492
J. Phys. Chem. 1980, 84, 492-495
hot PsOz is believed to dissociate quickly in comparison to both annihilation and gaseous third body collisions. However, it has been observed that, for oxygen dissolved in liquidsz9 or trapped in small diameter pores of silica gels,gthe quenching cross section is substantially greater than the cmz predicted on the basis of the conversion process. In consideration of the high collision frequency occurring in the condensed phase, it has been suggested that deexcitation of hot PsOz is likely. In support of the argument, high precision angular correlation measurements by Goldanskii and co-workersg have shown the existence of two components of positronium quenching by oxygen in silica gels. As the pore diameter decreases, the fraction of positronium which annihilates via a chemical reaction increases. For pore sizes greater than 75 A and gas phase work, conversion (with a cross section of W 9cmz) is the dominant mode of annihilation, consistent with the interpretation proffered in this investigation. Whereas formation of PsOz from addition of positronium to molecular oxygen is energetically impossible in the gas phase, formation of PsOz from ozone is very likely. In this case, the third oxygen atom can serve as the chaperone. The third and perhaps most likely possibility is that PsO is formed and the release of an oxygen molecule serves to deexcite the system, similar to the role of the odd oxygen atom above. Of the suggested reaction paths, this latter scheme seems most reasonable. Ozone is generally observed to oxidize via a one-atom transfer,30and the energy gained by the resulting oxygen-oxygen bond provides a considerable driving force for the reaction. Evidence for the stability of PsO was provided by Tao and Green,31who studied positronium quenching in a series of aqueous oxyacids, the results of which established a PsO bond strength of 2.2 f 0.5 eV. N
(2) Center for Naval Analyses, Ariihgton, VA 22209. (3) W. J. Madia, A. L. Nichols, and H. J. Ache, J . Am. Chem. SOC., 97, 5041 (1975). (4) 0. E. Mogensen, J . Chem. Phys., 60, 998 (1974). (5) W. Brandt and D. Spektor, Phys. Rev. Len., 38, 595 (1977). (6) S. Y. Chuang and S. J. Tao, Phys. Rev. A , 9, 989 (1974). (7) S. J. Tao and J. H. Green, J . Chem. SOC.A , 408 (1968). (8) D. P. Kerr, A. M. Cooper, and B. G. Hogg, Can. J . Phys., 43, 963 (1965). (9) V. I. Goldanskii, A. D. Mokrushin, A. 0. Tatur, and V. P. Shantarovich, Appl. Phys., 5, 379 (1975). (10) V. I. Goldanskii, At. Energy Rev., 6 (l), 1 (1968). (1 1) D. M. Schrader and C. M. Wang, J. Phys. Chem., 80,2507 (1976). (12) D. M. Schrader, Adv. Chem. Ser., No. 175 (1979). (13) W. J. Madia, J. C. Schug, A. L. Nichols, and H. J. Ache, J . Phys. Chem., 78, 2682 (1974). (14) P. J. Karol and R. L. Klobuchar, Nucl. Insfrum. Mefhods, 151, 149 (1978). (15) R. L. Klobuchar and P. J. Karol, J . Phys. Chem., preceding article in this issue. (16) J. G. Calvert and J. N. Pas, "Photochemistry of Gases", Dover, New York, 1966, pp 689-695. (17) W. A. Noyes, "The Photochemistry of Gases", Dover, New York, 1941, pp 38-40. (18) W. C. Pierce, E. L. Haenisch, and D. T. Sawyer, "Quantitative Analysis", Wiley, New York, 1967, pp 291-307. (19) The exact expression for W,, is given by
-
[ 1114X;
+ 4456KX2 + 4448(Xq - K)X, + 15.957K(hq - K) 4+ 4458KXt + 4460(Xq - K)Xt + 16K(Xq - K) + 4(hq - K)']
3.989(Xq - K)'] /[4458h:
and is the form used in all calculations. (20) G. Herzberg, "Spectra of Diatomic Molecules", Van Nostrand, New York, 1950, Appendix; "Electronic Spectra and Electronic Structure of Poiyatomic Molecules", Van Nostrand, New York, 1950, Appendix. (21) V. I. Gokhnskii and A. D. Mokrushin, Huh Energy Chem., 2, 77 (1968). (22) J. D. McNutt and V. B. Summerour, Phys. Rev. 5, 5, 2019 (1972). (23) J. D. McNutt, S. C. Sharma, M. H. Franklin, and M. A. Woodall, 11, Phys. Rev. A , 20, 357 (1979). (24) P. E. Osmon, Phys. Rev., 140, A8 (1965). (25) P. G. Coleman, T. C. Griffih, G. R. Heyland, and T. L. Kiileen, J. Phys. 5, 8, 1734 (1975). (26) S. J. Tao, S. Y. Chuang, and J. Wiikenfeld, Phys. Rev. A , 6, 1967 (1972). (27) C. J. Celitans, S. J. Tao, and J. H. Green, Proc. Phys. SOC.,83, 833 (1964). (28) S. Y. Chuang and S. J. Tao, J . C b m . Phys., 52, 749 (1970). (29) G. J. Celitans and J. Lee in "Positron Annihilation", A. T. Stewart and L. 0. Roellig, Ed., Academic Press, New York, 1987. (30) M. Ardon, "Oxygen, Elementary Forms and Hydrogen Peroxide", Benjamin, New York, 1965, pp 50-60. (31) S. J. Tao and J. H. Green, J . Phys. Chem., 73, 882 (1969).
