Quenching of singlet oxygen by tertiary aliphatic amines. Structural effects on rates and products

Article abstract of DOI:10.1002/hlca.200690211

A kinetic and product study of the reaction of a series of α-methyl-substituted N-methylpiperidines with thermally generated ¹O₂ in MeCN was carried out. It was found that as the number of α-methyl groups (Me in α-position relative to the N-atom) increases, the rate of ¹O₂ quenching (physical plus chemical) slightly decreases. This finding shows that, with respect to the reaction rate, steric effects are much more important than electronic effects as the latter should have produced the opposite result. The opposite outcome was instead found for the chemical quenching that leads to the N-demethylation products and N-formyl derivatives. The same trend was observed for the ratio between N-demethylation and formation of the N-formyl derivatives (NH/NCHO ratio). All these results are consistent with the mechanism reported in Scheme 1 where an exciplex is first formed that by a H-atom transfer process produces an α-amino-substituted C-radical. The latter forms the product of N-demethylation by one electron oxidation, or affords the N-formyl derivative by radical coupling (Scheme 1). Similar results were obtained with N,N-dimethylcyclohexanamine. However, this 'acyclic' amine exhibited behaviors quite distinct from those of the N-methylpiperidines series, with respect to reaction rate, extent of chemical quenching, and NH/NCHO ratio.

Full text of DOI:10.1002/hlca.200690211

Helvetica Chimica Acta – Vol. 89 (2006)
2273
Quenching of Singlet Oxygen by Tertiary Aliphatic Amines. Structural Effects
on Rates and Products
a
b
a
by Enrico Baciocchi* ), Tiziana Del Giacco ), and Andrea Lapi )
a) Dipartimento di Chimica, Università ꢀLa Sapienzaꢁ, P.le A. Moro 5, I-00185 Rome
(
phone: +39 06 49913711; e-mail: enrico.baciocchi@uniroma1.it)
b) Dipartimento di Chimica and Centro di Eccellenza Materiali Innovativi Nanostrutturati (CEMIN),
Università di Perugia, via Elce di sotto 8, I-06123 Perugia
In memoriam Professor Hanns Fischer
Akinetic and product study of the reaction of a series of a-methyl-substituted N-methylpiperidines
1
with thermally generated O in MeCN was carried out. It was found that as the number of a-methyl
2
1
groups (Me in a-position relative to the N-atom) increases, the rate of O quenching (physical plus chem-
2
ical) slightly decreases. This finding shows that, with respect to the reaction rate, steric effects are much
more important than electronic effects as the latter should have produced the opposite result. The oppo-
site outcome was instead found for the chemical quenching that leads to the N-demethylation products
and N-formyl derivatives. The same trend was observed for the ratio between N-demethylation and for-
mation of the N-formyl derivatives (NH/NCHO ratio). All these results are consistent with the mecha-
nism reported in Scheme 1 where an exciplex is first formed that by a H-atom transfer process produces
an a-amino-substituted C-radical. The latter forms the product of N-demethylation by one electron oxi-
dation, or affords the N-formyl derivative by radical coupling (Scheme 1). Similar results were obtained
with N,N-dimethylcyclohexanamine. However, this ꢀacyclicꢁ amine exhibited behaviors quite distinct
from those of the N-methylpiperidines series, with respect to reaction rate, extent of chemical quenching,
and NH/NCHO ratio.
Introduction. – The ability of tertiary aliphatic amines to efficiently quench O
D ), henceforth indicated as O , is well recognized [1]. There is substantial agreement
g 2
2
1
1
(
that this reaction involves the reversible formation of an exciplex with partial charge-
transfer character that can proceed to the starting amine and the ground triplet state
3
À
O ( S ), henceforth indicated as O (physical quenching), or to products (chemical
2
g
2
quenching) [1a,c]. In some cases, an electron-transfer mechanism has been proposed
[
2], but this seems unlikely in view of its high endergonicity.
1
While the kinetic aspects of the reactions of amines with O have been investigated
2
in some detail [1], much less information is presently available concerning the products
formation, although it is well established that chemical quenching accompanies physi-
cal quenching in a substantial number of amines [2][3]. Few pioneering studies are
available [3a,b,d], but information is almost completely lacking with respect to the
role of structural effects on the extent of chemical quenching and, most particularly,
with respect to the mechanistic aspects of the process by which the exciplex is converted
into the reaction products.
ꢂ 2006 Verlag Helvetica Chimica Acta AG, Zürich

