Increasing the steric hindrance around the catalytic core of a self-assembled imine-based non-heme iron catalyst for C-H oxidation

Article abstract of DOI:10.1039/d0ra09677f

Sterically hindered imine-based non-heme complexes4and5rapidly self-assemble in acetonitrile at 25 °C, when the corresponding building blocks are added in solution in the proper ratios. Such complexes are investigated as catalysts for the H₂O₂oxidation of a series of substrates in order to ascertain the role and the importance of the ligand steric hindrance on the action of the catalytic core1, previously shown to be an efficient catalyst for aliphatic and aromatic C-H bond oxidation. The study reveals a modest dependence of the output of the oxidation reactions on the presence of bulky substituents in the backbone of the catalyst, both in terms of activity and selectivity. This result supports a previously hypothesized catalytic mechanism, which is based on the hemi-lability of the metal complex. In the active form of the catalyst, one of the pyridine arms temporarily leaves the iron centre, freeing up a lot of room for the access of the substrate.

Full text of DOI:10.1039/d0ra09677f

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
RSC Advances
View Article Online
View Journal | View Issue
PAPER
Increasing the steric hindrance around the catalytic
core of a self-assembled imine-based non-heme
iron catalyst for C–H oxidation†
Cite this: RSC Adv., 2021, 11, 537
a
a
Federico Frateloreto,
Carla Sappino,
Osvaldo Lanzalunga
Giorgio Capocasa,
Giorgio Olivo,a Karim Abdel Hady,a
Marika Di Berto Mancini,a Stefano Levi Mortera,b
a
a
a
*
and Stefano Di Stefano
Sterically hindered imine-based non-heme complexes 4 and 5 rapidly self-assemble in acetonitrile at 25 ꢀC,
when the corresponding building blocks are added in solution in the proper ratios. Such complexes are
investigated as catalysts for the H2O2 oxidation of a series of substrates in order to ascertain the role and
the importance of the ligand steric hindrance on the action of the catalytic core 1, previously shown to
be an efficient catalyst for aliphatic and aromatic C–H bond oxidation. The study reveals a modest
dependence of the output of the oxidation reactions on the presence of bulky substituents in the
backbone of the catalyst, both in terms of activity and selectivity. This result supports a previously
hypothesized catalytic mechanism, which is based on the hemi-lability of the metal complex. In the
active form of the catalyst, one of the pyridine arms temporarily leaves the iron centre, freeing up a lot
of room for the access of the substrate.
Received 13th November 2020
Accepted 15th December 2020
DOI: 10.1039/d0ra09677f
rsc.li/rsc-advances
popular amine-based non-heme iron catalysts5 appearing in the
literature so far, both in the oxidation of aliphatic and aromatic
Introduction
C–H bonds.
The reaction mechanism adopted by complex 1 in the H2O2
C–H functionalization is a prominent topic in contemporary
research due to the opportunity of a late stage derivatization of
organic compounds. The chance of a direct installation of
a heteroatom on non-activated positions has opened new
synthetic routes and methodologies.1 In particular, our atten-
tion was ultimately oriented to the sustainable oxygenation of
aliphatic and aromatic C–H bonds carried out with green
terminal oxidants such as hydrogen peroxide (H2O2) in the
presence of non-heme iron complexes2 based on the imine
function.3,4 The choice of this kind of catalyst was guided by
a series of needs: (i) use of an abundant and environmentally
friendly metal, (ii) easy and (iii) cheap preparation of the
catalyst.
oxidation of both aliphatic and aromatic substrates was deeply
studied2b,d but some features remain still poorly understood
mostly due to the elusive character of the involved active
species. The hemi-lability of at least one of the pyridine arms on
the metal centre is believed to be essential for the iron activa-
tion of the H2O2, however, while the metal-based nature of the
active species is given for certain, its precise structure is still
unknown. An iron ve oxo-complex N5FeV]O was hypoth-
esized2b,d although an iron four (N5FeIV]O) based pathway
cannot be excluded.6
Complex 1, which can be assembled in acetonitrile by simple
addition of the cheap and commercially available Fe(CF3SO3)2,
2-picolylaldehyde and 2-picolylamine added in a 1 : 2 : 2 ratio,
respectively (see Scheme 1), was demonstrated to satisfy the
above requirements. In a number of reports it was shown that
complex 1 competes on equal terms with many of the most
aDipartimento di Chimica and Istituto CNR per i Sistemi Biologici (ISB-CNR), Sezione
`
Meccanismi di Reazione, Universita di Roma La Sapienza, P. le A. Moro 5, 00185
Rome, Italy. E-mail: stefano.distefano@uniroma1.it
Scheme
1 In situ preparation of complex 1 from cheap and
b
`
Area of Genetics and Rare Diseases, Unit of Human Microbiome, Bambino Gesu
commercially available precursors. Within 5 min from the addition of
the precursors the complex is completely formed and the solution
becomes deep violet.
Children's, Italy
† Electronic supplementary information (ESI) available: Full characterization of
complexes 4 and 5. See DOI: 10.1039/d0ra09677f
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 537–542 | 537

