Chemically triggered C-ON bond homolysis in alkoxyamines: Regioselectivity and chemoselectivity

Article abstract of DOI:10.1039/c3ob41549j

Recently, we examplified the activation of the C-ON bond homolysis by protonation, alkylation, benzylation, acylation, oxidation and complexation with a Lewis acid of the nitrogen atom of the 1-(pyridin-4-yl)ethyl fragment (Chem. Commun., 2011, 4291 and Org. Lett., 2012, 358) and of the 1-(pyridin-2-yl)ethyl fragment (J. Org. Chem. ASAP Doi:10.1021/jo401674v) of (N-(2-methylpropyl)-N-(1- diethylphosphono-2,2-dimethylpropyl)-N-oxyl) SG1-based alkoxyamines. The quaternization of the 1-(pyridin-3-yl)ethyl fragment by the aforementioned reactions was investigated for the corresponding SG1-based alkoxyamines. In sharp contrast to the quaternization at ortho and para positions of the pyridyl moiety, the effect of the quaternization at the meta position was weak. The effects of quaternization at ortho, meta and para positions were investigated through natural bond orbital and Mulliken charges, HOMO-LUMO interactions in the starting materials and the radical stabilization energy of the released 1-puridylmethyl radicals using DFT calculations with the B3LYP/6-31G(d) and UBMK/6-311+G(3df,2p)//R(O)B3LYP/6-31G(d) methods, respectively. The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c3ob41549j

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Chemically triggered C–ON bond homolysis in
alkoxyamines: regioselectivity and chemoselectivity†‡
Cite this: Org. Biomol. Chem., 2013, 11,
738
7
Gérard Audran,* Paul Brémond,* Matisse Bim Batsiandzy Ibanou,
Sylvain R. A. Marque,* Valérie Roubaud and Didier Siri
Recently, we examplified the activation of the C–ON bond homolysis by protonation, alkylation, benzylation,
acylation, oxidation and complexation with a Lewis acid of the nitrogen atom of the 1-(pyridin-4-yl)ethyl frag-
ment (Chem. Commun., 2011, 4291 and Org. Lett., 2012, 358) and of the 1-(pyridin-2-yl)ethyl fragment
(
J. Org. Chem. ASAP Doi:10.1021/jo401674v) of (N-(2-methylpropyl)-N-(1-diethylphosphono-2,2-dimethyl-
propyl)-N-oxyl) SG1-based alkoxyamines. The quaternization of the 1-(pyridin-3-yl)ethyl fragment by the
aforementioned reactions was investigated for the corresponding SG1-based alkoxyamines. In sharp con-
trast to the quaternization at ortho and para positions of the pyridyl moiety, the effect of the quaterniza-
tion at the meta position was weak. The effects of quaternization at ortho, meta and para positions were
investigated through natural bond orbital and Mulliken charges, HOMO–LUMO interactions in the
starting materials and the radical stabilization energy of the released 1-puridylmethyl radicals using DFT
calculations with the B3LYP/6-31G(d) and UBMK/6-311+G(3df,2p)//R(O)B3LYP/6-31G(d) methods,
respectively.
Received 29th July 2013,
Accepted 11th September 2013
DOI: 10.1039/c3ob41549j
www.rsc.org/obc
development of alkoxyamines (initiators) which are rather
stable and suitable for selective chemical activation is one way
Introduction
1
,2
The discovery 3 decades ago of nitroxide mediated polymer- to fulfil the aforementioned requirements. Recently, we
1
4
15,16
ization (NMP) has generated a large number of investigations showed that the C–ON bond homolysis in 1a and 3a
into the synthesis of initiators/mediators (alkoxyamines), (Fig. 1) was chemically activated by protonation, oxidation,
kinetic investigations, and preparation and application of new alkylation (for 3a), acylation (for 3a), benzylation (for 3a), and
3
–7
materials.
As discussed in several reviews and articles, the complexation with a Lewis acid (for 3a). The activation was an
initiation stage plays a significant role in the quality of the interplay of several effects: stabilization of the released alkyl
5
,8–11
polymer prepared by NMP.
Hence, it has been shown that radical, charge distribution (polar effect) in the starting
1
4,16
15–18
the faster the C–ON bond homolysis, the better the control of materials,
the NMP. Another interesting aspect of the preparation of tigations on 1a and 3a and their derivatives revealed very
such activated molecules is related to the safety of their use.
steric effect, and solvent
effect. These inves-
1
1
1
2
strong similarities (charge distribution) and clear differences
Indeed, fast decomposing compounds have drastic shipping (unsuccessful activation, and stabilization of the released
1
3
15,16
and storing rules. Then, two antagonist properties for the radical).
Thus, this prompted us to investigate the effect of
initiator must be combined: fast homolysis for NMP and the activation on the meta position in 2a (Fig. 1). Again simi-
high stability for shipping and storage. Consequently, the larities and differences of 2a with 1a and 3a were observed;
that is, as for 3a all positions of 2a were available for the acti-
vation, while charge distribution in starting materials and
stabilization of the released alkyl radical were different com-
Aix-Marseille Université, CNRS, ICR UMR 7273, case 551, Avenue Escadrille
pared with 1a, 3a and their derivatives. To get deeper insight
Normandie-Niemen, 13397 Marseille Cedex 20, France.
into the various effects involved in the activation of the C–ON
E-mail: sylvain.marque@univ-amu.fr, g.audran@univ-amu.fr,
bond homolysis in 1a–3a, DFT calculations were performed to
paul.brémond@univ-amu.fr; Fax: +33-(0)4-91-28-87-58
Part 8: G. Audran, L. Bosco, P. Brémond, S. R. A. Marque, V. Roubaud, and determine the natural bond orbital (NBO) and the Mulliken
†
D. Siri, J. Org. Chem., accepted.
Electronic supplementary information (ESI) available: Optimization of geome-
charge distribution, the HOMO–LUMO energy gap in the start-
ing materials and the NBO interactions, as well as the radical
stabilization energy (RSE) and the shape of the SOMO for
benzylic models of the released alkyl radicals.
