One-electron oxidation of sterically hindered amines to nitroxyl radicals: Intermediate amine radical cations, aminyl, α-aminoalkyl, and aminylperoxyl radicals

Article abstract of DOI:10.1021/jp980089j

Sterically hindered amines (2,2,6,6-tetramethyl-substituted piperidines) are easily oxidized (i) by electron transfer to parent cations in n-butyl chloride solution, (ii) by the sulfate radical anion in aqueous solution, and (iii) by sensitized electron transfer to carbonyl triplets. In nonpolar surroundings in the nanosecond time range, the radical cations of the tertiary piperidines have been directly observed by optical spectroscopy to exhibit absorption maxima below λ = 300 nm and around 550 nm. Subsequently, they deprotonate to α-alkylamine radicals, which are also the first observable products of oxidation with sulfate radical anions in water. In the case of secondary piperidines, the amine radical cations deprotonate to aminyl radicals in times < 10 ns. The triplet-sensitized electron transfer to the benzophenone as well as cyclohexanone triplet results in amine-derived and ketyl-type radicals formed at a nearly diffusion-controlled rate, which suggests an electron-and subsequent proton-transfer mechanism. In the presence of oxygen, the amine-derived radicals are oxidized to nitroxyl radicals by different pathways for secondary and tertiary piperidines. For the reaction of the nitroxyl radicals with other radicals, rate constants are found to be quite similar (about 5 × 10⁸ M^{-1 s-1}) for several alkyl radicals and for the tert-butyloxyl radical and less than 10⁵ M^-1 s^-1 for alkylperoxyl radicals. Because of the minor importance of radical reactions with the sterically hindered amines (HALS), the antioxidant effect of these compounds ought to be explained by oxidation, primarily via cationic and subsequently radical intermediates to the persistent nitroxyl radicals, which scavenge other radicals very efficiently.

Full text of DOI:10.1021/jp980089j

J. Phys. Chem. A 1998, 102, 1457-1464
1457
One-Electron Oxidation of Sterically Hindered Amines to Nitroxyl Radicals: Intermediate
Amine Radical Cations, Aminyl, r-Aminoalkyl, and Aminylperoxyl Radicals
O. Brede,* D. Beckert, C. Windolph, and H. A. Go1ttinger
UniVersity of Leipzig, Interdisciplinary Group for Time-ResolVed Spectroscopy, Permoserstrasse 15,
D-04303 Leipzig, Germany
ReceiVed: NoVember 5, 1997; In Final Form: December 26, 1997
Sterically hindered amines (2,2,6,6-tetramethyl-substituted piperidines) are easily oxidized (i) by electron
transfer to parent cations in n-butyl chloride solution, (ii) by the sulfate radical anion in aqueous solution,
and (iii) by sensitized electron transfer to carbonyl triplets. In nonpolar surroundings in the nanosecond time
range, the radical cations of the tertiary piperidines have been directly observed by optical spectroscopy to
exhibit absorption maxima below λ ) 300 nm and around 550 nm. Subsequently, they deprotonate to
R-alkylamine radicals, which are also the first observable products of oxidation with sulfate radical anions in
water. In the case of secondary piperidines, the amine radical cations deprotonate to aminyl radicals in times
< 10 ns. The triplet-sensitized electron transfer to the benzophenone as well as cyclohexanone triplet results
in amine-derived and ketyl-type radicals formed at a nearly diffusion-controlled rate, which suggests an electron-
and subsequent proton-transfer mechanism. In the presence of oxygen, the amine-derived radicals are oxidized
to nitroxyl radicals by different pathways for secondary and tertiary piperidines. For the reaction of the
nitroxyl radicals with other radicals, rate constants are found to be quite similar (about 5 × 108 M-1 s-1 ) for
several alkyl radicals and for the tert-butyloxyl radical and less than 105 M-1 s-1 for alkylperoxyl radicals.
Because of the minor importance of radical reactions with the sterically hindered amines (HALS), the
antioxidant effect of these compounds ought to be explained by oxidation, primarily via cationic and
subsequently radical intermediates to the persistent nitroxyl radicals, which scavenge other radicals very
efficiently.
Introduction
Steady-state aging experiments with polyolefin matrixes
containing HALS compounds resulted in complicated consid-
erations concerning the antioxidant and photostabilizing effect
of the sterically hindered amines that included the quenching
of excited polymer-oxygen complexes.7
None of these indications gave a detailed picture of the
stabilizing effect of HALS, and the following questions
remained: Are the HALS themselves the stabilizers, and if so,
what is the effective mechanism? Or do the HALS generate
any products that are the species interrupting the oxidative
degradation chain of the organic material?
Although sterically hindered amines have long been used as
stabilizers mainly against the light-induced degradation of
polyolefins, the detailed mechanism of their action is still under
discussion.1 In a review of the mechanistic effects of photo-
stabilization, Chirinos2 emphasized the central role of alkyl-
peroxyl radicals and alkylhydroperoxides within the process of
the photooxidation of polymers. Hence in accordance with the
classical philosophy of stabilization, the termination of the
oxidation chain is explained to proceed by the deactivation of
the alkylperoxyl radicals under amine radical formation (1) or
by direct nitroxyl radical formation in a synchronous mechanism
(2), as proposed by Geuskens3 on the basis of a steady-state
kinetic study.