Acknowledgment. This research was supported in part by the National Science Foundation through Grant No. GP-38390. References and Notes (1) Research supported in part by National Science Foundation Grant GP38390.
Photochemistry of o-Nitrobenzaldehyde and Related Studies M. V. George* and J. C. Scaiano" Radiation Laboratory,' University of Notre Dame, Notre Dame, Indiana 46556 (Received July 30, 1979) Publication costs assisted by the U S .Department of Energy
Photolysis of solutions of o-nitrobenzaldehyde leads to the formation of o-nitrosobenzoic acid with quantum yields in the neighborhood of 0.5. The reaction proceeds via the intermediacy of a triplet state with a lifetime of 0.6 ns. A laser flash photolysis study of the reaction reveals the sensitivity of the transient enol to the presence of hydroxylic molecules in the media. The reaction is proposed to proceed via a short-lived biradical which leads to an enol (13), which is responsible for product formation.
Introduction As early as 1901, Ciamician and Silberz had observed that o-nitrobenzaldehyde (1) undergoes an interesting phototransformation to give o-nitrosobenzoic acid (2) (Scheme I). Since then, several investigators have examined this reaction in detail, in attempts to understand the mechanism of this t r a n s f ~ r m a t i o n . ~Thus, - ~ it is known that this reaction proceeds intramolecularly and that it can 0022-3654/80/2084-0492$0 1.OO/O
take place in both the solid state7 and in s o l u t i ~ n . ~ , ~ Further, it has been shown that this type of rearrangement is characteristic of other aromatic nitro compounds containing an ortho formyl functionality as in the case of 2,4-dinitrobenzaldehyde (3)5 and nitroterephthalaldehyde (7).9 Besides o-nitroaldehydes, even substrates containing a suitable C-H functionality in the ortho position of the aromatic nitro compound have been found to undergo this 0 1980 American Chemical Society
The Journal of Physical Chemistry, Vol. 84, No. 5, 1980 493
Photochemistry of o-Nitrobenzaldehyde
Scheme I R2
R'
R2
0
Ay.,
2L
N=O
R'
O.OI0
2, R' = R 2 = H 4, R 1= NO,; R2= H 6,R' = FLz = NO, 8, R' = CHO; R2 = H
1 R' = LE2 = 14 3. R' = NO,;RZ= H 5, R'= RZ= NO,
7.R' = CHO; RZ= H
0.005
d
100
0
hv
."f
I
t
0.015
0-
H 5;x6 \
I
d
0.020
I
0.020
H@H5 \
.N=O
300
&
I
A h
0.015
10
0-
200
4
9
400
I
ACETONITRILE
0.010
0.02 1
I 0.00 5
20
d
0
a
0.01 -
300
40
60
80
TIME (ns) Figure 2. Decay of the transient from 1 and calculated first-order fit.
