Solvent effect on the sensitized photooxygenation of 2,3-dihydropyrazine derivatives

Article abstract of DOI:10.1021/jo026303o

Detection of O₂(¹Δ_g) phosphorescence emission, λ_max = 1270 nm, following laser excitation and steady-state methods was employed to determine the total rate constant, k_T, and the chemical reaction rate

Full text of DOI:10.1021/jo026303o

Solven t Effect on th e Sen sitized P h otooxygen a tion of
2,3-Dih yd r op yr a zin e Der iva tives
Else Lemp,* Antonio L. Zanocco, German Gu¨nther, and Nancy Pizarro
Universidad de Chile, Facultad de Ciencias Qu´ımicas y Farmace´uticas, Departamento de Quı´mica
Orga´nica y Fisicoquı´mica, Casilla 233, Santiago-1, Santiago, Chile
elemp@ciq.uchile.cl
Received August 9, 2002
Detection of O2(1∆g) phosphorescence emission, λmax ) 1270 nm, following laser excitation and steady-
state methods was employed to determine the total rate constant, kT, and the chemical reaction
rate constant, kR, for reaction between 5,6-disubstituted-2,3-dihydropyrazines and singlet oxygen
in several solvents. Values of kT ranged from 0.26 × 105 M-1 s-1 in hexafluoro-2-propanol to 58.9
× 105 M-1 s-1 in N,N-dimethylacetamide for 5,6-dimethyl-2,3-dihydropyrazine (DMD) and from
5.74 × 105 M-1 s-1 in trifluoroethanol to 159.0 × 105 M-1 s-1 in tributyl phosphate for 5-methyl-
6-phenyl-2,3-dihydropyrazine (MPD). Chemical reaction rate constants, kR, for DMD are similar
to kT in polar solvents such as propylencarbonate, whereas for MPD in this solvent, the contribution
of the chemical channel to the total reaction is about of 4%. Dependence of the total rate constant
on solvent microscopic parameters, R and π*, for DMD can be explained in terms of a reaction
mechanism that involves formation of a perepoxide exciplex. Replacement of the methyl by a phenyl
substituent enhances dihydropyrazine ring reactivity toward singlet oxygen and modifies the
dependence of kT on solvent parameters, specially on the Hildebrand parameter. These results are
explained in terms of an additional reaction path, involving a perepoxide-like exciplex stabilized
by the interaction of the negative charge on the terminal oxygen of the perepoxide with the aromatic
π system.
In tr od u ction
tions in reactions of singlet oxygen with molecules having
carbon-nitrogen double bonds.6-10 The product distribu-
tion dependence on temperature and substituents and
comparison with the behavior of the alkene homologues
indicate that reactions of O2(1∆g) with imino compounds
involve several intermediates, i.e., dioxazetidines by [2
+ 2] concerted cycloaddition to the carbon-nitrogen
double bond,7,8 peroxide ions from electrophilic attack on
iminic carbon,7 and hydroperoxides from rearrangement
of pernitrones in reactions with substituted 2,3-dihydro-
pyrazines.11 The involvement of pernitrone intermediates
in this last reaction is explained through a mechanism
in which excited oxygen interacts with the iminic nitro-
gen lone pair.
Reactions of singlet oxygen, O2(1∆g), with olefins pro-
ceed through several mechanisms, among them ene-
reactions producing hydroperoxide intermediates in mol-
ecules with allylic hydrogens; [4 + 2] cycloadditions
giving relatively stable or labile endoperoxides, depend-
ing on the olefin substitution; and [2 + 2] cycloadditions
that mainly yield dioxetane intermediates.1-4 Extensive
work shows that the main process and its dominant
pathway in these systems is determined by olefin struc-
ture and solvent properties. Furthermore, it has been
pointed out that the solvent has only a minor role in
determining the total rate of interaction of singlet oxygen
with monoolefins or 1,3-dienes.5 Unlike the extensive
studies covering reactions of singlet oxygen with com-
pounds having carbon-carbon double bonds, only very
few data are suitable for rationalizing product distribu-
From these studies it is not possible to decide whether
imino group reactivity toward singlet oxygen is similar
(6) George, M. V.; Bhat, V. Chem. Rev. 1979, 79, 447-478.
(7) Ito, Y.; Kyono, K.; Matsuura, T. Tetrahedron Lett. 1979, 2253-
2256.
(8) Vaidya, V. K. J . Photochem. Photobiol. A: Chem. 1994, 81, 135-
137.
(9) Castro, C.; Dixon, M.; Erden, I.; Ergonec, P.; Keeffe, J .; Sukho-
vistsky, A. J . Org. Chem. 1989, 54, 3732-3738.
(10) Erden, I.; Griffin, A.; Keeffe, J .; Brinck-Kohn, V. Tetrahedron
Lett. 1993, 34, 793-796.
(11) Gollnick, K.; Koegler, S.; Maurer, D. J . Org. Chem. 1992, 57,
229-234.
