EPA-650/1-74-010
DECEMBER 1973
Environmental Health Effects Research Series
-------
EPA-650/1-74-010
REACTIVITY
OF POLYNUCLEAR
AROMATIC HYDROCARBONS
WITH 02 AND NO
IN THE PRESENCE OF LIGHT
by
Nicholas E. Geacintov
Chemistry Department
New York University
New York, New York 10003
Grant No. R801393
ROAP No. 21BLD
Program Element No. 1AA005
EPA Project Officer: David L. Coffin, D. V. M.
Human Studies Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
December 1973
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
ABSTRACT
The reactivity of 20 different aroiriatii- hydrocarbons adsorbed on solid
polystyrene fluffs with oxygen and nitric oxide in the presence of light has
been studied. The reaction conditions simulated those encountered in polluted
atmospheres. Among the compounds studied were anthracene, pyrene, naphthalene,
chrysene, benz(a)anthracene and coronene. The photoexcited triplet and
singlet states of the aromatic hydrocarbons react predominantly via the quench-
ing of the fluorescence and phosphorescence by the paramagnetic QS. and NO gases.
The quenching of the triplets by oxygen occurs via the formation of an inter-
mediate collision complex in which electrophilic exchange type interactions
appear to be important. The probability of quenching per collisional encounter
and the formation of singlet oxygen does not exceed 0.01 - 0.10 and depends on
spin selection rules, the triplet energy, and the electron density (in the case
of the monomethyl derivatives of benz(a)anthracene). In the case of NO
quenching of the triplets this probability is much lower, in the range of
0.0005 - 0.005 and appears to be a charge-transfer process for molecules with
a high triplet energy. The quenching of the singlet excited states by Q% and
NO is much more efficient than the quenching of the triplets and has a proba-
bility in the range of ~ O.JO - 1.0 and is not necessarily diffusion controlled.
The most important contribution of the photoexcited aromatic hydrocarbons (per
photon absorbed) to the photochemistry of atmospheres containing QZ and NO
appears to be the generation of singlet oxygen, since photochemical degradation
of the compounds studied was negligible compared to quenching. The quenching
probability of the triplets of the monomethyl derivatives of benz(a)anthracene
by oxygen is compared to their carcinogenic and photodynamic activities.
While the correlation with the carcinogenic activity is unclear, the relatively
low photodynamic activity of the 7 and 12 monomethyl derivatives can be explained
in terms of their low triplet lifetimes and lower probability of singlet oxygen
formation per collisional quenching encounter.
iii
-------
TABLE OF CONTENTS
PUBLICATIONS RESULTING FROM THIS GRANT — 1
RESULTS PRESENTED AT MEETINGS OR CONFERENCES- — 1
INTRODUCTION — 2
OBJECTIVES 2
SCOPE OF THE RESEARCH PERFORMED 3
SIGNIFICANCE 3
SUMMARY OF RESULTS - — - 3
REPORT ON WORK NOT PUBLISHED - 8
Section I. Quenching of Polynuclear Aromatic Hydrocarbon
Triplet States by Nitric Oxide 9
Section II. Quenching of Singlet Excited States of Polynuclear
Aromatic Hydrocarbons by Oxygen and Nitric Oxide-- 26
REFERENCES - --- 32
-------
-1-
PUBLICATIONS RESULTING FROM THIS GRANT
1. R. Benson, "Interaction of Excited States of Polynuclear Aromatic
Hydrocarbons with Oxygen and Nitric Oxide", Ph.D. Thesis, New York
University, October 1973-
2. "Interaction of Triplets of Aromatic Hydrocarbons with Oxygen and
Nitric Oxide", N.E. Geacintov, R. Benson and S. Pomeranz, Chem. Phys.
Lett. 1J_, 280 (1972).
3. "The Quenching of Excited Triplets of Aromatic Hydrocarbons by Molecular
Oxygen", R. Benson and N.E. Geacintov, J. Chem. Phys. 59, ^28 (1973)-
U." "Deuterium Effect on the Quenching of Photoexcited Aromatic Hydrocarbon
Triplets by Oxygen", R. Benson and N.E. Geacintov, J. Chem. Physics,
60, 3251 (197U).
5. In preparation (December 1975) "The Quenching of the Excited States of
Benz(a)anthracene by Oxygen and Nitric Oxide", by N.E. Geacintov, R.
Benson and J. Khosrofian, to be submitted to Chemical Physics Letters.
RESULTS PRESENTED AT MEETINGS OR CONFERENCES
1. International Conference on Luminescence, Leningrad, August 1972.
2. American Physical Society Meeting, New York, January 1975-
-------
-2-
IKTRODUCTION
This report summarizes the work performed under Grant Numbers AP ES
01 395-01, APO 1595-02, HEW - Public Health Service and later referred to as
Grant 80193 from the Environmental Protection Agency. The period of support
was about two years, from 06/01/71 to 05/31/72, and from 07/01/72 to 08/31/73.
The findings are described in full in the Ph.D. thesis of Robert Benson who
graduated in October 1973 and who was a graduate student supported by the EPA.
The thesis is available from University Microfilms, Ann Arbor, Michigan.Q) In
this report the main results of the research are summarized. The work is
described in greater detail in the published papers.(2,3,4,5) Also, the results
not yet written up as a paper are included in this report starting on page 8.
The summary gives an overall view of the results obtained in this study,
while the published papers, preprints, draft of the forthcoming paper and
R. Benson's thesis contain all the details.
OBJECTIVES
A detailed understanding of the physico-chemical properties and reactivities
of polynuclear aromatic hydrocarbon carcinogens and allied non-carcinogens in
simulated polluted atmospheres was sought. Of particular interest was the
interaction of the aromatic hydrocarbons with components of polluted air such
as oxygen, nitric oxide (NO), nitrogen dioxide (NOa), and sulfur dioxide (SOa)
in the presence of light. The stability of these compounds and the eventual
degradation products were to be assessed.