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
RSC Advances
Paper
In the recent literature supramolecular, non-covalent inter-
actions have been employed in order to direct the selective
oxidation of particular C–H bonds among many others present
in the same substrate.7 In one case7a–c two crown-ether receptors
were implanted in the ligand of a catalyst based on the Mn or
Fe(pdp) core,1i,5b for the recognition of an ammonium
anchoring group8 in the substrate (see Fig. 1, complex 2).
Recognition of the ammonium head of the substrate by the
crown-ether receptors allowed the selective oxidation of C–H
bonds located at the right distance from the anchoring group
(8–9 simple bonds). Application of the same strategy to the
imine-based catalytic core 1 (see Fig. 1, complex 3), did not
afford comparable results in terms of selectivity.9 In the latter
case, the difference between the selectivity properties of
complexes 1 and 3, devoid of and endowed with the crown-ether
receptors, respectively, towards the oxidation of the C–H bonds
present in the tested substrates was not evident enough, and
when appreciable, it was ascribed to a mere steric hindrance
rather than to recognition by the crown-ether moieties. In fact,
steric hindrance around the catalyst is known to affect the
reaction selectivity, favouring oxygenation of the most acces-
sible sites (C–H or C]C bonds) on the substrate.1k,10,11
Scheme 2 In situ preparation of complexes 4 and 5. Within 45 and
75 min, respectively, from the addition of the precursors, the
complexes are completely formed and the solution becomes deep
violet.
based on 1H and 13C NMR monitoring, HSQC 2D NMR, HR-MS
and UV-Vis spectroscopy, showed the high similarity of
complexes 4 and 5 with the parent complex 1 (compare, for
example, trace at t ¼ 44 min in Fig. S6† with the 1H NMR
spectrum of complex 1 reported in Fig. S1†). The UV-Vis spectra
reported in Fig. S22† show the near resemblance among
complexes 1, 4 and 5. Apart from a modest bathochromic effect
due to the TIPS groups, the shape of the bands related to the
iron core from 400 to 700 nm remains practically unchanged.
Eventually, the 2 : 2 : 1 stoichiometry of the new complexes 4
and 5 is denitely demonstrated by the Job's plots reported in
the ESI (Fig. S11 and S21,† respectively). Thus, the only differ-
ence among complexes 1, 4 and 5 should lie in the increasingly
limited access to the metal centre.
The results obtained in the H2O2 oxidation of a series of
aromatic compounds carried out in the presence of complexes 4
and 5 are reported in Table 1 together with those obtained with
catalyst 1 under the same conditions for the sake of compar-
ison. In all cases hardly detectable trace amounts or no trace of
products with oxidized lateral chain were found, in accordance
with the well-known preference of this imine based catalytic
core for aromatic C–H oxidation with respect to aliphatic C–H
oxidation.2d
When catalyst 1 is taken into account, total yield of oxidation
generally increases on increasing the size of the lateral alkyl
chain of the substrate as previously observed.2d When
complexes 4 and 5 are used as catalysts under the same
conditions, quite astonishingly and in contrast with what was
found with catalyst 3,9 the yields of phenol products remain
more or less the same for each substrate in comparison with
catalyst 1 in the limit of experimental errors. Even more
surprising is the fact that the presence of two or four, very bulky
TIPS groups in the catalyst backbone does not inuence at any
extent the (meta + para)/ortho ratio in the reaction products.
Such ratio remains the same along each series 1, 4 and 5.
Although the exact nature of the active species involved in
the reaction is still unknown, a series of clues collected in
previous investigations2 prompted us to hypothesize a mecha-
nism in which, aer an initial outer-sphere oxidation of FeII to
FeIII, the temporary detachment of one of the pyridine arms
would unmask the iron centre, which could be available to take
In order to shed light on the role and importance of steric
hindrance on the action of catalyst 1 we investigated in detail
the effect of the presence of bulky substituents in the catalyst
structure. Here below we report the results of such
investigation.
Results and discussion
Like complexes 1 and 3, new complexes 4 and 5, which are
characterized by an increased steric hindrance around the
catalytic core due to the presence of triisopropylsilyl (TIPS)
groups, were prepared in situ by self-assembly of the parent
compounds added in solution in the proper ratios (see Scheme
ꢀ
2), in acetonitrile at 25 C.
1H NMR monitoring of the related solutions showed that
complexes 4 (see ESI, Fig. S6†) and 5 (Fig. S18†) are completely
formed within 45 min and 75 min, respectively, in contrast with
complex 3, whose assembly requires 60 h at the same temper-
ature. Formation of complex 1 under the same conditions is
completed in 5 min.2a Characterization of the new complexes,
Fig. 1 Catalysts endowed with crown-ether receptors for substrate
decorated with ammonium head. Catalyst 2 is based on the M(pdp)
catalytic core (amino-pyridine catalyst) while catalyst 3 on complex 1
(imino-pyridine catalyst).
538 | RSC Adv., 2021, 11, 537–542
© 2021 The Author(s). Published by the Royal Society of Chemistry