2
274
Helvetica Chimica Acta – Vol. 89 (2006)
Recently, we have reported a detailed kinetic and products study of the reaction of
1
N,N-dimethylbenzylamine with O [4]. The mechanism depicted in Scheme 1 has been
2
suggested where the key step is a H-atom transfer (HAT) process, within the exciplex,
from the CH ÀN group to oxygen (Path b). The formed C-radical and HOC can then
3
2
1
undergo radical coupling, to eventually give a formamide derivative (Path c) ), or elec-
tron transfer to produce an iminium cation leading to the N-dealkylation product (Path
d).
Scheme 1
To have more detailed information concerning this mechanism and the structural
factors at the basis of the competition between physical and chemical quenching as
well as between the paths leading to N-demethylation and formamide formation, we
carried out a kinetic and product study of the reaction of a series of a-methyl substi-
1
tuted N-methylpiperidines (i.e., of 1–4) with O in MeCN. For the purpose of compar-
2
ison, an ꢀacyclicꢁ amine, N,N-dimethylcyclohexanamine (5), was also investigated. As in
1
our previous study [4], O was generated by thermal decomposition of 1,4-dimethyl-
2
naphthalene endoperoxide [5].
Results and Discussion. – The reactions of the amines 1–5 with thermally generated
O were carried out with 0.01M substrate and 0.01M –0.1M 1,4-dimethylnaphthalene
1
2
endoperoxide (depending on the extent of chemical quenching), at 408 for 4 h.
1)
It is, however, possible that some formamide also derives from reaction of the C-radical with O2.

Helvetica Chimica Acta – Vol. 89 (2006)
2275
Under these conditions, all the 1,4-dimethylnaphthalene endoperoxide decomposes to
1
give 1,4-dimethylnaphthalene and O in 70% yield, referred to the initial amount of
2
endoperoxide [6]. Products were characterized (comparison with authentic com-
pounds) and quantified by GC and GC/MS. With piperidines 2–4, the reaction prod-
ucts were the secondary amine (N-demethylation product) and the N-formyl derivative
(i.e., the piperidine-1-carboxaldehyde as shown in Scheme 2 for the specific case of
piperidine 2) that formed in substantial yields. With 1, however, practically only the
N-formyl derivative was formed. The ꢀacyclicꢁ amine 5 exhibited the same behavior
as 2–4, affording the N-demethylated product (N-methylcyclohexanamine) and the
N-formyl derivative (N-methyl-N-cyclohexylformamide).
Scheme 2
1
Under the above conditions, all O generated by the endoperoxide is quenched by
2
1
8
À1 À1
M
the 0.01M substrate, since the latter quenches O at a rate in the order of 10
s
2
1
4
À1
(
vide infra) that is much higher than the decay rate of O in MeCN (10 s [1b]).
2
Thus, the fraction of chemical quenching (Q ), that is the fraction of exciplex that
c
undergoes product formation with respect to the total quenching (chemical plus phys-
ical), can easily be calculated as the total product yield vs. the amount of singlet oxygen
that has been generated. Products, yields, and fractions of chemical quenching (Q ) are
c
reported in the Table. In the same table are also displayed the rates of total quenching,
kQ, measured by time-resolved luminescence at 1270 nm, together with the peak-poten-
tial (E ) values of the amines that were measured in MeCN by cyclic voltammetry.
p
Considering first the kinetic results for the N-methylpiperidines, the data in the
Table show that the progressive addition of a-methyl groups in the piperidine ring
1
Table. Amine Oxygenations by O
2
b
À1 À1
Products yields [%]a)
Q [%] ) NH/NCHOd)
c
Amine E ) [V] kQ [M
s
]
p
c
Secondary amine N-Formyl derivative
e
8
1
2
3
4
5
)
1.04
0.98
0.94
0.76
0.90
1.7·10
traces
12 (Æ1)
2
7
9
35
10
0
1.9
11
19
2.2
f
f
7
7
7
8
)
)
9.3·10
8.8·10
8.9·10
2.7·10
15 (Æ1)
28 (Æ2)
23 (Æ1)
9.2 (Æ0.5)
8.0 (Æ0.4)
2.5 (Æ0.1)
1.2 (Æ0.1)
4.2 (Æ0.2)
g
)
)
h
a) Referred to the initial amount of the starting amine and determined by GC analysis. ) E vs. Ag/AgCl
b
p
c
1
d
in MeCN at 258. ) Total product yield vs. the amount of O
2
generated. ) Molar ratio between the sec-
e
f
ondary amine and the N-formyl derivatives, see text. ) Amine/endoperoxide 10 :1. ) Amine/endoperox-
ide 5 :1. ) Amine/endoperoxide 1:1. ) Amine/endoperoxide 2 :1.
g
h