The long-range objective was to learn if there are any specific physical
properties and chemical reactivities of polynuclear aromatic carcinogens which
distinguish them from structurally similar non-carcinogens. The monomethyl
derivatives of benz(a)anthracene provide an interesting system of carcinogenic
molecules whose activity strongly depends on the position of the methyl group
on the aromatic nucleus and their reactivity was studied from this point of
view.
-------
SCOPE OF THE RESEARCH PERFORMED
The originally proposed grant period of three years was reduced to two
years in 1972. Consequently not all of the objectives were achieved. In
particular the study of the reactivities of the polycyclic aromatic hydro-
carbons with NOa and SOg were not performed.
The reactivity of the photoexcited states of 15 different aromatic
hydrocarbons and five perdeuterated analogs with Oa and NO were studied.
This series included benz(a)anthracene and eight of its monomethyl derivatives.
An effort was made to understand these reactivities from a theoretical point
of view and to relate them to structural parameters of the aromatic hydro-
carbons .
A technique was developed to study the reactivity of polycyclic aromatic
hydrocarbons adsorbed on solid organic particles with various gases in the
presence of light.
SIGNIFICANCE
Polycyclic aromatic carcinogens are present in urban polluted atmospheres
and offer a potentially serious health hazard. This study has contributed to
an understanding of the physico-chemical processes which occur when these
polycyclic aromatic hydrocarbons adsorbed on organic particulates are irradiated
in the presence of nitric oxide and oxygen. Neither of these components of
polluted atmospheres leads to a significant quantum yield of degradation of- the
polycyclic aromatic hydrocarbons in the presence of light. However, the aromatic
hydrocarbons act as efficient photocatalysts for the formation of the highly
reactive singlet oxygen molecule (quantum efficiency per triplet photoexcited
state formed^ 0.01 - 0.10) which is lethal to many living organisms.
SUMMARY OF RESULTS
1. The photoexcited triplet and singlet excited states of aromatic hydrocarbons
are much more reactive than the ground state. A method was devised to study
the reactivity of these states with gaseous oxygen and nitric oxide. It involves
the preparation of fluff-like high surface/weight ratio organic particulates
onto which the polycyclic aromatic hydrocarbons are adsorbed. The most suitable
matrix was polystyrene and the best methods of sample preparation were determined.
-------
-k-
2. The following aromatic hydrocarbons were studied (the structure is depicted
on the following page): naphthalene, anthracene, chrysene, a,h-dibenzanthracene,
pyrene, coronene, benz(a)anthracene and eight raonomethyl derivatives of the
latter (the 2,3,^,6,7,8,11 and 12 derivatives). Also perdeuterated anthracene,
benz(a)anthracene, chrysene, pyrene and naphthalene were studied.
3- The reactivity of the singlet and triplet photoexcited states of these
aromatic hydrocarbons with gaseous oxygen and nitric oxide was studied by
observing the quenching of the fluorescence and phosphorescence emissions of
the singlet and triplet excited states respectively. The singlet lifetime is
in the nanosecond range and was measured as a function of Oa and NO pressure
by the photon counting technique. The triplet lifetime was in the range of
10 msec to 15 sec and was measured by signal averaging techniques as a function
of Oa and NO gas pressure.
k. Both QS and NO give rise to radiationless transitions within the intermediate
aromatic hydrocarbon - Og (or NO) collision complex which leads to a shortening
of both the singlet and triplet lifetimes.
5. The probability of quenching by 02 per collisional encounter is in the range
of 0.01 - 0.10 for triplets and ~ 0.3 - 0.7 for singlet excited states. For
triplets the probability is decreased upon perdeuteration of the aromatic com-
pounds. The results are consistent with a mechanism in which excited states of
oxygen (singlet oxygen) are formed, which is in accord with the findings of other
workers. Since the efficiency of formation by light of triplets exceeds 30$,
these compounds appear to be efficient photocatalysts for the formation of singlet
oxygen in polluted atmospheres.
Singlet oxygen does not appear to be formed upon quenching of the singlets.
The quenching of the excited states by molecular oxygen was entirely
reversible and the photochemical oxidation of the aromatic hydrocarbons is of
minor importance in comparison to quenching of the singlet and triplet excited
states.
6. The probability of quenching of triplets by Og and NO depends on spin
selection rules, the triplet energy level (it decreases as the triplet energy
increases), and to a lesser extent on methyl substituents and their location
on the benz(a)anthracene nucleus. The quenching probability is due to an
electrophilic interaction between ©2 and the aromatic molecule. If a methyl
group is located at a site of high electron density, the 0% quenching is less
-------
-5-
Molecular Structures of Pol/nuclear Aromatic Hydrocarbons Studied
CO CCO
Naphthalene
Anthracene
Chrysene
a,h-Dibenzanthracene
Pyrene
Coronene
11
Benz(a)anthracene
-------
-6-
probable than when the methyl group is located at a site of low electron density.
With NO quenching the site of the methyl group appears to be unimportant. This
may be due to a charge-transfer quenching mechanism in the case of NO quenching,
while in the case of Oa quenching it appears to be mostly an exchange mechanism
with only some charge-transfer character.
T. The probability of quenching of triplets by NO per collisional encounter
is in the range of ~ 0.0005 - 0.005, thus much less efficient than in the case
of Oa quenching. The singlets on the other hand, with a probability of ~ 0.8 -
1.0 per encounter are somewhat more efficiently quenched by NO than by Oa« A
charge-transfer mechanism is the most likely mechanism.
8. In the case of the quenching of triplets by NO, the fate of the excitation
energy is much less clear than in the case of oxygen. In the case of oxygen
part of the triplet energy is transferred to oxygen, resulting in the formation
of the reactive singlet oxygen molecule. Nitric oxide does not possess any
low-lying electronic excited states to which the energy could be transferred.
Deuterium effect quenching experiments with NO indicate that only part of the
excitation energy is dissipated by the C-H vibrational modes of the aromatic
molecule. The rest is dissipated by the C-C modes, the matrix or the vibrations
of the NO molecule.
