Oxidation of 4-substituted TEMPO derivatives reveals modifications at the 1- and 4-positions

Article abstract of DOI:10.1039/c1ob05037k

Potenital pathways for the deactivation of hindered amine light stabilisers (HALS) have been investigated by observing reactions of model compounds - based on 4-substituted derivatives of 2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPO) - with hydroxyl radicals. In these reactions, dilute aqueous suspensions of photocatalytic nanoparticulate titanium dioxide were irradiated with UV light in the presence of water-soluble TEMPO derivatives. Electron spin resonance (ESR) and electrospray ionisation mass-spectrometry (ESI-MS) data were acquired to provide complementary structural elucidation of the odd- and even-electron products of these reactions and both techniques show evidence for the formation of 4-oxo-TEMPO (TEMPONE). TEMPONE formation from the 4-substituted TEMPO compounds is proposed to be initiated by hydrogen abstraction at the 4-position by hydroxyl radical. High-level ab initio calculations reveal a thermodynamic preference for abstraction of this hydrogen but computed activation barriers indicate that, although viable, it is less favoured than hydrogen abstraction from elsewhere on the TEMPO scaffold. If a radical is formed at the 4-position however, calculations elucidate two reaction pathways leading to TEMPONE following combination with either a second hydroxyl radical or dioxygen. An alternate mechanism for conversion of TEMPOL to TEMPONE via an alkoxyl radical intermediate is also considered and found to be competitive with the other pathways. ESI-MS analysis also shows an increased abundance of analogous 4-substituted piperidines during the course of irradiation, suggesting competitive modification at the 1-position to produce a secondary amine. This modification is confirmed by characteristic fragmentation patterns of the ionised piperidines obtained by tandem mass spectrometry. The conclusions describe how reaction at the 4-position could be responsible for the gradual depletion of HALS in pigmented surface coatings and secondly, that modification at nitrogen to form the corresponding secondary amine species may play a greater role in the stabilisation mechanisms of HALS than previously considered. The Royal Society of Chemistry 2011.

Full text of DOI:10.1039/c1ob05037k

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 4936
www.rsc.org/obc
PAPER
Oxidation of 4-substituted TEMPO derivatives reveals modifications at the
1
- and 4-positions†‡
a
b
c
c
a,b
David L. Marshall, Meganne L. Christian, Ganna Gryn’ova, Michelle L. Coote, Philip J. Barker and
a
Stephen J. Blanksby*
Received 7th January 2011, Accepted 7th April 2011
DOI: 10.1039/c1ob05037k
Potenital pathways for the deactivation of hindered amine light stabilisers (HALS) have been
investigated by observing reactions of model compounds—based on 4-substituted derivatives of
2
,2,6,6-tetramethylpiperidine-N-oxyl (TEMPO)—with hydroxyl radicals. In these reactions, dilute
aqueous suspensions of photocatalytic nanoparticulate titanium dioxide were irradiated with UV light
in the presence of water-soluble TEMPO derivatives. Electron spin resonance (ESR) and electrospray
ionisation mass-spectrometry (ESI-MS) data were acquired to provide complementary structural
elucidation of the odd- and even-electron products of these reactions and both techniques show
evidence for the formation of 4-oxo-TEMPO (TEMPONE). TEMPONE formation from the
4
-substituted TEMPO compounds is proposed to be initiated by hydrogen abstraction at the 4-position
by hydroxyl radical. High-level ab initio calculations reveal a thermodynamic preference for abstraction
of this hydrogen but computed activation barriers indicate that, although viable, it is less favoured than
hydrogen abstraction from elsewhere on the TEMPO scaffold. If a radical is formed at the 4-position
however, calculations elucidate two reaction pathways leading to TEMPONE following combination
with either a second hydroxyl radical or dioxygen. An alternate mechanism for conversion of TEMPOL
to TEMPONE via an alkoxyl radical intermediate is also considered and found to be competitive with
the other pathways. ESI-MS analysis also shows an increased abundance of analogous 4-substituted
piperidines during the course of irradiation, suggesting competitive modification at the 1-position to
produce a secondary amine. This modification is confirmed by characteristic fragmentation patterns of
the ionised piperidines obtained by tandem mass spectrometry. The conclusions describe how reaction
at the 4-position could be responsible for the gradual depletion of HALS in pigmented surface coatings
and secondly, that modification at nitrogen to form the corresponding secondary amine species may
play a greater role in the stabilisation mechanisms of HALS than previously considered.
Introduction
Nitroxyl radicals based on 2,2,6,6-tetramethylpiperidinyl-N-oxyl
a
ARC Centre of Excellence for Free Radical Chemistry and Biotechnology,
1
(
TEMPO) are highly stable, isolable species. The addition of
School of Chemistry, University of Wollongong, Northfields Ave, Wollon-
gong, NSW 2522, Australia. E-mail: blanksby@uow.edu.au; Fax: +61 2
TEMPO to alkyl and acyl radicals occurs below the diffusion-
controlled limit, but nonetheless these trapping reactions proceed
4
221 4287; Tel: +61 2 4221 5484
b
2
BlueScope Steel Research, PO Box 202, Port Kembla, NSW 2505, Australia
more readily than dimerisation or other self-reactions of TEMPO.
c
ARC Centre of Excellence for Free Radical Chemistry and Biotechnology,
Research School of Chemistry, Australian National University, Canberra,
ACT 0200, Australia
A variety of TEMPO derivatives are therefore used as spin labels
3
for supramolecular complexes; reversible inhibitors for nitroxide
4
mediated polymerisation; and as superoxide dismutase mimetics
†
This work is dedicated to the memory of Athel Beckwith. Two of us
P.J. Barker and M.L. Coote) have first hand experience of his mentorship
5
(
for the protection of biomolecules against oxidative stress.
as co-workers, while all of us have benefited from his many and varied
contributions to free-radical chemistry.
Similarly, derivatives of TEMPO can act as radical scavenging,
anti-oxidant stabilisers for polymers, improving durability and
aesthetic properties throughout their service lifetime. By using
a colourless precursor to generate the TEMPO-based nitroxyl in
situ, hindered amine light stabilisers (HALS) are widely used to
prolong the lifetime of automotive coatings, bulk polymers, and
‡
Electronic supplementary information (ESI) available: ESI-MS spectra
for the irradiation of 3 and 4, analogous to Fig. 5. Normalised energy
profiles for TEMPONE formation from each of 2–4 (see Schemes 4 and
5
). Optimised geometries in the form of GAUSSIAN archive entries,
corresponding total energies, thermal corrections, zero-point vibrational
energies, entropies, free energies and free energies of solvation. See DOI:
1
6
0.1039/c1ob05037k
thin films.
4
936 | Org. Biomol. Chem., 2011, 9, 4936–4947
This journal is © The Royal Society of Chemistry 2011