The chief experimental problem is the difficulty of clearly
detecting and identifying the radical intermediates. Most of the
stationary experiments are based on the EPR spectroscopical
observation of the stable nitroxyl radicals, which are involved
in the HALS action mechanism, more or less as the final
products of amine oxidation.3
ROO• + >NH f [RO2- + >NH•+] f RO2H + >N• (1)
To overcome the particular problems given in polyolefin
systems such as no uniform molecular weight, high viscosity,
and a multiphase system due to the semicrystallinity, time-
resolved experiments on the HALS oxidation have been
performed with low-molecular (liquid) systems.
(2)
Excited singlets act as quenchers without chemical conse-
quence,4 whereas in the case of triplet quenching radical
products have been observed, which indicate electron transfer
followed by proton transfer within the ion pair, or a hydrogen
abstraction reaction5,6 (eq 3).
Hence, triplet sensitized electron transfer (eq 3) enabled the
observation of the acceptor radicals AH• with optical spectros-
copy,8 whereas AH• and >N• have been detected by electron
spin resonance spectroscopy.5,9 Furthermore, amine radical
cations have been detected in low-temperature and matrix
isolation experiments.8,10-12 Our own pulse radiolysis experi-
ments on electron transfer from sterically hindered amines to
A
T1 + >NH f [A•-‚‚‚>NH•+] f AH• + >N• (3)
* Corresponding author.
S1089-5639(98)00089-9 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/07/1998

1458 J. Phys. Chem. A, Vol. 102, No. 9, 1998
Brede et al.
solvent parent ions resulted in the case of secondary HALS in
the optical detection of the corresponding amine radical
cations.6,8
The aim of this paper is to give a detailed picture of the
transient reactions and antioxidant action mechanism of steri-
cally hindered amines. For this purpose we examined polar
and nonpolar liquid and glassy model systems with stationary
and pulsed radiolysis and laser flash photolysis combined with
detection by both optical and electron paramagnetic spectros-
copy.
with a dose between 100 and 200 Gy. The optical detection
system consisted of a pulsed 900-W xenon lamp (XBO 900,
Osram), a Spectra Pro-500 monochromator (Acton Research
Corp.), an 1P28 (RCA) photomultiplier, and a TDS 640 (500-
MHz) Tektronics digitizing oscilloscope.
Laser Photolysis. Experiments were carried out with a
perpendicular probe excitation arrangement. The solutions were
photolyzed in a quartz cell connected to a flow system using
the fourth (266-nm) or third (355-nm) harmonic of a Quanta
Ray GCR-11 Nd3+:YAG laser (Spectra Physics Inc.). Pulses
3 ns in length (fwhm) had an energy of 10-30 mJ. Optical
detection was performed with equipment similar to that de-
scribed for pulse radiolysis.
Experimental Section
Chemicals. The following sterically hindered amines were
used for our investigations:
Steady-State Radiolysis Experiments. Solutions of the
HALS compounds in the alkanes in the presence or absence of
ketones and with well-defined oxygen concentration were
prepared by the freeze-pump-thaw technique, gas introduction,
and sealing. Irradiation was performed at room temperature in
a
60Co γ-ray source at doses up to 2 kGy. The nitroxyl yield
of the samples was then analyzed by electron spin resonance
spectroscopy.
Results and Discussion
To analyze the mechanism of the actual HALS antioxidant
action, we studied those elementary reactions in detail that could
be involved either in the radical reaction channel, the excited-
state pathway, or the one-electron oxidation (ionization).
Radical Reactions. Pulse radiolysis enables the generation
of different types of solvent radicals which could be considered
as reaction partners of HALS. Hence, we studied the reactivity
•
of the two basic HALS compounds 1 and 2 to OH, c-C6H11
,
c-C6H11OO•, •CH2OH, and other alkyl and alkylperoxyl radicals.
As more or less expected, only in the case of OH did we observe
a marked effect due to HALS radical formation (absorption tail
in the UV) as formulated in eqs 4 and 5.
1 + OH f >N• + H2O
k4 ) 5 × 109 M-1 s-1 (4)
2 + OH f >N-CH2 + H2O k5 ) 4 × 109 M-1 s-1 (5)
•
In neither the pulse radiolysis of cyclohexyl chloride nor
cyclohexane and methanol did we find any indication of the
above-mentioned reactions of the various alkyl and alkylperoxyl
radicals with HALS. Hence, only upper limits of the rate
constants could be estimated, which differ owing to the different
lifetime of the solvent radicals.
2,2,6,6-Tetramethylpiperidine (1), 1,2,2,6,6-pentamethylpiperi-
dine (2), N-methoxy-2,2,6,6-tetramethylpiperidine (3), and the
industrially used HALS compounds bis(2,2,6,6-tetramethyl-4-
piperidinyl)sebaceate, Tinuvin 770 (4), and an oligomeric ester
of butandioic acid with 4-hydroxy-2,2,6,6-tetramethyl-1-pip-
eridineethanol, Tinuvin 622 (5), were supplied by Ciba-Geigy.