* Corresponding author. Phone: 56-2-6782877. Fax: 56-2-6782868.
(1) Clennan, E. L. Tetrahedron 1991, 47, 1343-1382.
(2) Frimer, A. Chem. Rev. 1979, 79, 359-387.
(3) Prein, M.; Adam, W. Angew. Chem., Int. Ed. Engl. 1996, 35, 477-
494.
(4) Stratakis, M.; Orfanopoulus, M. Tetrahedron 2000, 56, 1595-
1615.
(5) Lissi, E. A.; Encinas, M. V.; Lemp, E.; Rubio, M. A. Chem. Rev.
1993, 93, 699-723.
10.1021/jo026303o CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/14/2003
J . Org. Chem. 2003, 68, 3009-3016
3009

Lemp et al.
F IGURE 1. Structures of 5,6-dimethyl-2,3-dihydropyrazine
(DMD) and 5-methyl-6-phenyl-2,3-dihydropyrazine (MPD).
to that of homologous alkenes or due to the nitrogen
heteroatom if the behavior departs significantly from the
well-known alkene-singlet oxygen chemistry. In this
work we report on reactions of singlet oxygen with 2,3-
dihydropyrazine derivatives (Figure 1) including different
substituents on the iminic carbon. Our aim is to account
for the reaction mechanism both from kinetics, including
measurements of rate constants for chemical and physical
reaction channels in several solvents, and product char-
acterization. Solvent effects are analyzed in terms of
linear solvation energy relationships (LSER)12,13 and
theoretical linear solvation energy relationships (TLS-
ER).14 These equations give very useful quantitative
evaluations of solvent effects and are powerful tools in
interpreting mechanisms of such processes.15-19
F IGURE 2. Stern-Volmer plot for deactivation of singlet
oxygen by DMD in acetone. (Inset a) Singlet oxygen phospho-
rescence decay at 1270 nm, following dye laser excitation at
532 nm, with Rose Bengal as sensitizer in acetone. (Inset b)
as in part a, but with 13.6 mM DMD.
TABLE 1. Va lu es of kT for Rea ction s of DMD a n d MP D
w ith O2(1∆g) in Differ en t Solven ts
Resu lts
kT/105 M-1 s-1
Tot a l R ea ct ion of DMD a n d MP D w it h Sin glet
Oxygen . The total quenching rate constant (physical and
chemical), kT, for reaction of O2(1∆g) with the 5,6-
disubstituted-2,3-dihydropyrazines DMD and MPD (Fig-
ure 1) in several solvents was obtained from the first-
order decay of O2(1∆g) in the absence (τo-1) and presence
of dihydropyrazine (τ-1) according to eq 1.
solvent
DMD
MPD
1
2
3
4
5
6
7
8
hexafluoro-2-propanol
chloroform
trifluoroethanol
n-hexane
0.26 ( 0.01
2.23 ( 0.13
3.04 ( 0.16
3.60 ( 0.17
3.69 ( 0.19
4.31 ( 0.18
4.48 ( 0.21
4.94 ( 0.22
5.23 ( 0.22
5.34 ( 0.21
5.44 ( 0.19
6.01 ( 0.33
6.08 ( 0.31
6.25 ( 0.41
7.36 ( 0.31
8.60 ( 0.38
9.87 ( 0.41
12.9 ( 0.48
13.6 ( 0.58
13.7 ( 0.43
14.6 ( 0.55
19.3 ( 0.72
20.3 ( 0.89
56.8 ( 2.34
58.9 ( 2.61
12.0 ( 0.06
5.74 ( 0.25
ethanol
22.1 ( 0.99
28.9 ( 1.15
23.3 ( 0.09
n-heptane
benzyl alcohol
n-butanol
τ-1 ) τo + kT[dihydropyrazine]
(1)
-1
9
methanol
26.4 ( 1.15
49.5 ( 1.98
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
benzene
n-pentanol
anisole
63.0 ( 2.29
37.7 ( 1.69
41.9 ( 1.65
37.1 ( 1.29
25.5 ( 1.09
67.3 ( 3.02
50.6 ( 2.85
51.6 ( 2.12
60.4 ( 2.63
60.8 ( 3.00
101.1 ( 4.19
68.1 ( 3.27
159.0 ( 4.38
Triplet decay of the sensitizer employed in most
solvents (TPP) was not affected by addition of substituted
dihydropyrazine, even at concentrations higher than
those used to quench excited oxygen. Linear plots of τ-1
vs [dihydropyrazine] were obtained in all solvents em-
ployed (Figure 2). Intercepts of these plots match closely
with reported20 and measured (in a large number of
independent experiments from our laboratory) singlet
oxygen lifetimes in the same solvent. For example, the
singlet oxygen lifetime in acetone, obtained from data in
Figure 2, was 50 µs, very close to the reported value of
51 µs.20
diethyl ether
benzonitrile
methylene chloride
n-propanol
acetonitrile
tetrahydrofuran
ethyl acetate
acetone
dioxane
N,N-dimethylformamide
propylencarbonate
tributyl phosphate
N,N-dimethylacetamide
Values of kT calculated from slopes of these plots are
given in Table 1 and were independent of the laser pulse
energy (between 2 and 5 mJ ). Possible rapid chemical
changes of samples during illumination or interference
of the O2(1∆g) luminescence with the scattered laser light
and the tail end of the sensitizer fluorescence21 can be
disregarded, since the rate constant measured in some
solvents (data not shown) by using competitive steady-
state methods, such as the inhibition of the autoxidation
rate of rubrene (λmax ) 520 nm)22 or by following the
consumption rate of 9,10-dimethylanthracene in absence
(12) Reichardt, C. Solvents and Solvent Effects in Organic Chemistry,
2th ed.; VCH: Weinheim, 1990.