View Article Online
The predominant mode of HALS action is believed to involve
activation of the 2,2,6,6-tetramethylpiperidine moiety (TEMP-X,
X = H, alkyl, or ether, Scheme 1) to the equivalent TEMPO
free radical, which subsequently scavenges otherwise reactive
polymeric alkyl or acyl radicals to form an amino-ether (TEMPO-
R). Upon further reaction with a peroxyl radical, the nitroxyl
radical is regenerated, converting harmful propagating radical
species to less reactive even-electron by-products, as in Scheme 1.
product of HALS possessing a secondary amine linkage at the
4-position (R ).
4
15
In order to complement such studies, simplified chemical
systems are employed in mechanistic studies of antioxidant action
to elucidate plausible reaction pathways. For example, simple
derivatives of TEMPO have been employed as putative models for
7
7c
studying chemical reactions at the active sites of HALS. Further
studies, undertaken from a medicinal perspective, reported that
4
-hydroxy-TEMPO (TEMPOL) is susceptible to modification
at the 4-position, leading to the formation of 4-oxo-TEMPO
TEMPONE), proposed to occur via the mechanism shown in
(
Scheme 2. In many currently utilised HALS, the active TEMPO
4
moiety is anchored to a larger molecular weight scaffold at R ,
and thus such an elimination could lead to deactivation of HALS
via volatilisation of the low mass TEMPONE fragment.
Scheme 1 A generalised stabilisation mechanism for HALS.
Scheme 2 Proposed mechanism for the hydroxyl radical induced
formation of 4-oxo-TEMPO (TEMPONE) from 4-hydroxy-TEMPO
16
The exact nature of the oxidation and propagation processes
described by Scheme 1 remains an area of active discussion,
(TEMPOL).
8
In this study we have undertaken a detailed experimental and
theoretical investigation of the reactions of TEMPO and selected
however, little is known about the fate of active HALS molecules
in-service. Testing HALS molecules for efficacy in surface coatings
involves empirically monitoring aesthetic properties in HALS-
doped coatings to demonstrate an extension of product lifetime
4
-substituted TEMPO derivatives (1–5, Fig. 1) with hydroxyl
∑
radicals (HO ), with an aim to identify pathways that might
lead to deactivation of these compounds, and thus by extension
deactivation of HALS in polymer coatings. Hydroxyl radicals
were generated by irradiating dilute aqueous suspensions of
9
in a specific application, compared to a control coating. Such
investigations demonstrate that the benefits afforded by HALS
are finite, as shown by the eventual discolouring, fading, or
degradation of the stabilised polymer, with the gradual decline
in physical properties mirrored by a concomitant decrease in
the concentration of available active HALS within the polymer
ꢀ
R
AEROXIDE P-25 nanoparticulate titanium dioxide, providing a
milder approach than direct photolysis or radiolysis of hydrogen
peroxide solutions. The conditions also allow a greater oppor-
tunity to observe the evolution of the reaction (and potential
reaction intermediates) by both electron spin resonance (ESR)
spectroscopy and electrospray ionisation mass spectrometry
10
over time. Current understanding is limited as to why HALS
do not stabilise polymers catalytically and indefinitely, as might
be expected from Scheme 1. Characterisation of deactivation
mechanisms could be crucial in selecting appropriate HALS for
specific applications.
(
ESI-MS), in contrast to traditional end-product analysis.
Moreover, this method is of direct relevance to the conditions
In principle, there are several general mechanisms by which
HALS molecules could be deactivated or decomposed. HALS
have been shown to migrate into lower coating layers where they
11
can no longer act as efficient stabilisers. Deactivation driven by
the inherent basicity of HALS could occur by chemisorption
onto activated particle surfaces within the polymeric matrix,
12
particularly in highly pigmented coatings, or by acid/base
interactions with the catalysts of acid-curing surface coatings.
Chemical or photochemical decomposition of HALS can lead
to the formation of non-oxidisable adducts whereby regeneration
13
of the active nitroxyl is inhibited. Moreover, the deactivation
of HALS need not depend on the modification of the nitroxyl
radical. Oxidative decomposition can cause HALS fragments to
be volatilised from the coating during high temperature processing
or in-service. Low molecular weight methoxy- and hydroxy-
piperidines were detected when ester-linked oligomeric HALS in
Fig. 1 TEMPO derivatives used as HALS model compounds:
(
(
(
1) 2,2,6,6-tetramethylpiperidine-N-oxyl ‘TEMPO’;
2) 4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxyl ‘TEMPOL’;
3) 4-methoxy-2,2,6,6-tetramethylpiperidine-N-oxyl ‘4-methoxy-
14
polypropylene were subjected to ultraviolet (UV) light. Similarly,
after prolonged ambient exposure of polyethylene films, 4-amino-
TEMPO’; (4) 4-carboxy-2,2,6,6-tetramethylpiperidine-N-oxyl ‘4-carboxy-
TEMPO’; (5) 4-oxo-2,2,6,6-tetramethylpiperidine-N-oxyl ‘TEMPONE’.
2
,2,6,6-tetramethylpiperidine was detected as a decomposition
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4936–4947 | 4937

View Article Online
encountered by HALS in situ, as titanium dioxide is a common in-
organicpigmentpresentinpolymer coatings. Bothpigmentaryand
nanoparticulate grades of titanium dioxide accelerate the degra-
dation of polymer films, with little benefit afforded by HALS due
and frequency counter. Instrument settings were as follows:
microwave power, 2.0 mW; modulation frequency, 100 kHz;
modulation amplitude, 0.2 mT. Measurements were conducted in a
◦
◦
climate-controlled laboratory at 20.0 C ± 0.5 C. For the solution
17
3
to their photocatalytic destruction. Perhaps the most astonishing
demonstration of the photocatalytic destruction of durable surface
coatings containing HALS is that caused by certain commercial
sunscreen formulations, in which the presence of photocatalytic
mixing experiments, 5 cm of a 100 mM TEMPOL solution was
placed in a small vial. A similar solution of TEMPONE was added
dropwise, with ESR spectra recorded after the addition of every
drop until the relative concentrations were in the range similar
to those observed in the photolysis experiments. The composite
spectra were scaled to a similar intensity to the experimental
spectra.
18
grades of titanium dioxide were confirmed spectroscopically.
While not quantitative in nature, this combined experimental
and theoretical study identifies potential mechanisms of TEMPO
degradation and deactivation that are not accounted for by current
descriptions of their major reactivity pathways (cf . Scheme 1).
The identification of thoroughly characterised mechanisms for
irreversible chemical deactivation of TEMPO derivatives, even if
minor compared to well-established nitroxyl radical chemistry,
provides important new understanding of long term HALS
deactivation over the service lifetime of surface coatings.
Mass spectrometry and chromatography
Positive ion mass spectra were recorded with a Waters QuattroMi-
cro (Waters, Manchester, U.K.) triple quadrupole mass spectrom-
eter equipped with an ESI source and controlled by Micromass
MassLynx software (version 4.1). The aqueous solutions were
diluted to ca. 15 mM in acetonitrile, and infused directly into the
-
1
ESI source at 10 mL min . The capillary voltage was set to 3.5 kV,
Experimental
◦
cone voltage 25 V, source temperature 65 C and desolvation
◦
temperature 100 C. Nitrogen was used as the drying gas, at a
Materials
-
1
flow rate of 320 L h . In all collision induced dissociation (CID)
scans, argon was used as the collision gas at a pressure of 3 mTorr,
and the collision energy was 15–25 eV, depending on the fragility
of the isolated ion.
TEMPO derivatives and the corresponding piperidine analogues
were purchased from Sigma Aldrich (Castle Hill, Australia), stored
◦
ꢀR
at 4 C, and used without further purification. AEROXIDE P-
2
5 TiO was a gift from Evonik (Essen, Germany). Deionised
2
TLC separation of analytes in irradiated samples was performed
on 2.5 cm ¥ 7.5 cm glass backed normal phase TLC plates (silica gel
-
6
-1
water (reverse osmosis, median conductivity 3–4 ¥ 10 S m ) was
employed throughout. Solvents employed for mass spectrometry
6
0 F254, Merck, Darmstadt, Germany). The plate was developed
in a mobile phase of 1 : 1 ethyl acetate:hexane containing 1%
w/w) triethylamine. A basic permanganate staining solution was
prepared by addition of KMnO (1 g), K CO (7 g), and NaOH
0.1 g) to water (100 cm ). The developed plates were dipped into
(
acetonitrile and methanol) were HPLC grade (Ajax Finechem,
Taren Point, Australia) and used as received. Solvents used for
thin layer chromatographic (TLC) analysis (ethyl acetate, hexane
and triethylamine) were reagent grade (Ajax Finechem, Australia)
and also used as received.
(
4
2
3
3
(
the permanganate solution and then completely dried using a heat
gun.
Photochemistry
Photolysis experiments were carried out in a custom-built reactor,
with a 1 kW Xe arc lamp (Oriel Corporation, Stratford, CT,
USA) producing a collimated parallel beam of 4.8 cm diameter.
The system was equipped with a 10 cm pathlength infrared
filter (water), a green glass filter passing only wavelengths above
Computational procedures
Standard ab initio molecular orbital theory and density functional
19
theory calculations were carried out using Gaussian 09 and
20
MOLPRO 2009.1. Calculations were performed at a high level
of theory, recently demonstrated to predict solution-phase bond
dissociation energies and associated equilibrium constants to
3
75 nm, and a neutral density attenuation filter passing 30% of
the incident light. The filtered beam entered a quartz reactor,
containing a stirring suspension of P-25 TiO (125 mg) and the
2
1
within chemical accuracy. Geometries of all species were fully
optimised at the B3-LYP/6-31G(d) level of theory. For all species,
full systematic conformational searches (at a resolution of 120 )
were carried out to ensure global, and not merely local, minima
were located. Frequencies were also calculated at this level and
scaled by recommended scale factors. Improved energies for
all species were calculated using the G3(MP2)-RAD level of
theory, a high level composite ab initio method that approximates
URCCSD(T) calculations with a large triple zeta basis set using
2
3
TEMPO derivative 1–5 (100 mM) in 200 cm water. The reactor
was enclosed within a dark box, fitted with an extractor fan
for temperature control, and openings for sampling and stirrer
◦
3
control. Samples (ca. 1.5 cm ) were taken at regular time intervals,
22
centrifuged, and the supernatant transferred to a quartz ESR flat
3
cell (active volume 0.15 cm ). After the ESR measurement, the
solutions were reserved for mass spectrometric analysis. Control
experiments were undertaken whereby TEMPO solutions were
2
3
additivity approximations. This method was recently shown to
reproduce a large test set of experimental gas-phase BDEs to
mixed with TiO maintained in darkness, and irradiated under
2
identical experimental conditions for 3 h in the absence of TiO .
2
2
4
within chemical accuracy.
Entropies and thermal corrections were calculated using stan-
ESR spectroscopy
25
dard formulae for the statistical thermodynamics of an ideal
gas under the harmonic oscillator approximation in conjunction
with the optimised geometries and scaled frequencies, and reaction
ESR data was obtained on a Bruker ESP300E spectrometer
Bruker GmbH, Rheinstetten, Germany) operating in the X-band
of the microwave spectrum, equipped with an NMR gaussmeter
(
4
938 | Org. Biomol. Chem., 2011, 9, 4936–4947
This journal is © The Royal Society of Chemistry 2011