2
276
Helvetica Chimica Acta – Vol. 89 (2006)
(
Me in a-position to the N-atom) lowers the rate of total quenching k . The effect, how-
Q
8
À1 À1
ever, is rather small since k goes from 1.7·10
8
M
s
for 1 (no a-methyl groups) to
Q
7
À1 À1
.9·10
M
s
for 4 (4 a-methyl groups). Apossible explanation is that, as previously
proposed by Monroe [7], the a-methyl groups reduce the rate by a steric effect since the
1
formation of the exciplex requires a close approach of O to the N-atom of the amine.
2
The entity of the steric effect is probably much larger than that appearing from the rate
data. Accordingly, the oxidation potential values reported in the Table show that the
progressive addition of a-methyl groups has also the effect of decreasing the oxidation
potential of the piperidine. This effect should determine an increase in the quenching
rate [8], which is a result opposite to what found. Clearly, steric effects must be very
significant as they have overcome electronic effects to a large extent.
It can also be noted that N,N-dimethylcyclohexanamine (5) exhibits a kQ value
higher than that of 4, even though the oxidation potential of the former is significantly
higher than that of the second (0.90 vs. 0.76 V). Model inspections clearly indicate that
2
steric hindrance around the N-atom is much less in 5 than in 4 ). Thus, again, steric
1
effects appear to outweigh electronic effects with respect to the rate of O quenching.
2
In contrast to what was observed for the k value, the presence of the a-methyl
Q
groups in N-methylpiperidines has a very positive effect upon the extent of chemical
quenching that significantly increases on passing from 1 to 4 (from 2 to 35%). Thus,
the a-methyl groups favor more the path leading to products than that leading to phys-
ical quenching. Interestingly, a plot of Q values for 1–4 against the oxidation potentials
c
is roughly linear (Fig. 1) indicating that the competition between physical quenching
and chemical quenching (a HAT process according to Scheme 1) is influenced by the
same (electronic) effects that influence the oxidation potential. Namely, the extent of
the chemical quenching increases as the oxidation potential of the amine becomes
lower. Since the exciplex is suggested to be a charge-transfer complex, the present
results would indicate that its tendency to undergo a HAT transfer increases as the
extent of charge transfer increases (the oxidation potential of the amine becomes
lower). This is in agreement with our previous hypothesis [4] that the H-atom transfer
3
in the exciplex has substantial character of proton transfer ).
Even though the correlation between Q and E for piperidines is very rough, it is
c
p
quite clear that the Q value for the ꢀacyclicꢁ amine 5 (statistically corrected) falls quite
c
significantly below the Q vs. E line of the piperidines (Fig. 1). At present, it is not pos-
c
p
sible to provide a reasonable explanation.
Coming to the product nature, the first notation is that with both N-methylpiperi-
dines 1–3 and N,N-dimethylcyclohexanamine (5), the chemical quenching concerns
only the CÀH bonds of the N-methyl groups. The ring CÀH bonds in a-position to
2)
3)
Agreater steric demand for 4 than for 5 has recently been suggested to explain the different catalytic
effect of the two amines in ketone enolization [9].
Clearly, dealing with a competition, it might also be suggested that the observed trend in the Q val-
c
ues is due to a physical quenching that becomes less efficient as the oxidation potential of the amine
decreases. However, this possibility is highly unlikely as it is well known that the opposite holds in
systems where only physical quenching is observed [8]. Steric effects could also be called into
play, but it is very difficult to envisage that an increase in the steric hindrance from 1 to 4 can
favor chemical quenching with respect to physical quenching.