‡
try structures of 1a, 2a, 3a and their derivatives as well as their NBO charges by
DFT, radical stabilization energy estimated using DFT, preparation of 2a–g, and
4. See DOI: 10.1039/c3ob41549j
7
738 | Org. Biomol. Chem., 2013, 11, 7738–7750
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 1 Alkoxyamines 1a–3a and their derivatives.
quantitative yields and purified by precipitation or solvent
removal in vacuo. Each alkoxyamine was obtained as a mixture
of two diastereoisomers.
Experimental section
Synthesis
All reactants and solvents were purchased and used as received.
Alkoxyamine 2a was prepared from 4-(1-bromoethyl)pyridine
and SG1 nitroxide (N-(2-methylpropyl)-N-(1-diethylphosphono-
Kinetic measurements
1
5–19
Rate constants were measured as previously described
2
,2-dimethylpropyl)-N-oxyl) as previously described (Scheme 1
using 2 equivalents of TEMPO as an alkyl radical scavenger in
and ESI‡). Alkoxyamines 2b–g were then prepared from 2a as
displayed in Scheme 1 (see ESI‡). Alkoxyamines 2b and 2d
were prepared in situ, and 2c, 2e, and 2g were prepared in
tert-butylbenzene (t-BuPh) as a solvent (Scheme 2) and using
P NMR on Bruker 300 and 400 Advance spectrometers. A few
examples of the plots for the first-order decay (eqn (1)) of the
3
1
major and minor diastereoisomers of alkoxyamines 2a–g are
displayed in Fig. 2. The activation energy E of each reaction
a
(
Table 1) was estimated using the averaged frequency factor
1
4
−1
20
A = 2.4 × 10
s
and eqn (2).
½
alkoxyamineꢀt
ln
¼ ꢁkdt
ð1Þ
ð2Þ
½
alkoxyamineꢀ0
kd ¼ Aeꢁ
Ea=RT
Scheme 1 Preparation of 2a–g.
d
Scheme 2 Scavenging experiment to determine k .
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 7738–7750 | 7739

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 2 Plot of ln([alkoxyamine]
t
/[alkoxyamine]t=0) vs. t for major (left) and minor (right) diastereoisomers of 2a (■, □) and 2g (★, ☆) at 85 °C and 2d (●, ○) and 2e
(
▲, △) at 61 °C.
d a
Table 1 Experimental temperature T (°C) and C–ON bond homolysis rate constant k , activation energies E and the relative rate constants krel,2 (krel = kd,2b–g/kd,2a)
at 120 °C for the minor and the major diastereoisomers of alkoxyamines 2a–g as well as krel,1 and krel,3 for the diastereoisomers of 1a–c and 3a–g in tert-butylben-
zene as a solvent unless otherwise mentioned
a,b
b,c
a,d
c,d
a,e
c,e
a
c
a, f
c, f
2
T (°C)
k
d
k
d
E
a
E
a
k
rel,1
k
rel,1
k
rel,2
k
rel,2
k
rel,3
k
rel,3
a
b
c
d
e
f
85
80
80
61
61
61
81
2.1
5.3
2.8
0.5
0.4
0.5
4.3
2.6
6.9
2.6
1.2
0.9
2.7
5.0
123.8
119.4
121.2
119.5
120.1
119.5
120.3
123.2
118.6
121.5
117.1
117.9
114.8
119.9
1
1
1
1
4.1
1.7
6.4
5.0
13.0
2.7
1
9.3
17.1
10.2
20.0g
65.5
10.2
1
9.5
16.1
12.2
31.5
209.6
9.5
10.2
19.2
11.8
11.8
3.9
2.2
3.7
3.1
3.7
2.9
g
h
h
g
g
g
—
—
h
h
—
—
g
h
h
g
g
—
—
h
h
g
—
—
a
Major diastereoisomer. Given in 10−4 s−1
b
c
Minor diastereoisomer. Given in kJ mol−1
d
e
f
g
.
.
Given in ref. 14. Given in ref. 16. In t-BuPh–CH
Cl
2 2
h
(v/v 1 : 1). Not determined.
pK determination
6-311+G(3df,2p) method. NBO and Mulliken charges as well
as NBO interactions were estimated using NBO 5.0 software.
For RSE calculations on benzyl radicals, the geometry optimi-
zations were performed using R(O)B3LYP/6-31G(d) and
vibrational frequencies were calculated at the R(O)B3LYP/
a
2
2
Experiments were performed in the D O–MeOH-d mixture as
2
4
1
2
a is not soluble in water, and H NMR signals of the aromatic
region were recorded to determine pK values. Each diastereo-
a
isomer was measured independently. pH was adjusted by
adding the corresponding acid and controlled using an
HI2211 pH/ORP meter from Hanna Instruments and a 4 mm
micro-titration electrode from Bioblock.