View Article Online
Gibbs free energies were computed using the Gibbs fundamental
equation. Free energies of solvation were computed using the
ESR spectra are presented as a function of irradiation time for
TEMPOL in Fig. 2. The initial spectrum is shown in Fig. 2(a).
As the irradiation proceeds, asymmetry is observed after 30 min,
indicative of a mixture of nitroxyl radicals. Distinctive features of
a second radical begin to emerge after 70 min, particularly on the
26
polarised continuum model, PCM-UAKS, also at the B3-LYP/6-
1G(d) level. Geometries of all species were re-optimised in
solution. Free energies of each species in water solution at
98.15 K were calculated as the sum of the corresponding gas-
3
2
I
M = -1 line, as in Fig. 2(b). After 150 min, these features of the
phase free energy and the obtained free energy of solvation.
The phase change correction term RT(lnV) was included in the
energy of all species. Gibbs free energies of activation for the
emerging radical become clear (Fig. 2c). The experiment continues
until after 360 min the new radical is the only species present in
the spectrum (Fig. 2d). Alongside Fig. 2(b) and 2(c), Fig. 2(e) and
2(f) are the spectra obtained by gradual addition of TEMPONE
5 to a solution of TEMPOL without irradiation. Fig. 2(g) is a
spectrum of pure TEMPONE. These pairs of spectra confirm the
nature of the radical mixture observed during the experiment to
be TEMPOL and TEMPONE.
reactions involving a hydrogen transfer have been corrected for
‡
the effects of tunnelling using the standard formulae: DG corr
=
‡
DG - RTln(k(T)), where k(T) is the tunnelling correction factor,
T is the absolute temperature, R is the universal gas constant and
‡
27
DG is the Gibbs free energy of activation.
Results and discussion
ESR characterisation of TEMPO derivatives
Initially, the ESR spectra of TEMPO 1 and derivatives containing
at least one hydrogen at the 4-position (2–4) are characterised by
fairly broad ESR lines as a result of the unresolved couplings to
28
the methyl hydrogens. In contrast, the spectrum of TEMPONE
1
3
5
shows much sharper lines with well resolved C-coupling, due
to the restrictions placed on the piperidine ring conformation by
29
the ketone double bond. Typical g-values and coupling constants
are given in Table 1, and show good agreement with the available
30
literature data.
Reaction monitoring by ESR
∑
While there is some debate about the exact role of free HO and
surface charge carriers (photo-generated electrons and holes) in
31
TiO photocatalysis, this approach provides mild conditions in
2
which to monitor TEMPO oxidation. The conditions employed
∑
here were modelled on previous ESR experiments in which HO
32
production was monitored by either spin-trapping, or nitroxyl
radical decay. By using low concentrations of TiO and the
33
2
TEMPO derivatives (1–5), combined with a controlled irradiation
regime to investigate purely photocatalytic reactions, extended
periods of irradiation (up to 6 h) were possible without unwanted
photochemical side reactions. Direct absorbance and subsequent
destruction of the nitroxyl moiety by photons of wavelength
below 320 nm was excluded by the use of the green glass filter,
Fig. 2 ESR spectra of TEMPOL 2 as a function of irradiation time in
the presence of P-25 TiO
2
. (a) No Irradiation; (b) 70 min; (c) 150 min;
(
d) 360 min; (e), (f) dropwise addition of TEMPONE to TEMPOL to
34
transmitting only photons of wavelength greater than 375 nm.
Under our conditions, no changes were observed in the spectrum
when a standard solution of TEMPOL was maintained with TiO
in darkness, nor when irradiated in the absence of TiO
reproduce (b) and (c); (g) standard TEMPONE 5.
2
This general pattern of results was repeated for the irradiation
of both 4-methoxy-TEMPO 3 and 4-carboxy-TEMPO 4, with the
observed formation of TEMPONE concurrent with decreasing
intensity of the parent nitroxyl radical. No other paramagnetic
product species were identified. When either TEMPO 1 or
TEMPONE 5 are exposed to the same conditions and subjected
to the same analysis, no changes to the coupling constants are
observed, the signal merely decays over time until no nitroxyl
radicals are detectable after 3 h.
2
.
Table 1 ESR properties of nitroxyl radicals 1–5 in water at 293 K.
Microwave power 2.0 mW; modulation frequency 100 kHz; modulation
amplitude 0.2 mT
4
14
R
g
N
a /mT
1
2
3
4
5
H
OH
OCH
COOH
O
2.00563
2.00570
2.00568
2.00566
2.00562
1.73
1.71
1.70
1.71
1.61
3
The total peak to peak signal intensity (DM = -1 line)
I
attributable to each of the TEMPO derivatives 2–4 as a function of
time is plotted in Fig. 3, along with the formation of TEMPONE
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4936–4947 | 4939