Helvetica Chimica Acta – Vol. 89 (2006)
2277
Fig. 1. Chemical quenching (Q ) of amines 1–5 vs. their oxidation potentials. The Q value for 5 is
c
c
statistically corrected on the basis of the number of N-methyl groups to be compared with that of
the piperidine series.
the N-atom are practically unaffected. In general, a greater reactivity of exocyclic
HÀC(a) bonds (particularly of MeN groups) with respect to the endocyclic ones
may be expected [1a], probably since only the former bonds can rotate around the
NÀC bond and assume the conformation most suitable for the cleavage (that with
4
the HÀC(a) bond syn-periplanar to the lone pair at the N-atom) ). For the case in
point, it should be considered that the equatorial HÀC(a) bonds of piperidine are
not syn-periplanar with the lone pair at the N-atom and should have a low reactivity
5
due to stereoelectronic effects ). On the other hand, no reactivity may be predicted
for the axial HÀC(a) bonds since they are anti-periplanar to the lone pair at the
N-atom and, in the exciplex, reside on the face of the ring opposite to that where oxy-
gen is located. This situation probably holds also for the ring HÀC(a) bond of 5 since, in
the most stable conformation of the amine (ab initio calculations), this bond too should
be anti-periplanar with respect to the lone pair at the N-atom. Moreover, it should be
added that in the case of 5, there is also a high (6 :1) statistical factor in favor of the
MeN H-atoms.
Coming to the secondary amine/formamide (NH/NCHO) product ratio, it can be
noted that it increases quite regularly in the order 1<2<3<4, that is by increasing
the number of a-methyl groups and decreasing the oxidation potential. As Q , also
c
4)
5)
This conformation provides the maximum stabilization to the formed C-radical (stereoelectronic
effect).
Dealing with competition reactions, also weak stereoelectronic effects (1 kcal mol or less) can
À1
deeply influence the product distribution.

2
278
Helvetica Chimica Acta – Vol. 89 (2006)
Fig. 2. Secondary amine/N-formyl derivative (NH/NCHO) molar ratio vs. amine oxidation potentials
the NH/NCHO ratio exhibits a roughly linear correlation when plotted against the Ep
values of the amines (Fig. 2). One possibility is that the factors that lower the oxidation
potential of the amine also lower the oxidation potential of the C-radical formed in
Path b of Scheme 1. Therefore, the electron-transfer path leading to N-demethylation
(Path d) may be favored with respect to Path c. Steric effects, however, should also
be considered given that they increase as well in the order 1<2<3<4. Thus, an addi-
tional suggestion might be that steric effects indirectly favor the secondary-amine for-
mation by making more difficult the radical coupling pathway (Path c in Scheme 1). It is
reasonable to think that very likely both electronic and steric effects can play a role in
determining the NH/NCHO ratio. Even though we are unable to establish the relative
importance of the two effects, there is no doubt that the observed trend in the NH/
NCHO ratio is consistent with the mechanism described in Scheme 1.
The ꢀacyclicꢁ amine 5 behaves differently than N-methylpiperidines also with
respect to the NH/NCHO ratio that, accordingly, is significantly lower than that
expected for an N-methylpiperidine with similar E value (see Fig. 2). Apossibility is
p
that with the ꢀacyclicꢁ amine, the radical coupling of Path c is favored by the same steric
factors suggested to operate with respect to the rate of quenching.
1
In conclusion, the results observed in the O -promoted oxidation of the N-methyl-
2
piperidine series indicate that electronic and steric effects may be differently felt in the
various steps of the mechanism that leads to products (Scheme 1). Thus, the overall
reaction rate appears to be affected much more by steric effects than by electronic
effects. Conversely, electronic effects appear to be the dominant ones in the conversion
of the exciplex into products. Accordingly, the extent of chemical quenching quite reg-
ularly increases as the amine oxidation potential becomes lower. This indicates that the
HAT inside the exciplex is facilitated by enhancing the charge-transfer character of the