(13) Abraham, M.; Doherty, R.; Kamlet, M.; Harris, J . M.; Taft, R.
J . Chem. Soc., Perkin Trans 2 1987, 913-920.
(14) Cronce, D. T.; Famini, G. R.; De Soto, J . A.; Wilson, L. Y. J .
Chem. Soc., Perkin Trans. 2 1998, 1293-1301.
(15) Zanocco, A. L.; Gu¨nther, G.; Lemp, E.; De La Fuente, J . R.;
Pizarro, N. Photochem. Photobiol 1998, 68, 487-493.
(16) Zanocco, A. L.; Lemp, E.; Pizarro, N.; De La Fuente, J . R.;
Gu¨nther, G. J . Photochem. Photobiol. 2001, 140, 109-115.
(17) Lemp, E.; Pizarro, N.; Encinas, M. V.; Zanocco, A. L. Phys.
Chem. Chem. Phys. 2001, 3, 5222-5225.
(18) Lissi, E. A.; Lemp, E.; Zanocco, A. L. Singlet Oxygen Reactions.
Solvent and Compartmentalization Effects. Molecular and Supramo-
lecular Photochemistry: Understanding and Manipulating Excited-
States Processes; Ramamurthy V., Schanze, K. S., Ed.; Marcel Dekker
Inc.: New York, 2001; Vol 8, Chapter 3.
(21) Michaeli, A.; Feitelson, J . Photochem. Photobiol. 1994, 59, 284-
288.
(19) Lemp, E.; Zanocco, A. L.; Lissi, E. Curr. Org. Chem. Submitted.
(20) Wilkinson, F. J . Phys. Chem. Ref. Data 1995, 24, 663-1021.
(22) Carlson, D. J .; Mendenhall, E. D.; Suprunchuk, T.; Wiles, D.
M. J . Am. Chem. Soc. 1972, 94, 8960-8963.
3010 J . Org. Chem., Vol. 68, No. 8, 2003

Solvent Effect on Sensitized Photooxygenation
and presence of dihydropyrazine derivatives,23 afforded
the same value as that obtained by time-resolved experi-
ments.
In eq 3, the bulk/steric term is described by the
Hildebrand parameter, FH. The parameter π1 corresponds
to the index of polarizability, accounts for the ease of
moving or polarizing the electron cloud, and is obtained
from the ratio between polarizability volume and molec-
ular volume. The hydrogen bond acceptor basicity (HBA)
involves covalent, ꢀb, and electrostatic, q-, terms. Simi-
larly, hydrogen-bond donor acidity (HBD) includes cova-
lent, ꢀa, and electrostatic, q+, terms.29,31
Data in Table 1 show that the total quenching rate
constant for reaction of DMD with singlet oxygen in-
creases by more than 2 orders of magnitude when the
solvent is changed from hexafluoro-2-propanol to N,N-
dimethylacetamide. Similarly, MPD rate constants in-
crease by about a factor of 30 on changing from trifluo-
roethanol to tributyl phosphate. Also, we found for all
solvents that the total rate constants for MPD are near
to 1 order of magnitude larger than those for DMD. As
expected, these results show that dihydropyrazines be-
have similarly to related olefins and reactivities toward
singlet oxygen depend on both dihydropyrazine structure
and solvent properties. However, solvent effects are
larger than those for diolefins5 and cannot be associated
only with changes in macroscopic properties such as
dielectric constant, suggesting the existence of specific
solute-solvent interactions. We have previously demon-
strated that in these cases it is useful to use linear
solvent free-energy relationships to correlate experimen-
tal rate constants with solvent properties.15-19
The coefficients of the LSER and TLSER equations (eqs
2 and 3) obtained by multilinear correlation analysis for
dependence of kT on solvent parameters are given in
Table 2 and are based on purely statistical criteria.