View Article Online
Table 2 Mass-to-charge ratio (m/z) and relative intensities (shown in
parentheses and uncorrected for isotopic contributions) obtained from
ESI-MS analysis of 1–5 in 4 : 1 CH
3
2
CN : H O
4
+
+∑
+
+
R
M
[M+H]
[M+H
2
]
[M-14]
1
2
3
4
5
H
OH
OCH
156.3 (13.0) 157.3 (84.5) 158.3 (56.6)
172.3 (38.9) 173.3 (14.8) 174.3 (100.0) 158.2 (48.1)
186.3 (11.3) 187.3 (17.8) 188.4 (100.0) 172.3 (37.5)
142.2 (100.0)
3
COOH 200.3 (20.7) 201.3 (11.6) 202.3 (100.0) 186.3 (32.9)
170.3 (16.3) 171.3 (16.8) 172.3 (100.0) 156.3 (45.0)
O
Characterisation of TEMPO derivatives by ESI-MS
ESI-MS analyses were undertaken to provide complementary
detection of diamagnetic products arising from photocatalytic
degradation of TEMPO derivatives 2–4. It should be noted that
under these conditions ESI-MS is not quantitative and cannot
be used in this manner to provide a comprehensive mass-balance
for the reaction. This is due to (i) the differences in ionisation
efficiencies among the different species present (e.g., Table 2)
and (ii) the possibility that extensive oxidative fragmentation may
result in low molecular weight products that can be volatilised from
the reaction vessel. Despite these stated limitations, ESI-MS is a
well recognised method for reaction monitoring, including those
Fig. 3 Photocatalytic consumption of TEMPO derivatives 2–4 (open
shapes) with the concurrent formation of TEMPONE 5 (filled shapes) as
a function of irradiation time as detected by ESR.
5
. For clarity, only the first 3 h of irradiation are shown. In
all cases, the initial nitroxyl radical concentration is depleted
open shapes) and an increase in abundance of TEMPONE is
37
(
employing photocatalytic TiO
effective at identifying reaction intermediates.
2
,
and has proven particularly
38
observed (corresponding filled shapes). Data for the formation
of TEMPONE are not present prior to 100 min, as the narrow
peaks are obscured by the broader peaks of 2–4. The rate at
which the nitroxyl radical signal diminishes follows the trend of
Prior to irradiation, the initial positive ion ESI mass spectra
of the nitroxyl radicals display a distribution of ions around
the molecular mass M. As redox processes can play a role in
electrospray ionisation, the observation of multiple ions in the
molecular ion region can be attributed to oxidation and reduction
4
> 3 > 2 and is thus clearly a function of the substituent in
4
the 4-position of the piperidine ring (R = CO
2
3
H > OCH >
3
9
OH). This is broadly consistent with the trend in redox potentials
of the nitroxyl moiety under positive ion ESI conditions.
◦
+
+∑
of these nitroxyl radicals (e.g., corrected E (4) = 0.924 V versus
The M oxoammonium cation, [M+H] protonated nitroxyl
radical cation, and [M+H ] protonated hydroxylamine cation
◦
35
+
E (2) = 0.877 V) and, in-line with previous studies, suggests
2
that oxidation to the diamagnetic oxoammonium cation may be a
major reaction pathway.
are observed in varying abundances, with a strong dependence
on the analyte, as well as the solvent and source conditions.
36
40
The rate and abundance of TEMPONE 5 formation is similarly
dependant on R , and does not form at all in the absence of a
Typical abundances of these ions obtained from ESI-MS of 1–5 are
provided in Table 2. Collision induced dissociation (CID) of these
ions gave rise to fragmentation patterns similar to those obtained
4
substituent in the 4-position, as demonstrated by the irradiation
of TEMPO 1. In every case, the yield of 5 is less than would be
expected if all of the reactant had been transformed to this product,
indicating that this is a minor pathway in the overall reaction
scheme, while the bulk of the products are diamagnetic (see
above). From the relative signal intensities of standard solutions of
TEMPOL 2 and TEMPONE it can be estimated that TEMPONE
formation from TEMPOL accounts for ca. 12% of the theoretical
yield. Similarly, for 4-methoxy-TEMPO 3, the formation of 5 is
only ca. 6% of the theoretical yield, and for 4-carboxy-TEMPO
41
for larger N-ether HALS analogues previously reported. For
example, one prominent product ion observed in the CID of the
+
+
[M+H
2
] ions is protonated acetone oxime [(CH
3
)
2
C
NHOH] ,
comprised of two hindering methyl groups and the piperidinyl
nitrogen at m/z 74. This is illustrated for TEMPOL 2 in Fig. 4(a).
From Table 2 it is seen that an abundant ion, 14 Da below
the molecular mass, is also observed in positive ion ESI-MS
+
spectra of 1–5. CID of the [M-14] peak of TEMPOL 2 at
m/z 158 is shown in Fig. 4(b), with the product ion at m/z 74
4
the yield is less than 1%. While TEMPONE formation in these
corresponding to protonated acetone oxime noticeably absent.
instances is clearly a minor product, its observation is significant
because only the hydroxyl radical driven conversion of TEMPOL
to TEMPONE has previously been reported. Indeed, previous
Rather, fragmentation to an equivalent propan-2-iminium ion
+
[(CH
3
)
2
C
NH
2
] is observed 16 Da lower at m/z 58 suggesting
that the loss of 14 Da from the precursor nitroxyl radical results
∑
+
studies of HO reactivity with 4 did not report TEMPONE
from reductive formation of [M+H
2
-O] , rather than the isobaric
36
+∑
formation. Overall the differing yields of TEMPONE from
[M+H-CH
3
]
radical ion. Indeed, the CID spectrum shown in
4
+
precursors 2–4 indicate that R has a pronounced effect on the
Fig. 4(b) is identical to that obtained from the [M+H] ion of
an authentic sample of 4-hydroxy-2,2,6,6-tetramethylpiperidine
(denoted 2H, data not shown). The presence of these piperidine
propensity for formation of this product. This observation is
not correlated to the relative trend for the disappearance of
the reactant (see above), perhaps unsurprisingly given that the
latter is predominantly driven by modification of the nitroxyl
moiety.
+
ions at [M-14] could be due to their presence in the samples
themselves or they might arise during the ESI process. To
assess these possibilities, the carboxylate functionality enabled
4
940 | Org. Biomol. Chem., 2011, 9, 4936–4947
This journal is © The Royal Society of Chemistry 2011

View Article Online
Fig. 4 CID mass spectra (obtained at a collision energy of 15 eV)
of selected ions arising from ESI-MS analysis of TEMPOL 2 in 4 : 1
+
+
CH
3
CN : H
2
O. (a) [M+H
2
] at m/z 174 and (b) [M+H
2
-O] at m/z 158.
-
4
-carboxy-TEMPO 4 to also be detected as an [M-H] ion in
negative ion ESI. In this case, where the charge is maintained
-
remote from the nitroxyl moiety, [M-H] is the only peak observed
-
in the molecular ion region at m/z 199, i.e., no [M-H-14] was
+
observed. Furthermore, the [M-14] peak can be eliminated from
Fig. 5 ESI-MS spectra of 2 (in 4 : 1 CH
irradiation in the presence of TiO
and piperidines 2H and 5H are highlighted. (a) Pre-irradiation; (b) during
3
CN : H
2
O) at various stages of
the positive ion spectra entirely by changing the solvent system
from a mixture of acetonitrile and water to acidified methanol.
Moreover, when the analytes 1–5 are developed by thin-layer
chromatography, only one spot is observed, indicating that the
2
. Oxidation products TEMPONE 5,
irradiation; (c) post-irradiation.
+
additional [M-14] peak is most likely a consequence of the
modification of the nitroxyl moiety by positive ion ESI, rather
than solution processes. ESI-MS and CID analyses of derivatives
stage are peaks attributable to dimers observed at multiples of the
molecular mass.
The initial series of ions in Fig. 5(a) are consistent with those
13a,33
3–5 exhibited similar results as described above for TEMPOL 2.
However, neither m/z 58 nor 74 are observed in the CID spectrum
of 1 or 1H, suggesting that this fragmentation process is dependent
reported for TEMPOL 2 in Table 2. During (Fig. 5b) and following
+
+
(
Fig. 5c) irradiation, the M and [M+H
2
] ions at m/z 172 and m/z
4
on R .
1
74 (open shapes) respectively, are diminished and predominantly
supplanted by an ion at m/z 170, along with an ion at m/z 156
and an increase in the relative abundance of the ion at m/z 158
ESI-MS analysis of photo-oxidation products
(filled shapes). The appearance of the ion at m/z 170 and its
Multiple studies have reported on the non-radical products arising
from the reaction of TEMPO derivatives with HO under different
companion at m/z 156 would appear to be consistent with the
formation of TEMPONE 5, corroborating the ESR observations.
The CID spectrum of m/z 170 however, is not an exact match
with that obtained from the equivalent ion derived from ESI of
an authentic sample of 5 (data not shown). This likely indicates
the presence of a second, isobaric, product perhaps arising from
dehydrogenation of the piperidine ring. Consistent with the ESR
data, a putative TEMPONE ion at m/z 170 was also observed as a
reaction product from 3 and 4 (ESI Fig. S1 & S2, respectively‡). In
the absence of isobaric ions in both these cases the CID spectrum
matches that of authentic 5. Thus, on the basis of both ESR
and ESI-MS observations, the formation of TEMPONE upon
∑
conditions. Oxidation of the nitroxyl moiety to the oxoammonium
cation was reported as one product, while another study reported
reduction of the nitroxyl to the hydroxylamine, where the nitroxyl
16b,c
was recovered upon addition of potassium ferricyanide.
These
species are indistinguishable from the nitroxyl radical in positive
ion ESI-MS, as all three oxidation states of the nitroxyl are
+
+
+
intrinsically present in the spectrum at M , [M+H] , and [M+H
2
]
(
Table 2). With this in mind, ESI-MS spectra were concurrently
taken during the course of the irradiation to complement the
ESR data and provide additional insight into the nature of the
non-radical oxidation products of the TEMPO derivatives. Fig. 5
shows a series of ESI spectra of TEMPOL 2 when irradiated in the
∑
reaction with HO is not limited to 2. Rather, TEMPONE was
detected following irradiation of any 4-substituted analogue 2–4.
In contrast, neither of the product ions at m/z 170 and m/z 156 are
observed by ESI-MS during the irradiation of TEMPO 1 (data not
shown), just as no ESR signal attributable to 5 is observed in these
experiments. This supports the conclusion that where TEMPONE
presence of TiO
Fig. 2). Consistent with ESR findings, no changes are observed
in the spectrum when TEMPOL is irradiated with UV light in
the absence of TiO , or when mixed with TiO in darkness. At no
2
over the same time course as the ESR analysis
(
2
2
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4936–4947 | 4941

