J. Phys. Chem. 1989, 93, 6059-6064
6059
Photodissociation of Formaldoxime and Its Methylated Homologues: Search for H,CN Fluorescence Paul J. Dagdigian,* Department of Chemistry, The Johns Hopkins University, Baltimore, Maryland 21 218
William R. Anderson, Rosario C. Sausa, and Andrzej W. Miziolek Ballistic Research Laboratory, Aberdeen Proving Ground, Maryland 21 005-5066 (Received: February 21, 1989) The photodissociation of formaldoxime (H2CNOH),formaldazine (HzCNNCH2),and their methylated homologues at 193 and 248 nm has been studied with emphasis on oxime photolysis at 193 nm. The primary objective was to search for laser-indud fluorescence (LIF) of the RlR2CN radical (Rl, R2 = H or CH3) which is expected to be produced as a primary product of the photolysis. No LIF was observed upon irradiation in the regions where the radicals are known to absorb, 280-295 nm. Assuming the density of RlR2CN equals that of OH produced by photolysis of the oximes at 193 nm, upper limits for the quantum yield of fluorescence of 0.01 are calculated for CH3CHN and (CH,),CN and of 0.1 for H2CN. We therefore conclude it is highly likely the upper states of these radicals probed in the 280-295-nm region are predissociative, which is consistent with results of previous absorption experiments. The LIF spectra of the OH fragments produced from photolysis of the oximes at 193 nm were also studied. The dependence of the OH fragment LIF intensity on excimer laser power indicates a one-photon dissociation process, as expected from energetic considerations. The nascent OH rotational distribution is much hotter than for a room temperature Boltzmann distribution; the ratio of densities for v = 1 to u = 0 vibrational levels is 0.02 & 0.01. Only 2.5 kcal/mol of the excess energy appears as OH product internal energy, thus leaving about 80 kcal/mol to be divided between the RlR2CN fragment and translation. CN in its ground X2Z+ state was also detected via LIF in its B-X band system upon 193-nm photolysis of CH3CHNOH. Results of a power dependence study were consistent with a single-photonprocess to yield this product, which is energetically possible if the products are CH4 + CN + OH. However, the amount of CN produced is much less than that of OH. In addition to these results, LIF of an unknown molecule was observed in the 280-285-11111 region upon 193-nm photolysis of all three oxime precursors and an unidentified prompt emission was observed upon 193-nm photolysis of CH3CHNOH.
Introduction There has been considerable interest in the development of sensitive laser spectroscopic probes for the detection of trace transient species in combustion Laser-induced fluorescence has been utilized for the observation of a number of free-radical combustion specie^.^ In this paper, we describe our attempt to observe laser-induced fluorescence of the methyleneamidogen radical, HzCN, which is prepared by excimer laser photodissociation of formaldoxime [H2CNOH]. We have also investigated the photodissociation of the methylated homologues of formaldoxime, namely acetaldoxime [CH3CHNOH] and acetoxime [(CH,),CNOH]. We also report the internal state distribution of the companion photofragment, hydroxyl, for these three precursors. Methyleneamidogen was first observed by Cochran et al.4 by ESR spectroscopy in an argon matrix and has been observed in other experiments by the same t e c h n i q ~ e . ~ -The ~ ultraviolet absorption spectrum of HzCN was first observed through flash and subsequently in photolysis of formaldazine [ (H2CN),] the photolysis of forma1d0xime.l~ The observation of the same bands from these two precursors supported their assignment to ( I ) Crosley, D. R., Ed. Laser Probes f o r Combustion Chemistry; ACS Symposium Series; American Chemical Society: Washington, DC, 1980; Vol. 134. (2) Bechtel, J. H.; Dasch, C. J.; Teets, R. In Laser Applications; Ready, J. F., Erf, R. K., Eds.; Academic Press: New York, 1983; Vol. 5, p 129. (3) (a) Crosley, D. R. Opt. Eng. 1981,20,51 I . (b) Crosley, D. R.; Smith, G. P. Opt. Eng. 1983, 22, 545. (4) Cochran, E. L.; Adrian, F. J.; Bowers, V. A. J . Chem. Phys. 1962, 36, 1938.
(5) Banks, D.; Gordy, W. Mol. Phys. 1973, 26, 1555. (6) Behar, D.; Fessenden, R. W. J. Phys. Chem. 1972, 76, 3945. (7) Fujiwara, M.; Tamura, N.; Hirai, H. Bull. Chem. SOC.Jpn. 1973,46, 701. (8) Symons, M. C. R. Tetrahedron 1973, 29, 615. (9) Morgan, C. U.; Beyer, R. A. Combust. Flame 1979, 36, 99. (IO) Ogilvie, J. F.; Horne, D. G. J . Chem. Phys. 1968, 48, 2248. (11) Ogilvie, J. F. Can. J . Spectrosc. 1974, 19, 89. (12) Horne, D. G.; Norrish, R. G. W. Proc. R. SOC.London, A 1970,315, 301. (13) Horne, D. G.; Norrish, R. G. W. Proc. R. SOC.London, A 1970,315, 287.