The piperidines were of analytical grade and in some cases
purified by distillation in the presence of citric acid. 2,2,6,6-
Tetramethylpiperidyl TEMPO (6•), and 2,2,6,6-tetramethyl-4-
hydroxy-piperidyl radical, HO-TEMPO (7•), were obtained from
Aldrich. n-Butyl chloride (spectroscopical grade, Merck) was
dried by chromatography on A4 molecular sieves and then
distilled under nitrogen flux. Cyclohexane (spectroscopical
grade, Merck) and n-hexadecane (Erckner) were purified by
treatment with oleum, washing with water, drying on an A4
molecular sieve, and distillation. Benzophenone, cyclohex-
anone, and potassium peroxidisulfate were used in commercial
quality (p.a., Aldrich).
1, 2 + R• f products
k6 < 106 M-1 s-1
k7 < 104 M-1 s-1 (7)
(6)
1, 2 + ROO• f products
Summarizing these observations, it seems very unlikely that
reactions of solvent or matrix radicals play an important role
in the antioxidant action mechanism of sterically hindered
amines during oxidative degradation.
Carbonyl Compounds as Sensitizers for HALS Reac-
tions: Stationary Radiolysis. As reported in ref 3, γ-radiolysis
of oxic solutions of some HALS compounds in cyclohexane
yielded considerable amounts of nitroxyl radicals. If not of
radical origin according to eq 2, the nitroxyls could also be
formed by a much more complex mechanism involving the
oxidation of cyclohexane to cyclohexanone,15 followed by
carbonyl triplet sensitized electron transfer with HALS and
subsequent radical reactions as illustrated by the following
Pulse Radiolysis. Experiments were performed with a pulse
transformer type accelerator ELIT (Institute of Nuclear Physics
Novosibirsk, Russia) delivering 1-MeV, 16-ns electron pulses

Sterically Hindered Amines
J. Phys. Chem. A, Vol. 102, No. 9, 1998 1459
TABLE 1: Normalized Yields of Aminoxyl Radicals
Determined by EPR Spectroscopy of γ-Irradiated (2-kGy)
Solutions of 1 mM of the HALS Compound 5 in
Cyclohexane under Various Oxygen Concentrations and the
Addition of Ketones (0.1 M)
normalized yield of >NO•
sensitizer (0.1 M)
deaerated
air-saturated
O2-saturated
without ketone
cyclohexanone
benzophenone
0.03
0.07
0.03
1.35
3.40
2.90
1.00
2.00
0.82
reaction sequence (eq 8).
+ O2
γ8 c-C6H11
8 c-C6H11OO• 2×8
•
c-C6H12
9
c-C6H10O, etc. (8a)
Figure 1. Yield of nitroxyl radicals (0) after γ-irradiation (4 kGy) of
samples consisting of 0.1 M cyclohexanone and 0.5 mM 4 dissolved
in n-hexadecane in the presence of different oxygen concentrations
(relative concentration 1 equals oxygen-saturated solution) determined
by EPR spectroscopy. (1) and the dotted line indicate the estimated
curve shape at essentially lower oxygen yield. The inset gives the EPR
spectrum of the nitroxyl radical.
radiation-chem. excitation
by different ways, ISC
c-C6H10O
8 c-C6H10OT1
(8b)
c-C6H10OT1 + >NH f
[c-C6H10O•-‚‚‚>NH•+]
+
H
transfer, diffusion8
the reaction mechanism.18 The amine species are not appropri-
ate for this purpose because of their unfavorable absorption
properties.
We performed 355-nm laser flash photolysis experiments with
polyethylene (PE) doped with benzophenone (about 1 mM). As
shown in Figure 2 after the flash we observed the first excited
triplet state of the benzophenone. This triplet decayed by H
abstraction from the surrounding PE, yielding the benzophenone
ketyl radical. The spectra show the typical behavior also
observed in the liquid state, and hence, particularly the spectral
c-C6H10OH• + >N• (8c)
>N• + O2 f [ ] f >NO•
(8d)
To prove this hypothesis we undertook stationary γ-radiolysis
experiments on solutions of 5 in cyclohexane (and n-heptade-
cane) with different oxygen concentrations and compared these
measurements with those taken in the presence of an added
ketone. The samples were irradiated with a dose of 2 kGy,
and then the nitroxyl radical yield was determined by EPR
spectroscopy integrating the three-line spectrum (Table 1). It
can be seen that oxygen, of course, is necessary for nitroxyl
formation and that in the presence of ketones the nitroxyl yield
is about twice as high. Furthermore, in the air-saturated sample
a markedly higher nitroxyl yield is found than in the oxygen-
saturated one. Therefore, we suggest that in the case of
cyclohexane-containing oxygen, its radiation-induced oxidation
to carbonyl compounds initially takes place (cf. eq 8a), which
then act as sensitizers for HALS oxidation (eqs 8b,c).
range at >530 nm could be used for kinetic analysis.
H
abstraction (eq 10) was found to proceed with k11 ) 106 s-1, as
was estimated from the time profile in Figure 2a.