9. Since NO is a minor component of polluted atmospheres, the photoexcitedj
states decay primarily by reacting with oxygen via quenching, since the proba-
bility of quenching per encounter is fairly high, and oxygen is the dominant
reactive molecule.
10. The reactivity of the benz(a)anthracene molecule and its monomethyl
derivatives with oxygen was compared to their known biological activity. Some
of these compounds are known to be carcinogenic. The most carcinogenic ones
are the 6,7>8, and 12 derivatives. Within experimental error, these compounds
also appear to display the lowest quenching constants, i.e. are least reactive
with respect to quenching of triplets by oxygen. No special significance can
be attached to this correlation, except that detoxifying reactions in vivo
may render the more chemically reactive compounds harmless much more quickly
than the chemically less reactive carcinogenic 6,7,8 and 12 derivatives. This
would be a reasonable assumption if it could be shown that these detoxifying
reactions are also electrophilic bimolecular reactions (as is oxygen quenching
of the triplets).
-------
-7-
S.S. Epstein and his co-workers have shown that there is a general
correlation between the carcinogenic activity and photodynamic activity of
close to 150 different compounds. An exception to this general rule are the
7 and 12 monomethyl derivatives of benz(a)anthracene, which are most carcino-
genic, but photodynamically least active. Our studies indicate that the
relatively low photodynamic activity of these two compounds is due to their
lower probability per encounter of quenching (of triplets) by oxygen, and
also to their faster triplet decay times. Short triplet decay times decrease
the probability of a collision with Og, and thus the probability of
generation of singlet oxygen. The latter is the photodynamic lethal agent
which causes the death of many living cells when they are irradiated in the
presence of aromatic hydrocarbons and air.
-------
-8-
REPORT ON WORK NOT PUBLISHED
Page
Section I: Quenching of Polynuclear Aromatic Hydrocarbon
Triplet States by Nitric Oxide 9
Section II: Quenching of Singlet Excited States of Polynuclear
Aromatic Hydrocarbons by Oxygen and Nitric Oxide 26
*Extracted (with minor editing) from the Ph.D. dissertation of R. Benson
(New York University, 1975 )• A short paper entitled "The Quenching of
the Excited States of Benz(a)anthracene by NO and Oa" based on this work
is now in preparation (Dec. 1973).
-------
-9-
SECTION I
Quenching of Polynuclear Aromatic Hydrocarbons Triplet States by Nitric Oxide
Introduction
Nitric oxide is paramagnetic and is a doublet in its ground state. It is well
known that NO quenches both singlet and triplet states of PAH* molecules (6).
The lowest energy excited state of NO is at 37,900 cm , which exceeds the low-
est triplet energy levels and most of the lowest singlet levels of all the PAH
molecules. This means that energy transfer to electronic states of NO is ruled
out and that the pathway of quenching is therefore enhanced intersystem crossing.
?
The quenching of a triplet PAH( M*) by NO can be represented by the following
scheme:
The quenching constant can be written as :
where all the symbols have the same meaning as in 0_ quenching. The only difference
in the expression for y between 0- and NO quenching is the spin statistical factor,
which is 1/3 for NO quenching by the doublet enhanced intersystem crossing pathway,
which is the only possible quenching pathway.
The Franck-Condon factors (F) involved in NO quenching of triplets, should be
the same or .somewhat smaller (if NO is vibrationally excited) than the Franck-Condon
"Polycyclic aromatic hydrocarbons
-------
-10-
factor for unimolecular decay of the triplet. Thus, F should be much smaller
for NO quenching than for 02 quenching, because quenching by oxygen molecules
involves partial energy transfer to the excited electronic states of oxygen
which results in a smaller amount of excess energy (AE) to be dissipated by the
2
aromatic molecule. Thus, if 3 ,, the electronic matrix element coupling initial
and final states, is similar in magnitude in the quenching of triplets by both
02 and NO, it is predicted that K2(NO) < K2(02).
The experimental NO quenching constants should be simpler to interpret
since there is only one quenching pathway, while there are three possible path-
ways in oxygen quenching. Calculations should also be simpler because the PAH-
NO complex can be treated as a three electron problem, while the PAH-02 complex
is a four electron problem.
We performed two separate studies on NO quenching: 1) we studied the ef-
fect of triplet energy and deuteration on y(NO) and 2) we studied the effect
of methylation on the quenching of benz (a) anthracene.
The Dependence of y(NO) on the Triplet Energy
We measured the quenching constants, with NO, of the following PAH molecules:
anthracene, benz (a) anthracene (BA) , a,h-dibenzanthracene(a,h-DBA) and chrysene.
The quenching constants, P values and K2/K_. values are listed in Table I. The
P values are obtained from the following equation:
Gijzeman et al (6) have measured y(NO) for a series of PAH molecules in various
solvents using the photolysis technique. Their results, for hexane, are listed
in Table 2.
-------
-li-
lt is obvious from our and Gijzeman et al's (6) results that NO quenching
triplets is quite different from 0_ quenching. The plot of Gijzeman et al's
*
value of y (NO) versus E_ is shown in figure 1. We see that Y(NO) increases
Table 1: Quenching Constants YJ Probability of Quenching P and values of K0/K
2
for Nitric Oxide Quenching of Aromatic Triplets in this Work.
Compound
Y X 103 (ys)"1
VK-1
Anthracene 23 3
Benza(a)anthracene 9.5 .7
a,h-dibenzanthracene 2.0 .2
Chrysene 5.2 .4
.017(.014-.020)
.0072(.0066-.0077)
.0015C.0014-.0017)
.0040C.0037-.0043)
.018C.015-.021)
.0072C.0067-.0078)
.0015C.0014-.0017)
.0040C.0037-.0043)
Table II: Gijzeman et al's values for YCNO)
Quenching of Aromatic Triplets in
Compound
Triphenylene
Phenanthrene
Chrysene
Coronene
a , h-dibenzanthracene
Benz (a) anthracene
Anthracene
3,4, 8, 9-dibenzpyrene
Tetracene
ET(cm"1) Y X
23,300
21,600
20,000
19,100
.17,800
16,500
14,700
12,000
10,300
and PCNO) for Nitric
Hexane Solution
lO'^liifV1)
72
20
10
4.6
1.8
1.3
0.84
2.1
7.0
Oxide
P
072
020
010
0046
0018
0013
00084
0021
007
with E when E < 15,000 cm and increases with decreasing E when E > 15,000 cm" .