View Article Online
formation is observed, the 4-substituent must be implicated in the
reaction mechanism.
+
The [M+16] ion, denoted by a solid cross at m/z 188 in Fig.
5
(b), is observed during the oxidation of TEMPOL 2 and an
+
analogous [M+16] species is also observed for the reactions of
and 4 (ESI Fig. S1 & S2, respectively‡). These observations
are consistent with a net increase of one oxygen atom of mass
6 Da. Multiple additions of oxygen at 16 Da intervals were not
3
1
+
observed at any point. The CID spectra of the [M+16] peaks
exhibit characteristic losses of water (-18 Da) and R -H (i.e.,
4
H
2
O, CH
3
OH or HCO H for 2, 3 and 4, respectively), indicating
2
the presence of a hydroxyl moiety possibly the result of hydrogen
∑
16b
abstraction followed by HO addition as previously proposed.
While the position of the hydroxyl group is not certain, it is
consistent with the 4,4-dihydroxylated species outlined in Scheme
2
for 2, however analogous hydroxylation at other sites cannot be
Fig. 6 CID mass spectra (obtained at a collision energy of 15 eV) of the
ion at m/z 156 arising from ESI-MS analysis of: (a) a standard solution of
excluded. Indeed calculations of the barriers for hydrogen atom
abstraction indicate a preference for initial radical formation and
thus possible hydroxylation at the 3-position and perhaps even the
methyl moieties (see below).
4
2
-oxo-2,2,6,6-tetramethylpiperidine 5H, and (b) irradiation of TEMPOL
in the presence of TiO
2
.
+
Although the [M-14] ion at m/z 158 (Fig. 5, filled diamonds)
moreover, that 5H is observed at m/z 156 from each of 2–4. This
deoxygenation reaction is consistent with previous observations
of the reduction of cyclic 5-membered nitroxyl radicals to amines
was shown to form during positive ion ESI-MS of TEMPOL
2, it is unexpectedly observed to increase in abundance upon
3
3
irradiation in the presence of TiO despite a concomitant decrease
2
when irradiated in the presence of TiO
hindered piperidines from active HALS in weathered polymer
2
,
and the detection of
in other signature ions at m/z 172 and 174. Furthermore, when the
irradiated solutions are analysed by TLC, a second spot is observed
with a retention factor equal to the corresponding 4-hydroxy-
42
coatings.
The formation of piperidines from N-oxyl piperidines 1–5
during photooxidation in the presence of TiO could arise via
2
,2,6,6-tetramethylpiperidine 2H, whereas only the nitroxyl spot
2
+
is observed prior to irradiation. CID of the [M-14] ions yield
the same spectrum as pre-irradiation, indicating that 2H remains
present. This suggests that while the ion at m/z 158 in Fig. 5(a)
arises exclusively from ionisation of TEMPOL 2, two different
sources contribute to this ion abundance in Fig. 5(b) and (c)
and thus authentic 2H is present in the irradiated solutions as
an oxidation product. The contribution of the latter dominates
in Fig. 5(c) and is consistent with ESR data suggesting a
decreased concentration of TEMPOL 2. Fig. 5 also reveals an
increase in abundance of an ion at m/z 156 upon oxidation
of TEMPOL under these conditions with this ion dominating
the post-irradiation spectrum at the expense of TEMPONE at
m/z 170 (Fig. 5c). The CID spectrum of m/z 156 that arises
from the irradiation of TEMPOL 2 is shown in Fig. 6(a) and is
identical to standard 4-oxo-2,2,6,6-tetramethylpiperidine 5H (Fig.
one or all of the mechanisms outlined in Scheme 3. Combination
∑
of an alkyl radical (R ) with the nitroxyl oxygen results in an
intermediate amino ether (Scheme 3a). Although the formation
of aminyl radicals is often considered unlikely, for particular R
groups it has been shown that cleavage of the N–OR bond is
thermodynamically favoured over NO–R bond cleavage, releasing
21
an aminyl radical rather than a nitroxyl. The formed aminyl
radical may then abstract an available hydrogen to form the
secondary piperidine. Given the expected low concentration of
suitable alkyl radicals under our experimental conditions this
pathway is unlikely to dominate. Alternatively, addition of a
peroxyl radical to the nitroxyl results in the formation of aminyl
radical along with an alkoxyl radical and molecular oxygen
6
b). The presence of the characteristic iminium product ion at
m/z 58 in both spectra confirms the presence of the secondary
amine functionality in this ion. The formation of 5H in these
experiments appears to be a result of conversion of TEMPOL
2
to TEMPONE 5 with subsequent deoxygenation (as described
below).
+
Increased abundances of equivalent [M-14] ions are also
observed during the irradiation of derivatives 3–5, but the trend
is most obvious for TEMPO 1, because of the absence of the
competing pathway for formation of TEMPONE 5. Reduction
of TEMPO 1 to the corresponding secondary amine 1H is the
predominant feature of the ESI spectrum, resulting in an abundant
+
ion at m/z 142. Although [M-14] ions are initially present in the
mass spectra as a consequence of the ESI process, this combination
of data infers that the 2,2,6,6-tetramethylpiperidines 1H–5H are
diamagnetic products arising from photo-oxidation of 1–5, and
Scheme 3 Potential formation mechanisms of N–H piperidines from
nitroxyl radicals by: (a) alkyl, (b) peroxyl, or (c) hydroxyl radical addition
to TEMPO derivatives.
4
942 | Org. Biomol. Chem., 2011, 9, 4936–4947
This journal is © The Royal Society of Chemistry 2011