400
500
600
X,nm
Figure 1. Transierit spectrum observed upon excitation of a benzene solution of 1.
type of rearrangement.1° Thus, it has been observed that (2-nitropheny1)diphenylmethane(9) undergoes a photorearrangement t,o give (2-nitrosopheny1)diphenylmethanol (10) (Scheme ILIO In spite of the studies mentioned above which illustrate the generality of this type of rearrangement, the details of the methanism of the phototransformation of 1to 2 are yet to be understood. The present investigation was undertaken in an attempt to elucidate the mechanism and kinetics of the solution photochemistry of 1. Our laser flash photolysis studies, combined with quantum yield measurements, provide a closer insight into the mechanism of this reaction.
Results This section is divided according to the experimental technique used. Unless otherwise indicated, the experiments were carried out under oxygen-free conditions. Laser Flash Photolysis. The samples were excited with the pulses (337.11 nm, 8 ns, 3 mJ) from a nitrogen laser, and the resulting transient absorptions were monitored by using a system with nanosecond response. When a benzene solution of 1 was irradiated under the conditions mentioned above, we observed the formation of a transient species which showed relatively weak signals and had the spectrum of Figure 1. The decay of this species followed clean first-order kinetics and had a lifetime of 46 ns (Figure 2). Water had a dramatic effect on this lifetime: 0.1% (v/v) was enough to make this species short lived enough that it became undetectable in our instrument. In fact the lifetimes in benzene were somewhat dependent upon the batch of dry benzene used. In one occasion we measured a lifetime of 75 ns, which un-
[DIENE] , M Figure 3. Stern-Volmer plot for the quenching by trans-l,3-pentadiene of the transient signals generated from 1. The value of k q ~was T derived from the slope-to-intercept ratio.
fortunately could not be reproduced. Dry oxygen had no effect on the lifetime of the transient derived from 1. Addition of trans-1,3-pentadiene did not affect the lifetime of the transient, but instead reduced its yield, as indicated by the decrease in the "top OD" as measured from the experimental traces (Figure 2) before significant decay occurred. A Stern-Volmer plot of this data in benzene at 22 "C and in the 0-1.3 M diene concentration range led to hq7T = 3 M-l (Figure 3). This result is iindicative of a triplet state precursor; if the rate of triplet quenching can be taken as 5 X lo9 M-I s-l, the same value normally used for ketone triplets, then the lifetime of the triplet precursor is TT m.0.6 ns. Table I gives a summary of the lifetimes measured in benzene and other solvents. The effect of water on the lifetimes was as pronounced in acetonitrile as in benzene. In methanol we were unable to detect any transient signalls. When we examined the photochemistry of 2,6-dinitrobenzaldehyde (11) in benzene we observed relatively intense signals with a lifetime of 650 ns (Figure 4). Water, oxygen, and trans-1,3-pentadienehad no significant effect
494
The Journal of
Physical Chemistry, Vol. 84, No. 5, 1980
TABLE I: Lifetimes of the Transient Species Produced in the Photolysis of Nitrobenzaldehydes substrate 1 1 1 1 11 11 11 3
solvent benzene n-decane acetonitrile methanol benzene acetonitrile (10% water) met han o 1 benzene
George and Scaiano TABLE 11: Quantum Yields of o-Nitrobenzaldehyde Disappearance at 3500 A in Different Solvents
7 ,ns
solvent
45 40 17
oa
benzene (dry) benzene (moist) acetonitrile methanol ace tone ligroin
< 5a 650 500 1000 < 5a
a
a These values are based on the critical assumption that the failure to detect the transient is due to a short lifetime and not an unusually low extinction coefficient.
At room temperature.
(0.41 i (0.44 ?: (0.50 i (0.39 i 0.46' 0.52' This work.
0.06)b 0.06Ib 0.06)b 0.08)b
' From ref
5.
Scheme I1 0
0-
0-
12
1
X,nm
Transient spectrum observed upon irradiation of benzene, and (insert)typical decay trace and first-order fit. Flgure 4.