>COT1 + PE f >C•-OH + PE•
>COT1 + >NH f [>CO•-‚‚‚>NH•+] f >COH• + >N•
(11)
(10)
Then we allowed both liquid piperidines 1 and 2 to diffuse into
the foil, removed the residue on the surface, and performed flash
photolysis. There was no indication for a benzophenone radical
anion. Only its ketyl radical could be found. The changes in
the time profile caused by the reaction of the piperidines (eq
11) analyzed at 600 nm can be seen in the insets of Figures 2
(b and c). Because of uncertainties in the distribution of the
piperidines within the semicrystalline matrix, only the first-order
rate was determined, k11) 1.2 × 107 s-1 for 1 and 6.2 × 106
s-1 for 2. If nearly the same concentration of the two different
amines is assumed, the reaction of the secondary amine seems
to be faster than that of the tertiary one.
Oxygen dependence was studied systematically in the case
of samples containing 0.5 mM 4 and 0.1 M cyclohexanone
dissolved in n-hexadecane. Figure 1 shows the integral nitroxyl
radical yield depending on the oxygen concentration of the
solution. Corresponding to the finding given in Table 1, a yield
dependence typical of triplet intermediates was observed, i.e.,
first increasing and at a higher oxygen concentration decreasing
due to triplet quenching.
c-C6H10OT1 + O2 f c-C6H10O + O2
(9)
Considering the normalized nitroxyl yields (see Table 1) and
the influence of oxygen (see Figure 1), the electron-transfer (eq
8) hypothesis seems to be favored.
Sensitization by Aromatic and Alicyclic Ketones: Time-
Resolved Optical Spectroscopy. Because of its pronounced
transient properties,16,17 benzophenone is a very useful sensitizer.
Hence its triplet has visible absorption with a maximum at λ )
530 nm, the radical anion absorbs with a maximum at around
600 nm, and the ketyl radical has an absorption maximum at
540 nm. Hence these species could be used as monitors for
Because of the more qualitative character of the polymer
photolysis experiment, we undertook liquid-state model experi-
ments with ketones as sensitizers and different HALS com-
pounds as electron donors. As a solvent we used benzene,
which does not react in an H abstraction reaction, and so kinetic
analysis becomes easier. The mechanism of the triplet sensitized
reaction 11 was the same as in the case of PE; that is only the
benzophenone triplet and the ketyl radical were observed as
transients and there was no indication of the existence of the
anion. Table 2 gives the rate constants of reaction 11 of various

1460 J. Phys. Chem. A, Vol. 102, No. 9, 1998
Brede et al.
Figure 2. Time profiles (at 600 nm) taken during 355-nm laser photolysis of 1-mm PE foils doped with 1 mM benzophenone (BP) in this pure
state (inset a) and after lying 10 h in the liquid piperidines 1 (inset b) and 2 (inset c). The optical absorption spectra were taken from the pure
sample (a) and show the BP triplet (b, 50 ns after the flash) and the resulting BP ketyl radical absorption (9, 150 ns; 2, 800 ns after the flash).
TABLE 2: Rate Constants of the Benzophenone Triplet
Sensitized Reaction 10b with Different HALS Compounds
Taken from Pulse Radiolysis29 and Laser Photolysis
Experiments in Benzene Solution at Room Temperature
electron donor
k/dm3 mol-1 s-1
4-hydroxy-2,2,6,6-tetramethylpiperidine
4-n-heptyl-2,2,6,6-tetramethylpiperidine
triacetonamine
5, Tinuvin 622 (oligomer, tertiary amine)
4, Tinuvin 770 (dimer, secondary amine)
1, 2,2,6,6-tetramethylpiperidine
1.0 × 109
1.5 × 109
7.5 × 108
1.3 × 108
1.3 × 109
2.0 × 109
<109
2, 1,2,2,6,6-pentamethylpiperidine
differently structured HALS compounds. Although the products
are radicals,5,6 owing to the nearly diffusion-controlled rate, the
mechanism should be electron transfer followed by proton
transfer within the ion pair. As a general observation, the overall
rate constant of eq 11 is higher in the case of secondary HALS
than in the case of tertiary piperidines.
Using a model system that is closer to the polymer than
benzophenone, analogous experiments were performed with
cyclohexanone as a sensitizer and acetonitrile as a solvent. Both
HALS model compounds 1 and 2 quench the cyclohexanone
triplet at a rate of k12 ) 108 M-1 s-1. The products could not
be seen because they absorb within the UV range. Figure 3
shows the spectrum of the cyclohexanone triplet. The Stern-
Volmer plot demonstrates the quenching influence of the
secondary amine 1 and exhibits a nonlinear behavior, arguing
for a complexed intermediate. In the case of the tertiary amine
2 (not shown here), this phenomenon was not observed and the
Stern-Volmer plot fits well linearly.
Figure 3. Optical absorption spectra taken after 266 nm flash photolysis
of a solution of 0.1 M cyclohexanone in acetonitrile: (b) 30 ns, (4)
100 ns, and (9) 800 ns after the flash. The two absorption bands are
assigned to the cyclohexanone triplet. The inset gives the Stern-Volmer
plot due to reaction 12, i.e., the acceleration of the triplet decay in the
presence of different concentrations of 1.