Current criteria used for derivation of the correlation
equations include the following: (a) Descriptor coef-
ficients accepted in the correlation equation are those
that have a significance level g0.95 [large t-statistic,
P(two-tail) < 0.05]. (b) The VIF statigraph,32 a measure
of parameter orthogonality that indicates independence
of the parameters, is defined by VIF ) 1/(1 - R2), where
R is the correlation coefficient for that particular param-
eter in terms of the other parameters. For a given
parameter, the closer VIF is to unity, the less cross-
correlation there is with the other parameters. When VIF
is too large, the least significant variable is eliminated.
We accept a low variable collinearity (VIF < 5). (c) The
correlation coefficient, R, is equal to or higher than 0.90
(variance, R2 > 0.80). (d) The sample size is as large as
possible; N must be at least three times the number of
independent parameters included in the correlation
equation. In summary, the sample size, N, the product
correlation coefficient, R, the standard deviation, SD, and
the Fisher index of equation reliability, F, indicate the
quality of the overall correlation equation. The reliability
of each term is indicated by the t-statistic (t-stat) and
the variance inflation factor (VIF). Good quality is
indicated by large N, F, and t-stat values; small SD
values; and R and VIF close to 1.
Results in Table 2 show that not all descriptors are
important. Descriptor coefficients accepted in the cor-
relation equation were those that have a significance
level g0.95. For this reason FH and π* parameters were
not included in the LSER correlation for DMD and MPD,
respectively. According to the LSER coefficients in Table
2, kT values for DMD increase in solvents with the largest
capacities to stabilize charges and dipoles and decrease
in strong HBD solvents. Furthermore, for MPD rate
constants increase in HBA solvents with high cohesive
energy and decrease in strong HBD solvents. Similarly,
in TLSER equations, only q- and q+ were included for
DMD and ꢀb, q+, and FH for MPD.
Therefore, to obtain insights on solvent effects on
interactions of singlet oxygen with dihydropyrazines, we
analyze the quenching rate constant dependence on
microscopic solvent parameters by using the semiempiri-
cal solvatochromic equation (LSER) of Kamlet et al. (eq
2).12,24
2
log k ) log ko + sπ* + dδ + aR + bâ + hFH (2)
where π* accounts for polarizabilities and dipolarities of
solvents;25-27 δ is a correction term for polarizability; R
is related to the hydrogen-bond donor solvent ability; â
indicates solvent capability as a hydrogen-bond acceptor;
and FH is the Hildebrand parameter, which corresponds
to the square root of solvent cohesive density and is a
measure of disruption of solvent-solvent interactions in
creating a cavity.12,28 Also, we analyzed the dependence
of kT on solvent by using a theoretical set of correlation
parameters determined solely from computational meth-
ods.14,29,30 The theoretical linear solvation relationship
(TLSER) descriptors have been developed to give optimal
correlation with LSER descriptors.14,29 The generalized
TLSER equation proposed by Famini et al.14,29 (eq 3) can
be used to analyze dependence of reaction rates on
solvent properties.
2
log k ) log ko + sπ1 + bꢀb + cq- + dꢀa + eq+ + hFH
Results of TLSER analyses accord with those obtained
with LSER equations, showing that for DMD kT increases
in HBA solvents and decreases in HBD solvents, and for
MPD the rate constant increases in HBA solvents with
high cohesive energy and decreases in HBD solvents.
Ch em ica l Rea ction of O2(1∆g) w ith Dih yd r op yr a -
zin es. Irradiation of aerated solutions of DMD or MPD
in the presence of TPP or RB, at the wavelength for which
(3)
(23) Zanocco, A. L.; Lemp, E.; Gu¨nther, G. J . Chem. Soc., Perkin
Trans. 2 1997, 1299-1302.
(24) Kamlet, M. J .; Abboud, J . L. M.; Abraham, M. H.; Taft, R. W.
J . Org. Chem. 1983, 48, 2877-2887.
(25) Marcus, Y. Chem. Soc. Rev. 1993, 22, 409-416.
(26) Barton, A. F. M. Chem. Rev. 1975, 75, 731-752.
(27) Abraham, M. H.; Doherty, R. M.; Kamlet, M. J .; Harris, J . M.;
Taft, R. W. J . Chem. Soc., Perkin Trans. 2. 1987, 913-920.
(28) Kamlet, M. J .; Carr, P. W.; Taft, R. W.; Abraham. M. H. J . Am.
Chem. Soc. 1981, 103, 6062-6066.
(31) Famini, G. R.; Wilson, L. Y. J . Chem. Soc., Perkin Trans. 2 1994,
1641-1650.
(32) Belsley, D. A.; Kuh, E.; Welsch R. E. Regression Diagnostics.
Identifying Influential Data and Sources of Collinearity; Wiley &
Sons: New York, 1980.
(29) Famini, G. R.; Penski, C. A.; Wilson, L. Y. J . Phys. Org. Chem.
1992, 5, 395-408.
(30) Cronce, D. T.; Famini, G. R.; De Soto, J . A.; Wilson, L. Y. http://
www.rsc.org/suppdata/perkin2/1998/1293.