Helvetica Chimica Acta – Vol. 89 (2006)
2279
exciplex itself. Both electronic and steric effects are suggested to operate with respect
to the competition between the two pathways that convert the intermediate a-posi-
tioned C-radical into the secondary amine and the N-formyl derivative, respectively
(Scheme 1). At present, however, it is not possible to establish their relative weight.
Finally, the ꢀacyclicꢁ amine 5 was found to present behaviors clearly distinct from
those of piperidines. Accordingly, with respect to the latter, the rate of exciplex forma-
tion is higher, while the extent of chemical quenching and the ratio between N-deme-
thylation and formamide formation are significantly lower. Aminor role of steric
effects in the reaction of the ꢀacyclicꢁ tertiary amine than in that of piperidines
might, at least in part, explain the observed effects.
MIUR, University ꢀLa Sapienzaꢁ of Rome, and University of Perugia are thanked for the financial
support.
Experimental Part
General. Reagents: 1-methylpiperidine (1), 1,2,2,6,6-pentamethylpiperidine (4) and N,N-dimethylcy-
clohexanamine (5) were commercially available (Aldrich) and passed through alumina before use. The
1
,2-dimethylpiperidine (2) and cis-1,2,6-trimethylpiperidine (3) were prepared by N-methylation of 2-
methylpiperidine (Aldrich) and cis-2,6-dimethylpiperidine (Aldrich), resp., with methyl iodide (Fluka).
The 1,4-dimethylnaphthalene endoperoxide was prepared according to [5a]. Products: all the secondary
amines as well as piperidine-1-carboxaldehyde (Aldrich) were commercially available. N-Cyclohexyl-N-
methylformamide was prepared by reaction of N-methylcyclohexanamine with formic acid. The 2-meth-
ylpiperidine-1-carboxaldehyde, cis-2,6-dimethylpiperidine-1-carboxaldehyde and 2,2,6,6-tetramethylpi-
peridine-1-carboxaldehyde were prepared according to [10]. Acetonitrile (Carlo Erba, HPLC-plus
grade) was used without further purification. GC: Varian CP 3800. GC/MS: Agilent 6800N equipped
with an Agilent-5973Network mass selective detector.
À2
Oxidation by Thermally Generated Singlet Oxygen. A2-ml soln. of amine (10 M) and 1,4-dimethyl-
À2
À2
À2
naphthalene endoperoxide (0.1M for 1, 5·10 M for 2 and 3, 10 M for 4 and 2·10 M for 5) in MeCN was
heated at 408 (water bath) in the dark for 4 h. An internal standard (4-methylbenzophenone) was added,
and the mixture was analyzed by GC and GC/MS. No reaction was observed when oxygen-saturated
amine solns. were kept at 408 for 4 h in the absence of 1,4-dimethylnaphthalene endoperoxide.
Amine Oxidation Potentials. Cyclic voltammetry was carried out in an N -purged soln. of the amine
2
(
2 mM) and tetrabutylammonium tetrafluoroborate (0.1M) in MeCN at 258. The cell contained an 1.5-mm
disc glassy carbon electrode, a Pt auxiliary electrode and an Ag/AgCl (in 3M KCl) reference electrode.
For each amine, at least four determinations were carried out.
1
1
Determination of O Quenching Rate Constants (k ). O was produced in MeCN by energy transfer
2
q
À5
2
to O2 from the triplet state of phenalenone (5.0·10 M), generated by excitation at 355 nm from a
Nd :YAG laser (Continuum, Surelite II, pulse width ca. 7 ns and energy <3 mJ per pulse). The phosphor-
1
escence emission of O emerging from the cuvette passed through a cut-off filter at 1050 nm (Oriel,
2
5
1362) and three pieces of gelatin cut-off filter at 870 nm (Kodak Wratten, 87C) and was detected by a
germanium diode detector (Judson, J16 SP, 5-mm diameter) [11]. After amplification with a two-stage
home-made amplifer (ca. 100 MHz bandwidth, 14 dB), the output of the diode was fed into a digital sig-
nal analyzer (Tektronik DSA602) and computer stored and analyzed (Tektronik PEP301). Rate constants
1
1
for the quenching of O (k ) were determined from the decrease of O emission lifetime in O -saturated
2
q
2
2
À3
MeCN, in the presence of various amounts of amine (0.15–4.0·10 M). All measurements were carried
out at 22Æ28.
Ab initio Calculations on 5. The minimum energy geometry for 5 was determined by ab initio calcu-
lations at the HF/6-31G* level on a Silicon-Graphics-O platform using a Spartan 5.0.1 program.
2

2
280
Helvetica Chimica Acta – Vol. 89 (2006)
REFERENCES
[
1] a) A. A. Gorman, Adv. Photochem. 1992, 17, 217; b) F. Wilkinson, W. P. Helman, A. B. Ross, J. Phys.
Chem. Ref. Data 1995, 24, 663; c) C. Schweitzer, R. Schmidt, Chem. Rev. 2003, 103, 1685; d) E. L.
Clennan, A. Pace, Tetrahedron 2005, 61, 6665.
[
2] W. F. Smith, Jr., J. Am. Chem. Soc. 1972, 94, 186; H. Tsubomura, T. Yagishita, H. Toi, Bull. Chem.
Soc. Jpn. 1973, 46, 3051; C. M. Haugen, W. R. Bergmark, D. G. Whitten, J. Am. Chem. Soc. 1992,
114, 10293; R. Bernstein, C. S. Foote, J. Phys. Chem. A. 1999, 103, 7244; G. Cocquet, P. Rool, C. Fer-
roud, J. Chem. Soc., Perkin Trans. 1 2000, 14, 2277.
[
3] a) M. H. Fisch, J. C. Gramin, J. A. Olesen, J. Chem. Soc., Chem. Commun. 1970, 13; b) M. H. Fisch,
J. C. Gramin, J. A. Olesen, J. Chem. Soc., Chem. Commun. 1971, 663; c) D. Bellus, H. Lind, J. F.
Wyatt, J. Chem. Soc., Chem. Commun. 1972, 1199; d) K. Gollnick, J. H. E. Lindner, Tetrahedron
Lett. 1973, 1903; e) K. Inoue, I. Saito, T. Matsuura, Chem. Lett. 1977, 607.
4] E. Baciocchi, T. Del Giacco, A. Lapi, Org. Lett. 2004, 6, 4791.
5] a) N. J. Turro, M. F. Chow, J. Am. Chem. Soc. 1981, 103, 7218; b) W. Adam, M. Prein, Acc. Chem.
Res. 1996, 29, 275; c) A. Greer, G. Vassilikogiannakis, K.-C. Lee, T. S. Koffas, K. Nahm, C. S.
Foote, J. Org. Chem. 2000, 65, 6876; d) S. Ben-Shabat, Y. Itagaki, S. Jockusch, J. R. Sparrow, N. J.
Turro, K. Nakanishi, Angew. Chem., Int. Ed. 2002, 41, 814; e) T. Poon, N. J. Turro, J. Chapman, P.
Lakshminarasimhan, X. Lei, S. Jockusch, R. Franz, I. Washington, W. Adam, S. G. Bosio, Org.
Lett. 2003, 5, 4951.
[
[
[
[
[
6] E. Baciocchi, T. Del Giacco, A. Lapi, Org. Lett. 2006, 8, 1783.
7] B. M. Monroe, J. Phys. Chem. 1977, 81, 1861.
8] K. Furukawa, E. A. Ogryzlo, J. Photochem. 1972, 1, 163; R. H. Young, D. Brewer, R. Kayser, R.
Martin, D. Feriozi, R. A. Keller, Can. J. Chem. 1974, 52, 2889.
[
9] P. Zhao, D. B. Collum, J. Am. Chem. Soc. 2003, 125, 14411.
[
10] Z. Blum, K. Nyberg, Acta Chem. Scand., Ser. B 1981, 35, 743.
[
11] F. Elisei, G. G. Aloisi, C. Lattarini, L. Latterini, F. DallꢁAcqua, A. Guiotto, Photochem. Photobiol.
1996, 64, 67.
Received April 21, 2006