The ion-molecule reaction of the HALS with radiolytically
generated solvent parent ions ought to generate solute radical
cations.19,20 Hence we also performed pulse radiolysis experi-
ments with solutions of the model HALS compounds 1, 2, and
3 in butyl chloride as solvent, reactions 13a,b.
radiolysis
•
n-C4H9Cl
8 n-C4H9Cl•+, Cl-, n-C4H9 , HCl,
and other fragmentation products (13a)
c-C6H10OT1 + >NH f [c-C6H10O•-‚‚‚>NH•+] f
n-C4H9Cl•+ + >N-X f n-C4H9Cl + >N•+-X,
X ) H, CH3, OCH3 (13b)
products (not observed) (12)
Amine Radical Cations as Intermediates of the HALS
Oxidation. As already reported, because of the very rapid
internal deprotonation, the ionic intermediates in the ion pair
were not detected in the nanosecond time range. Hence the
general question of the stability and reactivity of the HALS
radical cations arose.
In this case the origin of the solute species is very clearly defined
and radical reaction paths can be excluded for kinetic reasons
(much lower rate). Figure 4a-c shows spectra taken in the
pulse radiolysis of the mentioned three amines. All parts of
the figure include also the spectrum of the parent ions taken in

Sterically Hindered Amines
J. Phys. Chem. A, Vol. 102, No. 9, 1998 1461
pure n-butyl chloride. In the presence of the amines, these
species disappear. For all three amines, the rate constant k13b
amounts to 1.5 × 1010 M-1 s-1, i.e., slightly higher than
expected for diffusion control. This could be caused by an
inhomogeneous spikelike part of the parent ion n-C4H9Cl•+ time
profile.19
No amine radical cation absorption could be seen for 1 as a
secondary amine after the disappearance of the solvent ion
absorption in the vis range (Figure 4a). There is only very minor
absorption of aminyl radicals around 500 nm. The UV part
represents a superposition of solvent and solute radicals,
including those partially of the aminyl type. These observations
are interpreted in terms of rapid electron transfer (eq 13b),
followed by an immediate deprotonation (eq 14) either by
solvent proton stabilization or by Cl- or HCl, respectively.
-
>N•+H solvent, Cl or HCl8 >N• + H+(n-C4H9Cl), etc.
(14)
The tertiary amine 2 shows relatively stable cation absorption
in the red range (Figure 4b), while that of the N-methoxy-
substituted amine 3 is much more stable (Figure 4c). Consider-
ing the time profiles at 500 nm, the increasing part gives a
kinetic indication of the product side of reaction 13b. These
measurements argue for the actual existence of HALS radical
cations and demonstrate their different kinetic stability. Radical
cations were also visualized21 by reaction 13b in the case of
the technically used HALS compounds.
Subsequent Radical Reaction Resulting in Nitroxyl Radi-
cals. Because of the different kinds of radicals produced in
the pulse radiolysis of organic systems and their spectral
superpositions, studying the fate of the HALS species is very
complicated. Therefore we undertook oxidation of the amines
•-
1-3 with the photolytically formed one-electron oxidant SO4
,
ref 22. Hence we were able to selectively form the radicals
derived from the HALS radical cation by deprotonation in the
aqueous surrounding.
•-
S2O82- + hν f 2SO4
(15a)
(15b)
SO4•- + >N-X f SO42- + >N•+X
>N•+X + H2O f H3O+ + radicals
•
•
>N• for 1, >N-CH2 for 2, >N-OCH2 for 3 (15c)
As shown in Figure 5, the 266-nm flash generates the known
sulfate radical absorption with λmax ) 450 nm, which reacts
with the amine 2 under the formation of solute radicals absorbing
at the UV border. Apart from the substitution of the amine
nitrogen, the spectrum for the solute N- and R-C-radicals looks
quite similar. The reaction rate k15b,c amounts to 1.1 × 109
M-1 s-1 for 1, 2.6 × 109 M-1 s-1 for 2, and 2.5 × 108 M-1 s-1
for 3, as was calculated from the time profiles. However, in
the presence of oxygen, the fate of the HALS aminyl and
R-aminomethyl radicals is very different. The decay of the
R-methyl radical of 2 is accelerated by oxygen such that for
low O2 concentration the time profile decays faster in pseudo-
first-order behavior (k16a ) 3 × 108 M-1 s-1), whereas with a
higher oxygen level the signal becomes stable and has a fixed
lifetime (k16b ) 7 × 104 s-1). This is shown in a Stern-Volmer
plot given as an inset of Figure 5. We interpret this phenomenon
Figure 4. Transient absorption spectra taken in the pulse radiolysis
of solutions of the piperidines in n-butyl chloride. The transient
absorption of the n-butyl chloride parent radical cation (60 ns after the
pulse) of the pure solvent is given as a solid line. (a) The sample
contains 10 mM of 1: (b) 60 ns and (0) 450 ns after the electron
pulse. (b) The sample contains 10 mM of 2: (b) 60 ns and (0) 1.5 µs
after the electron pulse. (c) The sample contains 10 mM of 3: (b) 60
ns and (0) 35 µs after the electron pulse. The insets give time profiles
of the vis transient at λ ) 500 nm and demonstrate the different stability
of the amine radical cations.
as reaction 16a with the formation of an R-aminoalkylperoxyl,
which has a shorter lifetime than the R-aminoalkyl radical and
decays by fragmentation (eq 16b). The uncertainties in the
Stern-Volmer plot and the apparently too slow rate constant
k16a could be explained by the superposition of the R-aminoalkyl

1462 J. Phys. Chem. A, Vol. 102, No. 9, 1998
Brede et al.