At low £„, NO quenching is thus similar to 02 quenching, but at high £„, it is dif-
ferent.
*ET... Energy of the triplet excited state.
-------
-12-
30
in
-i
'e
i-H
v_^
*x
o
I-l
X
10
10
2%
Et X 10~3 (cm"1)
Figure 1 - Plot of Gijzeman et al's y(NO) values versus
Triplet Energy (in hexane solution).
-------
-13-
The plot of y obtained in the polystyrene matrix, versus ET is shown in
figure 2. It is evident by comparing figures 1 and 2 that there are differences
in
between NO quenchingAhexane and in polystyrene, in contrast to oxygen quenching (3).
The minimum value of y in PS appears to be between 18,000-
19,000 cm" , about 3,000 to 4,000 cm" greater than hexane.
Another difference between NO quenching in PS and in hexane are the values
of P and K2/K , the values obtained in PS are larger by a factor of 2 - 20.
This difference in part may be due to K_, which obviously is not the same in the
two media. The values of P(NO) in PS are obtained by assuming K.. is the same for
02 and NO in the PS; this assumption may be in error. However, it is clear that
the values of P are small enough that K - > K , and thus y itself is a measure of
K2/K_r
Gij zeman et al have interpreted their results in terms of an exchange mechanism
at low ET, which is similar to the quenching mechanism. As ET increases, however,
the charge transfer character of the intermediate complex increases and the CT
mechanism appears to account for the triplet energy dependence of y(NO) better than
2
the exchange mechanism. 6 ., the electronic matrix element coupling initial and
final states, for a charge transfer mechanism is inversely proportional to the
square of the difference in energy between the triplet (E_) and the charge transfer
state (ECT): , |
2
BCT will increase with a decreasing energy gap, E - E . Gij zeman et al show that
the gap E_T - £„ decreases as £„ increases, thus accounting for the increase in y
(NO) as E increases. At very low £„, y decreases with £„, which is similar to
oxygen quenching, indicating that Franck Condon factor dominates the quenching in
this triplet energy domain.
-------
-14-
in
to
o
X
5-
i
X 10
"3
Figure 2 - Plot of our Y(NO) values versus Triplet Energy
(in polystyrene matrix).
-------
-15-
2 2
At high E_,, it is 3__ which dominates the variation in y. The increase in &_„
appear to be large enough so that y is increasing very rapidly despite the de-
creasing Franck-Condon factor with increasing ET Csee figure 1). This indicates
the importance of the energy gap denominator in equation (4).
Gijzeman et al also studied solvent effects. They observed large decreases
of one order of magnitude, for high E_ PAH compounds upon changing the solvent
from hexane to acetonitrile. The decreases in y with increasing polarity is
smaller at lower E_. In the very low ET range ( < 12,000), polarity has no ef-
fect on the y. These solvent effects substantiate Gijzeman et al's interpretations
of the variation of y with £„ given above. The decrease in y with increasing
polarity has not been adequately explained, since in a charge transfer mediated
quenching mechanism, one would expect y to increase with increasing polarity of
the solvent, because £_„ - ET should decrease.
The difference of 3,000 - 4,000 cm" in t
E curves (see fig. 1 and 2) between hexane and PS media can be explained by
assuming that the average gap E
can be approximately written as:
The difference of 3,000 - 4,000 cm" in the two minima of the y(NO) versus
assuming that the average gap ECT - ET is 3,000 - 4,000 cm" larger in PS. ECT-ET
ECT - ET ' 'P - EA - ET - C C5>
where: I_ is the ionization potential of the aromatic hydrocarbon
EA is the electron affinity of NO (.9ev)'2'
C is the coulomb stabilization plus solvation energies of the complex.
The only difference between PS and hexane, in ECT - £„ will be found in the C
term. Gijzeman et al, estimate that C = 4.1 ev, based on the analysis of their
data. Based on a charge transfer mechanism Gijzeman et al write (in our notation):
-------
-16-
F .
All the factors in equation (6) are known except a and C. By plotting equation
(6), the best fit is obtained when C = 4.1 ev. Subtracting .43 ev (3500 cnf ,
the difference in minima in fig. 1 and 2) from 4.1 ev, we estimate a value of
3.7 ev for C in our PS matrix.
The coulomb term for the iodine-benzene complex has been estimated to be
3 - 3.5 ev (3). Birk's (7) recently measured a value for C of 2.9 ev for the
benzene -02 complex in the vapor phase. In the vapor phase where no solvation
stabilization is present, C is smaller than in a non-gaseous state. Thus, we
can say that our value of C for the polystyrene matrix is not unreasonable.
Therefore, it is possible that the differences in the minima between the curves
shown in figures 1 and 2 is due to the C term in equation (6).
-------
-17-
From our own and Gijzeman et al's (6) results, the following can be said
about NO quenching of triplets: At low triplet energies, the quenching ap-
pears to be similar to 0_ quenching in that Y a 1/E ; at high triplet energy
NO quenching is dissimilar to 0- quenching in that Y a £„. From these results
it appears that the NO- PAH complex has charge transfer character which domin-
ates the variation of the quenching constant, Y* when the triplet energy E >
15,000 cm in hexane, or E_ > 18,500 cm in PS, while at lower energies the
Franck-Condon factors dominate the variation of Y with ET.
Effect of Deuteration on Y (NO) in Polystyrene
* -3-1
We measured Y for deuterated BA. The result is Y = 4.8 ± .4 X 10 (us)
and K2/K_1 = 3.7 X 10"3 (3.4 X 10"3 - 4.0 X 10~3) . Thus, there is a deuterium
effect (see table III). It will be recalled that the analogous ratio for 0_
quenching is 1.2.