0022-3654/89/2093-6059$01.50/0
H2CN,and deuterium-substitution studies confirmed the presence of two hydrogen atoms in the molecule.I0 N o rotational structure could be resolved in these bands. Recently, the infrared spectra of H2CN and its deuterated counterparts have been observed by Jacox in an argon matrix.14 All the vibrational bands have been assigned. The ultraviolet absorption spectrum was also observed, ' with only a small matrix shift. The methylated homologues of methyleneamidogen, namely CH3CHN and (CH3)2CN,were also observed in the gas Dhase bv flash Dhotolvsis. in this case. of the precursors acetaldazine [(CN3CHN),] -and dimethylketazine [((CH3)2CN)2] .12 Methyleneamidogen is believed to be formed in the early stages of the decomposition of nitramines, a class of important highenergy molecule^.^^^'^ Using ESR spectroscopy, Morgan and Beyerg observed H2CN, along with nitrogen dioxide, as one of the species present in the vapors produced by the slow pyrolysis of cyclotetramethylenetetranitramine(HMX) near its melting point. The electron spin resonance spectrum they obtained for the matrix-deposited vapors from this pyrolysis was essentially identical with spectra of H2CN obtained by other methodse This transient molecule is also believed to be formed by the dissociation of methylenenitramine [CH2NN02],which has been experimentally identified as a primary decomposition product in the molecular beam infrared multiphoton dissociation of cyclotrimethylenetrinitramine (RDX).I7 Methyleneamidogen may also be important as an intermediate in the reaction of hydrogen atoms with HCN.18 Further studies of the role of methyleneamidogen in nitramine decomposition would clearly benefit from a sensitive diagnostic tool for the detection of this species. The most prominent features (14) Jacox, M. E. J. Phys. Chem. 1987, 91, 6595.
( I 5 ) Fifer, R. A. In Fundamentals of Solid-Propellant Combustion; Kuo, K. K,,Summerfield, M., Eds. Prog. Aeronaut. Astronaut.; American Institute of Aeronautics and Astronautics: New York, 1984; Vol. 90, p 177. (16) Melius, C. F.; Binkley, J. S. Proceedings of the 21st International Symposium on Combustion; Combustion Institute: Pittsburgh, PA, 1986; p 1953. (17) Zhao, X.;Hintsa, E.; Lee, Y . T. J . Chem. Phys. 1988, 88, 801. (18) Bair, R. A.; Dunning, T. H., Jr. J . Chem. Phys. 1985, 82, 2280.
0 1989 American Chemical Society
6060 The Journal of Physical Chemistry, Vol. 93, No. 16, 1989 in its electronic absorption spectrum occur near 280 nm.l&I3 There actually appear to be two electronic transitions occurring in this spectral region."$'9 Quantum chemical calculations have also been carried out in order to estimate vertical excitation energie~.'~ These calculations give a reasonable explanation of the electronic spectrum of this radical and provide assignments for the electronic states involved in the bands near 280 nm. In this paper, we report a search for laser-induced fluorescence of methyleneamidogen and its methylated homologues upon excitation of the bands previously observed by flash photolysis. In view of the fact that no rotational structure was observed with a 10.7-m spectrograph with 0.3 cm-l spectral resolution,'' the excited states of H2CN may in fact be predissociated. Jacox14 observed that the ultraviolet absorptions of a matrix-isolated sample decreased upon exposure to light from a medium-pressure mercury arc, while they remained unchanged when the sample was irradiated with light from an arc of wavelengths greater than 280 nm. If the excited state were strongly predissociated, it would, of course, lead to a negligible fluorescence quantum yield. It is interesting in this regard to make the comparison with the isoelectronic formyl radical, HCO. Most of the excited levels of HCO are diffuse, except for those with K' = 0.20 A laser optogalvanic study2' revealed, in fact, a strong variation of the line width in the A2A''(0,9,0) K' = 0 manifold. In spite of this predissociation, H C O :as been succe2sfully detected by fluorescence excitation in the A2A"(0,9,0)-XzA'(O,O,O) band.22sz3 More recently, the formyl radical has been detected both in a and in a flamez5 by two-photon-resonant multiphoton ionization near 390-400 nm via its 3s and 3p Rydberg states. In view of the importance of developing a diagnostic for H2CN, it was deemed worthwhile to investigate whether fluorescence could be observed with this molecule upon excitation of its known ultraviolet band systems. We have chosen in this study to prepare methyleneamidogen by excimer laser photolysis of formaldoxime since the other photofragment is the hydroxyl radical, which is easily detected by fluorescence excitation. Formaldoxime exists at room temperature as a polymer and must be heated to produce the monomeric vapor.26 Detection of the hydroxyl radical allows indirect verification of the presence of the oxime precursor and also provides a way to estimate our detection sensitivity for methyleneamidogen. Because of the difficulty in producing formaldoxime vapor, we have also investigated the photodissociation of its methylated homologues, acetaldoxime [CH3CHNOH] and acetoxime [(CH,),CNOHJ. These precursors are expected to yield CH3CHCN and (CH3)*CN,whose absorption spectra have been previously observed by flash photolysis of the corresponding azines.'* These molecules were also investigated in the hope that their predissociation rate may be significantly less than that of H2CN. Finally, the photodissociation of formaldazine and acetaldazine was also briefly studied.