Figure 5. Transient absorption spectra taken in the 266-nm laser photolysis of a solution of 10 mM of 2 and 50 mM of potassium peroxidisulfate
in water: (b) after 20 ns, caused by SO4•-, and after (4) 750 ns and (9) 35 µs caused by R-alkyl radicals. In the presence of oxygen after 35 µs
(0) this radical absorption is completely depleted. The inset shows the rate dependence of reaction 16 on the oxygen concentration.
and the R-aminoalkylperoxyl radicals within the UV spectral
range.
tetroxide mechanism (eq 19), which has been frequently
proposed in analogy with hydrocarbon chemistry.
In contrast to this, the kinetics of the aminyl radical of 1 is
practically unaffected by oxygen (k17 e 2 × 106 M-1 s-1). This
corresponds to previous findings about the low reactivity of
aminyl radicals to oxygen, which has been estimated to have a
rate constant of k17 e 107 M-1 s-1, refs 23, 24. Low-
temperature matrix isolation experiments with γ-irradiation and
EPR detection confirm this assumption, identifying N-centered
peroxyl radicals and giving grounds for interpreting the oxygen
reaction as a loose equilibrium11 (eq 17).
>N• + ROO• f [>N-OO-R] f >N-O• + RO•,
R ) alkyl (18a)
>N• + >NOO• f [>N-OO-N<] f 2 >N-O•
(18b)
2>NOO• f [>N-OOOO-N<] f 2>N-O• + O2 (19a)
>NOO• + ROO• f [>N-OOOO-R] f
>NO• + RO• + O2 (19b)
k16a
k16b
>N-CH2 + O2 8 >N-CH2OO•
8
•
>N-O• + CH2O (16)
Nitroxyl Radicals as Efficient Scavengers of Alkyl, Alkox-
yl, and Alkylperoxyl Radicals. Generally speaking, nitroxyl
radicals are very persistent species with only a slight dimer-
ization tendency, as formulated in equilibrium 20.
>N• + O2 h >N-OO•
(17)
2 >N-O• h >N-OO-N<
(20)
Whereas reaction 16 gives a good hint for the mechanistic
interpretation of nitroxyl radical formation (normally, alkyl-
peroxyl radicals ought to survive much longer than alkyl
radicals), the explanation of the formation of nitroxyls from
aminyl radicals is somewhat speculative. CIDEP experiments
on the triplet-sensitized electron transfer with carbonyls and
HALS facilitated the time-resolved observation of the aminyl
radicals by FT-EPR spectroscopy.25 As expected, the aminyl
radical of 1 exhibits spin polarization. By contrast, the nitroxyl
radicals subsequently formed in the presence of oxygen do not
show this polarization. This suggests that during the course of
the transformation mechanisms a diamagnetic intermediate is
crossed and the polarization is lost in this state.
Because of the above-mentioned low reactivity of aminyls
with respect to oxygen (reaction 17), recombination reactions
with alkyl peroxyl radicals have a chance to compete. There-
fore, the reaction of the relatively persistent aminyl radicals with
alkylperoxyl or other peroxyl radicals (eq 18) ought to be the
main source of the nitroxyls rather than the more exotic-looking
In the case of sterical hindrance, this reaction can be completely
neglected. In the presence of other radicals, however, they act
as very efficient scavengers, forming neutral products.27,28 To
determine rate constants of the reactions of nitroxyl radicals,
we performed pulse radiolysis experiments with the stable
radicals TEMPO (6•) and HO-TEMPO (7•) and generated
transient radicals as alkyl, alkoxyl, and alkylperoxyl radicals
radiolytically. All these radicals have only small absorptions
at λ e 300 nm, in the same spectral range in which the nitroxyl
radicals also slightly absorb. Thus kinetic analysis of the time
profiles must take this fact into account. Figure 6 (upper part)
shows time profiles at λ ) 250 and 280 nm taken in pure and
oxygen-free cyclohexane compared to those taken in the
presence of TEMPO (lower part). Considering the time axis,
it can be seen that there is an important acceleration of the decay
of the cyclohexyl radicals as well as a depletion effect caused
by the consumption of the added nitroxyl radical. Analogous

Sterically Hindered Amines
J. Phys. Chem. A, Vol. 102, No. 9, 1998 1463
Figure 6. Time profiles of the cyclohexyl radical at λ ) 250 and 280 nm obtained in the pulse radiolysis of pure deaerated cyclohexane (upper
profiles) and in the presence of 6• (lower part).
TABLE 3: Rate Constants of the Reaction of the Stable
Nitroxyl Radicals 6• (TEMPO) and 7• (HO-TEMPO) with
Alkyl, Alkoxyl, and Alkylperoxyl Radicals
The situation is much more complicated for alkoxyl radicals,
because of their limited lifetime due to their fragmentation. We
generated the relatively stable tert-butoxyl radical in aqueous
solution according to reaction 23.
k (dm3 mol-1 s-1
)
6•
7•
•
c-C6H11
i-C8H17
6.3 × 108
2.6 × 108
8.0 × 108
4.6 × 108
<5 × 105
<105
6.0 × 108
3.1 × 108
(CH3)3COOH + eaq- f (CH3)3CO• + OH-
(CH3)3CO• + >NO• f products
(23)
(24)
•
•
n-C17H35
t-C4H9-O•
5.8 × 108
n-C17H35OO•
other ROO•
Under our experimental conditions, we found a lifetime of
about 600 ns for the tert-butoxyl radical. The decay of the
alkoxyl radical was accelerated by adding increasing concentra-
tions of 6•, and hence, for reaction 24 a rate of k24 ) 4.6 × 108
M-1 s-1 was determined, i.e., a value quite similar to that for
the alkyl radicals. Table 3 summarizes the rate parameters of
the reaction of the different radicals with nitroxyl radicals.