6
-31G(d) level using a scale factor of 0.9806 to provide zero
point vibrational energy. In order to obtain accurate energies,
single point calculations were performed at the UBMK/
6
-311+G(3df,2p) level of theory on the R(O)B3LYP/6-31G(d)
2
3
geometries.
DFT calculations
All calculations were performed using the Gaussian package
2
1
0
9, revision A02. The geometries of the alkoxyamines were
Results
Chemical modes of activation
optimized at the B3LYP/6-31G(d) level of theory. Vibrational
frequencies were calculated at the RB3LYP/6-31G(d) level of
theory to ensure that the obtained geometries are minima Alkoxyamine 2a was prepared from 4-(1-bromoethyl)pyridine
(
no imaginary frequency). The vibrational frequencies were and SG1 nitroxide (N-(2-methylpropyl)-N-(1-diethylphosphono-
1
5,16
scaled by a usual factor of 0.9608 to provide zero point 2,2-dimethylpropyl)-N-oxyl) with copper salts as a catalyst,
vibrational energy. The corresponding thermal corrections (Scheme 1 and ESI‡). Alkoxyamines 2b–g were then prepared
were included to obtain the enthalpy and Gibbs free energy from 2a using trifluoroacetic acid (2b), m-chloroperbenzoic
values under the standard conditions (p = 1 atm and acid (2c), methyl tosylate (2e), benzyl bromide (2f),
T = 298.15 K). The energies were calculated using the UBMK/ acetyl chloride (2d) and borane–THF complex (2g). The
7
740 | Org. Biomol. Chem., 2013, 11, 7738–7750
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
1
Fig. 3 (a) H NMR signal in the aromatic zone for the major (left, pH = 2.36, 4.14, and 6.00 from top to bottom) and the minor (right, pH = 2.20, 4.16, and 6.09 from
2 3
top to bottom) diastereoisomers of 2a/2b (0.02 M), at room temperature in D O–CD
OD (v/v 1 : 1). (b) Titration curves (■ and ▲ for major and minor diastereoi-
somers, respectively) obtained using the signal displayed in (a). pH was set with DCl and NaOD.¶
configurations were tentatively ascribed as RS/SR and SS/RR for
Table 2 Mulliken and NBO charges at the ipso carbon of the pyridyl moiety for
the major and minor diastereoisomers, respectively, using 1a–c,g, 2a–c,g and 3a–c,g, the HOMO and LUMO energies, and HOMO–LUMO
3
1
P NMR.§
energy gaps ΔEHL calculated using the B3LYP/6-31G(d,p) method for the RR/
SS diastereoisomer, and SOMO energies for a’ –b’ calculated at UBMK/6-311+G-
• •
(
3df,2p)//R(O)B3LYP/6-31G(d) level
Determination of pK
From the pH dependence of the H NMR signal recorded at
room temperature in D O–MeOH-d (1 : 1 v : v), a significant
shift was observed for the aromatic protons from pH 7 to 2.5
a
1
Charges
Energy
a
a
a
a,c
2
4
Mulliken
NBO
HOMO
LUMO
ΔEHL
SOMO
d
2
4
1a
1
1
0.26
0.35
0.31
0.33
0.11
0.09
0.07
0.08
0.14
0.16
0.14
0.15
0.20d
0.26
−577
−946
−581
−694
−590
−868
−63
−544
−107
−160
−69
−596
−105
−171
−66
−514
−402
−474
−534
−521
−272
−497
−445
−529
−246
−501
−458
−611
−1131
−63
(
(
Fig. 3a).¶ Using the Henderson–Hasselbach equation (eqn
b
c
2
5
3)), the titration curve for 2a (Fig. 3) afforded pK
a
values of
d
0.15
e
4
.27 and 4.13 for the major and minor diastereoisomers of 3a, 1g
0.26
−681
−603
−1079
−647
−672
−645
−1132
−605
−692
2
2
2
a
b
c
−0.10
−0.05
−0.07
−0.08
in sharp contrast with the value given for the meta-ethyl pyri-
2
6–31
dine (pK
a
= 5.68).k
These low pK
a
values for the diastereoi-
b
−602
somers of 2b can be ascribed to the electron withdrawing 2g
−616
−595
−846
−603
−622
f
3
3
3
3
a
b
c
−0.03
properties of the nitroxyl fragment. Such a lowering of pK
also been reported for 1b and 3b.
a
has
g,h
0.08f
b
−600
−102
−164
1
4
18,32
−0.08
f
g
0.01
δ2b ꢁ δ2a
δpH ¼ δ2a þ
ð3Þ
a
In kJ mol−1
b
c
pKaꢁpH
. HOMO−1, see text. SOMO are given for the benzylic
1
þ 10
• • • •
radical models a′ , b′ , c′ , and g′ at positions ortho (alkoxyamine
family 1), meta (alkoxyamine family 2), and para (alkoxyamine family 3),
d
e
see Fig. 8. Given in toluene in ref. 14. Not calculated in ref. 14.
Kinetic measurements
f
g
h
Given in ref. 16. In ref 32. It was given for 3e as 0.069 in ref. 16.