Table III. The Deuterium Effect on Triplet Lifetimes, 0_ and NO Triplet
Quenching Constants of Benz (a) anthracene
Process AE
(l/t(H) / (1/T (D) = 4.8 ET
YH(02) / YD(02) = 1.2 ET - E( \ )
YH(NO)/YD(NO) =2.0 ET
This result (i.e. YH(NO)/Y( No)> YH(02)/YD(02) ) can be explained'in terms of
the partial energy transfer from the excited aromatic triplet to 0-. This re-
sults in a lower AE (excess electronic energy) for 02 quenching than in NO
* We attempted to measure Y(NO) for other deuterated compounds also, but the
results were erratic and are thus not reported.
-------
-18-
quenching. This accounts for the larger deuterium effect in NO quenching.
The effect of deuteration on the triplet lifetime (in which AE = ET)
of benz(a)anthracene is still larger and is about 4. (i.e.' (l/T(H))/l/-r(D)=4).
The deuterium effect on NO quenching is thus similar to 0_ quenching
and unimolecular decay. This similarity indicates that in all three cases a
substantial amount of energy remains on the aromatic molecule after quenching,
(see table III).
According to our reasoning the Franck-Condon factor should be the same for
unimolecular decay and NO quenching of the triplet state (AE values are approx-
imately the same). Thus, the deuterium effect should be the same for both pro-
cesses. However, we observe that the deuterium effect on the lifetime is larger
than on y(NO). It is possible that nitric oxide is vibrationally excited during
the quenching process. This would reduce AE (excess electronic energy) and con-
sequently the F factor and deuterium effect would also be reduced. The exact
amount of energy that is transfered to NO is not known but if we assume that
three quanta of vibrational energy is transferred then the deuterium effect is:
Y(H)MD) = 3.35*
Though the calculated deuterium effect is less than the deuterium effect on
the triplet lifetime, it is still larger than the observed effect. We have no
explanation for this anomalous small deuterium effect. The energy degradation
mechanisms in the quenching process are thus not simply analogous to unimolecular
decay. The F factors may not be the same in the two process, although in the
case of oxygen quenching, this assumption appears to explain the deuterium effect
in a reasonable manner.
AE = E - 3 (1876 cm" ) where 1876 cm" is the fundamental frequency of NO.
-------
-19-
Effect of MetKyl Substituents on the Triplet State Quenching of Benz(a)anthracene
by Nitric Oxide
A study of the effect of methyl groups substituents on 0_ quenching allowed
us to characterize to some extent the actual physical interaction between the
0_ molecule and the triplet PAH.
We expect that a similar study with NO should also shed some light on the
quenching mechanism.
In the following table the quenching rate constant, y» for NO quenching of
the triplet states of five compounds of the methyl BA family are listed.
Table IV. Quenching Constants for Nitric Oxide Quenching of the Methylated
Benz(a)anthracene Triplets.
Compound y(yS)" X 10
8 -Me-BA 4.8 ± .7
7 -Me-BA 5.6 ± .3
6 -Me-BA 5.0 ± .6
2 -Me-BA 5.1 ± .3
BA 8.3 ± .9
Discussion
The NO quenching constants of all of the BA derivatives studied are more than
two orders of magnitude smaller than the quenching constants with oxygen.
This difference can be explained by differences in K. or K2. NO is physically
about the same size as 02, so that K. (NO) = K^ (0 ). Thus the difference
between YCNO) and Y(02) is probably due to differences in K_. As mentioned
before this difference is due to smaller Franck-Condon factors for NO quenching.
-------
-20-
Also there may be differences in g .. between 02 amd NO quenching.
Assuming that K is the same for NO and 0., P and K2/K values for NO
quenching can be obtained from the'0. quenching data by use of equation (3).
In table V values of y, P and K_/K . are tabulated.
Table V. Quenching Constants (y), Probability of Quenching (P) and Values
of K2/K_- for Nitric Oxide Quenching of the Methylated Benz (a)anthra-
cene Triplets
Compound y(ys.)"1 X 103 P X 103 KJK , X * "3
2
6
7
8
- MeBA
- MeBA
- MeBA
- MeBA
BA
(5.
(5.
(5.
(4.
C8.
1 ±
0 ±
6 ±
8 ±
3 ±
• 3)
• 6)
• 3)
.7)
.9)
4.
4.
4.
3.
6.
2(4.5
2(4.8
6(4.9
9(4.5
9(7.6
- 4.0)
- 3.8)
- 4.4)
- 3.4)
- 6.1)
4
4
4
4
7
.2(4.5
.2(4.8
.6(4.9
.0(4.6
.0(7.7
- 4.
- 3.
- 4.
- 3.
- 6.
0)
8)
4)
4)
2)
Besides being much smaller than y(09), y(NO) for the methylated BA compounds
£
show a much different methyl group effect. There does not seem to be a steric
effect, unlike 0_ quenching. This is supported by the fact that all of the
methylated BA's have roughly the same value of y, within experimental error. A
methyl steric effect would result in values of y(NO) which would be dependent on
the position of the methyl group, in analogy to the effects observed in oxygen
quenching. Thus the methyl effect must be due to a bulk property of the BA
molecule which changes upon methylation and which appears to be independent of
the position of the Me group.
lonization Potential
Since the quenching process with NO seems to involve a complex with considerable
CT character, a bulk property of interest is the ionization potential. In table
VI the ionization potentials of the different compounds studied are listed (2).
-------
-21-
Table VI. lonization Potential and Relative Energy Gap Between the Charge
Transfer State and the Triplet State
Compound Ip..{ev)* Ip—E (ev)
2 - MeBA
6 - MeBA
7 - MeBA
8 - MeBA
BA
7.39
7.41
7.29
7.39
7.50
5.
5.
5.
5.
5.
32
32
28
33
42
The Ip were estimated from CT absorption spectra
2 2
Equation 4 gives the expression for 3 , ~ BCT for a charge transfer state,
since
a
_2 -2
- ET) = (AE) t equation (5) is an expression for AE.