J . Org. Chem, Vol. 68, No. 8, 2003 3011

Lemp et al.
TABLE 2. LSER a n d TLSER Cor r ela tion Equ a tion s for th e Rea ction of Sin glet Oxygen w ith DMD a n d MP D
2
log k ) log ko + aR + bâ + sπ* + dδ + hFH
log ko
a
b
s
d
h
DMD
coeff
(
t-stat
P(two-tail)
VIF
5.577
0.081
68.802
<0.0001
-0.602
0.054
-11.053
<0.0001
1.006
0.439
0.125
3.507
0.0024
1.572
0.629
0.140
4.500
0.0002
1.536
-0.406
0.102
-3.964
0.0008
1.808
N ) 24, R ) 0.956, SD ) 0.144, F ) 50.029
MPD
coeff
(
t-stat
P(two-tail)
VIF
6.395
0.056
114.69
<0.0001
-0.692
0.050
-13.855
<0.0001
1.613
0.203
0.084
2.430
0.0291
1.379
0.003
0.001
4.561
0.0004
2.088
N ) 18, R ) 0.970, SD ) 0.075, F ) 74.475
2
log k ) log ko + sπ1 + bꢀb + cq- + dꢀa + eq+ + hFH
log ko
s
b
c
d
e
h
DMD
coeff
(
t-stat
P(two-tail)
VIF
5.701
1.862
0.277
6.729
<0.0001
1.087
-3.849
0.542
-7.102
<0.0001
1.087
0.078
72.630
<0.0001
-
N ) 20, R ) 0.900, SD ) 0.159, F ) 36.397
MPD
12.602
2.185
coeff
4.677
0.312
-5.148
0.538
0.005
0.001
t-stat.
P(two-tail)
VIF
14.993
<0.0001
5.767
<0.0001
1.009
-9.568
<0.0001
2.198
5.830
<0.0001
2.212
N ) 18, R ) 0.948, SD ) 0.104, F ) 41.143
TABLE 3. Ch em ica l Rea ction Ra te Con sta n ts, kR, for
Rea ction s of DMD a n d MP D w ith O2(1∆g) in Differ en t
Solven ts
kR/105 M-1 s-1
solvent
n-hexane
DMD
MPD
1
2
3
4
5
6
7
8
0.06 ( 0.01
0.09 ( 0.01
0.57 ( 0.02
1.70 ( 0.01
3.35 ( 0.02
5.12 ( 0.03
5.26 ( 0.04
21.0 ( 0.09
n-heptane
diethyl ether
tetrahydrofuran
ethanol
benzene
acetonitrile
propylencarbonate
2.18 ( 0.02
2.25 ( 0.02
1.53 ( 0.01
only the sensitizer absorbs, decreases the concentration
of the dihydropyrazine derivative, and the reaction rate,
r, follows a first-order kinetics, eq 4.
F IGURE 3. Dependence of the chemical reaction rate con-
stant, kR, for reaction of DMD with singlet oxygen on the
solvatochromic parameter π*.
r ) kR [1O2][dihydropyrazine] ) kobs
[dihydropyrazine] (4)
Rate constants for chemical O2(1∆g) quenching by
dihydropyrazine derivatives (kR) in various solvents,
obtained from slopes of first-order plots employing 1,3-
diphenylisobenzofuran (DPBF) or 9,10-dimethylanthracene
(DMA) as actinometers, are in Table 3.
Eventual quenching of excited states of sensitizers by
dihydropyrazine can be disregarded, since singlet or
triplet deactivation was not observed under our experi-
mental conditions. Data in Table 3 show that DMD
chemical rate constants are dependent on solvent proper-
ties. Thus, in n-hexane, the chemical channel of inter-
mediate decomposition is negligible, whereas in more
polar solvents, such as ethanol and propylencarbonate,
reactive quenching corresponds to the only path for O2-
(1∆g) deactivation. Solvent effects on chemical reaction
rate constants cannot be examined in terms of micro-
scopic solvent parameters employing empirical or theo-
retical linear solvation relationships, given that kR was
3012 J . Org. Chem., Vol. 68, No. 8, 2003

Solvent Effect on Sensitized Photooxygenation
F IGURE 4. (a) GC-MS chromatogram of 1 mM MPD in acetonitrile after 6 h of irradiation in the presence of TPP; (b) EI+ mass
spectrum of MPD; (c) EI+ mass spectrum of compound with retention time 6.33 min; (d) EI+ mass spectrum of compound with
retention time 5.56 min.
only measured in eight solvents. However, kR values have
an approximate dependence on the solvatochromic pa-
rameter π*, as shown in Figure 3, indicating that in
solvents with large π* values contributions of chemical
reactions to the total quenching increase. For MPD, the
contribution of the chemical channel to the total quench-
ing is minor (about of 4%) in the three solvents employed
in measuring kR. Also, for this compound kR values were
practically independent of the reaction media.