Rate constants were measured as previously described using
two equivalents of TEMPO as an alkyl radical scavenger in tert-
1
9
difference of ca. 2–3-fold in k between diastereoisomers has
d
butylbenzene (t-BuPh) as a solvent (Scheme 2). Except for 2b
and 2d generated in situ, all k values were measured on pure
33–36
already been reported
tion effect or to a remote steric effect.
and ascribed either to a hyperconjuga-
d
37
36,38
It does not deserve
compounds (Fig. 2 and Table 1). Interestingly, a 2–3-fold
weaker effect was observed for 2a–g than for 3a–c. Almost no
more comments. For 2f and 3f, a 5-fold difference between the
diastereoisomers might be ascribed to a remote strain.**
d
difference in krel or k was observed for the diastereoisomers of
2
a–c,g whereas a 2-fold difference was observed for 2d and 2e
and a 5-fold difference was observed for 2f. Such differences
were also observed for the diastereoisomers of 3a–g. Such a
DFT calculations
For the sake of simplicity, calculations were performed using the
B3LYP/6-31G(d) method only for the RR/SS diastereoisomer.
For comparison, Mulliken and NBO charges at the ipso
§
¶
0
The same approach was applied for 1a. See ref. 14.
All pH measured in D O–MeOH-d were re-estimated using pH = 0.929pH* +
.42. pH* is the pH measured in D O–MeOH-d solutions using a pH-meter cali-
2
4
2
4
brated with non-deuterated water solutions. See ref. 24.
kpK = 5.58 is also given in ref. 30 and 31. Non-corrected value.
**No calculations were performed because of a small energy gap and the pres-
ence of many conformers.
a
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 7738–7750 | 7741

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 4 HOMO for 1–3 (left to right) and a–c (top to bottom).
positions were calculated for 1a–c,g, 2a–c,g, and 3a–c,g display the expected shape which is observed for their
Table 2). To perform a frontier molecular orbital analysis, the HOMO−1†† – and all LUMOs (Fig. 5) exhibit the expected shape
(
HOMO (Fig. 4) and LUMO (Fig. 5) orbitals, as well as the whatever the type of alkoxyamine – ortho, meta, or para – and
HOMO–LUMO energy gap ΔEHL, were calculated for 1a–c,g, the activation mode – protonation or oxidation. This is expected
2
a–c,g, and 3a–c,g (Table 2). Taking into account that re- to hold for the other modes of activation, i.e., acylation, methyl-
formation and homolysis cross at the same TS, involving an ation, benzylation, and complexation with a Lewis acid.
interaction between the nitroxide π* SOMO and the alkyl Although HOMO and LUMO exhibit the right shape,
radical SOMO (Fig. 6b), which resembles the TS of the C–ON neither their own values nor the values of their energy gap
bond homolysis (Fig. 6a) leading to the formation of the σC–ON ΔEH–L (Table 2) describe qualitatively and quantitatively the
bond, one would expect, in the alkoxyamine, an interaction trends experimentally observed, except for family 2 for which
between the π-type LUMO and a σ-type HOMO (Fig. 6a) exhibit- the expected trend is observed.
ing a high electron density on both the nitrogen atom, due to
its lone pair, and the C–ON bond. As expected, all HOMOs
(Fig. 4) – except for 2c and 3c, for which the HOMOs do not ††For 2c and 3c, their HOMOs are mainly centered on the aromatic ring.
7
742 | Org. Biomol. Chem., 2013, 11, 7738–7750
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 5 LUMO for 1–3 (left to right) and a–c (top to bottom).
Scheme 3 Isodesmic reaction used to determine RSE.
benzylic radical models were also calculated using the UBMK/
-311+G(3df,2p)//R(O)B3LYP/6-31G(d) method (Fig. 8‡‡ and
6
Table 2). They agree with the RSE values and the mesomeric
forms (Fig. 7) as they confirm the delocalization of the odd
electron on the oxygen atom at the ortho and para positions in
•
c′ leading to the nitroxide mesomeric forms V and IV, respect-
ively (Fig. 7), whereas such a delocalization is not observed for
•
•
the meta position in c′ . However, the SOMO of a
o
′ is lower
•
than that of c ′ although less stabilized, while the SOMOs of
o
•
•
Fig. 6 Interaction expected at TS (a) for the C–ON homolysis and (b) for the
re-formation of the C–ON bond.
a
′ and c
′ exhibit close energies. Moreover, the least stabil-
′ radical exhibits the lowest SOMO. Furthermore, the
same mesomeric forms are expected for g ′ , g ′ , g ′ as for
p
p
•
ized c
m
•
•
•
o
m
p
Using the isodesmic reaction displayed in Scheme 3, RSE
for all positions – ortho (1), meta (2), and para (3) – were calcu-
lated for the benzylic radical models a′–c′,g′ (Fig. 7) of the
released alkyl radicals a–c,g (Fig. 1). The SOMOs of the gp
‡
‡As seen in Fig. 8, the absence of odd electron density on the nitrogen atom in
•
m
g ′ forbids any SOMO→σB–H interaction. Thus, the stabilizing two-electron–two
•
orbital interactions are the same in 2g and g
m
, in sharp contrast with 3g and
•
.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 7738–7750 | 7743

View Article Online
Paper
Organic & Biomolecular Chemistry
•
•
•
•
•
•
•
Fig. 7 Mesomeric forms (I–V) of the non-activated (a’ ) and activated (b’ , c’ and g’ ) benzylic models of the released alkyl radicals a –c , and g (Fig. 1) at the
ortho (o), meta (m), and para (p) positions and their corresponding RSE values.