If we make the reasonable assumption that only Ip changes significantly for the
methyl benz(a) anthracene family, then we can represent AE by Ip - £„. In table
VI the values of Ip - £„ are tabulated.
According to the Ip - ET values in table VI and invoking a charge transfer
mechanism, BA should have the smallest quenching constant, and the methylated
BA compounds should all have roughly the same value of y which should be larger
than Y (BA). This is not the case, and it is evident that this analysis does
not explain the experimental observations.
If E_T is located below E_, then the experimental results make more sense.
In this case AE would be smallest for BA, since it should have the highest energy
charge transfer state. Again AE would be similar in size for the methylated BA
compounds and larger than AE for BA. Thus Y for the methylated benz (a)anthracene
-------
-22-
would be smaller than y for BA according to equation (7).
This explanation of the observed results of NO quenching of triplet BA
and its methyl derivatives assumes that E_ is located below £„. We can make
an estimate Eprr according to eq. (9). The only unknown in this equation is C.
Li 1
Taking a value of C of 3.7ev as in the previous section, £_, for BA is 2.94ev
which is .86ev above the £„ level of 2.08ev. Because of this result it seems
unlikely that E_. would be below £_,. It is not impossible that C could have
a value greater than 4.5ev; as shown in the previous section, Gijzeman et al
(6) arrive at a value of C of 4.1 ev.
The inverse effect on y of increasing the polarity of the solvent could also
be explained by assuming that E is smaller than ET. Increasing the polarity
L*l I
of the solvent would decrease ECT because of increased stabilization due to sol-
vation. This would result in AE increasing as the polarity increases.
In summary, the experimental results can be explained by assuming that E
lies below £_,, but it is difficult to justify placing the E_T level below E_,
since the coulomb energy term C does not appear to be sufficiently large. This
hypothesis can be verified or dismissed only after an independent determination
of £„_ becomes available.
Promoting Modes - Another possible explanation of the methyl substituent effect
is that a promoting mode causes the transition from the neutral complex to the
charge transfer complex (M NO) •*• (M ....N0~). This hypothesis assumes that
ECT is higher than E_. Thus the presence of the methyl groups on the BA ring
system would result in a decreased probability of thermal population of the pro-
moting mode. This is because the more modes there are in the molecule, the less
the probability that any particular mode is excited. The promoting mode hypothesis
can also explain the deuterium effect. This is because the C-D modes being
"softer" (lower nw ) would have more energy in them than C-H modes. This results
-------
-23-
in the promoting mode having lower probability of being excited.
It is hard to see how a promoting mode mechanism could explain the variation
of Y with E_, as shown in figure 2, or the solvent effect observed by Gijzeman et
al. Because of these deficiences the promoting mode mechanism is not considered
a likely one.
Other Mechanisms - Another possible explanation of the methyl substituent effect,
which could also explain the solvent effect, is to be found in the energy indepen-
2 2
dent part of 3r_ . 3~,, can be expressed as:
Ll LI *y
ft'
j BCT
3cr = C7)
2 o
It can be shown that 3'2 is proportional to 4S /R , where S is the overlap
CT
integral between appropriate wave functions of the quencher and PAH in the CT
2
complex, R is the distance between the charge centers. 3iT should decrease
Li 1
with increasing R.
The methyl groups, because of the inductive effect, should cause the PAH
cation in the CT complex to be less positive than in the unsubstituted PAH cation.
Thus the electrostatic attraction is reduced resulting in an increased R value,
2
and thus to a decreased 3' value.
CT
The solvent effect (a decrease in y(NO) with increase in polarity of the
solvent), could be explained in the same way. Increasing the polarity of the sol-
vent would result in greater solvation and a slight increase in the distance
between charge centers in the complex, which would result in decreased values of
3' . As shown before, increasing the polarity would also decrease £_„, and should
CT CT
2
thus increase 3' . It is hard to say which of these two effects, which are due to
CT
an increased solvent polarity, are more important than the other.
The analysis of our results of the methyl substituent effect on NO quenching
of triplet BA is not satisfactory. The best explanation of the methyl effect,
-------
-24-
EfT < £_,, does not seem reasonable. The other explanations are speculative.
More experiments must be done to elucidate the quenching process. The most
important would be to try to locate the charge-transfer state energy level.
The promoting mode mechanism could be checked by temperature experiments.
-------
-25-
SECTION II
Quenching of Singlet Excited States of PAH's by Oxygen and Nitric Oxide
In an effort to characterize, the polystyrene matrix we'have measured the
quenching constants by 0_ and NO, of the following singlet excited aromatic
molecules: benz(a)anthracene (BA), 2-Me-benz(a)anthracene (2 MeBA)>7-MeBA,
12-MeBA, 4 - MeBA, naphthalene (0- only), and pyrene. The quenching constants
K , are listed in Table VII.
Table VII. Lifetimes in vacuum, oxygen (1 atm.), and nitric oxide (1 atm.),
and quenching constants K (02)» K (NO) of some aromatics
Compound T(VAC) r(0) t(NO)
kq(NO)
-1
-1
BA
2 - MeBA
7 - MeBA
12-MeBA
4 - MeBA
Naphthalene
Pyrene
(ns) (ns) (ns) (ys) (ys)
43.6 ± .1 41.7 ± .1 40.5 ± .1 1.5(1.2-1.6) 2.4(2.1-2.6)
38.0 ± .1 36.9 ± .1 34.6 ± .04 1.1(0.8-1.2) 3.4(2.8-3.7)
47.2 ± .04 45.8 ± .1 43.6 ± .1 0.8(0.5-0.9) 2.2(2.0-2.4)
33.1 ± .1 32.4 ± .1 30.6 ± .1 0.9(0.7-1.2) 3.3(3.0-3.6)
38.2 ± .1 36.9 ± .1 33.9 ± .1 1.2(1.1-1.5) 4.4(4.2-4.6)
95 ± 2 2.3 ± .1
306 ±4 0.8 ± .1 1.0 ± .1
From the data in Table VII it is obvious that singlet quenching is not diffusion
controlled since the quenching constants are not identical to each other. Thus
the porous polystyrene matrix experiments appear to be more similar to vapor
phase (8) than to liquid phase (9) experiments with regards to quenching. Thus
it can be concluded that P (the probability of quenching per encounter) for sing-
let quenching is less than unity. The value of P for singlet quenching by 0_
can be estimated from the following expression:
-------
-26-
P(S) = 1/9.(Y(S)/YCT))PCT) (8)
The analogous expression for P(S) for NO quenching of singlets is:
PCS) = 1/3(V(S)/Y(T))PCT) (9)
Using equation (8) and (9) we obtain for benz(a)anthracene -
P(S) = .7P(T) for 09
CIO)
PCS) = 88.6PCT) for NO
Thus the probability of quenching of singlets by oxygen is less than for triplets,
while the opposite is true for nitric oxide.