Mechanism of HALS Stabilizer Action. In the foregoing
sections we demonstrated that HALS compounds are much more
succeptible to ionization than to undergoing radical or energy-
transfer reactions. Now the question of the actual HALS
antioxidant action mode should be discussed.
It is quite clear that the low reactivity against alkyl and similar
radicals found in the pulse radiolysis experiments do not
recommend the sterically hindered amines as stabilizers. By
contrast, all ionization procedures such as direct biphotonic
ionization, electron transfer with parent ions (eq 13b), one-
electron oxidation with appropriate oxidants (eq 15b), and
triplet-sensitized electron transfer (eq 3) first lead to amine
radical cations, which were directly observed in certain cases.
Particularly in polar surroundings and also under the conditions
of sensitized electron transfer within the intermediate ion pairs,
the piperidine radical cations rapidly deprotonate under the
experiments were also performed with n-heptadecane and
isooctane and showed similar effects. The rate constants k21a
are on the order of some 108 M-1 s-1, evidently exhibiting the
impact of the structure of R• (cf. Table 3). This corresponds
well with data determined by other methods.26,27
RH irradiation8 [R• + H•] f 2R• + H2
>N-O• + R• f >N-O-R
(21a)
(21b)
To check the case of the reaction of alkylperoxyl radicals,
pulse radiolysis was carried out with oxic alkane samples in
the absence and presence of 6•. Under the given limitations of
sensibility and long-term resolution for heptadecylperoxyl
radicals (not of completely uniform structure), a rate constant
of k22 ) 5 × 105 M-1 s-1 was estimated in an analogeous
manner as described in the preceding paragraph. This rate seems
to be similar for other alkylperoxyl radicals, but could hardly
be detected.
>N-O• + ROO• f products
(22)

1464 J. Phys. Chem. A, Vol. 102, No. 9, 1998
Brede et al.
the formation of radical cations. This reaction seems to be a
key reaction step in explaining the antioxidant activity of HALS
compounds. Depending on the surroundings, the amine radical
cations deprotonate to secondary amines under the formation
of aminyl radicals and to tertiary amines of R-aminoalkyl
radicals, which in the presence of oxygen are transformed via
different peroxyl radical intermediates into nitroxyl radicals.
These very persistent radicals react efficiently with alkyl and
alkoxyl radicals, but less efficiently with alkylperoxyl radicals.
This reaction of the aminoxyls interrupts the radical oxidation
chain of organic material and therefore causes the actual
antioxidant phenomenon of HALS. Under technical and natural
conditions, the oxidation sequence of the HALS compounds
seems to be mainly initiated by ketone triplet sensitized electron
transfer.
References and Notes
(1) (a) Wiles, D. M.; Carlsson, D. J. In New Trends in the Photochem-
istry of Polymers; Allen, N. S., Rabek, J. F., Eds.; Elsevier Applied Science
Publ.: New York, 1985; p 147. (b) Scott, G. In Atmospheric Oxidation
and Antioxidants; Scott, G., Ed.; Elsevier: Amsterdam, 1993; Vol. I, p 121;
Vol. II, p 385. (c) Gugumus, F. Polym. Degrad. Stab. 1993, 40, 167; 1994,
44, 299 and citations. (d) Step, E. N.; Turro, N. J.; Klemchuk, P. P.; Gande,
M. E. Angew. Makromol. Chem. 1995, 232, 65. (e) Gugumus, F. In Current
Trends in Polymer Chemistry; Allen, N. S., Edge, M., Bellobaro, I. R.,
Seli, E., Eds.; Ellis Horwood: Hampstead, 1995; p 257.
(2) Chirinos Padron, A. J. J. Photochem. Photobiol., A: Chem. 1989,
49, 1; JMS-ReV. Macromol. Chem. Phys. 1990, C30 (1), 107.
(3) Geuskens, G.; Nedelkos, G. Polym. Degrad. Stab. 1987, 19, 365.
Geuskens, G.; Kanda, M. N. Polym. Degrad. Stab. 1996, 51, 227.
(4) Bortolus, N.; Flamingni, L.; Minto, F.; Monti, S.; Faucitano, A. J.
Photochem. Photobiol., A: Chem. 1992, 68, 239.
(5) Step, E. N.; Turro, N. J.; Gande, M. E.; Klemchuk, P. P. J.
Photochem. Photobiol., A: Chem. 1993, 74, 203.
Figure 7. Reaction diagram describing the path of the (one-electron)
oxidation of HALS compounds to nitroxyl radicals via intermediate
amine radical cations, radicals and peroxyl radicals. In addition to the
ionic pathway, the checked excited state and radical reaction channels
are also mentioned.
(6) Brede, O. Radiat. Phys. Chem. 1997, 49, 39.