•
•
•
•
a
o
′ , a
m p p
′ and a ′ , implying that the SOMO should vary by the slightly longer N→B dative bond (lN→B = 1.62 Å) than in g ′
•
•
same amount as observed when going from a ′ to g ′ and (lN→B = 1.60 Å).
from a ′ to g ′ (ΔE = −70 kJ mol ) whereas the SOMO of gp′
o
o
•
•
−1
•
m
m
−
1
Multiparameter correlations
is lowered by 47 kJ mol (Table 2). In fact, the stabilization of
•
•
c ′ and c ′ involves the nitroxide mesomeric forms V and IV, The rate constant k of the C–ON bond homolysis can be corre-
o
p
d
respectively (Fig. 7). These forms are due to a three-electron– lated to several parameters: σRS, for the radical stabilization
two-orbital interaction between the SOMO and the lone pair effect, σ , for the polar effect, and ν, for the steric effect. It can
I
on the oxygen atom of the N-oxide moiety. This SOMO→nO be assumed that the pyridinyl moiety activated at the para
interaction stabilizes the whole radical by stabilizing the lone position as well as at the meta position is not more sterically
pair and, in turn, raises the energy of the SOMO (Fig. 9). The demanding than a phenyl ring. It can be assumed that the
•
energy for the SOMO of gp being higher than expected is also pyridyl moiety activated by protonation or oxidation at the
due to the three-electron–two orbital interaction involving the ortho position is sterically not more demanding than a phenyl
SOMO and a B–H bond. This SOMO→σB–H interaction (NBO ring. A change in polarity due to the activation can be esti-
−
1
interaction, E(σB–H→SOMO) = 19.0 kJ mol ) stabilizes the radical mated for protonation and methylation although the original
and raises the SOMO whereas the two-electron–two orbital fragments are not available and need to be estimated using a
interaction (σB–H→π*, NBO interaction E = 21.5 kJ mol− ) does set of equations and several assumptions affording only rough
1
not stabilize the radical (Fig. 9). The SOMO→σ
interaction values of σ . In the seminal article of this series, it was
B–H
I
•
in g ′ is less efficient, which is likely due to the distribution of assumed that the stabilization for the pyridinyl-based released
o
the odd electron density on the two ortho positions and a radical was very close to that reported for the styryl radical.
7
744 | Org. Biomol. Chem., 2013, 11, 7738–7750
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
•
•
•
Fig. 8 SOMOs of radicals a’ –c’ , g’ (from left to right) for the ortho (o), meta (m), and para (p) positions, from top to bottom. Italicized values are for SOMOs ener-
−
1
gies (kJ mol ).
However, recent calculations on a model showed that the activated by the same modes as for 3a, in sharp contrast with
stabilization energy is different depending on the activation 1a for which only protonation and oxidation are effective due
and the position (Fig. 7). Unfortunately, the needed α and β to the steric strain at the ortho position. However, the acti-
hydrogen hyperfine coupling constants of the activated methyl vation observed for 2a does not depend too much on the mode
(
or alkyl methyl) pyridyl radical are not available in the litera- of activation, i.e., 2–4-fold and 3–6-fold for the major and the
2
0,39
ture to determine σRS values.
Due to drastic assumptions minor diastereoisomer,§§ respectively, in sharp contrast with
to estimate σ and σRS, the multiparameter correlation cannot the 10–30-fold observed for 3a (Table 1).¶¶ Indeed, the effects
I
be applied to discuss quantitatively the results reported in involved in the activation of 2a are not so clear-cut (vide infra)
Table 1.
as for 3a; that is, all released radicals are not stabilized (posi-
tive RSE values in Fig. 7) and the NBO chargeskk at the ipso
position do not change too much depending on the activation
Discussion
§
¶
§Solvent effects were included.
Recently, alkoxyamine activation was applied to families 1 and
¶Solvent effects were not included.
3
and proved to be dependent both on the alkoxyamine family
kkNBO charges are used to describe the effect of the activation on the electrone-
and on the activation modes. Interestingly, alkoxyamine 2a is gativity χ of the carbon atom of the C–ON bond. See ref. 16.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 7738–7750 | 7745

View Article Online
Paper
Organic & Biomolecular Chemistry
•
•
•
•
o p p m
Fig. 9 Orbital interactions involved in the stabilization of c ’ and c ’ (left side) and in the position (right side) of the SOMOs for g ’ (blue) and g ’ (red).
Fig. 10 Expected position of reactants, products, and TS states depending on (a) the polar effect, (b) no effect, and (c) the stabilization effect. Double head arrows
are for the activation energy E of the C–ON bond homolysis.
a
mode (Table 2), implying balanced effects affording only a low radical (stabilization of the product and hence of the TS, as
activation at the meta position. predicted by Hammond’s postulate, see Fig. 10c) and by the
Recently, we showed that the reactivity generated by the difference in electronegativity χ between the oxygen and
various modes of activation of the pyridinyl moiety of alkoxy- carbon atoms of the C–ON bond (eqn (4b)), as given by
amines belonging to families 1 and 3, as well as 2 (vide supra), eqn (4a) describing the strength of the bond as the sum of
was mainly controlled by the stabilization of the released alkyl enthalpic (difference in bond dissociation energy (BDE)) and
7
746 | Org. Biomol. Chem., 2013, 11, 7738–7750
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
•
•
•
Fig. 11 (a) Radical stabilization energy (RSE) for alkyl radicals a’ –c’ and g’ (from left to right) activated at ortho (o), meta (m), and para (p) positions and (b) NBO
charges at the ipso position for alkoxyamines 1–3 (ortho, meta, and para, respectively) for various modes of activation (from left to right, non-activated, protonated
+
−
(
3
H ), oxidized (O ), and complexed (BH )).