11 in
on the yields or lifetimes observed; in fact 1 M trans-1,3pentadiene caused a slight increase in the lifetime ( 10%) which is perhaps due to a change in solvent composition. In the case of 2,4-dinitrobenzaldehyde (3), all efforts to detect transient signals similar to those observed for 1 and 11 were fruitless. If similar species are involved5they must be considerably shorter lived. A few experiments were carried out in which we added the electron acceptor paraquat (l,l'-dimethyL4,4'-biM to solutions pyridilium) in concentrations of up to of 1 in methanol and methanol-acetonitrile mixtures. In all cases we observed the formation of small yields [a(PQ'.) 0.003-0.02)] of the radical ion PQ'.. The formation of the reduced products was too fast to allow a time-resolved study of their formation, but they should probably be taken as indicative of the intermediacy of a short-lived (7 I10 ns) electron donor. Quantum Yields. These were measured by monitoring the decrease in concentration of 1 by gas chromatography. These measurements were usually performed after 15-min irradiation; the length of time was chosen so that it would result in significant changes in concentration and yet be short enough to avoid the inevitable precipitation of dimeric nitroso derivatives. This precipitation also made it more convenient to monitor reagent disappearance rather than product formation. Table I1 gives a summary of quantum yields which in general are in good agreement with values reported earlier. In comparison with the dramatic effect of water on the transient lifetimes, the quantum yields are relatively insensitive.
k -I H
15
N
-
Discussion Our main observation is that, while the quantum yields of photoreaction of 1 are rather insensitive to the nature
2
and water content of the solvent, the transient signals observed in the same systems are extremely sensitive to the presence of water in the media. The quantum yields given in Table I1 are all of about the same magnitude. In addition, we have observed that the photorearrangement of 1 proceeds via a short-lived triplet state. Perhaps the clue to the mechanism is the effect of water on the lifetime of the transient species which is not reflected by the quantum yields. Scheme I1 illustrates the mechanism proposed. The transient species detected in the flash experiments has been tentatively assigned to 13. In fact, one could in principle suggest that either 13 or 14 are the species observed in our pulsed experiments. We prefer 13 because it seems to provide a better explanation for our observations. For example, hydrogen abstractions leading to enols closely resembling 13 are common in the case of carbonyl compounds and usually occur from the triplet state.'lJ2 By contrast, in the photochemistry of o-phthalaldehyde, where a transient resembling 14 is formed, it is singlet rather than triplet derived.I3 The effect of water on the lifetime of 13 is attributed to the catalysis of its transformation into 2. Two alternative explanations are possible: (i) step a in Scheme I1 is faster in hydroxylic media, and (ii) in the presence of water the transformation of 13 2 involves the intermediacy of 15. The yields are not affected because 13 would transform into 2 even in the absence of water. The triplet state is extremely short lived, and the overall inefficiency of the reaction (a < 1) could result from a low yield of intersystem crossing or, alternatively, a fast mode of radiationless decay. For example, 12 could partition between 13 and 1 and therefore lead to overall yields lower than
-
Photocherriistry of o-Nitrobenzaldehyde
unity. Biradical 12 is probably very short lived, and the observation that paraquat radical cations will trap inefficiently a species with T < 10 ns is consistent with this observation. If the phototransformations of Scheme I occur by a general mechanism, then one would expect similar processes to be involved in the photochemistry of 3 and 11. In the case of 3 we do not observe any transient signals, and we can only speculate that the additional nitro group makes the transient shorter lived. Since the lifetime of the transients observed in the case of 1 are already quite short, a clhange by a factor of 5 would be sufficient to make detection impossible. In the case of 11 we have observed that the transient signals are considerably longer lived than those attributed to 13, and its lifetime is not affected by addition of water. If 16 is involved in the reaction, then the absence of detectable interaction with water would have to be attributed to steric effects.