According to Gollnick,11 dye-sensitized photooxygen-
ations of DMP and MPD yield the corresponding 1-iso-
cyano-2-(acylamino)ethane and formaldehyde as main
J . Org. Chem, Vol. 68, No. 8, 2003 3013

Lemp et al.
SCHEME 1
reaction products. We did not attempt to isolate reaction
products for spectroscopic characterization, although
evidence of product distribution was obtained by GC-
MS analysis. When 0.001 M MPD in acetonitrile was
irradiated for 6 h in the presence of TPP, the chromato-
gram shown in Figure 4a was obtained with the mass
spectrometer in the electron impact (EI+) operation
mode. There were only three major peaks in the chro-
matogram. Unreacted MPD is the main one with a
retention time of 5.02 min. Figure 4b shows that the EI+
mass spectrum is that of MPD. The positive chemical
ionization (CI+) (not included) and EI+ mass spectra
corresponding to peaks at retention times of 6.33 and 5.56
min indicate that 1-isocyano-2-(benzoylamino)ethane and
1-isocyano-4-phenyl-4-carboxaldehyde-3-aza-3-butene are
the probable main products of the photooxidation of MPD.
Figure 4c,d shows mass spectra ionization patterns and
corresponding structures. Additionally, in a separate
experiment in benzene, we confirmed formation of form-
aldehyde as the photooxidation product by comparison
of a signal at retention time of 0.39 min, with that
obtained for formaldehyde in benzene.
tion but not to the extent observed with tertiary amines.5,16
This result could be explained if nitrogen is not the
reactive center but is close to it. Also, for the same
compound we found that the coefficient associated with
the π* parameter is larger than that for reactions of O2-
(1∆g) with amines. In addition, the TLSER treatment
shows that HBA solvents assist reaction between singlet
oxygen and DMD, whereas HBD inhibits them. Influ-
ences of both HBA and HBD solvents are mainly elec-
trostatic, because only the parameters q- and q+ are
included in the correlation equation. This result can be
understood by considering that π1 is not included in the
correlation and it models the solvent’s ability to stabilize
dipole-induced dipole and induced dipole-induced dipole
(dispersive) interactions. No adequate TLSER model for
dipolarity itself has been found. The dependence of total
reaction rate on the q+ parameter could be explained if
an appreciable charge separation is developed in the N-C
double bond. Stabilizing interactions with HBD solvents
can sterically block access of singlet oxygen to a reactive
bond. Furthermore, if the positive charge on the perep-
oxide is delocalized throughout the cyclic system formed
after singlet oxygen addition, no formal bonding interac-
tions with HBA solvents can be expected. Conversely,
there should be electrostatic stabilization of an exciplex
by such solvents.
Solvent dependence on kR for reaction of DMD with
O2(1∆g) is more significant than that observed for kT. The
chemical rate constant increases by more than 2 orders
of magnitude when the solvent is changed from nonpolar
(e.g. hexane) to polar (e.g., propylencarbonate). Then, in
nonpolar solvents, contribution of the chemical reaction
can be considered to be negligible, whereas in polar ones
their contribution is the main deactivation channel. Also,
it is clear from data in Table 3 and Figure 3 that solvents
with large π* values increase the contribution of the
chemical reaction relative to physical quenching. Solvent
dependence of kR can be explained if intermediates along
the reaction coordinate have more localized charge than
the initial complex. This result is compatible with
involvement of ion pairs in controlling the final product
distribution, as depicted in Scheme 2.
Analogous reaction products were observed in the
photooxidation of DMD. Formaldehyde, 1-isocyano-2-
(acylamino)ethane, and 1-isocyano-4-carboxaldehyde-3-
aza-3-pentene were the main reaction products.
Discu ssion
The quenching of singlet oxygen by cyclic R-diimines
has been discussed by Gollnick et al.11 Product distribu-
tion was explained in terms of formation of an hydrop-
eroxide intermediate resulting from the rearrangement
of a primary reaction intermediate, a “pernitrone” or
“nitrone oxide”. Other possible reaction pathways, involv-
ing interaction of singlet oxygen with C-5 of the dihy-
dropyrazine or addition of O2(1∆g) to the C-N double
bond giving peroxaziridines, were disregarded on the
basis of final product distribution.
Kinetic results obtained in this work indicate that kT
values for singlet oxygen quenching by dihydropyrazines
are highly dependent on solvent properties (Table 1). A
meaningful interpretation of kT solvent dependence was
obtained by using LSER and TLSER solvatochromic
equations. The LSER correlations listed in Table 2
indicate that singlet oxygen reactions with DMD or MPD
have different dependence on solvent properties. For
DMD, the reaction rate increases in solvents with higher
capabilities to stabilize charges and dipoles and decreases
in strong hydrogen-bond-donor solvents.
Data are compatible with an exciplex formation with
considerable charge separation. This result could be in
agreement with formation of both perepoxide and zwit-
terionic intermediates as proposed by Gollnick et al.11
(Scheme 1).