polar terms (difference in electronegativity χ) as proposed by
Pauling:
The RSE calculated for all positions possible for the pyridyl-
methyl radical are all positive, meaning that these radicals are
all less stabilized than the corresponding benzylic radical
(Fig. 7 and Scheme 3). However, the meta position is a little bit
less destabilized than the ortho and para positions (Fig. 11)
because the odd electron is never localized on the nitrogen
atom, as highlighted by its SOMO (Fig. 8) avoiding any electro-
static repelling effect between the nitrogen lone pair and the
odd electron. On the other hand, the NBO charge is more
important in the ortho position as the nitrogen atom is
nearer to the C–ON bond than the meta and para positions
1
2
2
BDEðA ꢁ BÞ ¼ ðBDEðA ꢁ AÞ þ BDEðB ꢁ BÞÞ þ 96:23ðχA ꢁ χBÞ
ð4aÞ
1
2
À
Á
1
2
3
4
5
1
2
1
2
BDEðR R NO ꢁ CR R R Þ ¼ BDEðR R NO ꢁ ONR R Þ
1
À
Á
3
4
5
3
4
5
þ
BDEðR R R C ꢁ CR R R Þ
2
2
þ 96:23ðχO ꢁ χCÞ
(
Fig. 11b). Consequently, polar and stabilization effects
values for
ð4bÞ balanced each other, leading to very similar k
d
−
1
1
a–3a and the styryl homologue (E = 124.5 kJ mol in ref. 20
a
Consequently, any increase*** in χ
C
would lead to a
and 33).
decrease in the square of the difference between χC and χO
As for the non-activated pyridinylmethyl radical, all RSE for
the protonated homologues are positive, meaning that the rad-
icals are destabilized upon protonation. In contrast with the
affording
a decrease in BDE compared to a reference
(Fig. 10b). In general, this polar effect is ascribed to the desta-
bilization of the reactant state (Fig. 10a). Changes in χ are due
•
•
C
non-activated a′ radical, the meta position for radical b′ is
less stabilized (Fig. 11a). On the other hand, 1b and 3b exhibit
the highest positive NBO charges (Fig. 11b) whereas this value
is negative for 2b because the protonation at positions ortho
to the strength of the electron withdrawing group attached to
the carbon atom of the C–ON bond, and are taken into
account by the NBO charges, in our case, at the ipso position
of the pyridyl moiety (Table 2). The enthalpic term in eqn (4a)
and (4b) is more closely related to the stabilization of the
released radicals. As given by Hammond’s postulate, an
endothermic reaction, such as the C–ON bond homolysis in
alkoxyamines, exhibits a late TS, i.e., a product-like TS. Con-
sequently, any stabilization of the products stabilizes TS
(
1b) and para (3b) generates a positive charge at the ipso posi-
tion (mesomeric forms III and IV for 3b and 1b, respectively,
Fig. 12). Thus the increase in k for 1b and 3b is ascribed to
d
the polar effect as all released radicals are destabilized.
In sharp contrast with the non-activated and the protonated
•
•
released benzylic-type radicals, RSEs for c
o
′ and c
p
′ are largely
(Fig. 10c) implying a decrease in BDE. This enthalpic effect is
•
negative (Table 2 and Fig. 11b) although still positive for c ′ ,
m
nicely accounted for by the RSE values for the released alkyl
radical.
meaning that the activation at the ortho and meta positions
stabilizes the radical more than the activation at the meta posi-
tion. This stabilization is nicely accounted for by an extra
Thus, to get a deeper and clearer insight into the effect of
various modes of activation at the ortho, meta, and para posi-
tions, calculations were performed to determine NBO charges
in alkoxyamines 1a–c,g, 2a–c,g, and 3a–c,g (Table 2) as well as
the RSE of the benzylic models of the corresponding released
radicals (Fig. 7).
•
•
stabilization mesomeric form (4 for c
m
′ against 5 for c
o
′ and
•
c
p
′ , Fig. 7), i.e. the nitroxide mesomeric forms V and IV
•
•
•
(
Fig. 7) for c ′ and c ′ , respectively. The ortho position in c ′
o p o
•
p
is less stabilized than the para position in c ′ (Table 2)
because the odd electron density is distributed at two ortho
•
positions in c ′ whereas the maximum density is observed at
o
•
*
**Assuming χ
C
< χ
O
.
the para position in c
p
′ as displayed in their respective SOMOs
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 7738–7750 | 7747

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 12 Mesomeric forms (I–IV) of the pyridyl fragments for 1b,c, 2b,c, and 3b,c.