Experimental Section These experiments were carried out in a large vacuum chamber normally used for molecular beam scattering s t ~ d i e s . ~ ' The photolysis source was an excimer laser (Lambda Physik EMGIOIMSC) usually operated at 193 nm with ArF; typical pulse energies of the unpolarized output were 10-20 mJ in a 1 cm X 3 cm rectangular area at the photolysis zone. A few experiments with acetaldazine precursor were carried out at 248 (19) Adams, G. F.; Yarkony, D. R.; Bartlett, R. J.; Purvis, G. D. Inf. J . Quantum Chem. 1983, 23, 437. (20) Johns, J. W. C.; Priddle, S. H.; Ramsay, D. A. Discuss. Faraday Soc. 1963, 35. 90. (21) Vasudev, R.; Zare, R. N. J . Chem. Phys. 1982, 76, 5267. (22) Konig, R.; Lademann, J. Chem. Phys. Left. 1983, 94, 152. (23) Stone, B. M.; Noble, M.; Lee, E. K. C. Chem. Phys. Lett. 1985,118, 83. (24) Tjossem, P. J. H.; Goodwin, P. M.; Cool, T. A. J. Chem. Phys. 1986, 84, 5334. (25) Bernstein, J . S.; Song, X.-M.; Cool, T. A. Chem. Phys. Left. 1988, 145, 188. (26) Scholl, R. Berichfe 1891, 24, 573. (27) Dagdigian, P. J. J . Chem. Phys. 1989, 90, 2617
Dagdigian et al. nm with KrF. The tunable ultraviolet probe laser beam, obtained by frequency doubling the output of a Nd:YAG pumped dye laser (Quantel), crossed at right angles to the excimer laser beam along the long dimension. Typical probe pulse energies at the apparatus were 200 pJ in a 4 mm diameter beam for laser fluorescence detection of O H in the A2Z+-X211 (1,O) and (2,l) bands at 280-290 nm. Slightly higher power (500-800 pJ) was employed in the search for fluorescence from HzCN and its homologues. Fluorescence detection of C N through excitation of its BzZ+-X2Z+ (0,O) band near 388 nm was accomplished by mixing the dye laser output with the residual 1.06 pm Nd:YAG fundamental. The incident pulse energy of the probe radiation was approximately 100 pJ. Fluorescence excited by the probe laser was collected with a three-lens telescope and was detected with a photomultiplier (EM1 9813QB), whose output was directed to a gated integrator (Stanford Research Systems). For experimental convenience, the fluorescence telescope has a 90' bend in it. To reduce the scattered light background, a dichroic mirror, with reflectivity peaked over 300-350 nm, was used to make this bend. Below 300 nm, the reflectivity dropped rapidly, reaching 50% at 285 nm. For fluorescence detection of CN, the dichroic mirror was replaced with an aluminized one, and excimer laser scattered light was eliminated with the insertion of a 390-nm center wavelength, 10-nm fwhm filter. In many runs, excitation spectra were acquired under computer control (DEC LSI-l1/23), and the spectra were stored on magnetic diskettes for later analysis on another laboratory computer (Apple Macintosh). The precursors acetaldoxime and acetoxime were obtained from Aldrich Chemicals. The stated purities were 95 and 98%, respectively. Acetaldazine was synthesized by the reaction of acetaldehyde and hydrazine hydrate according to the procedure of Curtius and Zinkeisen.28 The reaction was carried out at the interface of a solution of acetaldehyde in ethyl ether with a solution of hydrazine hydrate in water. This method of carrying out the reaction avoids a serious explosive boiling problem due to the strong exothermicity of the reaction and the low boiling point (20.8 "C) of acetaldehyde, in contrast to our first seemingly simpler attempted procedure which involved dripping the hydrazine hydrate solution directly into the acetaldehyde. The acetaldazine was driven into the ether phase by addition of potassium carbonate and was purified by fractional distillation. Checks of the infrared and mass spectra against literature infrared29and mass30spectra were made to ensure the identity of the product. The latter were run on a gas chromatograph/mass spectrometer and indicated high purity (>98%) of the product. Formaldazine polymer was prepared simply by the addition of formaldehyde to hydrazine hydrate using the procedure of Pul~ermacher.~'The alternative procedure of Hofmann and Storm3z yields a ring compound, tetraformaltrisazine, C4HIZN6.rather than the desired polymer of formula (CzH4Nz),.33 A check of the elemental c o m p o ~ i t i o nof~ ~ our product showed the correct elemental stoichiometry. An infrared spectrum of the polymeric sample was also taken and, in general, agreed with literature spectra of the monomer deposited in a low-temperature m a t r i ~ . ~ ~ , ~ ~ Some comments about the synthesis of formaldoxime are in order. This precursor was first prepared by the procedure given by Horne and Norrish.I2 A saturated solution of formaldoxime trimer hydrochloride (Aldrich) was neutralized with an equivalent of sodium carbonate. The formaldoxime trimer was allowed to precipitate out by keeping the solution in the cold, and the resulting solid was filtered. Final purification was carried out by subli(28) Curtius, Th.; Zinkeisen, E. J . Prakf. Chem. 1898, 58, 325. (29) Tabacik, V.; Pellegrin, V. Spectrochim. Acra 1979, 35A. 961. (30) Zeeh, B.; Beutler, R. Org. Mass Spectrom. 1968, I, 791. (31) Pulvermacher, G. Berichre 1893, 26, 2360. (32) Hofmann, K. A.; Storm, D. Berichfe 1912, 45, 1725. (33) Neureiter, N. P. J . Am. Chem. SOC.1959, 81, 2910. (34) Analytical and Physical Measurements Service, E. I. duPont de Nemours & Co., Inc. (35) Ogilvie, J. F.; Cole, K. C. Spectrochim. Acfa 1971, 27A, 877. See also Ogilvie, J. F.; Cyvin, S. J.; Cyvin, B. N. J . Mol. Sfruct. 1973, 18, 285. (36) Bondybey, V. E.; Nibler, J. W. Spectrochim. Acta 1973, 29A, 645.