However, there are several results supporting forma-
tion of a perepoxide as the primary intermediate arising
from the interaction between singlet oxygen and DMD.
In the LSER equation the coefficient associated with the
R parameter is smaller than that generally observed for
electrophilic attack of singlet oxygen on a nitrogen lone
pair in amino compounds, implying that steric hindrance,
due to hydrogen-bonding interactions between strong
HBD solvents and the nitrogen atom, inhibits the reac-
Values of kT for reaction of MPD with singlet oxygen
are larger than for DMD, and solvent effects are char-
acterized by dependence of kT on the Hildebrand param-
eter, as found in both LSER and TLSER analyses. The
dependence of the rate constant on the Hildebrand
parameter has been ascribed to O2(1∆g) reactions that
involve a concerted or partially concerted cycloaddition
of singlet oxygen to an activated diene.33 In these reac-
(33) Aubry, J . M.; Mandard-Cazin, B.; Rougee M.; Bensasson, R. J .
Am. Chem. Soc. 1995, 117, 9159-9164.
3014 J . Org. Chem., Vol. 68, No. 8, 2003

Solvent Effect on Sensitized Photooxygenation
SCHEME 2
tude larger than for its dimethyl-substituted analogues.
These data are consistent with the mechanism of sensi-
tized photooxidation of MPD in Scheme 3. The increased
reactivity of this compound toward singlet oxygen, rela-
tive to the dimethyl-substituted R-diimine, implies that
the main path of reaction involves the interaction of O2-
(1∆g) with the phenyl-substituted N-C double bond, to
give a perepoxide-like exciplex that exclusively evolves
by intersystem crossing due to there being no accessible
R-hydrogen. Values of kR relative to kT are greatly
reduced when comparing MPD results with DMD results
in polar solvents. For DMD, kR is very close to kT, due to
singlet oxygen interaction with the methyl-substituted
N-C double bond to generate a perepoxide through path
b, allowing possible hydrogen abstraction from a methyl
substituent.
SCHEME 3
In conclusion, 5,6-disubstituted cyclic R-diimines are
moderate to good singlet oxygen quenchers, being most
effective in polar solvents. A reaction mechanism involv-
ing a perepoxide intermediate that evolves to a hydro-
peroxide from which products arise seems to be the main
reaction path. Replacement of a methyl by a phenyl
substituent opens an additional reaction path involving
a perepoxide-like exciplex in which a stabilizing interac-
tion of a negative charge on the free oxygen of a
perepoxide with an aromatic π system contributes to the
increased singlet oxygen quenching ability of cyclic
R-diimines.
tions, dependence on the Hildebrand parameter is ex-
plained in terms of formation of an encounter complex
of smaller molecular volume than the parent compounds.
These reactions are also dependent on the dipolarity-
polarizability parameter. Furthermore, examination of
LSER and TLSER equations for MPD shows that the
reaction is also assisted by HBA solvents and inhibited
by HBD solvents. These results can be explained if phenyl
substitution opens a further reactive channel in which
the perepoxide is stabilized by interaction with the π
system of a phenyl group such as proposed by Orfanopo-
ulos et al.4 to explain the solvent-dependent syn selectiv-
ity in reactions of O2(1∆g) with â,â-dimethylstyrene. A
reaction mechanism compatible with our results is
depicted in Scheme 3.
Reaction path a leads only to physical quenching
through intersystem crossing to produce oxygen and the
parent R-diimine. The geometry of the exciplex hinders
the intramolecular hydrogen abstraction processes that
stabilize it. The perepoxide shown in path a of Scheme 3
allows explanation of the observed solvent dependence.
Decrease of the total reaction rate constants in HBD
solvents can be understood in terms of interactions with
the reactive center (the phenyl-substituted N-C double
bond). Furthermore, an increase in sensitivity to HBA
solvents may be due to electrostatic-stabilizing interac-
tions with a positive charge on the complex. In addition,
the dependence of kT on the Hildebrand parameter is
easily understand in terms of a phenyl group-perepoxide
interaction. This interaction would be favorable in sol-
vents with high cohesive energy, because interaction of
negatively charged oxygen with the neighboring phenyl
disrupts solvent-phenyl group interactions in the sub-
strate. Furthermore, this hypothesis permits us to ex-
plain the low values of the chemical reaction constant
measured for MPD even in polar solvents. As was
mentioned above, the only reaction products detected
were formaldehyde, 1-isocyano-2-(benzoylamino)ethane,
and 1-isocyano-4-phenyl-4-carboxaldehyde-3-aza-3-bu-
tane, arising from singlet oxygen attack on a methyl-
substituted N-C double bond, and the total reaction rate
constant for this compound is at least 1 order of magni-
Exp er im en ta l Section
5,10,15,20-Tetraphenyl-21H,23H-porphine (TPP), 9,10-di-
methylanthracene (DMA), and 1,3-diphenylisobenzofurane
(DPBF) were used without further purification. Rose Bengal
(RB) was recrystallized twice from ethanol prior to use. All
solvents were of spectroscopic or HPLC grade.