View Article Online
(
Scheme 3b), however for the present species, the first step in this
and at 298.15 K. Both computed values are in agreement with
-
1
45
process is slightly endothermic (8.5 kJ mol ) and has a significant
literature values. Comparison of these values with the data
-
1
43
∑
∑
energy barrier (72 kJ mol ). Similarly, addition of HO to
the nitroxyl oxygen, with subsequent cleavage of the nitrogen–
oxygen bond can result in formation of the secondary amine and
release of the hydroperoxyl radical (Scheme 3c), however our
unpublished calculations indicate that the decomposition step in
this process is also significantly endoergic (146.7 kJ mol at 298 K).
Further pathways might also be possible via the intermediary
oxoammonium cations formed by electron transfer, but the precise
mechanism of these is currently unclear. Such pathways may
be favoured in aqueous media due to the stabilisation of the
in Table 3 demonstrates that abstraction of hydrogens by HO
4
3
will be exothermic from all positions (i.e., H , H or CH
3
).
By contrast, abstraction of all hydrogens in the experimental
33
∑
compounds by HO
2
would be thermodynamically disfavoured,
∑
consistent with our recent study of HO
2
abstractions in other
organic compounds. Although H generally has a lower BDE
-
1
46
4
3
than either H or CH in the TEMPO compounds studied, these
3
thermodynamic differences alone do not provide an explanation
for the reaction at the 4-position observed by ESR and ESI-MS
analysis. Therefore, a study of the transition states of hydrogen
44
43
∑
charged species afforded by solvation. While several mechanistic
possibilities are suggested in the literature there is no definitive
explanation for secondary amine formation.
abstraction by HO was undertaken, and the resulting gas-phase
activation energies are displayed in Table 4. The corresponding
solution-phase data follow the gas-phase trends and are provided
in Table S2 of the ESI.‡
The gas phase activation Gibbs free energies of CH hydrogen
abstractions from 1 and 2 were calculated to be ca. 25 kJ mol ,
3
Computational investigation into the effects of the 4-position
substituent
-
1
and are assumed to be approximately constant with changes to
The observed oxidative modification at the 4-position of 2–4
4
3
the remote R substituent in both 3 and 4. H abstraction barriers
warranted further investigation. The influence of this substituent
-1
varied from ca. 24–30 kJ mol and were consistently lower for
4
on the adjacent carbon–hydrogen (H ) bond dissociation enthalpy
the substituted derivatives 2–4 than the corresponding activation
(
BDE)—a parameter related to the stability of the intermedi-
3
energies for abstractions of H from TEMPO 1 itself (ca. 34 kJ
ate carbon-centred radical resulting from the initial hydrogen
abstraction—was compared to the C–H BDEs of the methylene
-1
4
mol ). Conversely, the adjacent functional group (R ) does not
4
significantly lower the activation barrier of H abstraction relative
to TEMPO 1 with a barrier of ca. 35–38 kJ mol in all cases except
3
ring hydrogens (H ) and the hindering methyl hydrogens (denoted
-1
CH ). Calculations were undertaken using the G3(MP2)-RAD
3
that of 2. Whilst 2 appears to be a significant outlier, it should be
approach at 298.15 K and are presented in Table 3. Consideration
of reaction temperature and bulk solvent does not appear to have
any significant effect on the observed trends (see Table S1 in
ESI‡). The data in Table 3 demonstrate that for the 4-substituted
noted that for this reaction, standard transition state optimisations
for H abstraction led to geometries in which the attacking OH
4
∑
instead interacts with the covalently bound 4-OH substituent. For
this system, a series of constrained optimisations were performed.
While the final “transition structure” contains a single imaginary
4
TEMPO derivatives 2–4, the C–H BDE is significantly lower
than the C–H BDEs of either the 3-position or in the hindering
4
frequency consistent with H abstraction, we expect that the
4
methyl groups. Furthermore, the C–H BDE in the substituted
resulting structure is not fully optimised and the reported barrier
is thus likely to be significantly overestimated. Even so, in contrast
to the thermochemistry (Table 3), calculation of the activation
barriers of hydrogen abstraction (Table 4) suggests a clear kinetic
-
1
analogues 2–4 is significantly lower (by at least 16 kJ mol ) than
that in the archetypal TEMPO 1 system indicating stabilisation
4
of the carbon-centred radicals resulting from H -abstraction. In
contrast, the BDEs of the hindering methyl group hydrogens
are consistent with those of TEMPO and are not significantly
3
4
preference for abstraction at H and CH
3
over H . The increasing
3
4
abstraction barrier heights in the order CH
3
< H < H can be
4
influenced by variations in the remote R substituent. Similarly,
attributed to polar inductive effects overriding the thermodynamic
preference for the formation of a-radicals and is consistent with
previous computational and experimental observations of other
3
C–H BDEs are constant, except in the case of TEMPONE
1
5
(ca. 35 kJ mol- lower than 1–4) due to the stabilising
effect afforded by the conjugation with the neighbouring C
bond.
Considering hydroxyl (HO ) and hydroperoxyl radicals (HO
O
47
systems. Based on these calculations, it is plausible to conclude
that the formation of TEMPONE 5 in the experiments described
above does not arise from any preferential hydrogen abstraction
at the 4-position. Indeed hydrogen abstraction at the 3-position
∑
∑
2
)
are likely reactive species arising from aqueous TiO
2
photocatal-
-
1
ysis, the relevant BDE(HO–H) = 493 kJ mol and BDE(HOO–
-
1
H) = 366 kJ mol were calculated at the same level of theory
Table 4 Gas phase activation barriers for hydroxyl radical-induced
hydrogen abstraction from 4-substituted TEMPO species, calculated at
298.15 K by the G3(MP2)-RAD method
Table 3 Gas phase carbon–hydrogen BDEs for 4-substituted TEMPO
species, calculated at 298.15 K by the G3(MP2)-RAD method
Free energy barrier/kJ mol-
1
BDE/kJ mol-
1
4
4
3
R
H
H
CH
3
4
4
3
R
H
H
CH
3
1
2
3
4
H
OH
OCH
37.4
105.5
38.5
33.6
30.7
25.0
23.9
24.9
26.0
—
a
1
2
3
4
5
H
OH
OCH
COOH
O
412.4
392.1
396.2
378.0
—
412.1
414.1
414.9
415.5
377.9
422.5
427.5
430.4
427.7
428.0
3
3
COOH
35.3
—
a
Approximate barrier only (see text for further details).
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4936–4947 | 4943

View Article Online
and from the methyl groups is kinetically preferred, though given
the generally low barriers abstractions at all positions are likely
to take place to some extent. In order to rationalise TEMPONE
formation, plausible reactions following formation of the carbon-
centred radicals were considered in further detail. Although we
Under the conditions of the experiments described above
hydroxyl radicals and dioxygen are both present, with the latter
estimated to be more concentrated by at least an order of
33
magnitude. Addition of hydroxyl radical to carbon-centred
4
3
radicals arising from H , H or CH hydrogen abstraction from
3
4
do not exclude the possibility of H abstraction from TEMPOL
the 4-substituted TEMPO derivatives 2–4 would give rise to an
even-electron alcohol. These isomeric species are all plausible in-
termediates and/or products of this reaction and could contribute
2
based on the reported barrier, an alternate mechanism to
TEMPONE formation unique to TEMPOL via an alkoxyl radical
is also considered below.
+
to the observation of [M+16] ions upon ESI-MS analysis of the
4
∑
Scheme 4 After formation of a biradical by H abstraction, (a) the product from recombination with HO can rearrange to TEMPONE through a
6
2
-membered cyclic transition state with a water molecule to facilitate hydrogen transfer. (b) Alternatively, addition of O to the biradical yields a peroxyl
-1
radical, which forms TEMPONE by intramolecular hydrogen transfer. Values in italics (in kJ mol ) correspond to gas-phase Gibbs free reaction energies
and activation barriers (denoted by an asterisk).
4
944 | Org. Biomol. Chem., 2011, 9, 4936–4947
This journal is © The Royal Society of Chemistry 2011