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
Paper
RSC Advances
Table 1 H2O2 Oxidation of ethylbenzene, cumene and tert-butylbenzene carried out in acetonitrile at room temperature in the presence of
catalysts 1, 4 and 5a,b
Cat.
Substratec (recov.%)
orthoc (yield%)
meta + parac (yield%)
(m + p)/od
1
4
5
537
524
639
7.10.1
7.20.1
7.70.9
8.50.1
8.30.1
7.60.9
1.200.02
1.100.04
0.980.22
1
4
5
699
649
743
8.10.9
7.40.3
7.40.1
162
131
111
20.5
1.80.3
1.50.1
1
4
5
501
582
601
8.21
7.90.1
8.00.3
181
171
161
2.20.4
2.20.1
2.00.2
a
Reaction conditions are: substrate 0.20 M, H2O2 2.5 mol eq. added in 30 min with a syringe pump, catalyst 2% mol, 25 ꢀC, acetonitrile/H2O, 1 h 30
min total reaction time. b Yields from GC measurements calculated using nitrobenzene (0.5 mol eq.) as the internal standard. c Error is calculated
from at least three independent runs. d Error from propagation applied to at least three independent runs.
part to the oxidation process, generating an initial FeIII–OOH become completely unimportant when the monosubstituted
based species (Fig. S23†).
substrates reported in Table 1 are considered.
A possible evolution of the latter could be the heterolytic
In an attempt to exasperate the steric hindrance of the
cleavage of the peroxidic bond leading to a FeV]O species substrate, 1,3-di-tert-butylbenzene (Fig. 3) was employed in the
competent for the oxidative properties of the system.5a–d On the study and oxidized in the presence of complex 1 and complex 5
other hand, the homolytic cleavage of the same bond would which are the extreme limits of the series. In this case, with both
originate an in-cage couple {FeIV]O + HOc}. The HOc radical catalysts phenol B was found to be the largely predominant
would then attack the aromatic ring leading to a hydrox- product, with trace amounts of C and no trace of A. However,
ycyclohexadienyl radical, which in turn, would be oxidized by also in this case yield of phenol B, remains the same for cata-
FeIV]O to give the product upon deprotonation and to restore lysts 1 and 5, 7.0 ꢁ 0.3% and 7.3 ꢁ 0.1%, respectively. It has to
the FeIII species.6 However, whichever the active species is, be mentioned that under the same conditions but in the
many pieces of evidence have been collected pointing to absence of any ligand (Fenton-like conditions), although the
a controlled oxidation mechanism where free radical oxidations total yield was extremely low, also phenol A was found in the
do not play any role.2
reaction crude.
The fact that in the series reported in Table 1 the (meta +
Thus, to sum up, in the case of oxidation of simple aromatic
para)/ortho selectivity remains practically untouched for each substrates, we were not able to detect any effects due to the
substrate no matter what the catalyst is, is a clear-cut proof of presence of the TIPS groups on the backbone of the catalyst.
the absence of any steric interaction between the alkyl chain of
the substrate and the TIPS groups mounted on the ligand. A
little but appreciable substrate steric effect can be noted in the
series ethylbenzene, cumene, tert-butylbenzene. Independently
of the nature of the catalyst the (meta + para)/ortho ratio steadily
increases along the series (1, 1.5, 2, respectively), however the
access to the active core must be the same for all of three
catalysts 1, 4 and 5.
A possible explanation for the above, quite unexpected
results is that, although the catalytic core of 4 and particularly of
5 appears to be highly hindered in the close form (Fig. 2, le,
CPK model of complex 5), when one of the pyridine arms is
opened by H2O2 (see Fig. 2 right where a CPK model of the open
FeV]O putative active species is shown), the steric hindrance
Fig. 2 CPK models of complex 5 in its resting state (FeII, left) and in its
offered by the TIPS group decreases in such an extent as to
putative active form (FeV]O, right).
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 537–542 | 539

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
RSC Advances
Paper
Fig. 5 Oxidation of cyclohexanol and 2,6-dimethylcyclohexanol. The
reported yields are obtained by 1H NMR analysis of the reaction crude
after work-up and addition of bibenzyl (0.5 mol eq.) as the internal
standard. The oxidant was added over the course of 30 minutes, after
which the mixture was allowed to react for 1 hour at 25 ꢀC.
Fig. 3 Oxidation of 1,3-di-tert-butylbenzene. The reported yields
derive from GC analysis of the reaction crude after work-up and
addition of nitrobenzene (0.5 mol eq.) as the internal standard. The
oxidant was added over the course of 30 minutes, after which the
mixture was allowed to react for 1 hour at 25 ꢀC.
complex 5, seems to moderately enhance the catalyst sensitivity
to the steric bulk of the substrate.
Given that complex 1 is also known to be active in the
catalysis of the oxidation of aliphatic C–H bonds,2a,b complexes
1 and 5 were also compared in the oxidation of menthyl acetate
(see Fig. 4). Menthyl acetate is indeed oen used as a bench-
mark substrate in order to evaluate steric effects on the regio-
selectivity of C–H oxidation. In both cases alcohols D and E were
the major products. In the case of catalyst 1 total yield D + E was
about 20% with a D/E ratio of 0.25. When catalyst 5 was used,
total yield D + E was about 12% with a D/E ratio of 0.33. Thus, it
is evident that for this substrate the steric effect of the TIPS
groups is not unimportant and the D/E selectivity increases on
increasing the steric hindrance on the catalyst, the peripheral
isopropylic C–H bond being sterically more accessible than the
tertiary endocyclic one.
Conclusions
In this report the synthesis of new complexes 4 and 5 is
described. Self-assembly of both complexes is a relatively fast
process taking place within 45 and 75 min, respectively, in
CH3CN at 25 ꢀC when the starting building blocks are added in
solution in the proper ratios. Complexes 4 and 5 essentially
differ from the parent complex for an increased steric
hindrance around the metal ion.
When complexes 4 and 5 were tested as catalysts for the H2O2
oxidation of
a series of monoalkyl-substituted aromatic
Eventually, since complex 1 was proved to be an active
catalyst also for alcohol oxidation2c (although this reaction
cannot be properly considered a C–H oxidation but more
properly an oxidation, from a steric standpoint it can serve to
our purpose), we tested complexes 1 and 5 in the oxidation of
cyclohexanol and its more sterically congested analogue 2,6-
dimethylcyclohexanol (cis + trans mixture, see Fig. 5) to cyclo-
hexanone and 2,6-dimethylcyclohexanone F, respectively.
While no difference appears in the oxidation of cyclo-
hexanol, complex 5 seems to be less active in the oxidation of
2,6-dimethylcyclohexanol with respect to complex 1. The effect
is modest, but the origin of such difference has to be steric in
nature. It should be noted that catalyst 1 was already found to
be quite sensitive to steric effects in the oxidation of a-(poly)
substituted cyclohexanols.2c Thus, the increased volume of
compounds, their catalytic activity was found to be nearly
superimposable to that of the parent complex 1, within the
experimental error. This behaviour was ascribed to a strong
decrease of the steric hindrance when passing from the resting
state to the active state of the catalyst, where one of the pyridine
arms has temporarily le the iron centre, freeing up space for
the access of the substrate. The absence of any steric effect was
also found in the oxidation of a disubstituted aromatic
substrate like 1,3-di-tert-butylbenzene.
Although moderate, appreciable differences between the
efficiencies of catalysts 1 and 5 in term of yields and selectivity
instead arise when substrates endowed with hindering groups
around non-aromatic C–H oxidizable bonds are considered.
In conclusion it could be stressed that the insertion of large
and hindering groups in the catalyst backbone does not
necessary lead to the insurgence of steric effect on the catalysed
reaction. The data collected in this report reinforce the mech-
anistic hypothesis depicted in Fig. S23,† remarking the impor-
tance of the hemi-lability of the complex in the explication of
the catalytic action.
Experimental
Methods and material
Fig. 4 Oxidation of menthyl acetate. The reported yields are obtained
by 1H NMR analysis of the reaction crude after work-up and addition of
bibenzyl (0.5 mol eq.) as the internal standard. The oxidant was added
over the course of 30 minutes, after which the mixture was allowed to
react for 1 hour at 25 ꢀC.
GC analyses were carried out on a Varian CP-3800 gas chro-
matograph equipped with a capillary methylsilicone column (30
m ꢂ 0.25 mm ꢂ 25 mm) Chrompack CP-Sil 5 CB. NMR spectra
were recorded on a Bruker AC300 (300 MHz) spectrometer and
540 | RSC Adv., 2021, 11, 537–542
© 2021 The Author(s). Published by the Royal Society of Chemistry