Intuitively one would expect in general that PCS) > PCT) because the Franck-
Condon factors for singlet quenching should be much larger than for triplet
quenching, Ci-e., Ec - £„ , (or E_ ) < E_, - E0 ) if one assumes, as Potashnik
Sl Tl T2 Tl S0
et al have indicated (10) that triplet states are produced in quenching. Thus,
the larger nitric oxide quenching constants of singlets as compared to triplets
is understandable in terras of Franck-Condon factors. The smaller 0, quenching
constants of singlets as compared to triplets on the other hand, is not under-
standable in terms of Franck-Condon factors.
It has been shown that the likely quenching pathway for singlets is enhanced
intersystem crossing (10). If the T2 state is lower in energy than the S. state,
it is likely that the T2 state, rather than the Tj state, will be the final state
of the quenching process. This is because the energy gap (E_ - £„, ) will be
Sl ^2
smaller and consequently the Franck-Condon factor will be larger than if T. is
the final state. In Table VIII the quenching constants of three of the aromatic
compounds studied is compared with the energy gap E_- ET where T. is the triplet
Sl Ti 1
state closest to S, in energy but still being below the Sj state.
-------
-27-
Table VIII. Singlet Quenching Constants K (00), K (NO) and E ' r £„
q i q j>j K
Compound
Pyrene
BA
Naphthalene
Kq(02)
0.75
1.6
2.2
-Kq:(NO)
0.99
2.51
—
E - E
Sl ' ' Ti
10,000
9,500
1,500
T..
Tl
Tl
T2
There is a rough correlation between both K (00) and KfNO) and E_ - £„
4 * H 1 i
but
not enough compounds were studied to check the validity of the correlation.
(Although we have an oxygen quenching constant for DBA, no data on the T~ level
could be found.) If the correlation is true it would indicate that variation in
the Franck-Condon factor dominates variation in the quenching constants of
singlets as well as of triplets. However, since this dominance did not hold
true for NO quenching of triplets in those cases where charge transfer inter-
action in the intermediate complex was shown to be important, it is not obvious
that Franck-Condon factors will dominate in singlet quenching by either CL or NO
in such cases. With a charge transfer mechanism there are two important energy
gaps to consider. These are either Ec - ET which determines the Franck-Condon
Sl Ti 2
factor, and £_„, - Ec which determine the magnitude of 3™, (see equation (4)).
Ll D-i LI
E__ - £„ can be estimated from the equality Ip - Eg - E. - C, where E. is .5ev
for 0- and .9ev for NO, and we take 3ev as the value for C for both the 02 and
2
NO complex. In table IX values of K (02), Kg (NO) and (1/ECT-ES ) are listed.
Table IX. Quenching Constants K (NO), K (OJ and values of (E_T - E )"2 for
q q 4 LI &.
the aromatic compounds studied
Compound
Kq(02)(ys.)-1 (I
oxygen
Benz (a) anthracene 1.6
Pyrene
Naphthalene
0.78
2.2
: _E I'2
*f*rr* *"*o J
• V» i O -I *'
1.6
1.98
0.77
K CNQ)
nitric
2.51
0.99
-
.(ECT..-.E
'oxide '
6.57
9.18
-
Si)-2
-------
-28-
_2
There seems to be no correlation between K (09) or K (NO) with (£__, - Ec )
q e. q LI o^
as one would expect. It is obvious that our models for singlet quenching are
too simple or missing some crucial factor or concept since we cannot obtain
any correlation between K (NO) or K (0_) and Ec or Ip of the PAH. Such correlations
q q z t>j
were found with triplet quenching. The reason for this difficulty may be that the
Franck-Condon factors are large in singlet quenching, so that subtle changes in
2
p (density of final states factor) or B may dominate variations in the singlet
6 \.
quenching constants.
The Effect of Methyl Substituents on the Quenching of Singlet States of Benz(a)-
anthracene by Oxygen and Nitric Oxide
Oxygen
The methyl group substituent effect appears to be the same for singlet
quenching as for the triplet quenching by 0_. Therefore, the methyl effect on
singlet quenching appears to be steric in nature. Thus the quenching constant is
a function of the position of the methyl group. The 7 and 12 methyl derivatives
of BA again have the lowest quenching constants indicating that oxygen acts as an
electrophilic agent. Thus most of the conclusions made from the methyl effect on
triplet state quenching by 0_ can also be made for singlet state quenching by 02<
In summary these conclusions are 1) singlet quenching is not similar to a 1,4-
cycloaddition reaction, 2) or to free radical attack, 3) the effect of the methyl
groups is mainly steric, thus 02 must attack the singlet excited molecule around
the perimeter of the molecular skeleton, and 4) oxygen acts as an electrophilic
agent.
Nitric Oxide
The methyl effect on singlet quenching by NO does not appear to be similar
to the methyl effect on triplet quenching by NO. The methyl effect on NO quenching
of triplet states was not steric in nature but electronic. The methyl group
reduces the ionization potential of the BA molecule; this decrease is independent
-------
-29-
of the methyl groups position. With singlet quenching by NO the quenching con-
stants in Table VII do not appear to be steric in the same sense as with 02
quenching. If this were true then K (NO) for BA and 7 - Me - BA should be very
different, yet the observed K (NO)'s are similar in magnitude. A similarity
of K (NO) magnitudes for BA and 7-Me-BA and a 30% range in K (NO) for the other
methyl derivatives studies indicates that the ionization potential is not impor-
tant in determining the magnitude of K (NO). Thus a CT mechanism does not appear
likely from our results. We come to the conclusion that the data for singlet
state quenching by NO of the methyl derivatives of BA does not allow us to
make any conclusion about the methyl effect.