Photodissociation of Formaldoxime mation. The procedure gave low and erratic yields of the polymer. We found that the procedure of Scho11,26involving the reaction of formaldehyde with hydroxylamine, gave much higher yields and purer product. Water was removed from the pastelike product by pumping under vacuum. The identity and purity of the solid product was checked by infrared and gas chromatograph/mass spectra.,' Unfortunately, the only infrared spectrum available in the literature is for the gas-phase monomer.38 Also, no literature mass spectrum could be found. However, our observed mass spectrum exhibits approximately the fragmentation pattern expected for H2CNOH, and the physical properties match known values (see below in this section). Two types of photodissociation experiments were carried out for the study of both nascent and thermalized products. In the former, the vacuum chamber was evacuated with a baffled diffusion pump. Typical pressures for the precursors were 2-5 mTorr, with pump/probe delays of 2-4 ps. For the study of thermalized products, the chamber was evacuated with a roughing pump, and nitrogen was added to give a total pressure of 0.6-1 Torr; pump/probe delays were 4-50 ps and usually greater than 20 MS. Acetaldoxime, acetoxime, and acetaldazine, which are monomeric precursors, were admitted into the vacuum chamber by means of a needle valve on an evacuated flask containing the degassed material. Under normal conditions (298 K and 1 atm), formaldazine exists as a solid polymer, while formaldoxime is a solid trimer. In our first experiments, the desired monomeric precursors were obtained by depolymerization of the solids in a heated crucible located just under the photolysis/detection zone. This arrangement generated a significant amount of particulate material, which led to laser scattering that interfered with the taking of laser fluorescence excitation spectra. The problem was exacerbated when an inert diluent was added to the chamber (even at only 1 Torr pressure) because the buoyancy of the diluent gas caused a large amount of particulates to float in the laser beams, or the condensation rates of the generated vapors were affected. To avoid the light-scattering problem, we first attempted to make monomeric formaldoxime liquid, as has been done before.'3,26*38The liquid was prepared by depolymerization of the solid at 135 "C and collection at 70 "C in a simple trap-to-trap distillation apparatus. (In our first attempt at this procedure, we tried to collect the liquid at 50 "C, which Horne and NorrishI3 reported as their storage temperature for the liquid, but the liquid quickly polymerized to the solid.) This procedure was beset by two difficulties: (1) the high depolymerization temperature caused the decomposition of a large fraction of the starting material and the collection of the decomposition products in the liquid, as evidenced by the appearance of brown/black solids, and (2) the collection temperature at 70 "C is near enough to the boiling point (1 09 "C) of monomeric formaldoxime to lead to a safety risk since the procedure must take place in a closed system. We found that the best way to handle the gaseous formaldoxime monomer was to depolymerize the solid polymer in a test tube and inject it directly through a long, mildly heated section of metal tubing into the photolysis chamber. This procedure avoids the formation of particulates in the chamber. Also, there is very little decomposition because the temperature required is much lower than for formation of the liquid. We found that a depolymerization temperature of 80-90 "C produced a pressure of several milliTorr in the chamber with the pumps on. This is an adequate pressure to perform photolysis experiments. We did not devise a very satisfactory technique to inject formaldazine monomer into the chamber. Use of a technique similar to that which worked well for formaldoxime (depolymerization in a heated side arm to the chamber) was thought to be difficult because the depolymerization temperature (ca. 200-220 oC1o~") is high enough to cause materials problems since the heated components would be exposed to room air. However, in retrospect, this technique might work well because a lower temperature could (37) Infrared and mass spectra of formaldoxime will be furnished by W. R. A. upon request. (38) Califano, S.; Liittke, W. 2. Physik. Chem., Neue FoIge 1956, 6, 83.
The Journal of Physical Chemistry, Vol. 93, No. 16, 1989 6061 still be sufficient to produce several milliTorr of monomer. (Lower pressures of monomer are required in these experiments than were used in the flash photolysis experiment^.'^'^) In any case, it is clearly probable that the formaldazine monomer could be produced nicely in the photolysis chamber by heating a side arm contained under vacuum. Such apparatus was not readily available to us. Therefore, most of the experiments on azines were performed with acetaldazine.