(7) Gugumus, G. Makromol. Chem., Macromol. Symp. 1989, 27, 25.
(8) Brede, O. Macromol. Symp. 1995, 100, 71.
(9) Beckert, D.; Fessenden, R. W. J. Phys. Chem. 1996, 100, 1622.
(10) Werst, D. W.; Trifunac, A. D. J. Phys. Chem. 1991, 95, 1268.
(11) Go¨ttinger, H. A.; Zubarev, V. E.; Brede, O. J. Chem. Soc., Perkin
Trans. 2 1997, 2167.
(12) (a) Malatesta, V.; Ingold, K. U. J. Am. Chem. Soc. 1973, 95, 6400.
(b) Shida, T.; Nosaka, Y.; Kato, T. J. Phys. Chem. 1978, 82, 695. (c)
Eastland, G. W.; Rao, D. N. R.; Symons, M. C. R. J. Chem. Soc., Perkin
Trans. 2 1984, 1551. (d) Qin, X.-Z.; Williams, Ff. J. Phys. Chem. 1986,
90, 2292. (e) Shiotani, M.; Sjo¨nqvist, L.; Lund, A.; Lunnel, S.; Erikson,
L.; Huang, M.-B. J. Phys. Chem. 1990, 94, 8081.
(13) Gugumus, F. In Current Trends in Polymer Chemistry; Allen, N.
S., Edge, M., Bellobaro, I. R., Seli, E., Eds.; Ellis Horwood: Hempstead,
1995; p 257.
(14) Das, S.; Schuchmann, M. N.; Schuchmann, H. P.; v. Sonntag, C.
Chem. Ber. 1987, 120, 319.
(15) Wojnarovits, L. In Radiation Chemistry of Hydrocarbons; Fo¨ldiak,
G., Ed.; Elsevier Publ.: New York, 1981; p 201.
(16) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochem-
istry; Marcel Dekker: New York, 1993.
(17) Brede, O.; Helmstreit, W.; Mehnert, R., Z. Phys. Chem. (Leipzig)
1975, 256, 505, 513.
(18) Brede, O.; Hermann, R.; Helmstreit, W.; Taplick, T.; Stephan, L.
Makromol. Chem., Macromol. Symp. 1988, 18, 113.
(19) Mehnert, R.; Brede, O.; Naumann, W. Ber. Bunsen-Ges. Phys.
Chem. 1982, 86, 525.
(20) Brede, O.; Mehnert, R.; Naumann, W. Chem. Phys. 1987, 115, 279.
(21) Go¨ttinger, H. A.; Brede, O. To be published.
formation of aminyl radicals (for secondary HALS) or R-ami-
noalkyl radicals (tertiary HALS). These radicals, however,
exhibit different decay behavior: (i) the R-aminoalkyl radicals
rapidly react with oxygen and fragment under the elimination
of formaldehyde to nitroxyl radicals, whereas (ii) the aminyl
radicals react slowly with oxygen and seem to prefer recom-
bination with other radicals, which leads to aminoxyl and
alkoxyl radicals in the case of the combination with alkylperoxyl
radical (eqs 18a,b) via peroxides.
Because of the sterical hindrance, the 2,2,6,6-tetramethylpi-
peridyloxyl radicals are rather stable against recombining
themselves. But with other types of radicals such as alkyl and
alkoxyl and to a lesser extent alkylperoxyl radicals, they react
relatively rapidly under product formation (eq 21). This has
been demonstrated in the pulse radiolysis of TEMPO solutions
in organic solvents, under which conditions the mentioned
radicals are generated.
Hence, the conclusion seems to be justified that neither the
HALS compounds nor the ionic and radical transients with
amine function are the active antioxidants. The actual stabilizing
species are the nitroxyl radicals formed in a sequence of
reactions starting with amine ionization. This conclusion is
illustrated by a reaction diagram (Figure 7) that shows a
mechanism that ought to be generally valid for nitroxyl radical
formation and its stabilizing effect.
(22) (a) Deeble, J. D.; Schuchmann, M. N.; Steenken, S.; v. Sonntag,
C. J. Phys. Chem. 1990, 94, 8186. ( b) Brede, O.; David. F.; Steenken, S.
J. Chem. Soc., Perkin Trans. 2 1995, 2, 23.
(23) Hepp, M. G.; Kramer, H. E. A. J. Photochem. Photobiol., A: Chem.
1994, 78, 19.
(24) Schuchmann, H.-P.; v. Sonntag, C. In Peroxyl Radicals; Alfassi,
Z. B., Ed.; Wiley: New York, 1997; p 439.
(25) Sa¨uberlich, J. Ph.D. Theses, University of Leipzig, 1996.
(26) Arends, J. W. C. E.; Mulder, P.; Clark, K. B.; Wayner, D. D. M.
J. Phys. Chem. 1995, 99, 8182.
Conclusions
(27) Nigam, S. N.; Asmus, K.-D.; Willson, R. L. J. Chem. Soc., Faraday
Trans. 1 1976, 72, 2324.
(28) Brede, O.; Mehnert, R.; Naumann, W.; Teply, J. Ber. Bunsen-Ges.
Phys. Chem. 1985, 89, 1036.
By using direct transient detection and characterization
techniques, in this paper we have demonstrated that the sterically
hindered amines easily undergo one-electron oxidation under