is valid for
the iodine(III)-promoted oxidative dearomatizing hydroxylation
and, with a high degree of confidence, alkoxylation of phenols. As
far as the other dearomatizations of phenols, such as the
spirolactionization, are concerned, we believe that further
experimental studies are needed to draw definite conclusions.
Nevertheless, the presented results provide important new
insights and a framework for understanding and improving the
iodine(III)-mediated reactions involving phenols. They carry also
major implications for the design of novel processes employing
hypervalent iodine reagents and catalysts, including asymmetric
ones.
[4]
[5]
a) T. Dohi, A. Maruyama, M. Yoshimura, K. Morimoto, H. Tohma, Y. Kita,
Angew. Chem. Int. Ed. 2005, 44, 6193-6196; b) R. D. Richardson, T.
Wirth, Angew. Chem. Int. Ed. 2006, 45, 4402-4404; c) T. Dohi, Y.
Minamitsuji, A. Maruyama, S. Hirose, Y. Kita, Org. Lett. 2008, 10, 3559-
3562; d) T. Dohi, Y. Kita, Chem. Commun. 2009, 2073-2085.
Reviews: a) M. Ngatimin, D. W. Lupton, Aust. J. Chem. 2010, 63, 653-
658; b) A. Parra, S. Reboredo, Chem. Eur. J. 2013, 19, 17244-17260; c)
A. M. Harned, Tetrahedron Lett. 2014, 55, 4681-4689; d) F. Berthiol,
Synthesis, 2015, 47, 587-603; selected examples: e) T. Dohi, A.
Maruyama, N. Takenaga, K. Senami, Y. Minamitsuji, H. Fujioka, S. B.
Caemmerer, Y. Kita, Angew. Chem. Int. Ed. 2008, 47, 3787-3790; f) S.
Quideau, G. Lyvinec, M. Marguerit, K. Bathany, A. Ozanne-Beaudenon,
T. Buffeteau, D. Cavagnat, A. Chénedé, Angew. Chem. Int. Ed. 2009, 48,
4605-4609; g) M. Uyanik, T. Yasui, K. Ishihara, Angew. Chem. Int. Ed.
2010, 49, 2175-2177; h) T. Dohi, N. Takenaga, T. Nakae, Y. Toyoda, M.
Yamasaki, M. Shiro, H. Fujioka, A. Maruyama, Y. Kita, J. Am. Chem. Soc.
2013, 135, 4558-4566; i) M. Uyanik, T. Yasui, K. Ishihara, Angew. Chem.,
2013, 125, 9385-9388; j) K. A. Volp, A. M. Harned, Chem. Commun.
2013, 49, 3001-3003; k) M. Uyanik, N. Sasakura, M. Mizuno, K. Ishihara,
ACS Catal. 2017, 7, 872-876; l) T. Dohi, H. Sasa, K. Miyazaki, M. Fujitake,
N. Takenaga, Y. Kita, J. Org. Chem. 2017, 82, 11954-11960; m) M.
Ogasawara, H. Sasa, H. Hu, Y. Amano, H. Nakajima, N. Takenaga, K.
Nakajima, Y. Kita, T. Takahashi, T. Dohi, Org. Lett. 2017, 19, 4102-4105;
n) N. Jain, S. Xu, M. A. Ciufolini, Chem. Eur. J. 2017, 23, 4542-4546; o)
T. Hashimoto, Y. Shimazaki, Y. Omatsu, K. Maruoka, Angew. Chem. Int.
Ed. 2018, 57, 7200-7204; p) Q. Ding, H. He, Q. Cai, Org. Lett. 2018, 20,
4554-4557.
Acknowledgements
We acknowledge the financial support from the National Science
Centre Poland (2016/22/E/ST5/00566). We thank the Faculty of
Physics of the University of Warsaw for granting generous access
to NMR spectrometer. Computer time was provided by the
Interdisciplinary Centre for Mathematical and Computational
Modelling of the University of Warsaw (G75-0).
[6]
H. Zheng, Y. Sang, K. N. Houk, X.-S. Xue, J.-P. Cheng, J. Am. Chem.
Soc. 2019, 141, 16046-16056.
Keywords: phenol dearomatization • hypervalent iodine •
mechanistic investigations • radical-chain pathway
[7]
[8]
[9]
A. M. Harned, Org. Biomol. Chem. 2018, 16, 2324-2329.
T. Tang, A. M. Harned, Org. Biomol. Chem. 2018, 16, 8249-8252.
B. Ganji, A. Ariafard, Org. Biomol. Chem. 2019, 17, 3521-3528.
[1]
a) L. Pouységu, D. Deffieux, S. Quideau, Tetrahedron 2010, 66, 2235-
2261; b) S. P. Roche, J. A. Porco Jr., Angew. Chem. Int. Ed. 2011, 50,
4068-4093; c) W. Sun, G. Li, L. Hong, R. Wang, Org. Biomol. Chem.
2016, 14, 2164-2176; d) W.-T. Wu, L. Zhang, S.-L. You, Chem. Soc. Rev.
2016, 45, 1570-1580.
[10] G. Wells, A. Seaton, M. F. G. Stevens, J. Med. Chem. 2000, 43, 1550-
1562.
[11] The involvement of radicals in the mechanism of the iodine(III)-promoted
dearomatization of phenols was proposed before, but this idea has not
been further explored, see: B. D. Gates, P. Dalidowicz, A. Tebben, S.
Wang, J. S. Swenton, J. Org. Chem. 1992, 57, 2135-2143.
[2]
Reviews: a) A. Pelter, R. S. Ward, Tetrahedron 2001, 57, 273-282; b) H.
Liang, M. A. Ciufolini, Tetrahedron 2010, 66, 5884-5892; selected
examples: c) S. Wang, B. D. Gates, J. S. Swenton, J. Org. Chem. 1991,
56, 1979-1981; d) S. Quideau, M. A. Looney, L. Pouységu, Org. Lett.,
1999, 1, 1651-1654; e) L. H. Pettus, R. W. Van De Water, T. R. R. Pettus,
Org. Lett. 2001, 3, 905-908; f) S. Canesi, D. Bouchu, M. A. Ciufolini, Org.
Lett. 2005, 7, 175-177; g) M. A. Marsini, Y. Huang, R. W. Van De Water,
T. R. R. Pettus, Org. Lett. 2007, 9, 3229-3232; h) L. Pouységu, S.
Chassaing, D. Dejugnac, A.-M. Lamidey, K. Miqueu, J.-M. Sotiropoulos,
S. Quideau, Angew. Chem. Int. Ed. 2008, 47, 3552-3555; i) M.-A.
Beaulieu, C. Sabot, N. Achache, K. C. Guérard, S. Canesi, Chem. Eur.
J. 2010, 16, 11224-11228; j) A. N. Flyer, C. Si, A. G. Myers, Nat. Chem.
2010, 2, 886-892; k) J. L. Wood, J. K. Graeber, J. T. Njardarson,
Tetrahedron 2003, 59, 8855-8858; l) I. R. Baxendale, J. Deeley, C. M.
Griffiths-Jones, S. V. Ley, S. Saaby, G. K. Tranmer, Chem. Commun.
2006, 2566-2568; m) C. Zheng, L. Wang, J. Li, L. Wang, D. Z. Wang,
Org. Lett. 2013, 15, 4046-4049.
[12] G. Litwinienko, K. U. Ingold, Acc. Chem. Res. 2007, 40, 222-230.
[13] a) L. R. Mahoney, Angew. Chem. Int. Ed. 1969, 8, 547-555; b) G. W.
Burton, K. U. Ingold, Acc. Chem. Res. 1986, 19, 194-201; c) Methods in
Enzymology, Vol. 186 (Eds. L. Packer, A. N. Glazer), Academic Press,
1990; d) M. C. Foti, J. Pharm. Pharmacol. 2007, 59, 1673-1685; e) P.
Gijsman in Handbook of Environmental Degradation of Materials (Ed. M.
Kutz), William Andrew Publishing, Oxford, 2012, pp 673-714.
[14] a) J. Stubbe, W. A. van der Donk, Chem. Rev. 1998, 98, 705-762; b) S.
Yamamura in PATAI'S Chemistry of Functional Groups (Ed. Z.
Rappoport), John Wiley & Sons, Ltd., 2009; c) T. R. Blum, Y. Zhu, S. A.
Nordeen, T. P. Yoon, Angew. Chem. Int. Ed. 2014, 53, 11056-11059; d)
M. N. Hopkinson, A. Gómez-Suárez, M. Teders, B. Sahoo, F. Glorius,
Angew. Chem. Int. Ed. 2016, 55, 4361-4366.
[15] a) H. Togo, M. Katohgi, Synlett 2001, 565-0581; b) L. Wang, J. Liu, Eur.
J. Org. Chem. 2016, 1813-1824; c) R. Budhwan, S. Yadav, S. Murarka,
Org. Biomol. Chem. 2019, 17, 6326-6341; d) X. Wang, A. Studer, Acc.
Chem. Res. 2017, 50, 1712-1724.
[3]
a) A. W. G. Burgett, Q. Li, Q. Wei, P. G. Harran, Angew. Chem. Int. Ed.
2003, 42, 4961-4966; b) P. Wipf, S. R. Spencer, J. Am. Chem. Soc. 2005,
127, 225-235; c) S. P. Cook, A. Polara, S. J. Danishefsky, J. Am. Chem.
8
This article is protected by copyright. All rights reserved.

10.1002/chem.202002026
Chemistry - A European Journal
FULL PAPER
[16] a) Y. Kita, H. Tohma, K. Hatanaka, T. Takada, S. Fujita, S. Mitoh, H.
Sakurai, S. Oka, J. Am. Chem. Soc. 1994, 116, 3684-3691; b) I. Colomer,
C. Batchelor-McAuley, B. Odell, T. J. Donohoe, R. G. Compton, J. Am.
Chem. Soc. 2016, 138, 8855-8861.
[17] The EPR spectra measured in pure 1,4-dioxane and toluene (Figures S2
and S3) are qualitatively similar to this in Figure 4A, indicating that the
formation of radicals from 1 and PIDA is not unique for the 1,4-
dioxane/H2O mixture.
[18] A. Maity, B. L. Frey, N. D. Hoskinson, D. C. Powers, J. Am. Chem. Soc.
2020, 142, 4990-4995.
[19] Theoretical predication of EPR spectra for iodine-containing radicals is
troublesome, due to the relativistic effects present in heavy atoms.
[20] a) D. Griller, K. U. Ingold, Acc. Chem. Res. 1980, 13, 317-323; b) M.
Newcomb, Tetrahedron 1993, 49, 1151-1176.
[21] While the cyclization of a primary 5-hexenyl radical has a rate constant
of 2.3·105 s-1 at 25 °C (see, C. Chatgilialoglu, K. U. Ingold, J. C. Scaiano,
J. Am. Chem. Soc. 1981, 103, 7739-7742), for the corresponding
benzylic 1-phenyl-5-hexenyl radical, the cyclization rate constant is
already as low as 0.14 s-1 (see, J. A. Franz, R. D. Barrows, D. M.
Camaioni, J. Am. Chem. Soc. 1984, 106, 3964-3967). Phenoxyl radical
is expected be even more stable than the benzylic radical, thus the rate
at which it cyclizes should be still lower.
[22] Typical rate constants for the opening of cyclopropylcarbinyl radical are
in the range of 107-109 s-1 at 25 °C, see: V. W. Bowry, J. Lusztyk, K. U.
Ingold, J. Am. Chem. Soc. 1991, 113, 5687-5698.
[23] For examples of acetamidations of radicals with MeCN, see: a) O.
Hammerich, V. D. Parker, J. Chem. Soc., Chem. Commun. 1974, 245-
246; b) P. S. Engel, W. K. Lee, G. E. Marschke, H. J. Shine, J. Org. Chem.
1987, 52, 2813-2817.
[24] For a recent study on the sensitivity of mechanistically different radical
reactions, see: L. Pitzer, F. Schäfers, F. Glorius, Angew. Chem. Int. Ed.
2019, 58, 8572-8576.
[25] V. W. Bowry, K. U. Ingold, J. Am. Chem. Soc. 1992, 114, 4992-4996.
[26] B. C. Schardt, C. L. Hill, Inorg. Chem. 1983, 22, 1563-1565.
[27] H. W. Richter, B. R. Cherry, T. D. Zook, G. F. Koser, J. Am. Chem. Soc.
1997, 119, 9614-9623.
[28] L. F. Silva, R. S. Vasconcelos, N. P. Lopes, Int. J. Mass Spectrom. 2008,
276, 24-30.
[29] A slight shift toward a less negative potential may be due to a hydrogen
bonding between PIDA and H2O, similarly but to a lower extent as has
been observed for the binding between PIDA and HFIP (see Ref. 16b for
details).
[30] The oxidative dearomitizing acetoxylation of phenols with PIDA requires
heating in neat acetic acid and it is considerably slower than the
hydroxylation. For example, see: M. Novak, S. A. Glover, J. Am. Chem.
Soc. 2004, 126, 7748-7749.
[31] The equilibrium isotope effect for the conversion of PIDA to 8 is expected
to be negligible, as the fractionation factors for the distribution of
deuterium between different OH groups are close to 1, see: V. Gold, Adv.
Phys. Org. Chem. 1969, 7, 259-331.
[32] The considerably large quantity of the presumed iodanyl(II) radical
detected by the EPR (Figure 4; ~1:1 ratio vs. 3) is in a contradiction with
the kinetic data. Namely, the accumulation of this radical implies that its
fragmentation and the subsequent reaction with phenol 1 have a
comparable rate to that for the reaction between radical
3 and
intermediate 8. This would in turn result in the appearance of the
concentration of 1 in the rate law, which is not the case. However, the
kinetics was measured in MeCN/H2O, while the EPR spectra was
registered in 1,4-dioxane/H2O, wherein the relative rates of the
mechanistic steps may be somewhat different. Alternatively, the EPR
signal may have another, yet unknown origin.
9
This article is protected by copyright. All rights reserved.

10.1002/chem.202002026
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents
A radical way: All-around experimental investigations demonstrate that the hypervalent iodine-promoted dearomatizing hydroxylation
of phenols proceeds via a radical-chain pathway.
Institute and/or researcher Twitter usernames: @Kalek_group
10
This article is protected by copyright. All rights reserved.