View Article Online
reaction mixture (see above). Of these alcohols, only those arising
from hydroxyl addition at the 4-position can reasonably give rise
to formation of TEMPONE 5. Calculations for reactions arising
from the addition of HO to an alkyl radical centred on the 4-
position have been undertaken and are summarised in Scheme
thermodynamically reversible O
2
addition to an alkyl radical, a
process which also involves a potentially prohibitive activation
barrier to hydrogen abstraction. This alternate mechanism may
also explain the higher yield for 5 arising from 2 compared to 3 or
4, as observed by ESR (Fig. 3).
∑
4
(a). The corresponding solution-phase free energies follow the
gas-phase results and are provided in Table S3 of the ESI.‡
These data indicate that the combination reaction of hydroxyl
radical and HALS biradical is significantly exoergic, with free
Conclusions
This work demonstrates the conversion of TEMPOL 2 to TEM-
PONE 5 from reaction with hydroxyl radicals generated by photo-
-
1
-1
energies ranging from -316 kJ mol for 4 to -364 kJ mol for
relative to the nascent biradical in the gas phase. However,
2
catalytic TiO and further shows the broader applicability of this
2
transition states for subsequent hydrolysis of the a-hydroxy
intermediates to form TEMPONE were located and represent
significant kinetic barriers to this reaction, ranging from 121 kJ
mol- for 3 to 282 kJ mol for 4. While these barriers to
direct hydrolysis mediated by a single water molecule may seem
prohibitive, other pathways—especially those involving acid or
base catalysis—may provide lower energy alternatives.
reaction to other 4-substituted TEMPO derivatives (3, 4). Several
putative mechanisms for the formation of TEMPONE initiated
by hydrogen abstraction at the 4-position have been investigated
by high level ab initio calculations and suggest both hydroxyl
radical and dioxygen addition to the biradical intermediate may
play a role in this conversion. A second TEMPONE formation
mechanism unique to TEMPOL 2 via an alkoxyl radical was
shown to be a plausible alternative to the peroxyl radical pathway.
The quantitative loss of nitroxyl radicals observed by ESR during
the oxidation (Fig. 3) can be rationalised by ESI-MS observations
of formation of a secondary amine at the 1-position. This is
broadly consistent with previous estimates of the sites of reaction
in TEMPO derivatives that suggest ca. 40% of reactive encounters
result in reaction at the nitroxyl moiety, with the remaining
1
-1
In contrast, addition of ground state molecular oxygen to the
carbon centred radicals (Scheme 4(b)) was found to be mildly
exoergic for 2 and 3 and negligibly exoergic for 4 due to the
greater relative stability of the radical in the latter case (cf .
Table 3). Significantly however, the barriers to unimolecular
rearrangement of the peroxyl radicals to form TEMPONE were
-
1
moderate in all cases, ranging from 56 kJ mol for 2 to 82 kJ
-
1
44b
mol for 3. These calculations suggest that Scheme 4(b) provides a
viable mechanism for the experimental observation of TEMPONE
described above, consistent with previous proposals of carbonyl
formation from a-hydroxy, a-methoxy, and a-carboxy peroxyl
chemistry initiated by hydrogen atom abstraction at other sites.
It is further consistent with our own observations of the conversion
of functionalised HALS to the corresponding secondary amine in
polymer coatings and detailed investigations of the mechanisms of
this pathway and its consequences for the Denisov cycle (Scheme
48
radical intermediates. It should be noted however that O
2
42b
addition to the intermediate biradicals is clearly reversible with the
direct dissociation in the reverse direction entropically favoured
1) are currently underway.
Experiments similar to those described herein with authentic
HALS molecules are limited by their insolubility in aqueous
media. It is not unreasonable however, to extend the current
discussion to speculation of a plausible mechanism for HALS
consumption in pigmented surface coatings, where TiO2 is
ubiquitously employed. Active HALS are recorded at highest
concentrations in the top 30 mm of automotive clearcoats, near
over forward rearrangement to products. Therefore even with the
∑
greater abundance of O
2
compared to HO under the reaction
conditions the irreversible trapping of the biradical as an alcohol
is likely to be competitive, as evidenced by ESI-MS observations
+
of [M+16] ions.
For TEMPOL 2, an alternative pathway to TEMPONE 5, exclu-
sive to this system, needs to be considered (Scheme 5). Abstraction
of the hydrogen from the hydroxyl group itself would give rise to
a nascent alkoxyl radical that could then donate a hydrogen atom
to dioxygen (or another acceptor) to form the carboxyl moiety
directly. Calculated kinetic and thermodynamic parameters (see
also Table S3 in the ESI‡) are in favour of this pathway—
both steps are exoergic and the activation barrier of hydrogen
abstraction is plausible under the experimental conditions. Thus,
49
the air to polymer interface. Coincidentally, this is the region
50
of highest photocatalytic activity in TiO -pigmented materials,
2
given the necessity of moisture and oxygen to photocatalytic
oxidation. It is plausible that a transient encounter between
∑
photocatalytically generated HO and a HALS molecule may
lead to formation of TEMPONE by elimination of the HALS
backbone (added for the very purpose of thermal stability). The
active HALS moiety would subsequently be volatilised from
the coating at service temperatures. Given that the elimination
5
is likely to form from 2 via an alkoxyl radical rather than by
∑
Scheme 5 Alternative pathway to TEMPONE 5 from TEMPOL 2 via an alkoxyl radical, which subsequently adds to O
in italics (in kJ mol ) correspond to gas-phase Gibbs free reaction energies and activation barriers (denoted by an asterisk).
2
and eliminates HO
2
. Values
-1
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4936–4947 | 4945

View Article Online
described herein is not the major pathway of HALS activity, and
19 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. Montgomery, J. A., 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, N. J. 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,
∑
furthermore, that HO production is suppressed by TiO surface
2
modifications in durable pigments, the chances of these encounters
would be low. However, over the entire lifetime of the polymer it
may provide a significant, irreversible pathway for depletion of
these otherwise demonstrably prophylactic additives.
Acknowledgements
We acknowledge the generous allocation of time on the Na-
tional Facility of the National Computational Infrastructure
in Canberra, Australia, which is supported by the Australian
Commonwealth Government. S.J.B. and D.L.M. acknowledge
funding from the Australian Research Council (ARC) and
BlueScope Steel (LP0775032). S.J.B and M.L.Coote are funded
through the ARC Centre of Excellence for Free Radical Chemistry
and Biotechnology (CE0561607). D.L.M. is the holder of an
Australian Postgraduate Award, and M.L.Coote acknowledges
an ARC Future Fellowship.
¨
P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas,
J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09,
Revision A.02, Wallingford, CT., 2009.
2
0 H.-J. Werner, P. J. Knowles, R. Lindh, F. R. Manby, M. Sch u¨ tz, P.
Celani, T. Korona, A. Mitrushenkov, G. Rauhut, T. B. Adler, R. D.
Amos, A. Bernhardsson, A. Berning, D. L. Cooper, M. J. O. Deegan,
A. J. Dobbyn, F. Eckert, E. Goll, C. Hampel, G. Hetzer, T. Hrenar,
G. Knizia, C. K o¨ ppl, Y. Liu, A. W. Lloyd, R. A. Mata, A. J. May, S.
J. McNicholas, W. Meyer, M. E. Mura, A. Nicklaß, P. Palmieri, K.
Pfl u¨ ger , R. Pitzer, M. Reiher, U. Schumann, H. Stoll, A. J. Stone, R.
Tarroni, T. Thorsteinsson, M. Wang and A. Wolf, MOLPRO 2009.1,
a package of ab initio programs, see http://www.molpro.net.
1 J. L. Hodgson, L. B. Roskop, M. S. Gordon, C. Y. Lin and M. L. Coote,
J. Phys. Chem. A, 2010, 114, 10458–10466.
2
2
2
2 A. P. Scott and L. Radom, J. Phys. Chem., 1996, 100, 16502–16513.
3 D. J. Henry, M. B. Sullivan and L. Radom, J. Chem. Phys., 2003, 118,
4849–4860.
References
1
2
R. G. Hicks, Org. Biomol. Chem., 2007, 5, 1321–1338.
(a) J. Chateauneuf, J. Lusztyk and K. U. Ingold, J. Org. Chem., 1988,
24 M. L. Coote, C. Y. Lin, A. L. J. Beckwith and A. A. Zavitsas, Phys.
5
3, 1629–1632; (b) A. L. J. Beckwith, V. W. Bowry and K. U. Ingold, J.
Chem. Chem. Phys., 2010, 12, 9597–9610.
Am. Chem. Soc., 1992, 114, 4983–4992.
25 J. I. Steinfeld, J. S. Francisco and W. L. Hase, Chemical Kinetics and
Dynamics, Prentice Hall, Englewood Cliffs, N.J., 1989.
26 J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999–
3094.
27 M. L. Coote, Macromol. Theory Simul., 2009, 18, 388–400.
28 R. Chiarelli and A. Rassat, Tetrahedron, 1973, 29, 3639–3647.
29 S. R. Burks, M. A. Makowsky, Z. A. Yaffe, C. Hoggle, P. Tsai, S.
Muralidharan, M. K. Bowman, J. P. Y. Kao and G. M. Rosen, J. Org.
Chem., 2010, 75, 4737–4741.
30 H. Fischer and K. H. Hellwege, ed., Landolt-B o¨ rnstein: Numerical Data
and Functional Relationships in Science & Technology, Springer-Verlag,
Berlin, 1979, vol. 9c.
31 (a) K. Ishibashi, A. Fujishima, T. Watanabe and K. Hashimoto, J.
Photochem. Photobiol., A, 2000, 134, 139–142; (b) Q. Xiang, J. Yu and
P. K. Wong, J. Colloid Interface Sci., 2011, 357, 163–167.
32 M. A. Grela, M. E. J. Coronel and A. J. Colussi, J. Phys. Chem., 1996,
100, 16940–16946.
33 P. F. Schwarz, N. J. Turro, S. H. Bossman, A. M. Braun, A.-M. A.
Abdel Wahab and H. D u¨ rr, J. Phys. Chem. B, 1997, 101, 7127–7134.
34 C. F. Chignell, A. G. Motten, R. H. Sik, C. E. Parker and K. Reszka,
Photochem. Photobiol., 1994, 59, 5–11.
35 J. P. Blinco, J. L. Hodgson, B. J. Morrow, J. R. Walker, G. D. Will, M.
L. Coote and S. E. Bottle, J. Org. Chem., 2008, 73, 6763–6771.
36 K. Takeshita, K. Saito, J.-i. Ueda, K. Anzai and T. Ozawa, Biochim.
Biophys. Acta, Gen. Subj., 2002, 1573, 156–164.
3
4
V. Chechik and G. Ionita, Org. Biomol. Chem., 2006, 4, 3505–3510.
(a) M. K. Georges, R. P. N. Veregin, P. M. Kazmaier and G. K. Hamer,
Macromolecules, 1993, 26, 2987–2988; (b) M. Conte, Y. Ma, C. Loyns,
P. Price, D. Rippon and V. Chechik, Org. Biomol. Chem., 2009, 7, 2685–
2
687.
5
6
C. S. Wilcox and A. Pearlman, Pharmacol. Rev., 2008, 60, 418–469.
(a) F. Gugumus, in Plastics Additives, ed. R. G a¨ chter and H. M u¨ ller,
Carl Hanser Verlag, Munich, 4th edn, 1993, ch. 3, pp. 129–270; (b) C.
Schaller, D. Rogez and A. Braig, J. Coat. Technol. Res., 2009, 6, 81–88.
(a) P. P. Klemchuk and M. E. Gande, Polym. Degrad. Stab., 1988, 22,
7
8
2
41–274; (b) F. Gugumus, Polym. Degrad. Stab., 1993, 40, 167–215;
(
c) E. N. Step, N. J. Turro, P. P. Klemchuk and M. E. Gande, Angew.
Makromol. Chem., 1995, 232, 65–83.
(a) P. P. Klemchuk, M. E. Gande and E. Cordola, Polym. Degrad. Stab.,
1
990, 27, 65–74; (b) G. Geuskens and M. N. Kanda, Polym. Degrad.
Stab., 1996, 51, 227–232; (c) O. Brede, D. Beckert, C. Windolph and H.
A. G o¨ ttinger, J. Phys. Chem. A, 1998, 102, 1457–1464; (d) I. Rossi, A.
Venturini and A. Zedda, J. Am. Chem. Soc., 1999, 121, 7914–7917.
(a) F. Gugumus, Polym. Degrad. Stab., 1995, 50, 101; (b) J. L. Gerlock,
C. A. Peters, A. V. Kucherov, T. Misovski, C. M. Seubert, R. O. Carter
and M. E. Nichols, J. Coat. Technol., 2003, 75, 34.
9
1
0 (a) J. L. Gerlock, D. R. Bauer and L. M. Briggs, Polym. Degrad. Stab.,
1
986, 14, 53–71; (b) J. L. Gerlock, A. V. Kucherov and C. A. Smith,
Polym. Degrad. Stab., 2001, 73, 201–210.
1
1
1
1 P. Yaneff, K. Adamsons, N. Cliff and M. Kanouni, J. Coat. Technol.
Res., 2004, 1, 201–212.
37 R. M. Alberici, M. A. Mendes, W. F. Jardim and M. N. Eberlin, J. Am.
Soc. Mass Spectrom., 1998, 9, 1321–1327.
38 L. S. Santos, ed., Reactive Intermediates: MS Investigations in Solution,
2 G. Haacke, E. Longordo, F. F. Andrawes and B. H. Campbell, Prog.
Org. Coat., 1998, 34, 75–83.
Wiley-VCH, Weinheim, 2010.
3 (a) K. Murayama and T. Yoshioka, Bull. Chem. Soc. Jpn., 1969, 42,
39 (a) J. L. Hodgson, M. Namazian, S. E. Bottle and M. L. Coote, J.
Phys. Chem. A, 2007, 111, 13595–13605; (b) S. Manda, I. Nakanishi,
K. Ohkubo, H. Yakumaru, K. I. Matsumoto, T. Ozawa, N. Ikota, S.
Fukuzumi and K. Anzai, Org. Biomol. Chem., 2007, 5, 3951–3955.
40 (a) J. O. Metzger and J. Griep-Raming, Eur. J. Mass Spectrom., 1999,
5, 157–163; (b) C. D. Smith, J. P. Bartley, S. E. Bottle, A. S. Micallef
and D. A. Reid, J. Mass Spectrom., 2000, 35, 607–611.
2
307–2309; (b) G. Moad, E. Rizzardo and D. H. Solomon, Tetrahedron
Lett., 1981, 22, 1165–1168; (c) G. S. Ananchenko and H. Fischer, J.
Polym. Sci., Part A: Polym. Chem., 2001, 39, 3604–3621; (d) A. Nilsen
and R. Braslau, J. Polym. Sci., Part A: Polym. Chem., 2006, 44, 697–717.
4 Y. Taguchi, Y. Ishida, H. Ohtani and H. Matsubara, Anal. Chem., 2004,
1
7
6, 697–703.
1
1
5 N. Haider and S. Karlsson, Polym. Degrad. Stab., 2001, 74, 103–112.
6 (a) U. Deffner and W. Schimmack, Int. J. Radiat. Biol., 1976, 29, 71–
41 T. A. Lowe, M. R. L. Paine, D. L. Marshall, L. A. Hick, J. A. Boge, P.
J. Barker and S. J. Blanksby, J. Mass Spectrom., 2010, 45, 486–495.
42 (a) J. L. Gerlock, T. Riley and D. R. Bauer, Polym. Degrad. Stab., 1986,
14, 73–84; (b) M. R. L. Paine, P. J. Barker and S. J. Blanksby, Analyst,
2011, 136, 904–912.
43 J. L. Hodgson and M. L. Coote, Macromolecules, 2010, 43, 4573–4583.
44 (a) R. L. Willson, Int. J. Radiat. Biol., 1972, 21, 401–403; (b) S.
Goldstein and A. Samuni, J. Phys. Chem. A, 2007, 111, 1066–1072.
7
5; (b) K. Saito, J. I. Takeshita and T. Ozawa, J. Pharm. Sci., 2003, 92,
75–280; (c) W. Kudo, M. Yamato, K. Yamada, Y. Kinoshita, T. Shiba,
2
T. Watanabe and H. Utsumi, Free Radical Res., 2008, 42, 505–512.
7 N. S. Allen, M. Edge, A. Ortega, G. Sandoval, C. M. Liauw, J. Verran, J.
Stratton and R. B. McIntyre, Polym. Degrad. Stab., 2004, 85, 927–946.
8 P. J. Barker and A. Branch, Prog. Org. Coat., 2008, 62, 313–320.
1
1
4
946 | Org. Biomol. Chem., 2011, 9, 4936–4947
This journal is © The Royal Society of Chemistry 2011

View Article Online
4
4
5 S. J. Blanksby and G. B. Ellison, Acc. Chem. Res., 2003, 36, 255–263.
6 G. Gryn’ova, J. L. Hodgson and M. L. Coote, Org. Biomol. Chem.,
Chem. Rev., 1986, 86, 69–201; (c) P. Ulanski, E. Bothe, K. Hildenbrand,
J. M. Rosiak and C. von Sonntag, J. Chem. Soc., Perkin Trans. 2, 1996,
23–28; (d) E. M. Collins, H. W. Sidebottom, J. C. Wenger, S. Le Calv e´ ,
A. Mellouki, G. LeBras, E. Villenave and K. Wirtz, J. Photochem.
Photobiol., A, 2005, 176, 86–97.
2
011, 9, 480–490.
4
7 (a) A. Samuni, S. Goldstein, A. Russo, J. B. Mitchell, M. C. Krishna
and P. Neta, J. Am. Chem. Soc., 2002, 124, 8719–8724; (b) K. D. Beare
and M. L. Coote, J. Phys. Chem. A, 2004, 108, 7211–7221; (c) Z. I.
Watts and C. J. Easton, J. Am. Chem. Soc., 2009, 131, 11323–11325.
8 (a) E. Bothe, M. N. Schuchmann, D. Schulte-Frohlinde and C. von
Sonntag, Photochem. Photobiol., 1978, 28, 639–644; (b) R. Atkinson,
49 J. L. Gerlock, A. V. Kucherov and M. E. Nichols, J. Coat. Technol.,
2001, 73, 45–54.
50 H. G. V o¨ lz, G. K a¨ mpf, H. G. Fitzky and A. Klaeren, ACS Symp. Ser.,
1981, 156, 163–182.
4
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 4936–4947 | 4947