Results and Discussion A . Fragmentation Energetics. There is considerable uncertainty in the R,R,CN-OH bond dissociation energy for the oximes. The bond dissociation energy for formaldoxime (R, = R2 = H ) can be estimated by using its heat of formation (AHf0298 = 0 kcal/mol) derived from bond additivity considerations by Benson and O'Nea139 and the H-HCN binding energy (19 kcal/mol) calculated by Bair and Dunning18for the H2CN species. With the aid of the well-determined heats of formation of H and HCN,40we estimate a bond energy of 74 kcal/mol for formaldoxime. The bond-additivity-corrected Merller-Plesset fourth-order perturbation theory (BAC-MP4) method has been used to calculate the thermochemistry of a number of combustion species.16 Using the BAC-MP4 values4' for AHf"298 of H2CNOH and H2CN (5.1 and 58.9 kcal/mol, respectively), we obtain a H2CN-OH bond energy of 63.0 kcal/mol, which is in moderate agreement with the previous cruder estimate. We would expect methylation not to change drastically the R,R,CN-OH bond energy. Benson and O'Nea139 estimated a value of 48.4 kcal/mol for acetaldoxime (R, = CH,, R2 = H) from available experimental unimolecular decay kinetic data. This value seems quite low in light of the estimates for formaldoxime. Indeed, Benson and O'Neal questioned the validity of the experimental data. We can obtain another value using their estimated39AHf"298for acetaldoxime (-7.3 kcal/mol) and a BACMP4 value4' for CH3CHN (50.1 kcal/mol). We obtain a CH,CHN-OH bond energy of 66.6 kcal/mol, which is quite similar to that for formaldoxime. At present, there is no information on acetoxime (R, = R, = CH,); however, we expect a similar bond energy. The ultraviolet spectra of the oximes contain two absorption regions, a weak diffuse band near 210-213 nm and an intense band with a maximum below 190 nm." Accordingly, our photodissociation studies were carried out only with an ArF excimer laser (193 nm) as the photolysis light source. This implies that there is approximately 85 kcal/mol of energy available to the photofragments, if a RIR2CN-OH bond energy of 63 kcal/mol is assumed. B . OH Fragment Internal State Distributions. The O H product was observed by laser fluorescence excitation in the A2Z+-X211 (1,O) and (2,l) bands. These particular bands were chosen since they lie very close to the wavelengths of the absorptions previously observed for H2CN and its homologues. The dependence of the O H fluorescence signals for several lines was investigated as a function of excimer laser pulse energy. On a log-log plot the slope was found to equal 1.25 f 0.1. This is roughly consistent with a one-photon photodissociation process. Under thermalized conditions, the O H fluorescence signal decreased with a half-life of approximately 100 ps with respect to the pump/probe delay in 0.7 Torr of nitrogen buffer gas. This is comparable with the expected diffusion time out of the photolysis zone. The OH(u=O) rotational distributions were taken from (1,O) R , and R2 branch intensities. The intensities were converted to rotational populations by using fluorescence excitation line strength factors calculated by the formulas of Greene and Zare.42 It was (39) Benson, S. W.; ONeal, H.E. Natl. Stand. ReJ Data Ser. ( U S . ,Natl. Bur. Stand.) 1970, 21. (40) Chase, M . W., Jr.; Davies, C. A,; Downey, J. R., Jr.; Frurip, D. J.; McDonald, R. A,; Syverud, A. N. J . Phys. Chem. Rex Data 1985, 14, Supplement 1. (41) McKee, M.; Melius, C. F., private communication. (42) Greene, C. H.; Zare, R. N. J . Chem. Phys. 1983, 78, 6741.
6062 The Journal of Physical Chemistry, Vol. 93, No. 16, 1989
Dagdigian et al. I
I
5
E
0.0
T
-
5
.-.... ~
10
-
1
'. . .
.
+ *
~
15 0
*
-A7
~
5
N
10
__
15 0
-
I
IQl 1
*..
--, ' I 10
5
N
I
3
15
N
Rotational state populations for OH(o=O) produced in the 193-nm photodissociation of (a) formaldoxime, (b) acetaldoxime, and (c) acetoxime. The circles and squares represent the F,(Q=3/2) and F2(Q = 1 /2) spin-orbit levels, respectively. The distributions are each separately normalized to a particular F, level. The populations are not divided by the ( 2 J + 1) degeneracy factor. Figure 1.
11
12
n
1
I
34400
P1
34500
34600
laser wavenumber (cm
34700
-'
34800
)
Excitation spectrum near the OH A-X (2,l) band origin for photodissociation of acetaldoxime under thermalized conditions. All the strong lines are due to OH. The QI and R, branches of the (2,l) band are marked. The splitting of the R, branch lines is due to strong satellite transitions near the main branch lines for these low N values. The pressure and pumpprobe delay were 0.66 Torr (nitrogen added) and 25 ps, respectively. Figure 3.
aldoxime1°J'J3 and formaldazine.12 Because of the ease of introducing the precursor into the apparatus, the photolysis of acetaldoxime and acetoxime were more extensively studied. For these molecules, scans were taken over the range 287-295 nm in the regions where CH3CHN and (CH3),CN have been found to absorb. I 2 From our lack of observation of R,R2CN fluorescence and our 34600 34700 34800 observed signal-to-noise ratio for O H lines, we can estimate an laser wavenumber (cm-' ) upper limit to the R1R2CNfluorescence quantum yield, provided we assume that the concentrations of RlR2CN and OH are equal Figure 2. Excitation spectrum near the OH A-X (2,l) band origin for photodissociation of acetaldoxime under nascent conditions. Lines of the in the photolysis zone. Ogilvie and Homelo have estimated the (1,O) band and of the (2,l) QI branch are marked. The pressure and integrated oscillator strength for the 280-nm band of H2CN to pump/probe delay were 4 mTorr and 3 c(s, respectively. be (4 f 2) X lo4. We shall assume that the oscillator strength of the corresponding bands in the radicals produced from aceassumed that the O H photofragments had an isotropic MJ distaldoxime and acetoxime are similar. Since the band is approxtribution. The nascent distributions, which are plotted in Figure imately 200 cm-' wide, this implies an average oscillator strength 1, were found to be somewhat hotter than for thermalized samples. per unit bandwidth of 2 X 104/cm-'. We can estimate the The distributions are essentially identical for all three oxime oscillator strengths of the OH (1,O) and (2,l) bands from the precursors and exhibit a preference for production of the F l ( Q known radiative lifetimes and fluorescence branching =3/2) over the F2(Q=l/2) spin-orbit component, particularly at andf,,, = 4.5 X IO4. j l , o= 3.1 X low N . Our most sensitive searches for RIR2CN fluorescence were The OH(u= I)/OH(u=O) population ratio was also estimated carried out with acetaldoxime and acetoxime and with nitrogen for acetaldoxime precursor since high-N PI and P2 lines of the added as a thermalizing buffer. Under these conditions, lines in (1,O) band overlapped the (2,l) band origin, as shown in Figure the O H (2,l) band could be observed with signal-to-noise ratio 2. We estimate that the nascent u = 1 to u = 0 ratio equals 0.02 of greater than 100 in some scans. The oscillator strengths given f 0.01. In deriving this ratio, a correction for p r e d i s s o ~ i a t i o n ~ ~ - ~in the previous paragraph and the previously determined OH@= 1) of the low N levels of u' = 2 was made. From the measured OH to OH(u=O) population ratio imply that the ratio of an R,R2CN internal state distribution, we conclude that an average of only signal to that of a low N line in the OH (2,l) band should equal 2.5 kcal/mol appears as excitation of the OH fragments for approximately unity if the fluorescence quantum yield @ of acetaldoxime precursor. Similar results apply to formaldoxime R,R2CN were unity, assuming also that the quantum yield for and acetoxime. Thus, the overwhelming majority of the energy O H excited fluorescence is also unity. This suggests that @ is less available to the fragments must be present as translational recoil than 1% for CH3CHCN and (CH3)2CN. and internal excitation of the R,R2CN fragment. The N2 pressure and delay time of these scans were sufficiently C. Search f o r H2CN Fluorescence. With O H detected as a low to prevent vibrational relaxation49in the OH(X211) state. In photofragment from the oximes, it can be assumed that H2CN, fact, the zero-pressure fluorescence quantum yield for OH(u'=2) or its homologues, will be present in the photolysis at the same is considerably less than unity because of prediss~ciation.~~*" The concentration as O H if no subsequent excimer-laser-induced measured radiative lifetime (ca. 450 ns, which is significantly decomposition or reactions of RlR2CN occur. (In subsection D, longer than the lifetime of low N levels in u' = 2 under colliwe address the possibility that this assumption may not be correct.) sion-free condition^^^^") of the laser-excited OH(u'=2) indicated Extensive searches for laser fluorescence signals from H,CN and its homologues were carried out, both unier nascent i n d ther(45) Dimpfl, W. L.; Kinsey, J. L. J . Quant. Spectrosc. Radiat. Transfer malized conditions. No fluorescence signals attributable to these 1979,2 1 , 233. species could be found. For the formaldoxime precursor, scans (46) McDermid, I. S.; Laudenslager, J. B. J . Chem. Phys. 1982,76, 1824. were taken over 280-286 nm in the region of the strongest ab(47) Crosley, D. R.; Lengel, R. K. J . Quant. Spectrosc. Radiat. Transfer sorption bands seen in the flash photolysis experiments on formi977,17,59. (48) Copeland, R . A.; Jeffries, J . B.; Crosley, D. R. Chem. Phys. Lett. (43) German, K. R. J . Chem. Phys. 1975,63, 5 2 5 2 . (44) Brzozowski, J.; Erman, P.; Lyyra, M. Phys. Scr. 1978, 17, 507.
1987,138, 425. (49) Rensberger, K . J.; Jeffries, J. B.; Crosley, D. R. J . Chem. Phys. 1989, 90,2174.
The Journal of Physical Chemistry, Vol. 93, No. 16, 1989 6063
Photodissociation of Formaldoxime 7
,
8 I
I'
7
6
I
D
'2 I 4
I
5
34050
P.
I1
I
I
341 50
34250
012
,
34350
laser wavenumber (cm-' )
Figure 4. Excitation spectrum around 34 150 cm-' for photodissociation of acetaldoxime under thermalized conditions, showing a band of an unidentified species. Lines of the OH A-X (2,l) band are marked.
that vibrational relaxationSoto the nonpredissociating lower vibrational levels was occurring in the A2Z+state; this would have the effect of substantially raising the fluorescence quantum yield over that for u' = 2 in the absence of collision^.^^^^ Our estimate for the upper bound on Q, would be even further reduced if the quantum yield of O H fluorescence were less than unity. Hence, we conclude Q, is certainly less than 1% for CH3CHCN and (CH3)2CN. The derived upper limit for Q, is approximately 1 order of magnitude larger for H2CN since the O H signals observed for formaldoxime precursor were smaller. While not observing fluorescence attributable to RlR2CN, we did observe weak fluorescence from an as yet unidentifiable species other than OH. Figure 3 shows an excitation spectrum of a thermalized sample of photolyzed acetaldoxime in the region of the O H (2,l) band head. In contrast to the nascent spectrum in Figure 2, the high N lines of the OH (1,O) band are very weak compared to those of the (2,l) band because of rotational relaxation in u" = 1. In addition to the strong O H lines, we can see a weak set of lines with closer spacing which appear to be converging to a band head degraded to the red near 34 650 cm-I. In addition to these lines, a second set of more closely spaced lines were observed around 34 150 cm-I, as shown in Figure 4. The spacing between lines in the latter region is much smaller than in the former. We do not believe that these lines are due to R1R2CN since (i) they do not match the wavelengths reported for RIR2CNfrom the flash photolysis studies and, more importantly, (ii) they appear at identical wavelengths with acetoxime as the precursor. (The lines near 34 650 cm-' also appeared, weakly, with formaldoxime.) In an attempt to identify the molecular carrier of these lines, we measured the radiative lifetimes of several of the lines. We estimate the fluorescence decay lifetime to be approximately 1 ps, roughly independent of the line excited and the nitrogen pressure over 0.3-1 Torr. The precision of our measurement is poor because the signal is small compared with the probe laser scattered light. One possible candidate for the molecular species is methoxy, C H 3 0 , which does have bands in this regions1 and whose radiative lifetime is reported to be 1.5 p s S 2 However, the rotalional structure for a room-temperature sample of methoxy would be expected to be much more dense than the spectra shown in Figures 3 and 4. At present, we do not have a suitable candidate species to attribute to these bands. We do, however, have several comments to make about the bands. First, there is so much interference from the strong O H lines obscuring the short wavelength, better resolved band that it is difficult even to attempt to fit the rotational structure of the band. The OH lines make it particularly difficult to observe the band origin which would facilitate fitting the spectrum. Second, the splitting between the two bands is about (50) Copeland, R. A,; Wise, M. L.; Crosley, D. R. J . Phys. Chem. 1988,
-
~7- ,v i n _ (51) Foster, S. C.; Misra, P.; Lin, T.-Y. D.; Damo,
C. P.; Carter, C. C.; Miller, T. A. J . Phys. Chem. 1988, 92, 5914. (52) Inoue, G.; Akimoto, J.; Okuda, M. Chem. Phys. Lett. 1979,63, 213.
450 cm-I. This mitigates against the possibility that the two bands are subbands of the same vibrational band resulting from spinorbit splitting because the splitting is so large. If the two bands belong to the same molecule, it would therefore seem that they are different vibrational bands. The form of the short-wavelength bands appears to be much simpler than the long-wavelength band. The former may consist of only P and R branches, with no apparent spin splitting, while the latter apparently has more branches. If the bands arise from the same molecular species, it therefore seems likely that the vibrational symmetries of the levels involved in the two bands are different. Alternatively, there could be a Kf-dependejt predjssociation in the short-wavelength band, as in the H C O A2Aff-X2Afspectrum.20s21 We also investigated the 193-nm photolysis of formaldazine and the 193- and 248-nm photolysis of acetaldazine. N o fluorescence signals attributable to any molecular species were observed with the probe laser tuned through the wavelength regions where the appropriate RlR2CN fragments are known to absorb.'&13 The sensitivity of these runs was less than those with oxime precursors. Moreover, our detection sensitivity could not be calibrated in situ since, of course, no OH is produced by the photolysis of these compounds. D. Detection of CN. In estimating the upper bound to Q,, it is possible that our assumption of equal concentrations of the OH and RlR2CN photofragments may not be correct. The flash photolysis work'*13 suggests that the latter is not rapidly consumed by chemical reactions since the RlR2CN absorptions were seen to persist at much higher reactant pressures than used in the present work for at least 80 ps. However, in our experiments, it is possible, at least in principle, that RlR2CN could itself be destroyed in the initial photodissociation since the energy available to the fragments considerably exceeds the R1-R2CN bond energy (see subsection A) and only a small fraction of this energy appears as O H excitation (see subsection B). Thus, both RlR2CN O H and R I C N R2 O H photolysis products could be formed in a one-photon process. The dissociation wavelength in the present experiment (193 nm) is somewhat shorter than in the flash photolysis studies since the Kr flashlamp output there dropped rapidly below 200 nm.I2 The effective wavelength range of the photolysis radiation is actually a convolution of the flashlamp output and the oxime absorption coefficient, which is rapidly rising below 200 nm. Hence, it is possible that fragmentation of RlR2CN is more likely with photolysis using 193-nm radiation, than with a Kr flashlamp. In fact, the appearance of successive absorption maxima in the continuum spectra of both the oxime and azine precursor^^^^^^^^^ as one goes toward shorter wavelengths may indicate the onset of some fraction of higher fragmentation photolysis. In an effort to determine the importance of such destruction of R1R2CN,we looked for production of C N by laser fluorescence excitation. Indeed, we observed C N through fluorescence excitation in its B2Z+-X2Z+ (0,O)band upon 193-nm photolysis of acetaldoxime. No attempt was made to measure quantitatively the C N concentration relative to that of O H because of the widely different wavelength range, and laser and filter bandwidths. However, we estimate that the C N concentration was significantly less than that of OH since comparable photomultiplier signals were observed; the oscillator strength of the C N B-X (0,O)band equals (3.11 i 0.05) X and is hence much larger than for the OH(Au=+l) sequence. The dependence of the C N fluorescence signal was investigated for several lines as a function of excimer laser pulse energy under thermalizing conditions (1 .O Torr of nitrogen added, 4 ps pump/probe delay). On a log-log plot, the slope was found to equal 0.78 i 0.1. This suggests that C N is formed by a one-photon dissociation process from the oxime. The process CH3CHNOH CH4 C N + O H requires approximately 110 kcal/mol. This dissociation pathway is clearly feasible in a one-photon process at 193 nm. In addition to the observation of ground-state C N by laser fluorescence detection, an emission signal coincident with the ArF
+
+ +
-
+
(53) DuriE, N.; Erman, P.; Larsson, M. Phys. Scr. 1978, 18, 39
J . Phys. Chem. 1989, 93, 6064-6069
6064
excimer laser pulse was seen through the 390-nm bandpass filter with the acetaldoxime precursor. We do not believe this is due to direct photolytic production of CN(B2B+)because the decay time of this signal (ca. 1 ps) is much longer than the CN(B) radiative lifetime (ca. 70 nsS3). The dependence of the emission signal on the excimer pulse energy was found to be approximately linear (slope of 0.99 f 0.1 on a log-log plot). However, production of CN(B) in a one-photon process at 193 nm is not energetically allowed. Unfortunately, it was not possible in the present apparatus to take a spectrum of this emission, thus elucidating the species responsible.