0-
16
The absence of any effect of dienes on the yields of 16 should pirobably be attributed to a short triplet lifetime; again, the lifetime measured in the case of 1 is only 0.6 ns and a decrease by a factor of 4 would be enough to put it beyond the range that can be measured with techniques of this type. In all cases, the triplet lifetimes are short enough that the absence of oxygen effect (typical concentrations 2-10 X M ) can be readily understood. In conclusion, the well-known rearrangement of onitrobenzaldehyde to yield o-nitrosobenzoic acid proceeds from the triplet manifold. The quantum yields are rather insensitive to the polarity and hydrogen bonding properties of the solvent. By contrast, the transient species detected and assigned to 13 interacts with water quite efficiently. The reaction is proposed to proceed via the intermediacy of short-llived biradical intermediates. Experimental Section Starting Materials. All nitro derivatives were Aldrich products. They were sublimed under high vacuum (Caution: while we did not experience any accidents, molecules 3 and 11 should be handled with care while subliming). All solvents were Aldrich Gold Label. Benzene was dried over calcium hydride, distilled, and stored over molecular sieves. Sample Preparation. Unless otherwise indicated all samples were deaerated by bubbling dry, oxygen-free ar-
The Journal of Physical Chemistry, Vol. 84, No. 5, 1980
495
gon; those for quantum yield studies were contained in matched tubes made of precision bore tubing (0.2500 f 0.0002 in. i.d., made of Corning 7740 glass, Lab. Crest Scientific). The samples for the laser experiments were contained in cells made of rectangular Suprasil tubing with an optical path of 3 mm. When oxygen was bubbled into the samples, it was dried on a molecular sieves column. This precaution proved to be extremely important; our preliminary experiments revealed an apparent oxygen effect which was later shown to be the result of moisture in the gas. Quantum Yields. The samples were irradiated in a Rayonet photochemical reactor fitted with 16 RPR 3500-w lamps. A merry-go-round was used to ensure that all samples received the same radiation dose. The photofragmentation of valerophenone in benzene to yield acetophenone was used as an actinometer, taking @(acetophenone) = 0.30.14 Gas chromatographic analyses were carried out by using a Beckman GC-5 instrument, equipped with flame ionization detectors. Laser Flash Photolysis. The samples were excited with the pulses from a Molectron UV-400 nitrogen laser. The transient signals from an RCA-4840 tube were terminated into 93 ohm and into a Tektronix R7912 transient digitizer. The system is fully interphased with a PDP 11/55 computer which controls the experiment and averages and processes the information. Typically between 15 and 50 shots were averaged in the case of 1 and between 5 and 10 in the case of 11. Further details on the instrument will be given e1~ewhere.l~ References and Notes (1) The research described herein was supported by the Office of Basic Energy Sclences of the Department of Energy. This is documernt No. NDRL- 2043 from the Notre Dame Radiation Laboratory. (2) Ciamician, G.;Silber, P., Berichfe 1901, 34, 2040. (3) For several examples of the rearrangement of o-nitro aldehydes and related systems, see Schonberg, A. "Preparative Organic Photochemistry"; Springer-Verlag: New York, 1968; p 266. (4) Calvert, J. G.; Pitts, Jr., J. N. "Photochemistry"; Wiley: New York, 1967; p 477, and references therein. (5) Leighton, P. A.; Lucy, F. A. J . Chem. Phys. 1934, 2, 756. Lucy, F. A.; Leighton, P. A. Ibid. 1934, 3, 760. (6) Bowen, E. J.; Hartley, H.;Scott, W. D.; Watts, H. G. J . Chem. SOC. 1924, 725, 1218. (7) Pitts, Jr., J. N.; Wan, J. K. S.;Schuck, E. A. J . Am. Chem. Sac. 1964, 86, 3006. (8) Tench, A. J.; Coppens, P. J . Phys. Chem. 1963, 67, 1378. (9) Sytda, H. J . Prakf. Chem. 1911, 84, 827. (10) Tanasescu, I. Bull. SOC. Chim. Fr. 1926, 39, 1443. (11) (a) Wagner, P. J. Pure Appl. Chem. 1977, 49, 259. Wagner, P. J.; Chen, C.-P. J . Am. Chem. Sac. 1976, 98, 239. (b) Haag, H.; Wirz, J.; Wagner, P. J. Helv. Chim. Acta 1977, 60, 2595. (12) (a) Small, Jr., R. D.; Scaiano, J. C. J . Am. Chem. SOC. 1977, 99, 7713. (b) Das, P. K.; Encinas, M. V.; Small, Jr., R. D.; Scaiano, J. C. Ibid. 1979, 707, 6965. (13) Scaiano, J. C.;Encinas, M. V.; George, M. V. J. Chem. Soc., Perkin Trans. 2, in press. (14) Wagner, P. J ; Kesb, P. A.; Kemppainen, A. E.; McGrath,J. M.; Schott, H. N.; Zepp, R. G. J . Am. Chem. SOC. 1972, 94, 7506. (15) Encinas, M. V.; Scakno, J. C. J. Am. Chem. SOC 1979, 701,2146.