5,6-Dimethyl-2,3-dihydropyrazine (DMD) and 5-methyl-6-
phenyl-2,3-dihydropyrazine (MPD) were synthesized as de-
scribed;34 typically a solution of 1-phenyl-2-propanedione (5
g) in ethyl ether (10 mL) was added to ethylendiamine (2 g)
in 10 mL of ethyl ether maintained at 0 °C. The mixture was
refluxed for 30 min and the ethereal yellow solution was dried
over sodium sulfate and the solvent removed in a vacuum. The
remaining yellow oil was cooled in a freezer during several
hours, producing a solid from which pure 5-methyl-6-phenyl-
2,3-dihydropyrazine was obtained as yellow needles by recrys-
tallization from ethyl ether-hexane in 50% yield.
UV-vis absorption spectra and steady-state competitive
kinetic experiments were performed. A GC-MS system with
a Hewlett-Packard Ultra-2 capillary column (25 m) was used
to obtain electron impact and chemical ionization mass spectra.
Chemical reaction rate constants were determined in several
selected solvents using a 10 mL double-wall cell, light-
protected by black paint. A centered window allowed irradia-
tion with light of a given wavelength by using cutoff filters.
Circulating water maintained the cell temperature at 22 (
0.5 °C. Sensitizer irradiation (RB or TPP) was performed with
a visible, 200 W, Par lamp. A gas chromatograph equipped
with a NPD detector and a Hewlett-Packard Ultra-2 capillary
column was used to monitor substrate consumption. DMA and
DPBF were used as actinometers. Fresh DPBF solutions
prepared in a dark room and appropriate cutoff filters were
used. Autoxidation of this compound, followed by UV-vis
spectrophotometry, was <1% under our experimental condi-
tions.
(34) Flamet, I. F.; Stoll, M. Helv. Chim. Acta 1967, 50, 1754-1758.
J . Org. Chem, Vol. 68, No. 8, 2003 3015

Lemp et al.
Time-resolved luminescence measurements were carried out
in 1 cm path fluorescence cells. TPP or RB were excited by
the second harmonic (532 nm, nominal power 9 mJ per pulse)
of a 6-ns light pulse of a Q-Switched Nd:YAG laser. A liquid
nitrogen cooled germanium photodiode detector with a built-
in preamplifier was used to detect infrared radiation from the
cell. The detector was at a right-angle to the cell. An interfer-
ence filter (1270 nm) and a cutoff filter (995 nm) were the only
elements between the cell face and the diode cover plate. The
preamplifier output was fed into the 1 MΩ input of a digitizing
oscilloscope. Computerized experiment control, data acquisi-
tion, and analysis were performed with a LabView-based
software developed in our laboratory.
Laser flash photolyses were performed with a Q-switched
Nd:YAG laser (532 nm, ca. 35 mJ /pulse). A 150-W Xe lamp
mounted in a lamp housing system was employed as monitor-
ing light beam. The lamp beam was passed through a water
filter before impinging on the entrance of the cell holder. An
electronic shutter, controlled by a shutter driver/timer, was
placed between the water filter and the cell holder. The shutter
was triggered by the Q-switch from the laser. Two lenses and
slits were used to collimate and focus the monitoring light to
the cell holder and to the entrance slit of the monochromator.
A photomultiplier detector mounted in a homemade housing
was fitted to the monochromator exit slit port. PMT signals
oscilloscope. The signals can be stored and averaged in the
same scope at the repetition rate of the laser pulse (10 Hz) or
can be fed to a personal computer equipped with home-
designed software for data acquisition and treatment.
The determination of the total rate constants by steady-state
competitive techniques in nonpolar solvents was done using
the autoxidation rate of rubrene (λmax ) 520 nm).22 The rate
constants were obtained from the initial rate of rubrene
consumption. In more polar solvents, TPP was employed as
sensitizer (λmax ) 414 nm), and kT values were determined by
following the consumption of 9,10-dimethylanthracene upon
addition of the dihydropyrazine derivative.23 The irradiation
was performed with a visible Par lamp (150 W) using a cutoff
filter at 400 nm.
Equation coefficients and statistical parameters of LSER
and TLSER correlations were obtained by multilinear correla-
tion analysis with STAT VIEW 5.0 (SAS Institute Inc.). Results
agreed with the t-statistic of descriptors.
Ack n ow led gm en t. Financial support from FOND-
ECYT (grant 2990096), Postgraduate Department, Uni-
versity of Chile (grants PG/037/98 and PG/51/99), and
CEPEDEQ-Faculty of Chemical and Pharmaceutical
Sciences, University of Chile is gratefully acknowledged.
were monitored with
a 500-MHz digital oscilloscope. An
external PTI optical beam divider was employed to trigger the
J O026303O
3016 J . Org. Chem., Vol. 68, No. 8, 2003