Conclusion In this work, we have investigated the 193-nm photodissociation of the simplest oximes, namely formaldoxime, acetaldoxime, and acetoxime. The O H photofragment was detected, and its internal state distribution determined. An unsuccessful search was made to observe laser-excited fluorescence from the other photofragments R,R2CN, where R1 and R2 equals H or CH3. On the basis of the assumption of an equal concentration of O H and RlR2CN upon photolysis of the oxime, we were able to estimate a rough upper bound of 1% for the fluorescence quantum yield CP of the RlR2CNabsorption bands in the 280-295-nm region. This implies that electronic states excited in these transitions are predissociative. Through the detection of CN, there is some evidence that the R,RzCN is itself destroyed to some extent in the initial event of the 193-nm photodissociation of the oxime. This could have the effect of increasing our estimate on the upper bound to CP. The present study indicates that fluorescence excitation of the electronic bands near 280 nm of methyleneamidogen is not a feasible laser diagnostic for this species. The infrared absorption
multiphoton ionization through a Rydberg state, as has been
successfully carried out for HC0.24925 Recently, Marston et al.54 have shown that H,CN can be produced in high yield from the N + CH, reaction. Using the discharge flow/mass spectrometry technique, this group has measured the elementary reaction rate constant for the N H2CN reaction. These results imply that mass spectrometric detection of H2CN may be a viable alternative diagnostic for this species. In addition, it would be interesting to compare the production of H2CN by this chemical method with our photolytic approach.
+
Note Added in Proof. Recently, Duxbury and co-workersss have obtained high-resolution spectra in the 3- and 10-pm and O H overtone regions for formaldoxime and have carried out rotational analyses of these bands. Acknowledgment. We appreciate the assistance of M . A. Schroeder in the synthesis of several of the precursors. We also acknowledge M. A. Schroeder, R. A. Fifer, and K. L. McNesby for their aid in identification of these precursors. We thank B. J. Gaffney for the loan of a distillation apparatus used in the syntheses. P.J.D. acknowledges the support of the U S . Army Research Office under contract No. DAAL-03-88-K-003 1 and the National Science Foundation under grant No. CHE-8700970. A.W.M. acknowledges partial support from the Air Force Office of Scientific Research, Directorate of Aerospace Sciences, under contract No. 89-0017. Registry No. H2CNOH, 75-1 7-2; H2CNNCH2,503-27-5; (CHJ2CNOH, 127-06-0;CH,CHNOH, 107-29-9;OH, 3352-57-6;CN, 207487-5; H2CN, 15845-29-1;formaldazine polymer, 53351-18-1. (54) Marston, G.;Nesbitt, F. L.; Stief. L. Branching Ratios in the Reac-
393.
The n 3s Rydberg Transition of Jet-Cooled Tetrahydropyran, 1,CDioxane, and 1,CDioxane-d, Studled by 2 4- 1 Resonance-Enhanced Multiphoton Ionization -+
Timothy J. Cornish, Tomas Baer,* and Lee G . Pedersen Department of Chemistry, University of North Carolina, Chapel Hill, North Carolina 27599-3290 (Received: February 21, 1989)
The lowest electronic transitions of tetrahydropyran (THP), 1,4-dioxane, and 1,4-dioxane-d8cooled in a free jet expansion have been examined by 2 + 1 resonance-enhanced multiphoton ionization spectroscopy. This two-photon transition in dioxane has not been reported previously since it is forbidden in one-photon absorptions. These n 3s spectra are extremely sharp and intense, consisting primarily of Q branches. The absorption intensity ratio for circularly and linearly polarized light is less than 0.2, indicating that the transition is between states of the same symmetry, A‘ A’ in the case of THP and A, A, for the dioxanes. Assignments were made with the aid of ab initio molecular orbital calculations. Additionally, these calculations show that the Rydberg state is comprised mainly of 3s atomic orbital contributions from the ring carbons in both THP and dioxane. Several of the low-frequency ring vibrations in the Rydberg spectra of THP and dioxane have been assigned.
-
-
-
I. Introduction Recently, a remarkable sensitivity of the n ---* 3s transition energy to subtle changes in the structure of saturated cyclic ketones has been observed. In several studies,’-’ this technique has been useful for the identification of configurational isomers in methyl-substituted cyclohexanones and cyclopentanones using resonance-enhanced multiphoton ionization (2 + 1 REMPI). Absorption peaks of the jet-cooled samples are quite narrow because ( I ) Cornish, T. J.; Baer, T.J . Am. Chem. SOC.1987, 109, 6915. (2) Cornish, T. J.; Baer, T. J . Am. Chem. SOC.1988, 110, 3099. (3) Cornish, T.J.; Baer, T. J . Am. Chem. SOC.1988, 110, 6287.
0022-3654/89/2093-6064$01.50/0
of the long-lived Rydberg state, allowing the detection of slight changes in molecular geometry. These results are of potential utility in structure determination because the n 3s transition is often sharp in many other h e t e r ~ y c l e s .For ~ example, we have found that the Rydberg spectra of cyclic ethers, a class of compounds which serve as the building blocks for numerous biologically important molecules (such as hexose sugars), can be examined in a similar fashion to that performed on the cyclic ketones. That is, changes in the molecule’s lowest lying Rydberg spectrum
-
(4) Robin, M. B. Higher Excited States of Polyatomic Molecules; Academic press: New York, 1975; Vol. I.
0 1989 American Chemical Society