(
3
Fig. 8). The non-delocalization of the odd electron on the positive RSEs are observed when the BH group is carried by
•
•
N-oxide moiety in c
m m
′ is nicely confirmed by the shape of its the nitrogen atom at the meta and ortho positions for g ′ and
•
SOMO (Fig. 8) which shows the absence of spin density on g ′ , respectively, denoting non-stabilized radicals. The same
the nitrogen atom. In alkoxyamine 3c, the oxidation at the mesomeric forms are expected for g′ as for a′ (Fig. 7). As dis-
para positions for 3a generates a negative partial NBO charge cussed above, the stabilization of g
o
•
•
•
p
′ is due to the 3 electron–2
(
Fig. 11 and Table 2) at the ipso position, as highlighted by its orbital σB–H→SOMO interaction between one bonding B–H
•
mesomeric form III (Fig. 12). In sharp contrast, the oxidation orbital and the SOMO (Fig. 9), in sharp contrast with g
at the ortho position in 1a generates a positive partial NBO extra stabilization of g ′ combined with the positive NBO
m
′ . This
•
p
charge (Table 2 and Fig. 10b) whereas a negative partial NBO charge at the ipso position in 3g affords an activation as large
charge is expected from its mesomeric form IV (Fig. 11). This as that for 3b.
implies that the electron withdrawing effect of the N-oxide
Only a few general trends can emerge from the above
moiety overmatches its resonance effect affording an increase thorough analysis: (i) the polar effect is always weak when the
in the electronegativity χ at the carbon atom of the C–ON activation is performed at the meta position (negative NBO
bond.††† The oxidation at the meta position of 2a affords nega- charges at the ipso position, Fig. 11b); (ii) the polar effect is
tive NBO partial charges at the ortho and para positions as the strongest when the activation is performed at the ortho
highlighted by the mesomeric forms of 2c (Fig. 12). Conse- position (positive NBO charges at the ipso position, Fig. 11b);
quently, the ipso position in 2c is not deactivated by the pres- (iii) the released alkyl radical is always destabilized (positive
ence of a negative charge, except that the remote meta position RSE values, Fig. 11a) when the activation is performed at the
for the N-oxide moiety decreases its electron withdrawing meta position.
effect which cannot balance the effect of the negative partial
charges at the two ortho positions (mesomeric forms III and IV
in Fig. 12).
Conclusion
•
The benzylic-type radical g
p
′ carrying the BH
3
group on the
nitrogen atom exhibits a negative value of RSE (Fig. 11a and DFT calculations showed that the effects triggered by the mode
Table 2), meaning that this radical is stabilized whereas of activation depend both on the mode of activation and on
the position targeted. This leads to a subtle interplay of polar
and stabilization effects which can be either additive/syner-
getic or antagonistic. The main consequence is the occurrence
†
††It is known that the electron withdrawing effect of the N-oxide moiety is as
large as the one of the NO group. See ref. 40.
2
of similar reactivities (the same kd or kd,rel) stemming from
7
748 | Org. Biomol. Chem., 2013, 11, 7738–7750
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Paper
different effects. Furthermore, cybotactic and salt effects
9 C. H. J. Johnson, G. Moad, D. H. Solomon, T. H. Spurling
also play significant roles depending on the mode and likely
and D. J. Vearing, Aust. J. Chem., 1990, 43, 1215–1230.
the site of activation. Consequently, this entanglement of 10 D. Gigmes and S. R. A. Marque, Encyclopedia of Radicals in
effects forbids the use of DFT calculations and of multi-
parameter correlations to discuss quantitatively the effects
investigated.
This paper highlights very well the versatility of the mode of
activation, that is, the occurrence of the activation depends on
Chemistry, Biology, and Materials, ed. C. Chatgilialoglu and
A. Studer, Wiley, Chichester, UK, 2012, pp. 1813–1850.
11 F. Chauvin, P.-E. Dufils, D. Gigmes, Y. Guillaneuf,
S. R. A. Marque, P. Tordo and D. Bertin, Macromolecules,
2006, 39, 5238–5250.
the regioisomer (6 modes of activation for the meta and para 12 Registration, Evaluation, Authorisation, and Restriction of
regioisomers against two for the ortho regioisomer) as does the
efficiency (strong activation for the ortho and para regio-
isomers against weak activation for the meta regioisomer).
Chemical Substances (REACH), EC 1907/2006 (available
at http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=
CELEX:32006R1907:en:NOT).
This clear difference in activation depending on the position 13 MSDS and FDS sheets from Arkema.Quick-FDS [15678-
opens up interesting opportunities with heteroaromatic rings 21954-01387-015407].
such as 1,2-, 1,3-, and 1,4-pyrazine, triazine, pyrimidine, etc. 14 Part 8: G. Audran, L. Bosco, P. Brémond, S. R. A. Marque,
For example, with 1,4-pyrazine-based alkoxyamines, selective V. Roubaud and D. Siri, J. Org. Chem., DOI: 10.1021/jo401674v.
alkylation of the nitrogen at the meta position by an activated 15 Part 1: P. Brémond and S. R. A. Marque, Chem. Commun.,
polymer chain might be envisioned, affording only a weak acti- 2011, 47, 4291–4293.
vation of the alkoxyamine C–ON bond, and then, using the 16 Part 3: P. Brémond, A. Koïta, S. R. A. Marque, V. Pesce,
nitrogen atom at the ortho position for late activation by either V. Roubaud and D. Siri, Org. Lett., 2012, 14, 358–361.
protonation or oxidation. In such a case, it would be possible 17 Part 5: G. Audran, P. Brémond, S. R. A. Marque and
to enhance the versatility of the NMP method and its combi-
nation with other methods.
G. Obame, J. Org. Chem., 2012, 77(21), 9634–9640.
18 Part 6: G. Audran, P. Brémond, S. R. A. Marque and
G. Obame, J. Org. Chem., 2013, 78(15), 7754–7757.
1
2
2
9 D. Bertin, D. Gigmes, S. Marque and P. Tordo, e-Polymers,
003, 2, 1–9.