Summary of Results and Conclusions of Singlet State Quenching of Aromatic
Molecules by Oxygen and Nitric Oxide
The most important result of the singlet state quenching studies is that
the quenching is not diffusion controlled. This result indicates that the porous
polystyrene matrix is more akin to the vapor state than the liquid state, where
singlet quenching is invariably diffusion controlled. A very tentative conclusion
about singlet state quenching is that the quenching constant is inversely pro-
portional to the energy gap between the S, state and the next lowest triplet
state (T.. or TO. The methyl effect on singlet quenching by 0_ appears to be
similar to the methyl effect on triplet quenching. On the basis of the exper-
imental results, no definite conclusions are possible about the methyl effect
on NO quenching of singlet states.
-------
30
REFERENCES
1. R. Benson, "Interaction of Excited States of Polynuclear Aromatic
Hydrocarbons with Oxygen and Nitric Oxide", Ph.D. Thesis, New York
University, October 1973. University Microfilms, Ann Arbor, Michigan.
2. "Interaction of Triplets of Aromatic Hydrocarbons with Oxygen and
Nitric Oxide", N.E. Geacintov, R. Benson and S. Pomeranz, Chem. Phys.
Lett. 17, 280 (1972).
3. "The Quenching of Excited Triplets of Aromatic Hydrocarbons by Molecular
Oxygen", R. Benson and N.E. Geacintov, J. Chem. Phys. 59_, 4428 (1973).
4. "Deuterium Effect on the Quenching of Photoexcited Aromatic Hydrocarbon
Triplets by Oxygen", R. Benson and N.E. Geacintov, J. Chem. Physics,
accepted for publication, December 1973.
5. In preparation (December 1973) "The Quenching of the Excited States of
Benz(a)anthracene by Oxygen and Nitric Oxide", by N.E. Geacintov, R.
Benson and J. Khasrofian, to be submitted to Chemical Physics Letters.
6. O.L.J. Gijzeman, F. Kaufman and G. Porter, J. Chem. Soc. Farad. Trans. II,
69, 708, 721 (1973).
7. J.B. Birks, E. Pantos and T.D.S. Hamilton, Chem. Phys. Lett. 20, 544 (1973)
8. T. Brewer, J. Am. Chem. Soc. 93_, 773 (1971).
9. I.E. Berlman, Handbook of Flourescence Spectra of Aromatic Molecules,
Academic Press, New York (1965).
10. R. Potashnik, C.R. Goldsmith and M. Ottolenghi, Chem. Phys. Lett. 9_,
424 (1971).
-------
31
TECHNICAL REPORT DATA
(Please read Itatatctions on the reverse before completing)
. REPORT NO.
EPA-650/1-74-010
3. RECIPIENT'S ACCESSIO*NO.
4. TITLE AND SUBTITLE
5. REPORT DATE
Reactivity of Pol/nuclear Aromatic Hydrocarbons with
and NO in the Presence of Light
1 073
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Nicholas E. Geacintov
8. PERFORMING ORGANIZATION REPORT NO.
. PERFORMING ORGANIZATION NAME AND ADDRESS
Chemistry Department
New York University
New York, New York 10003
10. PROGRAM ELEMENT NO.
1AA007, RQAP No. 21AFU-39
11. CONTRACT/GRANT NO.
R801393
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
National Environmental Research Center
Office of Research and Development
Research Triangle Park. NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final .
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
is. ABSTRACT xhe reactivity of 20 different aromatic hydrocarbons adsorbed on solid poly-
styrene fluffs with oxygen and nitric oxide in the presence of light has been studied.
The reaction conditions simulated those encountered in polluted atmospheres. Among thi
compounds studied were anthracene, pyrene, naphthalene, chrysene, benz(a)anthracene an<
coronene. The phot9excited triplet and singlet states of the aromatic hydrocarbons re
act predominantly via the quenching of the fluorescence and phosphorescence by the
paramagnetic 07 and NO gases. The quenching of the reiplets by oxygen occurs via the
formation of an intermediate collision complex in which electrophilic exchange type in
teractions appear to be important. The probability of quenching per collisional en-
counter and the formation of singlet oxygen does not exceed 0.01-1.10 and depends on
spin selection rules, the triplet energy, and the electron density (in the case of the
monomethyl derivatives of benz(a)anthracene). In the case of NO quenching of the trip
lets this probability is much lower, in the range of 0.0005-0.005 and appears to be a
singlet excited states by 07 and NO is nuch more efficient than the quenching of the
triplets and has a probability in the range of ~ 0.30-1.0 and is not necessarily dif-
fusion controlled. The most important contribution of the photoexcited aromatic hydro
carbons (per photon absorbed} to the photochemistry of atmospheres containing Q~ and
NO appears to be the generation of singlet oxygen, since photochemical degradation of
the compounds studied was negligible compared to guenching. The quenching pr9bability
of the triplets of the monomethyl derivatives of benz (a) anthracene by oxygen is com-
pared to their carcinogenic and photodynamic activities. While the correlation with
the carcinogenic activity is unclear, the relatively low photodynamic activity of the
7 and 12 monomethyl derivatives can be explained in terms of their low triplet life-
times and lower probability of singlet oxygen formation per collisional quenching
encounter.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Hydrocarbons, Carcinogens, Anthracene,
Pyrene, Naphthalene, Chrysene, Benz(a)-
anthracene, Coronene
Reactivity of polycyclic
aromatic hydrocarbons.
Polynuclear Aromatic
hydrocarbons. Quenching
of hydrocarbons by
oxygen and nitric acid.
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
35
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |