J . Phys. Chem. 1990, 9 4 , 2963-2966
fluorescence spectrum recorded. This serves as a reference, unrelaxed spectrum. Next an intense light source is applied, so that 100% of the dye molecules are excited. The sample is heated to an appropriate temperature under illumination and cooled down again to near 0 K. Finally, the intense light source is turned off and the fluorescence spectrum recorded again. If relaxation has taken place at the elevated temperature, the spectrum measured after recooling should be shifted accordingly. Pulse-Train (“Pumping”) Experiment. This suggested experiment is also modeled after a heme protein experiment.Isb Suppose the sample is irradiated by an intense laser emitting a pulse train with a repetition rate that is high compared to the average Stokes shift duration (e.g.. a 50-ps pulse emitted every 500 ps for a coumarin sample at 220 K). The transient fluorescence spectra are measured as a function of a pulse-train duration after the last pulse in the train. If a relaxation mechanism operates, we expect that each pulse in the train promotes partial relaxation of the solvent toward its excited-state configuration. Between pulses, the solvent does not completely relax in the opposite direction before an additional pulse in the train promotes additional relaxation. The expected effect is of larger relaxation for longer pulse-train duration. Gas-Phase Measurements. The line shape and decay kinetics in the gas phase should be measured and used as a reference for the solution-phase measurements.28
2963
Optical Hole Burning. I f the line shape is inhomogeneously broadened, it should be possible to burn a hole in the absorption spectrum by using a narrow excitation band.29 Such measurements could determine the relative magnitude of inhomogeneous vs homogeneous line broadeningM In addition, fluorescence decay could be monitored as a function of excitation wavelength. In the limit that only a single conformation is excited, the inhomogeneity is removed and the only mechanism for shift is relaxation. Acknowledgment. I am grateful to Mark Maroncelli and Graham R. Fleming for permission to use the data of ref 6. I thank Mark Maroncelli for suggesting the presentation in Figure 6 and Attila Szabo for suggestions of appropriate kinetic schemes. This work was supported in part by USA-Israel Binational Science Foundation Grant 86-00197. Registry No. Coumarin 153, 535 18-1 8-6. (28) Ernsting, N . P.; Asimov, M.; Schafer, F. P. Chem. Phys. Leu. 1981, 91, 231. (29) Hayes, J. M.; Jankowiak, R.; Small, G . J. In Persistent Spectral Hole Burning: Science and Application; Moerner, W. E., Ed.; Springer: New York, 1987. Friedrich, J.; Haarer, D. Angew. Chem., Int. Ed. Engl. 1984, 23, 113. (30) Ormos, P.; Ansari, A.; Braunstein, D.; Cowen, B. R.; Frauenfelder, H.; Hong, M. K.; Iben, 1. E. T.; Sauke, T. 8.; Steinbach, P.; Young, R. D. Biophys. J . , in press.
The 248-nm Photofragmentation of the CH,O, Radical D. Hartmann, J . Karthauser, Institut f u r Physikalische Chemie, Universitat Gottingen, 0-3400 Gottingen, FRG
and R. Zellner* Institut f u r Physikalische Chemie und Elektrochemie, Universitat Hannover, 0-3000 Hannover, FRG (Received: August 4 , 1989; In Final Form: November 9, 1989)
The products of the photodissociation of CH3O2 in the 248-nm (KrF) excimer laser photolysis have been investigated by using bcth emission spectroscopy and laser-induced fluorescence (LI F). Among the products OH(A22+),OH(X211),and CH30(X2E)have been identified with fractional yields of the latter two of p ( 0 H ) = 0.06 f 0.03 and p(CH,O) = 0.2 & 0.1. Formation of OH(A2Z+)is suggested as a diabatic process involving an excited state of C H 2 0 0 H . The observed emission is sufficiently strong to utilize the 248-nm photodissociation of CH302as a monitor of this species in kinetic experiments.
Introduction The C H 3 0 2radical has a well-known near-UV absorption with an absorotion coefficient at maximum near 235 nm of u = (4.8 f 0.4)x’ I 0-l8 cm2.1-) The corresponding electronic transition has not been fully characterize! but bajed on the analogy with H 0 2 probably corresponds to X2A” - B2A” which is both spin and orbital allowed. The structure of the spectrum suggests that the excited C H 3 0 2state is predissociative with the energetically possible fragmentation channels (with AHR/kJ mol-’ values in parentheses) given in Scheme I. As a consequence we may expect cleavage of the 0-0 bond as well as of the (weaker) C - 0 bond. The most exothermic products are C H 2 0 O H which may be formed by 0-0 bond cleavage from C H 2 0 0 H after initial isomerization of CH3O2. Among the electr_onically excited fragments 02(X,a,b), OH(A2Z+) and CH20(A1A”) are energetically possible. Formation of excited C H 3 0 ( A 2 A I ) and CH3(A2A”)is endothermic by 143 and 553 kJ mol-’, respectively, and can therefore not be obtained with 248-nm radiation in a one-photon process. Investigations of the photodissociation products of C H 3 0 2have
+
* Author to whom correspondence should be addressed. 0022-3654/90/2094-2963$02.50/0
SCHEME I CH302
+
h’(248 n m )
-
-
CH30(R2E)
CH3O(R2E)+ O(lD)
0(3p)
(-44.3)
CH3(R2A,”) + 02(X’Z,)
(-3 53 .O)
CH3(R2A2”)+ 02(a’A,)
(-258.7)
CH,(R2A2”) + 02(b’Z,)
(-196.1)
CH2O(%’A1)+ OH(X211)
(-569.5)
+ OH(A%+) CH20(A1A”) + OH(X211)
(-183.1)
--
CH,O(PA,)
(-234.0)
(-239.5)
to our knowledge not been performed previously. One of the interesting questions is whether the photodissociation of CH,02, ( I ) Jenkin, M. A.; Cox, R. A.; Hayman, G.; Whyte, L. J. J . Chem. SOC. Faraday Trans. 2 1988, 84, 913. (2) Simon, F.; Schneider, W.; Moortgat, G . Inr. J. Chem. Kiner., in press. ( 3 ) Moortgat, G.; Veyret, R.; Lesclaux, R J. Phys. Chem. 1989, 93, 2368.
0 1990 American Chemical Society
The Journal of Physical Chemistry, Vol, 94, No. 7, 1990
2964
Hartmann et al.
t
Y
c
,T,
t
I
=
tim
300~
193 nm
E mission
t
*
Figure 1. Schematic representation of laser trigger sequence in studies of the 248-nm photolysis of CH3O2. Ar is the width of the boxcar gate. PF = ground-state photofragments; PFg = electronically excited photofragments.
despite its higher complexity and larger product manifold, reflects the photodissociation of H 0 2 . For this species both experimental4 and t h e o r e t i ~ a linvestigations ~?~ have been performed. The present consensus from this work is that H 0 2 , contrary to the adiabatic expectation, dissociates into OH(X21’I) O(’D). The alternative fragmentation pathway ( H + 0,) is apparently prevented by a substantial dissociation barrier.5 In the present paper we report the first photodissociation study of CH302 using emission and LIF technique in the near-UV region. Although the complete product state distribution has not been determined we observe important product routes for CH30(X2E)and OH(X21’I). In addition the observation of excited OH(A2Z+) has led to the development of a new detection method for CH302 in kinetic and possibly other experiments.’
+
Experimental Section The present experiments were performed in a photolysis cell which has been fully described in previous publications from this l a b o r a t ~ r y . ~Briefly, ,~ it consists of a stainless steel cell with two larger coaxial side arms containing light traps and spectrosil quartz windows at Brewster angle. These two side arms serve as input directions for the laser beams. Light collection optics in perpendicular direction to that of the laser beams are attached at the center of the cell. Methylperoxy radicals are generated by the 193-nm excimer laser (Lambda Physik EMG 102) photolysis of azomethane (AM) in the presence of 02.With [AM] = 1.5 X I O L 3cm-,, u,!,% = 1.2 X IO-’’ cm2, and a laser photon density in the order of 10l6 the typical initial C H 3 concentration amounts to 3 X 10l2 cm-), which in the presence of a large excess of O2 ( 18 mbar) is rapidly ( T I5 M)converted to CH3O2. Although this method of CH3O2 generation is not entirely free of complications (i.e., formation of C H 3 0 , C H 2 0 , O H , and 0) arising from the interaction of hot CH, radicals with 0 , ’ O and from the direct
-
306
307
308
309
310
311
X/nm
Figure 2. Emission of OH(A2Z+,~”=0)as observed in the photodissociation of CH3O2at 248 nm.
A‘l
photolysis of 02,it still ensures the dominant conversion of CH3 to CH3O2. The photolysis of CH3O2 = (4.3 f 0.3) X an2), is achieved with a second excimer laser (Lambda Physik EMG 200) operating at 248 nm and triggered with a variable time delay ( A l l , typically 0.5-1 ms) after the first laser pulse. Both laser beams are directed through the reaction cell coaxially but counterpropagating. A schematic representation of the trigger sequence used in this technique is shown in Figure I . The emission resulting from excited photofragments in the center of the cell is collected via quartz optics and focused onto a photomultiplier (EM1 9786 Q). For the purpose of identification of the emitted radiation a small monochromator (McPherson 218,0.3 m, spectral resolution 0.05 nm) is placed in front of the photomultiplier. The analog signal from the photomultiplier is detected by means of a boxcar integrator with the gate width (At,) and gate delay set at 0.5 and 0.1 ks, respectively. For the detection of ground-state fragments ( C H 3 0 , O H ) an excimer laser pumped frequencydoubled dye laser system (Lambda Physik EMG 102/FL 2000) is used. The output from this laser system is directed into the reaction cell coaxially with the 248-nm photolysis laser using a dielectrically coated 300-nm reflection mirror as a coupling element. This laser system is triggered with a variable delay ( A t , > Ar1) relative to the initial photolysis laser. In this way the fluorscence resulting from ground-state photofragments can be observed independently and unperturbed from the emission from excited-state fragments. The fluorescence intensity is collected with the same optics as used for the emission measurements except that the monochromator is replaced by an interference filter ( I F 306) in the case of O H and a filter combination (WG 320/UG 1 1) in the case of C H 3 0 . The fluorescence excitation wavelengths are 307.97 nm (Q12line of the A2;+-X211(0,0) transition) for O H and 304.1 7 nm ((2,O) band of the A2A,-X2E transition) for CH30. For the calibration of the fluorescence intensities the 193-nm photodissociation of either HNO, or C H 3 0 N 0 has been used. In all cases necessary the absolute radical concentrations generated by laser photodissociation were calculated from the photophysical properties ( u , 4) of the precursor molecule and the number of incident laser photons according to the expression used previously.8.9
N
(4) Lee, L. C. J . Chem. Phys. 1982, 76, 4909. (5) Langhoff, S. R.; Jaffe, R. L. J . Chem. Phys. 1979, 71, 1475. (6) Vasguez, G . J.; Peyerimhoff, S. D.; Buenker, R. J. Chem. Phys. 1985, 99, 239. (7) Hartmann, D.; Karthauser, J.; Zellner, R. Proceedings of the 5th
European Symposium on Physico-Chemical Behaviour of Atmospheric Pollutants, Varese, September 1989. (8) Ewig, F.; Rhasa, D.; Zellner, R. Ber. Bunsen-Ges. Phys. Chem. 1987, 91, 708. (9) Zellner, R.; Hartmann, D.; Karthauser, J.; Rhasa, D.; Weibring, G. J . Chem. Sor. Faraday Trans. 2 1988,84, 549.
Results and Discussion 1 . Formation of O H ( A 2 Z + ) . Following the 248-nm pulse, emission in the wavelength region 306-31 1 nm is observed. Both the spectral resolution of this emission (Figure 2) and the time constant of its decay identify this emission as due to the O H A22+(c”=O) X 2 n , (u’=O) transition. More detailed analysis of the individual branches of this transition indicates high rotational excitation with T,,, = I100 300 K even at p = 2.8 mbar. The dependence of the total emission intensity upon the photolysis laser fluence (Figure 3) is linear. As a consequence the emission must be assigned to a one-photon process.
-
*
(IO) Hartmann, D.; Karthauser, mitted for publication.
J.; Saathoff, M.; Zellner, R., to be sub-
Photofragmentation of the C H 3 0 2 Radical
The Journal of Physical Chemistry, Vol. 94, No. 7, 1990 2965
// 05
0
15
t
r a
T
lg11
Figure 3. Dependence of the emission intensity ( I F ) on the fluence of the
248 nm photolysis laser (IL).
248 nm 1
.
1
02
-
600
OI
-.‘\_t-“.
-200
-
1 1 @I ~ ~ I.[HdiyA;l 91
---_ [U~OOHl~24’l
[H3O2
)i“
-L t*oii’r,l. o ~ i ~ l n i
IOouUeI Sloles in
[I
p e o m ~ r y only1
-
Figure 4. Energy and correlation diagram for CH302+ hv(248 nm) products. Products resulting from C-0 cleavage are shown to the left; products from 0-0cleavage are to the right. Only doublet states arising from interactions in C, configuration are shown.
Several tests have been performed in order to ascertain the origin of this emission. As a result from these tests the emission is found (i) to require the occurrence of both (193 and 248 nm) laser pulses; (ii) to be dependent on the delay time Ari, Le., to reflect the temporal behavior of an intermediate produced by the 193-nm laser; (iii) not to be caused by a subsequent chemical mechanism following the 248-nm photolyis, i.e, O(lD) azomethane (where O(lD) would be formed from O3 which builds up from 0 atoms generated in a side reaction at 193 nm); and (v) not to arise from the excitation of OH(X211) at 248 nm. As a consequence we conclude that the O H (A2Z+) emission is the result of a direct formation of this species in the 248-nm photolysis of C H 3 0 2 . According to thermochemistry the process
+
+ hv(248 nm)
CH302
-
I
I
/
06
04
,
,
,
08 t i m s
I
IO
Figure 5. Observed change of fluorescence intensities of OH and CH30 upon irradiation of the reaction mixture with a 248-nm laser pulse (see text).
[l@lli24,1 *OI’Pi
.-
I
C H 2 0 ( g i A l )+ OH(A2Z+) ( I )
is exothermic by 183 kJ mol-’. Simultaneous excitation of both C H 2 0 and O H is energetically impossible with 248-nm radiation in a one-photon process. The dynamical origin of the formation of rotationally hot OH(A2Z+) in the photofragmentation of C H 3 0 2is at present unresolved. A first assignment, however, can be obtained from consideration of the correlation diagram (Figure 4). Products originating from C-0 cleavage (CH, + 0,) are shown to the left and products arising from 0-0 cleavage (CH,O 0, CH,O O H ) to the right. Only doublet states originating from C, interactions of the fragments are shown. As can be seen, there are a_ large number of adiabatic correlations between the dissociating B2A” state and the fragments, allowing formation o-f excited product-states such as 02Ja’Ag), O(lD), CH20(a3A”), and C H 2 0 ( A i A ” ) . The formation of OH(A2Z+)on the other hand is-not allowed by orbital symmetry since OH(A22+) C H 2 0 (X’A,) in C, geometry resolve into a state of character A’ and not A”. As a consequence, the observed emission cannot be the result of an adiabatic dissociation process unless we assume the complete loss of symmetry elements. We propose therefore that OH(A2Z+)+ C H 2 0 ( R ’ A l )aresenerated via the decomposition of an excited state of CH,OOH(A) (not shown in Figure 4) which
+
+
+
itself is formed by isomerization from CH302(B2A”)in a diabatic process. This complex dissociation route is suggested here because the ground state of C H 2 0 0 H has been assigned by a b initio calculation to be of character 2A’i’and therefore an excited state of the same character may exist. The remaining requirements then are that C_H200H(A)is at energies less than or equal to that of CH302(B2A”)and that there is no substantial barrier to isomerization between the two excited species. Although the latter requirement is in contrast to the 160 kJ mol-’ presently accepted for the isomerization barrier between C H 3 0 2and CH,OOH in their ground states,* there is no reason to assume that this may also be the case for this process between these species upon electronic excitation. The present explanation of OH(A2B+)formation is somewhat speculative and certainly not unique. First, in the_ case that torsional and/or other out-of-plane motions of CH302(BZA”)were important, the C, correlation rule may not apply and OH(A2Z+) could be produced adiabatically. Second, OH(A2Z+) could also be formed as a result of a collision of an O(’D) fragment with a methyl group H atom during the dissociation process. Clearly, further experimental and theoretical investigations are required to differentiate between these possibilities. The quantum yield of OH(A2Z+) formation in the 248-nm photolysis of C H 3 0 2has as yet not been determined. However, the emission intensity is found to be strong enough to utilize this technique for the detection of C H 3 0 2in kinetic experiments. First applications have been described’ and will be further substantiated in a forthcoming publication.i0 It may be worthy of note that with this technique no interference with H 0 2 has been noted, excluding the possibility of OH(A2Z+)formation from this species with 248-nm radiation. 2. Formation of OH(J?II) and CH,O(*E). In addition to the emission measurements described above, a series of experiments to determine photofragments in their ground electronic states have been performed using laser-induced fluorescence (LIF). Both OH(X2H) and CH30(X2E) were detected. Although a systematic investigation of C H 2 0 , the coproduct of OH, has as yet not been made, CH,O and O H are assumed to reflect the total photofragmentation of C H 3 0 2by 0-0 cleavage, viz., C H 3 0 2+ hv
-
-
CH,02(B2A”)
-
[CH200H(2A”)]
-+
CH30(g2E)
+ 0(3P,’D)
OH(X211) + CH20(R(,H,A)
Figure 5 represents the observed temporal changes of the LIF signals for O H and C H 3 0 in 193 nm pulse laser irradiated mixtures of azomethane/O, upon irradiation with a 248-nm laser pulse. As can be seen, both signals increase instantaneously, indicating the net generation of O H and C H 3 0 from C H 3 0 2at ( 1 1 ) Kamiya, K.; Matsui, H.; Asaba, T.Private communication cited in: K. Saito, R. 110, T. Kakumoto, A. lmamura J . Phys. Chem. 1986, 90, 1422.
2966
J . Phys. Chem. 1990, 94, 2966-2912
248 nm. From a calibration of the fluorescence intensities using the 193-nm photolysis of HNO, and C H 3 0 N 0 , respectively, the quantum yields of O H and C H 3 0 formation can be determined as cp(0H) = A[OH]/A[CH,O2] = 0.06 f 0.03 (p(CH30) = A [ C H 3 0 ] / A [ C H 3 0 2 ]= 0.2 f 0.1 where A[CH302]represents the amount of CH3O2 dissociated per 248-nm laser pulsed. The error limits are estimates based on the accuracies of the laser pulse energy measurements (fl5%) and the accuracies of the absorption coefficients (*10%). The concentrations O H and C H 3 0observed prior to the 248-nm pulse are the result of a reaction of hot C H 3 radicals, generated in the 193-nm photolysis of azomethane, with O,.'O The amount of C H 3 lost in these channels relative to C H 3 0 2formation can also be estimated from the calibrated fluorescence intensities. We obtain 0.55% and 2.7% for O H and C H 3 0 , respectively. These figures clearly exclude the possibility that the increase of O H and C H 3 0 observed after the 248-nm pulse is also due to a reaction of hot C H 3 with 0 With the ratio of the optical densities [ A M ] a ~ 9 ~ / [ C H 3 0 0 2 ] ~12~ and a typical ratio of the laser photon densities of N,Id3)N 248 N 0.5 the total concentration of photofragments from C 4 0 2 can at best be 10% that from azomethane (-3 X IO" cm-, vs 3 X 10I2 cm-,). As a consequence we would expect a correspondingly small increase of the O H and C H 3 0 signals even with unit quantum yield for CH3 formation. The sum of the quantum yields of O H and C H 3 0 formation reflects the total quantum yield for 0-0bond cleavage in the 248
-
nm photolysis of CH302. From mass balance considerations we then conclude that the dominant products (74 f 13%) of the photodissociation are CH, + 02.The correlation diagram of Figure 4 suggests that the O2 product must be formed in the excited (ala,) state. Since the quenching of 0 2 ( a 'Ag) by O2 is cm3 s-')12 emission from this relatively inefficient ( k = 2 X state should be observable. However, in view of the low concentrations expected (-2 X IO" cm-,) we have not attempted this experiment. The photodissociation of CH302 in its first UV band has to our knowledge not been investigated before. Compared to the photodissociation of H 0 2 , however, there appears to be a distinct modification in the product routes. Whereas for H 0 2 the dominant products are OH(X211) + 0('D),4*6the corresponding channel for CH302, namely CH30(X2E) 0, only accounts for 20 f 10%. The dominant photodissociation channel of C H 3 0 2 , namely 0, elimination, is apparently prevented for H 0 2 due to the existence of a high barrier along this coordinate. Finally, formation of OH(A2Z+),which for CHJO2 is interpreted as a result of a complex dissociation pathway, is endothermic for H 0 2 with 248-nm radiation and only accessible at 1 1 9 0 nm.
+
Acknowledgmenr. Financial support of this work by Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 93, Schwerpunktsprogramm MAP) and by Fonds der Chemischen Industrie is gratefully acknowledged. (12) Becker, K. H.; Groth, W.; Schurath, U.Chem. Phys. Lett. 1971.8, 259.
Decomposltlon of Propanal at Elevated Temperatures. Experimental and Modeling Study Assa Lifshitz,* Carmen Tamburu, and Aya Suslensky Department of Physical Chemistry, The Hebrew University, Jerusalem 91 904, Israel (Received: August 7, 1989: In Final Form: October 16. 1989)
The thermal decomposition of propanal (CH3CH2CHO)highly diluted in argon was studied behind reflected shocks in a mol/cm3. pressurized-driver single-pulse shock tube over the temperature range 970-1300 K and overall densities of -2.5 X Under these conditions the following decomposition products were obtained (in order of decreasing abundance): CO, C2H4, CH,, C2H6, C3H8,C3H6,C4H6,and C2H2. Mixtures containing 0.1% propanal and I % toluene in argon showed a factor of -3 decrease in the production rate of ethylene, indicating the involvement of a free-radical mechanism. Two dissociation reactions initiate the decomposition of propanal: ( I ) C2H5CH0 C2H5' + CHO' and (2) C2H5CH0 CH,' + CH,'CHO. I t then proceeds via a free-radicalchain mechanism where the most important reactions are C2HSCH0+ R' CH3CH2CO' + RH, CH3CH2CO' C2H5' + CO, and CH3CH2CO' CH3' + CH2C0. The overall decomposition of propanal was computer modeled with a reaction scheme containing 22 species and 52 elementary reactions. A sensitivity analysis showed that the system is very sensitive to the two initiation reactions and to the two-channel dissociation of the radical CH3CH2CO'. The concentration ratio [C2H4]/[CH4]is highly dependent on k(C2H5+CO)/k(CH3+CH2=CO). A ratio of 100 in the rate constants accounts for a ratio of -30 in the concentrationsof ethylene and methane. There are four major free radicals in the system, C2H5, CH3, and H, and HCO. The modeling studies show that they reach steady-state concentrations after several hundred microseconds. The rate constant for the overall pyrolysis of propanal over the temperature and pressure ranges covered in this investigation is k,,,,, = 1016.39 exp(-73 X I03/RT)s-', where R is expressed in units of cal/(K mol).
-
-
-
--
-
Introduction One of the major reactions in the thermal decomposition of cyclic ethers is the isomerization to aldehydes and/or ketones. Crotonaldehyde (butenal) is a major product in the decomposition of 2.3-dihydrofuran' wheras ethylene oxide isomerizes to acetPropionaldehyde (propanal) and acetone are the major products in the isomerization of propylene ( I ) Lifshitz, A.; Bidani, M. J . Phys. Chem. 1989, 93, 1139. (2) Setser, D. W. J . Phys. Chem. 1966, 70, 826. ( 3 ) Lifshitz, A.; Ben-Hamou, H. J . Phys. Chem. 1983,87. 1782
Whereas the decomposition of acetaldehyde has been thoroughly studied both at low and at high temperature^,',^ as far as we are aware only low-temperature studies of propanal pyrolysis have been reported. (4) Hardwick, T. J. Can. J . Chem. 1968, 46, 2454. (5) Blades, A. T. Can. J . Chem. 1968, 46, 3283. (6) Flowers, M. C. J . Chem. SOC.,Faraday Trans. 1 1977, 73, 1927. (7) Colket 111, M. B.; Nageli, D. W.; Glassman, I. Inf. J . Chem. Kiner. 1975, 7, 223. (8) Leport, M. J. Chim. Phys. 1937, 34. 206.
0022-3654/90/2094-2966$02.50/00 1990 American Chemical Society