0 D. Bertin, D. Gigmes, S. Marque and P. Tordo, Macromol-
ecules, 2005, 38, 2638–2650.
2
Acknowledgements
The authors are grateful to the ANR (grant ANR-11-JS07-002-
1 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,
B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng,
J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao,
H. Nakai, T. Vreven, J. A. Montgomery Jr., J. E. Peralta,
F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin,
V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari,
A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi,
N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross,
V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts,
R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi,
C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma,
V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg,
S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman,
J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09 (Revision
A.02), Gaussian, Inc., Wallingford, CT, 2009.
0
1), to Aix-Marseille University, and to CNRS for financial
support. MBBI thanks the government of Gabon for his grant
and L. Bosco for the pH measurements. This work was sup-
ported by the computing facilities of the CRCMM, ‘Centre
Régional de Compétences en Modélisation Moléculaire de
Marseille’.
Notes and references
1
D. H. Solomon, E. Rizzardo and P. Cacioli, Eur. Pat. Appl.,
35280, 1985; D. H. Solomon, E. Rizzardo and P. Cacioli,
1
US Patent, 4,581,429, 1986; D. H. Solomon, E. Rizzardo and
P. Cacioli, Chem. Abstr., 1985, 102, 221335q.
2
3
4
5
6
M. K. Georges, R. P. N. Veregin, P. M. Kazmaier and
G. K. Hamer, Macromolecules, 1993, 26, 2987–2988.
W. A. Braunecker and K. Matyjaszewski, Prog. Polym. Sci.,
2
007, 32, 93–146.
22 J. P. Foster and F. Weinhold, J. Am. Chem. Soc., 1980, 102,
7211–7218.
23 A. S. Menon, G. P. F. Wood, D. Moran and L. Radom,
J. Phys. Chem., 2007, 111, 13638–13644.
A. Goto and T. Fukuda, Prog. Polym. Sci., 2004, 29, 329–385
and references cited therein.
J. Nicolas, Y. Guillaneuf, C. Lefay, D. Bertin, D. Gigmes and
B. Charleux, Prog. Polym. Sci., 2013, 38(1), 63–235.
24 A. Krężel and W. Bal, J. Inorg. Biochem., 2004, 98, 161–166.
P. Nesvadba, Chimia, 2006, 60, 832–840 and references 25 J. J. Henderson, Am. J. Physiol., 1908, 21, 173–179.
cited therein.
R. B. Grubbs, Polym. Rev., 2011, 51, 104–137.
26 M. I. Zaretskii, Chem. Heterocycl. Compd., 1997, 33, 1–16;
M. I. Zaretskii, Khim. Getero. Soed., 1997, 3–22.
7
8
D. Bertin, D. Gigmes, S. R. A. Marque and P. Tordo, Chem. 27 A. Albert and E. Serjeant, Ionization Constants of Acids and
Soc. Rev., 2011, 40, 2189–2198 and references cited therein. Bases, Wiley, 1962.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 7738–7750 | 7749

View Article Online
Paper
Organic & Biomolecular Chemistry
2
2
3
3
3
8 J. W. Wilson and I. J. Worrall, J. Chem. Eng. Data, 1968, 13, 35 D. Bertin, D. Gigmes, S. R. A. Marque, S. Milardo, J. Peri
5
37–538.
9 H. C. Brown and X. R. Mihm, J. Am. Chem. Soc., 1955, 77,
725.
0 S. Espinosa, E. Bosch and M. Rosés, J. Chromatogr., A,
002, 964, 55–66.
and P. Tordo, Collect. Czech. Chem. Commun., 2004, 69,
2223–2238.
36 L. C. Li, G. K. Hamer and M. K. Georges, Macromolecules,
2006, 39, 9201–9207.
37 E. Beaudoin, D. Bertin, D. Gigmes, S. R. A. Marque,
D. Siri and P. Tordo, Eur. J. Org. Chem., 2006, 1755–
1768.
1
2
1 D. Augustin-Nowacka and L. Chmurzynski, Anal. Chim.
Acta, 1999, 381, 215.
2 Part 2: E. Bagryanskaya, P. Brémond, M. Edeleva, 38 A. Blachon, S. R. A. Marque, V. Roubaud and D. Siri, Poly-
S. R. A. Marque, D. Parkhomenko, V. Roubaud and D. Siri,
Macromol. Rapid Commun., 2012, 33, 152–157.
3 S. Marque, C. Le Mercier, P. Tordo and H. Fischer, Macro-
molecules, 2000, 33, 4403–4410.
4 D. Bertin, D. Gigmes, S. Marque, R. Maurin and P. Tordo,
J. Polym. Sci., Part A: Polym. Chem., 2004, 42, 3504–3515.
mers, 2010, 2, 353–363.
39 J. J. Brocks, H.-D. Beckhaus, A. L. J. Beckwith and
C. Rüchardt, J. Org. Chem., 1998, 63, 1935.
40 I. A. Grigor’ev, G. I. Shchukin and L. B. Volodarsky, in
Imidazoline Nitroxides, ed. L. B. Volodarsky, CRC Press,
Boca Raton, 1988, p. 151.
3
3
7
750 | Org. Biomol. Chem., 2013, 11, 7738–7750
This journal is © The Royal Society of Chemistry 2013